How much can we earn in retirement without paying federal income taxes?

Update 11/22/2019: After I published a shorter version of this piece on MarketWatch and the story was picked up by YahooFinance as well I got a lot more readers! Thanks and welcome to my blog! Make sure you subscribe to be notified of future blog posts! Both on Yahoo and MarketWatch I saw the expected assortment of hate comments. They fall into two categories, see below plus my response:

  • “I’m a CPA and this doesn’t make any sense!” My response: You’re either not a CPA at all or you’re a really bad & incompetent one. The standard deduction and the 0% bracket for capital gains are all very well-known in the financial/tax planner community. The same goes for the taxability worksheet for Social Security.
  • “How unfair that you retired already and don’t pay taxes while I’m working so hard and pay a lot of taxes!” My response: I hear ya! I’ve paid a ton of taxes throughout my work life. A seven-figure sum, more than most people pay over their entire lifetime. Keep that in mind if you complain about the unfairness of the U.S. federal tax system!

* * *

The question “how much can we earn without paying federal income taxes” is relatively easy to answer for most people. The standard deduction for a married couple is $24,400 in 2019 (if both are under 65 years old) and the top of the no-tax bracket for capital gains is $78,750. So, we can make a total of $103,150 per year, provided that our ordinary income stays below the standard deduction and the rest (2nd bracket + any leftover from the std. deduction) comes from long-term capital gains and/or qualified dividends. With our daughter, we also qualify for the child tax credit ($2,000 p.a.), so we could actually generate another $13,333 per year in dividends or capital gains, taxed at a 15% so that the tax liability of $2,000 exactly offsets the tax credit for a zero federal tax bill.

Once people file for Social Security benefits, though, things become a bit more complicated. That’s due to the convoluted formula used to determine how much of your Social Security is counted as taxable income, see last week’s blog post! So, calculating and plotting the tax-free income limits is a tad more complicated. Oh, and talking about tax planning in retirement: as promised, I will also go through an update on the Roth Conversion strategy for the Becky and Stephen case study from two weeks ago.

Let’s get started…

First, a disclaimer: This exercise is for federal taxes only! That’s good enough for us personally because we live in Washington State, one of the few places without an additional state income tax. If you do have state income taxes, you will probably start owing state income taxes at much lower income levels! Also, all the other disclaimers apply here as well: contact a tax expert before you apply any of this yourself in real life! 🙂   I will also frequently mention capital gains and dividends as tax-advantaged income because long-term capital gains and qualified dividends are taxed at a lower rate. Below, I may sometimes drop the terms “long-term” and “qualified” because it doesn’t always fit into the chart axis labels. But keep in mind that short-term capital gains and non-qualified dividends will fall into the ordinary income bucket, taxed at a higher rate!

Tax-free income limits without Social Security

Just to warm up, here are the income limits for a married couple (both under 65 years old) who file a joint federal tax return. They face a $24,400 standard deduction in 2019 as well as a $78,750 upper limit on the cap-gains no-tax bracket to fill up with long-term capital gains. So, to stay tax-free we need to stay under the blue line in the chart below. That way we’re guaranteed to fill up the standard deduction with either ordinary income or long-term capital gains/qualified dividends and the top off everything with a maximum of $78,750 with capital gains in the first two federal brackets at a 0% marginal rate. Since we have a kid and are eligible for the child tax credit, which could theoretically offset $2,000 in taxes. So, we can push the income limits a little higher:

  • To just under $45k if relying purely on ordinary income, by completely maxing out the 10% brackets and briefly touching the 12% federal bracket,…
  • To over $116k by adding another $13,333 in capital gains/dividends taxed at the 15% rate (provided these are long-term capital gains and/or qualified dividends)
  • Or a combination of ordinary income and tax-advantaged income between the kink points.
Roth Tightrope Chart01
The zero-tax boundaries for a married couple filing jointly in 2019. The $2,000 child tax credit pushes out the boundary even more!

Tax-free income limits with Social Security

What about 25 years down the road, when we no longer claim our daughter as a dependent and we file for Social Security? With Social Security, things become a little bit more complicated. Instead of two, we now have three separate major income categories with their own distinct impact on the federal tax

Taxable Social Security = F1(Social Security , Ordinary Income + Capital Gains)

Tax = F2( Taxable Social Security + Ordinary Income , Capital Gains )

(side note: there’s even a fourth category: Municipal Bond interest income, because that enters the Social Security worksheet formula as well – see last week’s post – but stays tax-free otherwise. But I disregard Muni bond interest for the purpose of today’s post because it’s already messy enough as is)

So, it’s no longer feasible to display the tax-free income limits in a simple one-size-fits-all chart because our tax liability depends on three distinct variables and I can’t easily plot that zero-tax boundary in three dimensions. So, here’s how I propose displaying the boundary.

  • Start with Social Security on the x-axis. I use a range of $0 to $90k, which is probably close to the absolute maximum two spouses can haul home in combined benefits.
  • On the y-axis, plot the maximum of the “other income” to guarantee zero federal taxes. This is the combination of all ordinary income and dividends and capital gains (i.e., Line 3 in the SS worksheet).
  • How much “other income” is sustainable at zero taxes clearly depends on the composition: ordinary income vs. tax-advantaged income (LT CG and Dividends). So, I plot a line for three different cases: 100%/0%, 50%/50%, and 0%/100% in the two income buckets.
  • I also assume that this is for a couple where both spouses are 65 years or older to increase the standard deduction to $27,000 p.a. ($24,400 base plus $1,300 extra per spouse above age 65!)
Roth Tightrope Chart02
Zero-federal-tax contours in the Social Security Income (x-axis) and Other Income (i.e., Line 3 in the Social Security worksheet). The contour depends on the composition of that Line 3 income!

Let’s look at the results in the chart above:

  • The lowest tax-free income allowance prevails if all of the other income is ordinary income. Say, you get $50k in combined Social Security, then you can make around another $20k in other ordinary income. The sustainable amount of income gradually declines because more Social Security income will become taxable and limit the amount of other income you can make before hitting the $27k standard deduction. But you can still haul in a lot of income: $50k in SS and another $20k in ordinary income. Or $90k in Social Security plus another close to $11k in other ordinary income for a total of more than $100k! Not bad!
  • Not surprisingly, that boundary shifts up if part of the “other income” is long-term capital gains. That’s because less of the income pushes against the standard deduction limit. Capital gains and dividends become taxable only if we go beyond the second federal tax bracket! Of course, capital gains still impact the Social Security taxable calculation in the IRS worksheet, see last week’s post.
  • If all of the other income comes from capital gains, we get the slightly peculiar looking green boundary. I had to double-check and triple-check my math here because of that drop at around $31k Social Security income. But it’s legit! For Social Security low enough, even with a lot of capital gains income, 85% of SS is taxable and we simply stay below the standard deduction as long as 0.85*SS<$27,000 and we fill up the rest with capital gains to get to the top of the second federal tax bracket. But once we hit $27,000/0.85=$31,765, we have to instantly and drastically lower the other income to push the taxable portion of SS back to $27k.

Same data, sliced differently

Here’s another way to slice and dice the data: I created a chart with capital gains on the x-axis and other ordinary income on the y-axis for one fixed Social Security level at a time, going from $0 to $90k in $5 steps. The capital gains and other income levels go in steps of $1,000 and I plot the different ranges of tax levels with dots: green = no federal tax at all, blue = average tax rate of 0 to 5% and red dots for 5-10%. If you go beyond the red region, you guessed it, you’ll owe more than 10% on average.

Because I didn’t want to publish 19 different charts here, I just put this all into a gif animation, please see below:

Tax regimes for capital gains/dividends (x-axis) vs. Other Ordinary Income (y-axis) for different Social Security income levels ($0 to $90k). green = 0%, blue=5% average tax and below, red=10% and below.      (gif created by

(See this zip-file on my Google Drive if you like to check out the 19 underlying charts)

So, if you wonder how a math/finance geek like me spends early retirement, this is it! Making gif animations out of Matlab Charts! Never a dull moment in my home office here! 🙂

But anyway, I was positively surprised that in retirement, we should be able to keep taxes quite low. Social Security will likely be taxable at a much lower rate than the 85% maximum. So you can still have a total income in the high-five-figures, potentially even six-figures and still keep federal income taxes low or even at zero! And then on top of that, you got Roth distributions and the tax-free basis in the taxable account withdrawals. Even folks in the FatFIRE community will hardly pay federal income taxes or avoid them altogether!

A Roth Conversion case study

As promised, and as an application of some of the things I’ve learned here, I wanted to do the Roth Conversion plan for the Becky and Stephen case study from two weeks ago.

If you recall, the idea here is that on the one hand, if you do the Roth conversions too aggressively (likely converting into and all the way to the top of the 12% federal marginal rate) and you eliminate all other income in retirement to rely only on Social Security and Roth IRA distributions, you’ll likely leave money on the table because taxable Social Security will stay below the standard deduction and distributions from the Roth are completely tax-free anyway! On the other hand, if you don’t do enough of the Roth conversions, you might have to distribute from the 401k at a marginal tax of around 18.5% to 22.2% when withdrawing and facing the Tax Torpedo.

So, planning the Roth conversions becomes a real tightrope! But how do we find the sweet spot in the middle? Here’s one idea: First, determine how much in additional ordinary income we can generate each year before we hit the standard deduction (and the Tax Torpedo if we go beyond). One additional challenge when mapping out this Roth conversion strategy: the kink points in the Social Security worksheet are apparently not adjusted for inflation, while all of the other relevant tax parameters (standard deductions, tax bracket, etc.) are. Thanks to my blogging buddies Dr. David Graham ( and Harry Sit (The Finance Buff) for pointing this out!

In the table below, I factor in the Social Security benefits over time (as recommended by the calculator, see the post from two weeks ago). All numbers are real, CPI-adjusted numbers, so some of the tax parameters (standard deduction, tax brackets, etc.) stay constant in real terms, but the kink points in the Social Security worksheet are deflated by 2% p.a., see lines 9 and 11. I then solve for the amount we can generate in Line 3 (up to the closest $100) to ensure we don’t hit the standard deduction.

Roth Tightrope Table01
Planning the Roth conversions before the large Social Security benefits kick in (2020-2024) and avoiding the Tax Torpedo in years 2025 onward. How much can you withdraw from a 401k and still remain in the 0% tax range?     (all figures real, CPI-adjusted)

We also have to take into account that later in retirement, probably around the year 2040 or so, one of the spouses might die and the survivor will have to file taxes as a single. So, the tax-free ordinary income will be lower in that situation:

Roth Tightrope Table02
Starting in years 2040 and onward we also have to account for the possibility of one spouse passing away and the survivor having to file taxes as a single. This reduces the amount you can withdraw from the 401k!     (all figures real, CPI-adjusted)

So, let’s take the upper bounds of ordinary income we can generate in years 2025 and onward and plug them into an Excel sheet. Let’s assume the portfolio gives you a 1.5% p.a. real return, not a bad assumption for a bond portfolio. Again, as mentioned in the case study, I propose shifting the bond portion into the tax-deferred accounts and it may not be a bad idea to put the equities (high expected return) into the Roth and the bonds (lower expected returns) into the 401k to keep a lid on the growth rate in the 401k, while keeping equities in the Roth IRA where they can grow completely tax-free.

Anyway, the trick here is to work backward! Let’s start at the beginning of the year 2054. We withdraw $4,400 at the beginning of the year which is the maximum anticipated amount we distribute from the 401k before hitting the Tax Torpedo. So if we plan to exhaust the 401k we need exactly that amount at the beginning of the year 2054. That means we need $4,400/1.015+$4,500=$8,835 at the beginning of the previous year. And so on, all the way to 2025 where we get a $259,018 target 401k level.

(see this link for how the table below was calculated)

Roth Tightrope Table03
Planning the path of the 401k over time. It’s best to work backward! Start in the year 2054 with a final zero balance, plan for the maximum withdrawals and capital returns over time and reach a target balance of around $259k at the end of 2024. To reach that target balance, plan for a little over $51k of Roth conversions in the calendar years 2020-2024.   (all figures real, CPI-adjusted)

If we start with $490,000 today we simply use the PMT function in Excel to determine how much we withdraw over the first 5 years of retirement to exactly hit the 259,018 at the end of 2024:

0.015 = rate of return
5 = # of years
490k = today's value
259018 = target future value
1 = for cash flow at the beginning of the year

And this tells us to convert $51,410 over the years 2020-2024. I’m assuming the conversions occur at the beginning of the calendar year. If you want to switch to the end of the year (might be better to be able to respond to other tax events) you’d use the 0 instead of the 1 as the final input in the PMT formula and this will get you to $52,181. Not a huge difference. So, there you go, the ideal Roth strategy would do pretty aggressive conversions, but certainly not all the way to the top of the 12% bracket. Which is great, because you have a slightly lower tax liability for the first few years in retirement before Social Security starts.

Some caveats about this approach:

  • In this particular example, we never got into trouble with the Required Minimum Distributions (RMDs). That may not be true for all retirees! So, when doing this exercise we’d potentially have to also keep those pesky RMDs in check!
  • We don’t know the returns, much less the real returns that far into the future. Luckily, we’re talking about a bond portfolio with a lot less volatility than a stock portfolio. If you believe that 1.5% is too meager for a bond return (recall this is the real return over and above inflation!) and/or you have equities in the 401k with a higher expected return, you’ll probably have to walk up the initial Roth conversions to make a bigger dent in the 401k portfolio! Even at a 5% real return, though, the target Roth conversions go up to only about $74,600 for the first five years. Still manageable if you want to stay in the 12% bracket in the years 2020-2024.
  • There are reasons to believe that taxes will go up in the future both on the Federal and State level. But of course, there’s no telling when and how. As a hedge against that possibility, you probably want to err on the side of caution and do the Roth conversions a bit more aggressively. Becky and Stephen certainly have some extra space in their 12% tax bracket.
  • Related to the above, another reason for more aggressive Roth conversions: estate planning. There is no guarantee that the last survivor will even live until the year 2054. Chances are, the last survivor will pass away well before then and it’s likely better to leave a Roth IRA to your kids than the tax liability of an inherited 401k. Or more precisely, the 12% marginal you pay today for the Roth conversion will very likely be lower than what you kids pay in 2054!
  • So, I’d view the $51k in Roth conversion as the lower bound of what you want to convert per year from 2020 to 2024!

Is it really worth it to go through this whole computation? I had the impression that on the ChooseFI podcast two weeks ago we talked a little too much about the tax planning. I think that getting the safe withdrawal rate calculations right is a lot more important! If you are forced to withdraw some of the funds in retirement and face the tax torpedo (an 18.5% marginal rate vs. a 12% rate you could have gotten during the conversion period) on maybe $10,000 of 401k distributions, that’s not the end of the world. Sitting in a $150k a year nursing home at age 90 and running out of money, now, that’s a bigger concern!


What gets lost sometimes in the whole Safe Withdrawal Rate discussion is that all those calculations are normally based on gross numbers! You still have to account for taxes! But I’m encouraged by today’s exercise! At least on the federal level, there is a large range of income combinations that are absolutely, completely tax-free. And if you’re willing to part with a small portion of your withdrawals, maybe 5%, you expand that range significantly. You almost have to try really, really hard to even reach a 10% average tax rate on your taxable portion. And notice that you can still generate additional retirement cash flow because part of the withdrawals will come from the cost basis in taxable accounts and all of your Roth IRA distributions are tax-free. Same with your Health Savings Account.

So, after spending waaayyyy too much time on taxes in the last few weeks, I’d like to refocus on my favorite topic again: the finance portion of early retirement, the safe withdrawal rates. Most people can actually assume that your gross retirement income is equal to the net income, especially if you’re lucky enough to live in a state without income taxes. And if you do face a state income tax, maybe apply a haircut of a few % to account that. I’m not saying that tax planning is irrelevant. But it’s a lot of effort and even if you get it slightly wrong it’s not the end of the world. Concentrate on what’s really important in retirement: the withdrawal rate math!

So much for today! Thanks for stopping by and please leave your comments and suggestions below!

Title picture credit:

95 thoughts on “How much can we earn in retirement without paying federal income taxes?

  1. Karsten, I tried to do something similar for my situation but ran into challenges figuring out the bend points of social security payments. I figured I would just wait since social security for early retirees is decades away and tax planning now would be futile with unknown rates of payments, tax brackets and tax treatment.

    Or do you believe it is worthwhile to plan now to avoid a potential tax hit down the road?

        1. Where can we find someone to help us with our specific situation to plan for withdrawals and distributions to minimize our taxes?

            1. Yes, please.

              A fee-only advisor who could review withdrawal and Roth conversion strategies would be very, very helpful.

              Your endorsement would mean quite a bit as I’m unconvinced most are as familiar with the various strategies as you.

              Maybe you only need endorse one as geography doesn’t mean much?

              You’d be busy yourself if you’d offer such a service.

              Thanks in advance.

              1. Yes, and that would fill a void. Not yet ready to jump into the market myself. Pursuing some other passion projects right now.

                If some RIAs/advisors want to “apply” for an endorsement, I could consider that, though.

    1. great information and thank you for sharing your hard work and passion. I read some response regarding health care insurance. One simple way to minimize medical / heath costs is to go plant based. It will eliminate 80-90% of the most common diseases like heart disease.

  2. Great post. An additional “tax” is health insurance. High net worth people between 60 and 65 with appropriate planning can live off cash, work very little (up to 32k ordinary income for a couple), and get free or nearly free health insurance in our state. If you take advantage of the space to the top of the 12% bracket for Roth conversions in my state (NC) premiums plus expected usage costs 30-40K a year, an enormous effective tax on that 100,000 in come. Unless you get free health insurance from a former employer, or are very young so premiums are lower you pay either way unless you don’t work.

    1. Yeah, we plan to stay with the health care ministry.
      But it’s certainly an interesting exercise to figure out the implicit marginal tax due to phasing out health care subsidies! It’s on my to-do list!

  3. Hi Karsten –
    As a fellow WA state resident and early retiree (with side LLC income) I’m getting health insurance through the WA Healthplan Finder (ACA) so continually have to manipulate my AGI to remain below the upper income threshold to remain eligible for the premium tax credit. Is this an issue for you?

    1. We use Medishare, a health care ministry. The premium and benefits are income-independent.
      But I agree, one should do an exercise to check the implicit marginal tax from pahsing out ACA subsidies! It’s on the to-do list! 🙂

      1. This is a surprising choice coming from you.

        AFAIK, the healtshares are not technically insurance so they are less likely to cover the catastrophic cases (fat tales), which would be my primary purpose of carrying insurance.

        Anything non-catastrophic and non-urgent you can find a reasonable price for, whether in the US (cash only clinics) or SE Asia or Latin America.

        What was your reasoning to chose the healthshare?

        “Typical” healthcare cartel in the US is beyond repair. If you want to hear some crazy healthcare economics:

  4. great post but your spreadsheets are very complex. I don’t think I understand 5% of them and that’s on all posts

  5. While these articles don’t help us non-US readers, I’m still so happy that you’re writing has been so prolific since retirement and after your RTW trips. Keep it up! 🙂

  6. Karsten,

    The SS bend points are not indexed by an inflation measurement, they are indexed by changes in the Average Indexed Monthly Earnings.

    See for the formulas used to calculate the bend points.

    This table shows how the bend points have changed over the years:

    I suspect your yearly 2% real reduction in the bend points is too aggressive.

    It would be interesting to compare AIME to some of the other inflation measures, such as CPI and Chained CPI. Did AIME keep up with other inflation measures?

    This prescient 2014 article explains what happens when you switch from CPI to Chained CPI for bracket calculations: The TCJA of 2017 changed the tax bracket indexing to Chained CPI-U.

    The Social Security COLA uses yet another inflation metric, CPI-W.

    1. Thanks for your comment.
      The kink points in the IRS Worksheet to determine the portion of SS that’s taxable are NOT indexed for inflation.
      What you are writing about are the bend points for the benefit calculations. That’s well known and I don’t dispute that. But those are different bend points! 🙂

  7. There is a lot of heavy lifting here in walking the tightrope. However, the underlying assumption that tax regime will stay the same for next X years is unrealistic. Tax regimes change every 5-10 years.

    Higher taxes is a real possibility (even though debt monetization and the resulting inflation is a much more politically palatable) as a solution to the ever-growing government debt.

    Projecting 30 years in the future is pointless, as 30 years from now the entire economy and global order may change (let alone tax rates and income classifications).

    1. Thanks.
      What do you propose instead then? I proposed a plan to deal with Roth conversions and gave a recommendation what to do under the current law, and then proposed to be a little bit more aggressive to hedge against rising tax rates.
      But what do you want to do? Nothing? No Roth conversions? Do you believe that’s a better recommendation?

  8. Great post Karsten. Love it. My wife and I are still twenty years from taking Social Security, but we will soon be starting our Roth conversion ladder. I am retired but she is still working part-time as a self employed contractor. Based on the numbers you are showing it looks like we will be able to avoid federal taxes almost entirely during the Roth conversion period.

    This just goes to show what a good deal early retirees get in this country (especially those of us that live in states with no income tax.)

    Thanks again and keep up the great work!

  9. Big ERN, all your posts are very helpful. I’ve been reading through them over the last few weeks. I wish I’d been more aware of all the considerations a long time ago. Now I have most of my retirement money in a T-IRA and 401K and even with IRA to Roth conversions between retirement at 62 and starting SS at 70 (at least that is the plan) I’ll probably have too much income from RMDs to take much if any tax advantage. I guess it’s a good problem to have vs. poor retirement savings, but wish I’d thought about the Roth years ago.

  10. Hi Karsten, I really liked this 3-part series. First because I learned a lot, e.g. how SS pension benefits are taxed (Part 2) , how to use the ERN Google Sheet in this specific case (always helps to supplement the good generic explanations with a widely applicable example like this with lots of nuances that we can all use) (Part 1), the interaction between the 3 factors of SS Benefits, Other Income, and Other income mix btw Ordinary income and “Cap Gains” (Part 3) (Nice examples of how to deal with 3 variables (factors) in 2 dimensions 😊) . I also liked the way you plan out the withdrawal sequence for them so that it is tax efficient, and diagram out what a Glide path could do for them. And the final Schematic path of Portfolio value (Chart) in Part 1. When a person sees this schematic, then they can avoid bailing out at the low points or going crazy spending at the high points.

    And the practical results for Becky and Stephen case are quite remarkable in their optimization results achieving (based on $1.35 M) : almost 7% withdrawal rate (meeting their spending goals before SS benefits start and then continuing to meet goals and even LTC goals), only 4% tax on the partial Roth Conversions (<$20K on $490K T.IRA amount) (that makes taking their 401 Ks and T. IRA during their working years very worthwhile since the reduction in taxes must have been at a rate of at least 10 to 15%. And finally, a good likelihood of leaving a legacy to family or charity when they pass.
    Very complete series that should give a good model so that interested persons can make a plan (Spreadsheet model) now for accrual, cash flow, SWR, and the taxes for the rest of their Accumulation phase, their Roth Conversion phase, and their after SS Benefits start phase optimizing it. And adapt the plan as their circumstance change, and the tax laws change. That is the most that we can do, but is quite a lot in comparison with what most people have (even if they have a good financial advisor.)

    Two requests:
    1. Can you share the spreadsheet(s) that produced the Figures 4 and 5 and 6 in this Part 3 post of the series? These will be useful for those of us who want to adapt the results to our specific situation.
    2. This example covers well a “normal FIRE” situation with about $500 K in IRA. But what about a case like Dr. Gasem (at MD on FI/RE) where he as $1.5 M to convert (a FAT FIRE situation) leaving 500 K in the T. IRA at age 70 when SS Benefits start. The good Doctor as written extensively about his conversion plans (e.g. and plans to spend $338 K in the partial Roth Conversions. This is certainly a case where optimizing will be important. So, could you take this kind of a case (either Dr. Gasem’s or other) and use your tools again to see how to optimize?

    This will give readers a full spectrum of examples to work with and so we will be able to plan our own situation and thus be able to assure that the SWR calculation that you have studied so extensively takes into account an optimized tax plan also.

    1. Also Karsten, does the NPV need to be used to better interpret any of the amounts in this 3 part Series of posts?
      (Or is NPV not necessary due to your other adjustments already made.)

      1. The NPV can be used in general in this kinds of analyis, to trade off paying taxes today vs. in the future.
        A good approximation is to simply look at the marginal rates today vs. in the future and use that to determine whether it’s beneficial to do more/less conversions.

  11. I’m also curious about your choice in Medishare vs. other health ministries and really why them but not ACA insurance? I’m guessing the cost would have been very high and perhaps you wouldn’t have qualified for the subsidies either.

  12. Can you give some examples of investments that would be generating $78k in Long Term Capital Gains per year. Unless i’m missing something, that would require a pretty large portfolio, if you’re generating gains that are greater than a year that large every year…

    1. Most folks in retirement (whether traditional retirement or FI/FIRE) will have a budget much lower than that. But some people have a six-figure annual budget out of a $3m or more portfolio.
      If you assume a 2% dividend yield you generate $60k from dividends alone. Withdraw some more capital gains and very quickly you’re at $78k.

      1. Yes, but dividends are not Long Term Gains, they are ordinary income, taxed as income. To generate that much in Long Term gains, you would have to be selling out of large long term investments every year. I’m not sure how many people are trying to draw down their principal like that. I was trying to differentiate between short and long term gains. Also, you should mention the impact on AGI. Realizing additional capital gain still increases your AGI, which impacts certain deductions, such as medical expenses, or can phase you out of certain tax credits tied to AGI.

        1. Not true.
          “Qualified dividends are taxed at a lower capital gains rate.”

          Or from Kitces:

          See the column labeled “L/T Cap Gains & qual. dividends”

          Also, as I mentioned, both in the MarketWatch article and here: Shourt-term Cap gains and non-qualified dividends are indeed taxed as ordinary income.

            1. It does but not 100% right. Many REITs and some companies dividends are not qualified therefore taxed as ordinary income.
              I’m one that have them in a taxable account and pay taxes on them since I’m a non-resident in the US can’t have a tax deferred acct and my home country doesn’t have a tax treaty w/ USofA therefore I’m taxed 30% directly (thanks Uncle San)

              1. Disagree. My comment is 100% right. I said:
                “Qualified dividends are taxed at a lower capital gains rate.”
                REITs (and some other companies’ dividends) are non-qualifying dividends. So, you’d need to reserve some space in your ordinary income bucket for those. But in the average index fund that’s a tiny percentage.

  13. Big Ern, this is one of the best pieces I’ve read for a while! Keep up your good work – especially with the mathematical supplements.

  14. This is just what I’ve been looking for. Began taking SS this year and trying to figure steps to lessen my taxes this year. I’ll be reading and re-reading this. Thanks!

  15. Hi Karsten,

    New subscriber and thanks for the information you provided.

    Question, my wife and I have five adorable little monsters. Therefore, what is the max combination of ordinary income and long-term capital gains/qualified dividends in order to remain tax free?


    1. Interesting question. Haven’t mapped it out completely but you’d generate about $10,000 in federal taxes to be offset with the child tax credit under the following combinations of income:

      Column1=Long-term Cap Gains + Qual’d Dividends
      Column 2 = Other Income (Interest, 401 Distrib., etc.)

      column1 column 2
      $170,017 $0
      $145,617 $24,400
      $100,000 $51,200
      $50,000 $78,956
      $0 $107,500

      1. Thanks Karsten. This information gives me tremendous maneuver room in conducting IRA to Roth conversions in coordination with my earned income.

        I look forward to your next post.

        Semper FI!


          1. Would it be possible to send me the spreadsheet?

            Regardless, have a Merry Christmas and Happy New Year.

            Semper FI,

              1. Merry Christmas Karsten,

                The link above the large table goes to the Roth Conversion table and the calculation of how you generated the two columns pertaining to my five kiddos is not shown. If possible, can you provide the link in your reply.

                Hope you had a great Christmas and wish you a Happy New Year.

                Semper FI,

                1. Ah, I see. There’s no designated spreadsheet for that calculation. Just played around with the original spreadsheet to see what kind of income combinations will give me exactly $10,000 in federal tax to be offset with the 5*$2k=$10k in tax credit. No magic there! 🙂

  16. Glad I found you article; it provides reassurance of our retirement savings plans. We too live in WA. I want to keep more of our retirement savings from tax burdens. We plan to retire early with a paid-for mortgage so we will not need as much from our 401k and IRA to live on and plan to convert as much in that 12% bracket to a Roth before we begin taking SS.
    What I found very interesting is your diversification of having bonds in the tax deferred 401k and equities in the Roth. We will continue to max out the Roth contribution until we retire in 7.5 years while also taking advantage of the company match in 401k contributions – I will have to review what bond fund options we have available in the plan.
    We are also maxing out our HSAs to help bridge the gap before medicare eligibility.
    Now planning to read your Safe Withdrawl Rate article – thank you so much for sharing your knowledge!

  17. Thanks so much for this post. I actually went to a CPA looking for the answers to many of these questions, and he didn’t have many answers for me. I wish I had found your site a couple of years ago!

  18. I’ve been going through this exercise for the past 3 years. I have 2 more years of conversion to go, or 4 depending on the passage of the Secure Act which will shift RMD from age 70 to 72. My analysis started with an analysis of the tax code. The median size retirement account at FIDO is about 500K- 600K. If you RMD 500K using a 3% real return at age 70 the first year the IRA pays out 18K. The 10th year it pays out 23K and the 20th year it pays out 25K. This implies a 500K TIRA mostly in bonds will act as a quasi inflation adjusted annuity over the course of 20 years ages 70-90. This annuity will continue to payout well past age 100 but it’s payout starts to decelerate past about age 92. This TIRA is therefore a ready and stable source of “quasi inflation adjusted” ordinary income for the entire course of retirement. When mixed with SS even at the 50K or 60K level, by the time you account for the inflation adjusted 85% tax advantage income from SS, with these 2 vehicles you stay in the 12% bracket for nearly 20 years or your entire retired life! Once you cross the 12% bracket taxes rise dramatically and your “TIRA annuity” becomes simply a revenue source for the government. You still get to keep about 20K/yr and effectively the government takes a bigger and bigger bite of anything above the 12% limit. What this says is the government considers anybody above 103K ordinary income as “the rich” and the tax code is definitely designed to “soak the rich” despite all the rhetoric of millionaires getting tax breaks. This analysis gave me my first revelation. If you want a tax break YOU MUST LOOK MIDDLE CLASS from an ordinary income perspective. The give away is the fact that 500K is the median retirement account size. As you move up the bell curve less and less voters fit the median definition until you reach far into the tail where taxes are no longer progressive but normalize to a constant. A 100 millionaire and a billionaire pay the same rate but there are only a couple hundred K of those payers so only a couple hundred K of votes. The government is going to be less inclined to screw the median payer. The moral of the story is to live your retirement as a median payer. An additional advantage of a TIRA is you can withdraw tax free if you get a bad medical diagnosis. So in the face of a cancer diagnosis my TIRA (or at least my half) will be the first to go. A point to consider.

    Let’s say you have 1.5M in TIRA. I.5M @ 3% real RMD’s 55K the first year, 70K the 10th year and 80K the 20th year. If SS pays 50K the first year taxable is 42.5K, and 62K inflation adjusted in the 10th year, so the net taxable ordinary income in the 10th year is 62K + 70K or 132K in the tenth year. You left the safety of the 12% bracket 5 years ago and the ability to pay 0% cap gains. You’re still living virtually a 100K life style but paying an approximate extra 30K in taxes. It’s not quite that bad but close enough. So the decision you have to make is NOT how to pay no taxes, but how to extend your longevity in the 12% bracket till you die. If you have for example 1.5M, 1M has to go out of the TIRA. If you have 2M, 1.5M needs to exit the TIRA. So the amount that needs to exit the TIRA to make you look middle class for the duration of your retirement is the actual control variable not zero taxes. It’s a similar but different calculation to devise a plan of conversion. I did an extensive analysis and found 5 years of conversion to be about the best length of conversion starting at age 65. Age 65 puts you on Medicare so you have less to worry about when it comes to ACA problems. Medicare still soaks the high wage earner but not as bad as the loss of ACA subsidies. If you make 170K MAGI MFJ and you pay $20 in part C charges your medicare cost goes from $155.50/mo to $222/mo per person. At 214K MAGI MFJ you pay 321.80/mo per person. At $267K you pay $455/mo per person. At 320K you pay per person. This 320K level includes the “top of the 24%” conversion level often touted and the $524/mo does not include a supplemental insurance usually around $150/mo per person. That’s an additional 16K expense of conversion that needs to be considered if you decide to convert to the top of the 24% and you’re MFJ.

    My analysis showed if you have 1M to convert over 5 years a little more than 200K/yr will do, about 210K. You have to consider the growth in the TIRA as a part of the conversion. Most people won’t have 1.5M in bonds but more like a 70/30 which will have a greater growth than 3% real. My analysis showed that you should not take SS at 67 but at 70 and use the extra 8% to offset the cost of conversion later. A slightly bigger SS will result in a slightly smaller WR over the course of retirement and a smaller WR = less exposure to SORR after age 70 when you are no longer working or side gigging etc. 200K keeps down the medicare surcharge. The order of assets moved matters. In my accounts I simply move assets from a TIRA to a Roth without converting the asset to cash. So 200 shares of AMZN gets moved as 200 shares and gets taxed at what ever 200 shares of AMZN was worth on the day of transfer. This is a subtle but important point. If you expect a crash to occur in the year of transfer, wait for the crash and then transfer. If AMZN was 2K/share, a 200 share transfer would generate a tax bill on 400K of ordinary income. If the market crashes and AMZN drops to 1K/share that 200 shares is taxed as 200K of ordinary income. It’s a point to consider. Also transfer your most risky assets first in an up-trending market, and last in a crash. If the S&P starts the year at 2500 and ends the year at 3500 and you transfer on Dec 26th you pay tax on the 3500 price. It’s a judgement but one to consider. As assets get less risky the transfer cost is more stable and therefore the tax bill is less volatile. I’m down to bonds and gold in my accounts yet to be transferred and the gold will be gone starting in Jan, leaving only bonds to transfer leaving 500K in the TIRA in 2 years in bonds primarily.

    My TIRA once cleaned out will hold a 15/85 portfolio of total stock/various bonds. If the market drops in half in the first year total stock will drop 7.5% leaving me an account 92.5% intact, which will barely hit my RMD. If the crash happens in 5 years total stock will have added enough return that I can pull out 100% of my projected RMD despite the market falling in half. By 10 years if there is no crash I can pull out enough money beyond RMD to buy a new car. That compounding can come in handy and work to your advantage.

    My retirement income will be SS + TIRA RMD + stock sales from my taxable account at 0% CG. I will take no income from the Roth for 10 years. This allows the 1M to grow unmolested and serves as self insurance. Eventually we all die and a lot die over the course of a decade or more meaning long term care and if MFJ there are 2 deaths to plan for. In addition the Roth is a ready source of money to pay the increased taxes when one partner croaks. After a few years of good market you can probably take some dough for a car or a trip. After 10 years you’re well covered. That’s my optimized plan. I pay for conversions by cashing in some stocks from my brokerage which gives me money to live on and money to pay the taxes. Cash in a bank is not ordinary income so my tax bill is strictly due to the cost of conversion which optimizes what I pay in taxes. Post conversion my tax bill is projected to be $3600 the first year on 100K income, rising slowly and then more quickly in the out years after the death of a spouse. With 60K ordinary income I need only 40K/yr from my brokerage to make the nut. I have more than 2M in the brokerage so my WR is quite reduced to only 2%. Since the WR is reduced I expect gains in the brokerage to outstrip my rising tax bill. In order to make this work I DID NOT MAX OUT MY PRETAX but rather put some in pretax and some in brokerage allowing me the flexibility to remove money from the brokerage at a lower tax rate than ordinary income would provide. In addition I tax loss harvested the brokerage over 20 years, so I was able to cash in some stock tax free for living expense. Hopefully some of these ideas are useful to someone.

    1. Wow, thanks, Gasem, for a mini guest post! 🙂

      All really great points. Everywhere it’s best too look like the middle class (i.e., median voter) for the best treatment by politicians. Hence the $100k cliff. Or the $80k or so Obamacare cliff.
      I had noticed that exact cliff when planning our retirement budget. There’s no point going much above $100k annual budget. It’s not making me much happier but that’s exactly where the high marginal rates kick. Why work for 5 more years to raise our retirement budget from $100k to $130k? Not worth it! Somewhere between 70 and 100k appears to be the sweet spot!

      And yes, the T-IRA with RMDs certainly looks like a quasi annuity plus CPI-adjustment! Excellent point!

  19. This is great Big ERN! Can we please get a link to the spreadsheet that you are using? I would love to calc my own withdrawals, but for the life of me can’t find the link to the same spreadsheet here.

      1. Thanks. I wish you would share the calculations/spreadsheet for the above calculations if possible. They are not in part 28 unfortunately. For example the backup for “Roth-Tightrope-Table03.png” above. Would love to go through and essentially update for my own situation as I try to optimize roth conversions to keep RMDs at bay.

          1. Thanks! I really appreciate the good work that you’re doing here and your attention to detail in parsing these modeling exercises. It’s great for a finance blogger to, you know, know finance.

  20. Hi Karsten…..thanks for much for posting this thoughtfully laid out tax scenario, very nice. I’m not an accountant, but do our taxes (Turbotax) and am struggling with understanding how one would use your scenario there. If in fact it could be used to help when doing returns there, I think we would be eligible to file amended returns for years in the past.

    1. Everything I do here is already implemented in Turbotax. I didn’t spill any tax secrets here, I merely showed how the existing basic tax laws afford zero taxes if the composition of income is right.
      If your Turbotax tax bill showed taxes owed it’s likely because you had too much ordinary income. I can’t change that. But if you retire and then live mostly off your investments you can make use of this in the future.

  21. 21% tax on shareholders share of company profit and no tax credit to the shareholder = 21% tax. Add 0% tax on qualified dividends = 21% tax on received share of profits.

    0% tax on some capital gains = avoid dividends and take income as capital gains.

  22. “whether the company retains earnings or pays dividends, both have been taxed”:

    And no tax credit (‘dividend imputation’) for shareholder; 21% tax in both cases.

    However, if company eliminates profit, by deducting interest or depreciation as examples, no tax on earnings is payable. Any earnings which result in capital appreciation and realised as ‘qualified’ (no-tax bracket) capital gains by a shareholder are then 0% taxed.

    1. Yeah, I know. i certainly prefer profits to be invested in the company (and hopefully do it in a way to artificially lower taxable income), drive up the firm value and then we, the shareholders realize the gains tax-free. Very good point!

      1. “shareholders realize the gains tax-free”:

        USA tax of company earnings is at odds with taxation of other income. Avoiding it results in ‘excessive’ corporate debt – which may result in company valuation being sensitive to interest rates.

        Withheld taxes on employee pay (‘earnings per hour’) are credited to the employee. Taxes on shareholder pay (‘earnings per share’) are not credited to the shareholder – which is regressive for small shareholders.

  23. Great article. One question about the graph about “Tax-free income limits without Social Security.” Should the equation be for x > 78750?

    x + (y-78750)*0.15 = 24000

    If this is true, for people without ordinary income (y), the capital gain (x) can be much higher.

    1. No. that’s not how the tax formula works. $24k is the deduction, not a tax refund. So, you first subtract $24 k from the income and then you can get $78k more in tax-free dividend income. 🙂

  24. Fantastic article… so glad I remembered to catch back up on your blog posts. This post sums up an exercise I’ve been beating myself up to run every April for the past 3 years, but never seem to get motivated enough to see through. Thanks for and well done on boiling this multi-variable tax calc process down into its basic moving parts.

    I truly hope that early retirement is going well for you and your family. Hopefully someday I’ll have the internal fortitude to pull the trigger myself.

    Thanks again

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.