Constructing Tax Efficient Withdrawal Strategies for Retirees

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
Beliavsky
Posts: 927
Joined: Sun Jun 29, 2014 10:21 am

Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by Beliavsky » Fri May 17, 2019 3:21 pm

Skimming this paper, I conclude that determining the optimal withdrawal strategy is a non-trivial problem, even when future stock returns
are constant and known, as the authors assume.

Constructing Tax Efficient Withdrawal Strategies for Retirees with Traditional 401(k)/IRAs, Roth 401(k)/IRAs, and Taxable Accounts
41 Pages Posted: 17 May 2019
James DiLellio
Pepperdine University - Graziadio School of Business and Management
Daniel N Ostrov
Santa Clara University
Date Written: August 8, 2018
Abstract
We construct an algorithm for United States retirees that computes individualized tax-efficient annual withdrawals from IRAs/401(k)s, Roth IRAs/Roth 401(k)s, and taxable accounts. Our algorithm applies a new approach that generates an individualized strategy that results in consistent improvements over non-individualized withdrawal strategies currently advocated by financial institutions and academics. Among other results, we quantifiably demonstrate why retirees should avoid, not seek, dividend producing stocks in their taxable accounts. Our model, which can work to optimize either portfolio longevity or the bequest to an heir, accommodates many salient tax code features, including dividends, different taxable lots, conversions, and required minimum distributions.
Keywords: retirement income, tax efficiency, optimization
JEL Classification: G11, H21

From the conclusion:

Our algorithm starts by using current tax law to create ve guiding principles that govern
prioritizing consumption from the three accounts: 1) We prove that if the retiree consumes a
given amount of Roth money and given amounts of TDA money at various marginal tax rates,
the order/allocation in which these are consumed is irrelevant; 2) In contrast, taxable stock
and dividends are better spent earlier rather than spent later; 3) When consuming taxable
stock, the lot with the highest cost basis should always be tapped; 4) Dividends should always
be consumed before Roth money; and 5) Required Minimum Distributions (RMDs) should
always be taken out of any TDA. These five guiding principles help us develop desirability
factors for the three accounts. These desirability factors help us determine the most favorable
annual allocations among the accounts that satisfy the consumption needs of the retiree. Our
algorithm incorporates a number of standard features studied in the research literature, such
as working with RMDs from the TDA account and optimizing portfolio longevity, as well as
a number of less standard features, such as incorporating the eect of dividends, optimizing
a bequest to an heir, allowing for two types of additional xed sources of money for the
retiree, accommodating dierent taxable lots in the taxable stock account, and working with
conversions from the TDA to both the Roth account and the taxable stock account.

RadAudit
Posts: 3421
Joined: Mon May 26, 2008 10:20 am
Location: Second star on the right and straight on 'til morning

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by RadAudit » Fri May 17, 2019 3:44 pm

Thanks for the post. Ought to be an interesting discussion after a number of us wade through it.
Non-academic advice, coming from investment firms, financial advisors, and books on retirement, recommend strategies for retirees’ withdrawal choices that are often far from optimal and often in contradiction with each other, with the exception that there is general agreement that Required Minimum Distributions (RMDs) should be taken from TDAs. For example, one very common category of strategies, termed “na¨ıve” strategies by Horan (2006a and 2006b), recommends that retirees completely exhaust one account before moving to the next. The books by Solin (2010), Rodgers (2009), and Lange (2009), suggest sequencing withdrawals so that retirees drain taxable accounts first, then TDAs, and finally Roth accounts. This strategy is also endorsed by large retail investment firms Fidelity (see Fidelity (2014) and Fidelity (2015)) and Vanguard (Vanguard (2013)). In contrast, other financial authors, such as Larimore, Lindauer, Ferri, and Dogu (2011), recommend first draining taxable accounts, but then recommend draining Roth accounts followed finally by TDAs.
Are these BHs Larimore, Lindauer, Ferri, etc?
FI is the best revenge. LBYM. Invest the rest. Stay the course. - PS: The cavalry isn't coming, kids. You are on your own.

jebmke
Posts: 9330
Joined: Thu Apr 05, 2007 2:44 pm

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by jebmke » Fri May 17, 2019 3:48 pm

RadAudit wrote:
Fri May 17, 2019 3:44 pm
there is general agreement that Required Minimum Distributions (RMDs) should be taken from TDAs.
lol, so they all agreed that retirees should follow the law?
When you discover that you are riding a dead horse, the best strategy is to dismount.

Big Dog
Posts: 1180
Joined: Mon Sep 07, 2015 4:12 pm

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by Big Dog » Fri May 17, 2019 3:59 pm

Are these BHs Larimore, Lindauer, Ferri, etc?
ok, I'll bite and have to go back and read the tome as to why....I expect that my marginal rate will be lower than my heirs, so not sure why leaving the TDAs instead of Roths is a good thing.

User avatar
grabiner
Advisory Board
Posts: 24572
Joined: Tue Feb 20, 2007 11:58 pm
Location: Columbia, MD

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by grabiner » Fri May 17, 2019 9:51 pm

The summary here oversimplifies some issues, but the article itself has the correct details.
Beliavsky wrote:
Fri May 17, 2019 3:21 pm
From the conclusion:

Our algorithm starts by using current tax law to create ve guiding principles that govern
prioritizing consumption from the three accounts: 1) We prove that if the retiree consumes a
given amount of Roth money and given amounts of TDA money at various marginal tax rates,
the order/allocation in which these are consumed is irrelevant
The order matters because marginal tax rates are not constant. If you are in a 12% tax bracket this year, and decide to withdraw more from your traditional account and less from your Roth, you may avoid having to make withdrawals from the traditional account in a 22% bracket in a future year. (The article gets this right; the equivalence is only for money withdrawn at the same known rate.)
2) In contrast, taxable stock and dividends are better spent earlier rather than spent later
However, it may be better spent not at all than spent earlier. If you have $10,000 in stock with a $5000 basis, and sell it with a 15% tax on capital gains, you lose 7.5% of the value. If you keep the stock, you might lose 0.3% per year to dividend taxes and then when you die after ten years, your heirs have lost only 3% of the value. (Again, the article considers this.)
3) When consuming taxable stock, the lot with the highest cost basis should always be tapped
This is true even if you pay 0% tax on capital gains. If you have space to take 0% capital gains, sell more stock to harvest your gains, reducing future gains.
4) Dividends should always be consumed before Roth money
This is clear. Once you have paid the tax on the dividends, you can spend the remainder now with no additional tax cost, It is better to keep money growing tax-free than in your taxable account (and usually also better to keep it growing tax-deferred).

Similarly, anything you can spend tax-free from your taxable account, such as a stock or bond investment with no capital gain, should be spent in preference to tax-deferred money.
5) Required Minimum Distributions (RMDs) should always be taken out of any TDA
Given the 50% penalty for not taking them, this should also be obvious. You can plan your finances as if your RMD had already been taken and taxed, and then decide how much of it to spend. If your RMD plus any dividends exceed your financial need, you can either reinvest the remainder or use it to pay taxes on Roth conversions.
Wiki David Grabiner

User avatar
jeffyscott
Posts: 8002
Joined: Tue Feb 27, 2007 9:12 am
Location: Wisconsin

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by jeffyscott » Sat May 18, 2019 8:02 am

grabiner wrote:
Fri May 17, 2019 9:51 pm
The summary here oversimplifies some issues, but the article itself has the correct details.
Thanks for clarifying the the part about taxable stock and dividends are better spent earlier rather than spent later but that it may be better spent not at all than spent earlier.

We have only a small taxable account but it may grow with RMDs someday. I think the article and your summary of the summary confirms what I have been thinking makes sense for us. Spend dividends from taxable account first, then spend from TDA...and for us that is likely as far as we would ever need to go. And, of course, converting TDA to Roth at 12% as much as possible.

We'd pay 0% on dividends and 12% on TDA now, rising to 15% and 22% when there is a sole survivor. Heirs will likely pay 12% or 22% and same or similar state tax rate to us.
Time is your friend; impulse is your enemy. - John C. Bogle

The Wizard
Posts: 13355
Joined: Tue Mar 23, 2010 1:45 pm
Location: Reading, MA

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by The Wizard » Sat May 18, 2019 9:44 am

I feel that retiree tax efficiency is best done simply by managing your Adjusted Gross Income (AGI) from your sixties through your seventies so that it increases by just a few percent per year, the approximate inflation rate.

Details on how to do this will depend on your mix of income sources and tax-deferred account size vs taxable account size.
But for many of us, this involves doing properly sized Roth conversions in our sixties.

There are at least two wrinkles to this approach:
1) avoiding having AGI "just over" a Medicare IRMAA threshold; there are several of them.
2) married couple considerations for survivor's tax situation...
Attempted new signature...

Phil DeMuth
Posts: 139
Joined: Thu Mar 06, 2008 7:02 pm

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by Phil DeMuth » Sun May 19, 2019 11:39 am

An important implication of this article, which also agrees with the Reichenstein/Meyer, is that discovering the retiree drawdown strategy with the highest utility cannot be done by following simple rules of thumb. It involves a complicated calculation that also entails many assumptions. The odds of someone just stumbling into the right strategy are very low. It takes software, and also an understanding of how the assumptions affect the answers, so that a retiree can make his best estimate among competing future scenarios (tax rates, investment returns, longevity, bequests), with a lot of money hanging in the balance.

randomguy
Posts: 7815
Joined: Wed Sep 17, 2014 9:00 am

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by randomguy » Sun May 19, 2019 6:38 pm

Phil DeMuth wrote:
Sun May 19, 2019 11:39 am
An important implication of this article, which also agrees with the Reichenstein/Meyer, is that discovering the retiree drawdown strategy with the highest utility cannot be done by following simple rules of thumb. It involves a complicated calculation that also entails many assumptions. The odds of someone just stumbling into the right strategy are very low. It takes software, and also an understanding of how the assumptions affect the answers, so that a retiree can make his best estimate among competing future scenarios (tax rates, investment returns, longevity, bequests), with a lot of money hanging in the balance.
It was obvious that the simple rules of thumbs were flawed. I don't think anybody ever took them seriously as the right thing to do versus just examples of the differences between approaches.

The question is doing complex work like this going to be noticeably better than just winging it with some blended strategy (ie. ROTH conversions up to x, taxable to top of bracket y, roth after that) that people normally recommend. At a certain level the unknowns (how long will I live, what and when will I want to spend money, what returns will I get, what will be the tax laws, what will my kids taxes be,...) are big enough that going for more precision is an illusion.

User avatar
jeffyscott
Posts: 8002
Joined: Tue Feb 27, 2007 9:12 am
Location: Wisconsin

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by jeffyscott » Sun May 19, 2019 10:14 pm

randomguy wrote:
Sun May 19, 2019 6:38 pm
Phil DeMuth wrote:
Sun May 19, 2019 11:39 am
An important implication of this article, which also agrees with the Reichenstein/Meyer, is that discovering the retiree drawdown strategy with the highest utility cannot be done by following simple rules of thumb. It involves a complicated calculation that also entails many assumptions. The odds of someone just stumbling into the right strategy are very low. It takes software, and also an understanding of how the assumptions affect the answers, so that a retiree can make his best estimate among competing future scenarios (tax rates, investment returns, longevity, bequests), with a lot of money hanging in the balance.
It was obvious that the simple rules of thumbs were flawed. I don't think anybody ever took them seriously as the right thing to do versus just examples of the differences between approaches.

The question is doing complex work like this going to be noticeably better than just winging it with some blended strategy (ie. ROTH conversions up to x, taxable to top of bracket y, roth after that) that people normally recommend. At a certain level the unknowns (how long will I live, what and when will I want to spend money, what returns will I get, what will be the tax laws, what will my kids taxes be,...) are big enough that going for more precision is an illusion.
I agree, seems like too many unknowns, too many assumptions, too many variables. The general ideas that come out of it may be useful, but I don't know about actually trying to use this to come up with a complicated, precise plan. Maybe the next step will be plugging it all into a Monte Carlo simulation?

For my situation, I think paying up to 12% Federal by choice on as much of TDA as I can makes sense. Anything higher, I'm doing only when required.

To the extent I am wrong, it will most likely mean someone would've paid 10% or that I could have paid 22%, rather than 24%. So either error is only a loss of 2%.
Time is your friend; impulse is your enemy. - John C. Bogle

Topic Author
Beliavsky
Posts: 927
Joined: Sun Jun 29, 2014 10:21 am

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by Beliavsky » Mon May 20, 2019 3:49 am

A traditional IRA has an insurance aspect that a Roth IRA does not, because the tax rate you pay when making TIRA withdrawals depends on your income. If you are low income you will pay less tax on TIRA withdrawals. I wonder if this insurance feature affects the optimal withdrawal strategy. OTOH, increases in tax rates will hit your TIRA withdrawals and cost you money in other ways, so a Roth IRA provides a better hedge against tax increases (assuming the Federal government does not start taxing Roth IRA withdrawals).

User avatar
dodecahedron
Posts: 4433
Joined: Tue Nov 12, 2013 12:28 pm

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by dodecahedron » Mon May 20, 2019 5:20 am

The Wizard wrote:
Sat May 18, 2019 9:44 am
I feel that retiree tax efficiency is best done simply by managing your Adjusted Gross Income (AGI) from your sixties through your seventies so that it increases by just a few percent per year, the approximate inflation rate.

Details on how to do this will depend on your mix of income sources and tax-deferred account size vs taxable account size.
But for many of us, this involves doing properly sized Roth conversions in our sixties.

There are at least two wrinkles to this approach:
1) avoiding having AGI "just over" a Medicare IRMAA threshold; there are several of them.
2) married couple considerations for survivor's tax situation...
A few more wrinkles:

3) if you are going to pursue a strategy of ¨itemized deduction bunching¨ in some years and take the standard deduction in other years, it is generally better to move some amount of AGI from the standard deduction years into the years when you have chosen to have high itemized deductions. This is particularly true if you might otherwise wind up bumping into the 30% AGI limitation on donations of appreciated securities. Deduction carryforwards become worthless if they move the deductions into years when you won´t be itemizing. So it might be better to move somewhat in the direction of smoothing/levelizing real taxable income rather than real AGI. (I say ¨somewhat¨ because there are AGI concerns independent of taxable income concerns.)

4) those projecting themselves to be in the SS hump zone once they begin collecting it may want to move some AGI from their SS-collecting years into their pre-SS years.

5) under current law taxes are ¨on sale¨ so it is worth considering moving AGI and taxable income from years after 2025 to the years before then.

6) qualifying for senior citizen property tax breaks can be an important consideration in some locations

But of course generally I agree with the The Wizard (and general consensus on this forum) that Roth conversions prior to RMD age are very much worth considering.

User avatar
dodecahedron
Posts: 4433
Joined: Tue Nov 12, 2013 12:28 pm

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by dodecahedron » Mon May 20, 2019 5:24 am

Beliavsky wrote:
Mon May 20, 2019 3:49 am
A traditional IRA has an insurance aspect that a Roth IRA does not, because the tax rate you pay when making TIRA withdrawals depends on your income. If you are low income you will pay less tax on TIRA withdrawals. I wonder if this insurance feature affects the optimal withdrawal strategy. OTOH, increases in tax rates will hit your TIRA withdrawals and cost you money in other ways, so a Roth IRA provides a better hedge against tax increases (assuming the Federal government does not start taxing Roth IRA withdrawals).
The insurance aspect is very much worth bearing in mind. If you anticipate the possibility of needing to spend significantly out of pocket for deductible long term care expenses or other large medical expenses down the road, then preserving some tax deferred money for that eventuality is worthwhile. After 2025, tax law is scheduled to revert to allowing more broadly for deductible casualty losses again, which is another insurance aspect of tax-deferred funds.

megaroth
Posts: 75
Joined: Sun Nov 30, 2014 2:08 am

Re: Constructing Tax Efficient Withdrawal Strategies for Retirees

Post by megaroth » Sat May 25, 2019 6:19 pm

Hopefully, we'll see continued academic (or layperson) work on this topic. It would be useful to get to a very robust model that could account for a wider range of custom inputs and assumptions (e.g., ACA subsidies, Medicare/IRMAA, Social Security, extent of concern (or not) re: heirs/inheritances).

Beyond that, a couple questions on this paper specifically:

(1) Why was "no bonds/cash in TDA/Roth/taxable" a necessary limitation of the model?


One of the ramifications of our model being deterministic is that it will be optimal to have strictly stock, as opposed to bonds or cash, in the TDAs, Roth accounts, and taxable accounts. That is, because the stock returns are assumed known, the model cannot recommend having bonds or cash given their lower return rates.

Wouldn't it be possible to use risk-adjusted returns in the model to even the playing field?


(2) Is "sell lowest-cost-basis stock in taxable first" really always optimal?

When consuming taxable stock, the lot with the highest cost basis should always be tapped.

For example, what if you (a) don't care about optimizing for heirs, (b) are an early retiree who can and will be (or should be, for tax optimization purposes) in the 0% capital gains bracket in early retirement years, and (c) expect to be in a higher income tax (and higher capital gains tax) bracket later (e.g., with Social Security and RMDs coming into play)? In that case, it seems that doing maximum tax-gain harvesting starting with *lowest* cost basis might be valuable because it would maximize opportunities for tax-loss harvesting in later, higher-tax-bracket years.

Post Reply