As Judge Learned Hand famously stated, “Anyone may arrange his affairs so that his taxes shall be as low as possible.” In practice this usually means engaging in tax strategies that minimize or at least defer taxes as long as possible. But given the progressive nature of income tax brackets, there really is such thing as being too good at this.
The optimal strategy balances these two forces, seeking out an equilibrium rate that accelerates enough income to fill up lower tax brackets today, but defers enough income to fill up tax brackets tomorrow.
And while nobody knows with certainty what future tax rates will be, the very nature of planning — and especially retirement projections — presumes making some straightforward assumptions about future Social Security and pension payments, RMD calculations, and anticipated interest, dividends and capital gains to make a reasonable approximation of an individual’s future tax rates.
One of the fundamental reasons for saving and investing is to accumulate enough assets to achieve a point of financial independence, where there’s no longer any need to work for income. Instead the individual supports themselves solely from their available assets and other retirement resources.
Substantial wealth, however, can produce significant tax consequences. Highly appreciated investments can produce very significant capital gains — e.g., in the case of selling a business or simply a concentrated investment position that grew substantially over the years — but also because much of our savings is placed into tax-deferred retirement accounts.
Tax rates may be modest for years or even decades, thanks to the tax-sheltering of the retirement account itself, but when it’s time to take retirement withdrawals or distributions mandated by Uncle Sam’s Required Minimum Distribution obligation at age 70 ½, suddenly so much taxable income can spill out of the account that the retiree is vaulted into drastically higher tax brackets, never to see lower rates again.
Of course pre-tax retirement contributions can be made to a Roth IRA instead of a traditional one, or the traditional retirement account can be converted to a Roth to avoid build-up of tax-deferred income in the future. Yet doing so accelerates what might have been a distant future tax liability up to the present, as Roth contributions are made with after-tax dollars and Roth conversions trigger immediate taxation. And if the Roth conversion itself is large enough, the individual can immediately propel themselves into even higher tax brackets.
Consequently, the optimal strategy for managing tax-deferred growth — and the potentially substantial build-up of pre-tax assets from unrealized capital gains, traditional IRAs and 401(k) plans — is to defer enough to avoid high tax rates now, but not so much as to cause significantly higher tax rates in the future.
Viewed another way, the benefit of finding the tax bracket equilibrium point is itself a form of tax bracket arbitrage, where income that will be taxed someday is shifted from higher-tax-rate years into lower-tax-rate ones without changing the underlying nature of the income or risk of the investments themselves. It’s just changing where and how they’re held for tax purposes.
While the basic principle behind tax rate equilibrium is fairly straightforward — create enough in Roth-style contributions, conversions or other taxable events to fill the lower tax brackets now, without pushing higher than what tax rates would have been in the future — that determination is more complex. Where that equilibrium point lies will vary depending on an individual’s current and projected wealth, and how that compounding wealth will itself cause more taxable income — and potentially higher tax rates — in the future.
So while it’s relatively straightforward to determine someone’s marginal tax rate today by projecting out their known or reasonably approximated income and expenses for the year to know what the tax rate would be on the next dollar of Roth contribution, Roth conversion or other taxable income, knowing whether that rate is good or not also requires projecting out the trajectory of wealth to figure out what that likely future marginal tax rate might be.
The good news is that while some prospective retirees throw their hands up in the air and say, “Who knows what the tax rules will be in the future?” projecting future tax rates is not really so blind.
The starting point is to count known future income streams first, including most notably Social Security benefits and whether the maximum 85% of benefits will be included in income for tax purposes, along with any other fixed income streams like pensions or lifetime immediate annuity payments.
The second step involves adding on any anticipated distributions from those pre-tax retirement accounts, from the distributions required to cover retirement spending itself to the onset of RMDs when they begin at age 70 ½.
Third, project out the accumulation of taxable investment accounts, recognizing the contribution that they too will make to annually taxable income in the form of passive interest and dividends, along with at least some assumption for capital gains — as there will typically be at least some portfolio turnover, either to generate retirement cash flows due to occasional investment changes over time as markets and opportunities change, or simply for rebalancing purposes.
Notably, not all income considered for projecting future tax rates will necessarily be income the retiree needs for spending. While some spending to fund retirement will come in the form of withdrawals from pre-tax retirement accounts, RMDs almost by definition are forced withdrawals for tax purposes that weren’t needed for cash flow purposes.
And taxable accounts will produce whatever interest, dividends and capital gains they produce, which may or may not bear much relationship to what’s needed just to cover retirement spending. In other words, the point is not to project spending income — i.e., cash flow distributions — but taxable income instead, which literally is income for tax purposes, regardless of whether or how it’s used.
Example 1. Samuel and Diane are both 62 and have accumulated nearly $1.3 million in savings for retirement, comprised of a $400,000 taxable account, $700,000 in a traditional rollover IRA and $200,000 in a Roth IRA. Between the two of them they are eligible for nearly $34,000 per year in Social Security benefits if they start payments now, but have decided to wait until age 70 when their benefit will be increased to almost $64,000 per year, plus cost-of-living adjustments. In addition, Diane will receive a teacher’s pension of nearly $20,000 per year, but with no annual inflation COLA.
The couple’s goal is to spend $8,000 per month in retirement, or $96,000 per year, which will be almost entirely covered by Diane’s pension plus their Social Security benefits at age 70 — at least before taxes are considered —totaling $84,000 per year. However, the couple must plan for the lack of inflation adjustments on Diane’s pension. Indeed, their $96,000 annual living expense will be up to more than $120,000 per year in just eight years, while their combined inflation-adjusted Social Security and non-inflation-adjusted pension will be only $100,000 per year. That’s saying nothing of the $76,000 per year year shortfall to cover their spending between now and age 70, when Social Security begins.
The good news is that in the aggregate, the couple should have more than enough in assets to cover their retirement needs. Their tax situation, however, will be very volatile, with several years of relatively low income through their 60s followed by much higher taxable income in their 70s once both Social Security and RMDs kick in from their IRAs.
The couple’s initial plan is to draw $50,000 per year from their taxable account for the next eight years, largely depleting it, supplementing with $25,000 per year from their Roth IRA —mostly depleting it as well. Meanwhile, the taxable account is projected to produce just about 3% per year in bond interest from the $300,000 invested in bonds, and 8% per year in qualified dividends and capital gains — on average at least — from their $100,000 in stocks. The bulk of their remaining equities are in retirement accounts.
Once Samuel and Diane turn 70 their inflation-adjusted Social Security payments of nearly $80,000 per year will begin, on top of RMDs that by then are projected to be nearly $50,000 per year, with an 8% annual growth rate on their otherwise untouched IRA, which provides more than enough cash flow to support their retirement spending at that time.
Thus, while the couple’s spending would be by intentional design $96,000 per year adjusting for inflation, the couple’s taxable income profile in the coming years would be approximately as shown below.
As the graph shows, while the intended liquidation plan may literally be very tax-efficient — particularly in the early years — it is arguably a little too tax-efficient, relying heavily on principal liquidations and just a small portion of taxable income in the early years, followed by a significant increase in taxable income as Social Security benefits begin and are 85% taxable, on top of then-significant RMDs.
In fact, presuming a standard deduction of $24,000 for Tax Year 2018, the couple would be eligible for the lowest 10% tax bracket on most of their income from age 62 to 70, including a 0% rate on long-term capital gains and qualified dividends. Later, when their Social Security and RMDs begin, they’re thrust up into the 22% tax bracket, from which they would never leave.
As the above example reveals, while Samuel and Diane will be in a very low tax bracket for the next eight years, they won’t actually benefit from those lower rates because none of their IRA income is subject to them. Instead, those favorable brackets are wasted in the early years, while their income spikes to higher levels in later years because they are deferring too much.
Of course it’s unlikely that the couple’s exact income will perfectly align to the projections, given both the potential for changes in interest rates and yields, volatility of capital gains and uncertainty of returns themselves. Nonetheless, simply projecting known income as well as reasonable growth and income expectations makes clear that the couple is materially below the threshold for the 22% tax bracket today, and will be materially above it after they turn age 70 and begin both Social Security benefits and RMDs. This all suggests their current tax plan is very much out of equilibrium.
FILLING TAX BRACKETS
So what should retirees do when their projected taxable income in retirement doesn’t cleanly align to the projected tax brackets? Simply put, they should shift income from the projected high-income years into the low-income years.
As noted earlier, this can be done with any combination of income-creating strategies, including:
- Changing the retirement account being contributed to: For those still working, consider changing the retirement contribution from traditional pre-tax to Roth if the current tax bracket is lower than it’s projected to be in the future. Alternatively, shift from Roth contributions to traditional IRA or 401(k) contributions if it would be better to claim the deductions now — at high tax rates during the working years — and recognize the income later in retirement, after wages are gone.
- Taking unnecessary retirement distributions: While the classic rule of thumb for account sequencing in retirement presumes spending from the taxable accounts first and letting the pre-tax retirement accounts grow, this can result in retirement accounts that compound too much, thrusting the retiree into higher tax brackets by the time those retirement distributions actually begin. The alternative? Don’t wait so long to begin distributions from pre-tax accounts. Instead choose to take withdrawals earlier, even when not required as RMDs, to fill lower tax brackets now and reduce exposure to higher tax brackets later.
- Beginning partial Roth conversions: For those who otherwise have sufficient non-retirement-account resources to make the necessary liquidations to generate cash — or who haven’t retired yet and don’t need the dollars — another appealing option to shift income across the years is to engage in partial Roth conversions. Notably this doesn’t call for converting the entire retirement, which would both drive the individual into higher tax brackets immediately and completely eliminate future retirement-account income, potentially making future tax brackets so low it would have been better not to convert at all. Instead they should convert just enough to fill the lower current tax brackets without crossing into higher ones, while still whittling down future tax exposure.
Example 2. Samuel and Diane instead decide to fill the gap on their low-income years by doing partial Roth conversions of $60,000 per year for the next eight years, keeping them within the 12% tax bracket and whittling their IRA down to the point that once their Social Security payments begin, there is very little left of their IRA. This strategy allows them to remain in the 12% tax bracket even when their RMDs begin.
As the above graph shows, the net result of finding Samuel and Diane’s tax equilibrium is significant: They save the 10% tax rate difference — i.e., between the 22% future and 12% current tax brackets — on most of their $700,000 IRA withdrawals for the rest of their lives as those withdrawals occur over time, saving literally tens of thousands of dollars in cumulative taxes.
While the primary focus thus far has been finding the right tax rate equilibrium in retirement, there are actually two states of equilibrium to consider — one for ordinary income and a second for long-term capital gains and qualified dividends. Not only are there seven ordinary income tax brackets spread from 10% to 37%, but there are also four effective capital gains brackets — 0%, 15%, 18.8% including the 3.8% Medicare surtax, and 23.8%. Here the risk of over-deferring and causing unnecessarily high future tax rates on capital gains — while failing to use nearer-term lower capital gains rates today — is also present.
Technically the ordering rules for calculating income tax liabilities dictate that ordinary income and the associated tax brackets are counted first, and long-term capital gains and qualified dividends then sit on top at whatever their applicable rates will be. This means that even once an optimal ordinary income tax equilibrium rate is determined, it’s still necessary to also determine the long-term capital gains rate equilibrium as well.
The situation is complicated by the fact that the long-term capital gains tax brackets do not perfectly align with the ordinary income tax brackets — especially starting in tax year 2018 after legislation passed at the tail end of 2017, and have their own thresholds for determining the bracket thresholds.
Nonetheless the point remains that once ordinary income tax brackets are projected to determine where current and projected future income will fall, it’s necessary to then consider the trajectory of long-term capital gains and qualified dividends, and where they may fall as well. This in turn may entail other capital loss harvesting to reduce capital gains in high-rate years, or even engaging in capital gains harvesting to accelerate capital gains into lower-tax-rate years instead.
Example 3. Nick and Carla are both 66 years old and have a sizeable $7 million portfolio held entirely in a taxable account — representing the proceeds of selling their family business — alongside a pair of IRAs worth about $800,000.
Their portfolio is invested in a 40/60 conservative growth allocation, resulting in approximately $4.2 million held in bonds generating an average yield of $126,000 (3%), and $2.8 million in stocks that are producing $56,000 per year (2%) in qualified dividends and an average of $168,000 per year (6%) in capital gains turnover. After their seemingly modest standard deduction of $24,000 — with few other deductions since they also purchased their retirement home for cash with part of their business sale proceeds — their taxable income is $326,000 per year, based on a blend of ordinary income and long-term capital gains rates. On top of this they’ll be receiving nearly $50,000 per year in Social Security benefits and almost $39,000 per year in RMDs beginning in four years at age 70.
In practice Nick and Carla really have two tax rate equilibria to plan around. The first is based on their ordinary income of $126,000 per year (bond interest) + $42,500 (the 85% of their Social Security that is taxable), on top of which they’ll add $39,000 per year of RMDs at age 70, reduced by their $24,000 standard deduction. This means their net ordinary income of $144,500 is enough to keep them in the 22% tax bracket for now, but when their RMDs begin, they will be pushed up into the 24% bracket. They’ll move even higher as the growth rate on their wealth compounds, given that their goal is to spend just $120,000 per year.
The second equilibrium to consider is Nick and Carla’s long-term capital gains and qualified dividend income, which at a combined amount of $224,000 per year projected on average and stacked on top of their $144,500 of ordinary income puts them squarely in the middle of the 18.8% tax bracket, with still some room before they hit the top 23.8% capital gains rate, which begins at $479,000 for married couples.
Thus Nick and Carla can combine strategies, engaging in partial Roth conversions of approximately $25,000 per year — which they can do to slightly whittle down their IRA while remaining in the 22% ordinary bracket — and also harvesting any capital gains in years they don’t have them to ensure they’re still filling up the 18.8% capital gains bracket before their compounding wealth drives them into the 23.8% bracket.
Nick and Carla consequently have the opportunity to plan around both the 22% — and ultimately perhaps 24% — tax rate equilibrium for their ordinary income to avoid 32%+ future tax brackets, and around the 18.8% equilibrium for their long-term capital gains and qualified dividends to avoid ever being pushed into the 23.8% tax bracket.
Determining optimal tax rate equilibrium will depend on a retiree’s income sources and when they begin, plus overall wealth and projected spending — all while operating under the assumption that wealth in the aggregate will continue to grow and compound toward even higher brackets.
For mass affluent retirees, the optimal equilibrium will likely be the 12% tax bracket and the highly appealing 0% long-term capital gains tax rate, rather than jumping up to 22% and 15% rates, respectively. Indeed, even those with wealth of $1 million or more may still be able to remain in those brackets, especially if income is proactively managed to ensure they stay in equilibrium, and don’t compound or defer too much for the future.
For more affluent individuals, staying in the 22% or 24% brackets and avoiding the sizable jump to the 32% tax bracket that kicks in at $157,500 for individuals and $315,000 for married couples filing jointly may be optimal. Notably this also aligns with staying in the 15% long-term capital gains rate and trying to avoid the 3.8% Medicare surtax on capital gains that kicks in at $200,000 of AGI for individuals, or $250,000 for married couples — recognizing that it may be more difficult for those couples to avoid in the future given the 3.8% Medicare surtax threshold is not indexed for inflation.
For the wealthiest households, the ideal tax rate equilibrium may simply be anything that is not the top tax brackets of 37% for ordinary income and 23.8% for long-term capital gains. The caveat, as shown in Example 3, is that even households with $7 million-plus of net worth may actually be nowhere near those top tax brackets. It often takes wealth of $10 million or more to actually reach — and remain in — the highest brackets. Though for those with significant wealth who aren’t proactively spending or giving away all the growth, future compounding may still eventually lift them up to the top brackets.
The bottom line is simply to understand that there really is such thing as being too good at minimizing tax exposure. The solution is to balance current and future tax rates, finding the equilibrium point that allows retirees to avoid higher tax rates in the future without unnecessarily increasing their taxes today.