Proper way to factor in taxes to safe withdrawal rate

[Toddtheformeraccountant]

A very interesting article from Kitces that kind of made me rethink how to factor in taxes to the safe withdrawal rate.

https://www.kitces.com/blog/the-impact-of-taxes-on-the-safe-withdrawal-rate/

I have seen many here advocate simply adding the amount of taxes you will pay (along with your living expenses) into the amount that will need to be "covered" in terms of ongoing costs. I've done the same thing in my planning.

However, what that method fails to consider is that, in the very negative downside scenario, there will actually be very little taxable income and taxes to pay (ie., if you're just drawing down principal). And on the upside scenario, where values are increasing and income is increasing, the fact that you're paying more taxes is irrelevant because you're in "fat city."

So that "add tax expense to the cost" (let's call it the "add taxes" method) overstates the negative. Income taxes only hit if..well...you actually have income to tax. For example, for me, my expected withdrawal rate is 2.75% on the "add taxes" method, 1.7% excluding taxes.

Kitces instead displays here the computations that consider these upside and downside scenarios properly, and you are to use those percentages against an "adjusted" safe withdrawal rate scaling down from the standard 4% (with "no" taxes) to about 3.5% (when you are at top tax rates).

So my "cushion" is actually much higher than it appears with the "add taxes" method. 2.75% "add taxes" actual WR vs. 4% SWR is a 1.25% cushion....but reality is it should be 1.7% actual WR compared to 3.5% max tax rate effected cushion or a 1.8% cushion...55 bps higher cushion. All because, under the "add taxes" method, the downside scenario effectively imputes a tax that will not occur under that scenario. 55 bps more cushion may sound small but we all know that's a HUGE difference.

Wondering what other folks thoughts are, maybe this is a "duh" and here I am thinking I've found the holy grail and the rest of you may be "well no sh** sherlock."

 

[NW-Bound]

... "add tax expense to the cost" (let's call it the "add taxes" method) overstates the negative...

Overstating the negative is OK by me. I am a pessimistic and cautious person. I prefer to be pleasantly surprised than the other way around.

And of course as Koolau always signs off on his posts, YMMV.

 

[scrabbler1]

I have been using a skeleton version of a tax return for over 20 years, going back to my working days. I simply used it to estimate my income taxes in retirement so I wouldn't be surprised much after my income went from mostly wages to all investment, and from different sources. I then added that tax amount to my expense budget.


I have split my income taxes into 2 parts. The first part is a basic part, based on the more predictable portion of my income from monthly and quarterly bond and stock mutual dividends. This is the part I include in my expense budget. The second part is an excess part which comes from the more volatile and less predictable cap gain distributions. I am not worried about being able to cover these taxes because they will only exist if I am receiving those larger, non-recurring distributions (which will pay the taxes).

 

[Montecfo]

I estimate my spending needs excluding income taxes. Then I estimate taxes based on the expected funding source. So my estimates should be pretty close and I don't expect surprises.

Now, if you simply take your spending and multiply by a marginal rate, yes, you are going to overstate your taxes. What you really need is an effective rate, built from an actual tax calculation.

 

[Midpack]

Not disagreeing with the idea, but it sounds like using a scalpel where an axe is the right tool. There are so many variables (future tax rates for one), over such a long period (30 years give or take), that trying to account for “negative downside” sounds like an academic exercise. YMMV

 

[BeachOrCity]

Has anyone on this forum who has been retired for 10+ years actually had their finances (assets and spending) follow their initial model and plan? By this I mean spending is as planned and assets are as planned for 10 years running. Somehow I doubt it.

All of these models are designed to give confidence that it is safe to retire and something to go back to to ensure one is “on track”.

Life doesn’t follow the spreadsheet, the spreadsheet follows life.

 

[donheff]

Not disagreeing with the idea, but it sounds like using a scalpel where an axe is the right tool. There are so many variables (future tax rates for one), over such a long period (30 years give or take), that trying to account for “negative downside” sounds like an academic exercise. YMMV

+1 It is pretty easy to estimate taxes in the "expected" path. As Kitces says, it won't matter much if everything turns up roses, In bad times I would want some wiggle room - the axe rather than the scalpel. Even in a downturn I anticipate significant taxes on RMDs (even if unspent) on top of continuing pension and SS income taxes.

 

[GenXguy]

However, what that method fails to consider is that, in the very negative downside scenario, there will actually be very little taxable income and taxes to pay (ie., if you're just drawing down principal).

During retirement, you aren't just drawing down principal. When I sell funds in my brokerage, I get my principal, but I still have capital gains, plus I'll still have income from dividends, interest income, IRA, 401K, SS/pension. The reason why the "federal" taxes are so low (or 0) when selling those funds is because I can keep my long term capital gains (including qualified dividends) within the 0% capital gains tax bracket. It is not that they are just principal - they are not. And even then, I still have to pay 5% "state" income tax on those capital gains. And then, all distributions from my 401K and traditional IRA as well as non-qualified dividends and interest income are potentially taxable as ordinary income for federal taxes. What keeps "federal" income tax reasonably low there is that the $12K standard deduction applies, so I'm able to reduce the amount of taxable income at the federal level, but I still have to pay state income tax on most of that income. As soon as I start receiving SS, a significant amount of that (approaching 80% of it over time) will be taxable at about 12% federal rate (some people will pay less, but the taxable threshold is not indexed to inflation, so more people get caught with higher SS taxes each year, therefore you need to figure how much will be taxable based on future inflated dollars against current taxable SS thresholds that are fixed).

And on the upside scenario, where values are increasing and income is increasing, the fact that you're paying more taxes is irrelevant because you're in "fat city."

Taxes are not irrelevant. When you're basing your drawdown on the 4% rule, you prepare for the worst case scenario and should be planning your drawdown and state/federal taxes based on that scenario, not planning (assuming) the "best outcome" where you become a fat cat. That said, even fat cats are concerned with taxes, so there's no situation that taxes are irrelevant, and certainly not when using the 4% rule.

4% should include ALL of your expenses, and that includes federal AND state income tax and any investment expenses.

What I have done is created my own custom spreadsheet based on a few decades of retirement, and I have all of my "buckets" of drawdown that I can live off with fields under each retirement year, and based on the amounts entered to take from any particular buckets in any year, it automatically figures the tax (if any) and how much money is available for required expenses and discretionary spending each year. It calculates automatically using the appropriate tax rates and deductions for federal and state. I estimate the amount of money available in each bucket based on minimal real gains over the years (just as the 4% rule does) and will reevaluate each year during retirement depending on growth/loss/rebalancing in my investments and changes in my expenses. I also calculate everything in "today's" dollars, which makes it easier and keeps the figures more meaningful to me. At age 65, my federal tax finally jumps significantly because that's when I plan to draw SS. I certainly cannot ignore that tax based on the 4% rule. And if my investments do better than my spreadsheet assumes, that is fine and good, but I'll still be paying taxes, and my spreadsheet will be updated accordingly.

Future tax rates could change, and that's out of my control, but I can optimize my drawdown to minimize taxes. Currently, my planned optimization is based on current tax rates, and I will adjust in the future if need be, just as I might when I reevaluate each year the growth/loss/rebalancing of my investments and changes in my expenses.

Real life and investments are dynamic, as will my spreadsheet be.

 

[RetireBy90]

I've just retired last year, and large part of savings are in TIRA. So, starting this year I'll be trying to move as much from our TIRA to Roth as possible in 22% bracket. My thinking is that with the current tax rates due to expire in a few years there is a good chance that when my RMD kicks in (6 years from now) rates will be higher. So, thinking it makes good sense to take the hit from taxes now rather than wait to see what the tax rate will be in 6 years. I'll meet with a FA this fall now that we have first SS check from DW and can look at some real income numbers for this year. Lots of things to consider, IRMAA for Medicare, cap gains rate, etc.

 

[pb4uski]

+1 It is pretty easy to estimate taxes in the "expected" path. As Kitces says, it won't matter much if everything turns up roses, In bad times I would want some wiggle room - the axe rather than the scalpel. Even in a downturn I anticipate significant taxes on RMDs (even if unspent) on top of continuing pension and SS income taxes.

+2 My taxes are pretty stable.... about 8% of withdrawals currently and bumping up to 10-11% of withdrawals once RMDs begin (and that is assuming that RMDs are fully spent). Kitces article sounds to me like an elegant solution in search of a problem.

 

[hulurt1]

I've been retired 11 years. All in IRA. I also have a Roth. Last years I moved money to Roth when the market was down. I gain all my tax back and up 15%. I'm thinking that next time we are down 20% I will move big time to Roth. Up to 24 tax bracket.

 

[RetireBy90]

I've been retired 11 years. All in IRA. I also have a Roth. Last years I moved money to Roth when the market was down. I gain all my tax back and up 15%. I'm thinking that next time we are down 20% I will move big time to Roth. Up to 24 tax bracket.

Just had a discussion about that with a buddy. I just retired last year and plan to convert to top of 22% bracket each year till we hit RMDs. However, talking it through I asked why 22%, next bracket is 24% up to $390K, and only 2% more. Would that 2% be a game changer ? My plan is based on thesis that after 2025 when current rates are projected to revert they will be higher than today. In 2016, income above $150K was 28% so if they do revert then I’d save 4%. Right ?

 

[audreyh1]

Definitely for me in hot market years taxable income taxes are higher, and in negative years taxable income drops dramatically. As a consequence my withdrawals may vary widely (% of portfolio), but my after tax income varies much less. In other words, after a tough market year my pre-tax income may take a big hit, but taxes are often way lower, so my income available for spending doesn’t go down nearly as much. This is a nice ameliorative effect.

 

[RunningBum]

I think it's an astute observation, as I've come to expect from Kitces. My taxes are pretty stable now, and probably will continue to be as long as things are good. And I'll also easily make it to any age in that case. Thus, taxes are pretty irrelevant.

But, if things go bad, I'm going to be selling more fund than normal. And since things are bad, I'll be selling them for less, thus fewer taxes.

I'll give it more thought, but I'm not inclined to take any action based on this right now. I'll consider it "buffer". I'm not sure how else to factor that in. But it is reassuring to realize I'll have less tax liability in the worst cases.