Comments on TvsR edits below.
FiveK wrote: ↑
Wed Jul 24, 2019 8:38 pm
Don't understand the addition of "in most cases." In what case does paying a higher tax rate leave one with more spendable after tax money?
There are potentially lots of reasons - different investment choices in the different T and R accounts, estate planning issues, maxing out accounts, etc. They're discussed in the article. I just think it's important to emphasize that while a marginal rate comparison is best in the simple case and a great first step in complicated cases, there are other factors that can swing things in the other direction.
The t vs. R article is long enough without including general investment considerations. Simply direct people to "Prioritizing investments" for those.
I agree that content shouldn't be duplicated without good reason, but the general investment considerations aren't specific to the TvsR decision, and the two non-tax considerations are high-yield enough that they definitely provide enough value to the reader to justify their space. Any other opinions? FWIW, those two bullets are from the existing article.
It's not practical to list all the permutations in the "General guidelines" section that any given individual might encounter. Keep it simple.
Agreed on the general idea of keeping it simple. If we could give all readers (advanced and novice) only one high-yield recommendation, it would have to be "traditional during peak earning years, Roth in low-income years". I'm flexible on removing some of the other ones, like estate tax, but at least that one bullet needs to appear somewhere near the top of the article. Do you disagree with me on this point, or are there other bullets (state tax, estate tax) you're not crazy about?
What was there before ("consider doing 100% traditional, or 50/50") is, I think, not very helpful. It doesn't suggest any criteria for choosing either of these options, so it's not very different from telling readers to flip a coin. I think it's better to give some simple rules of thumb that beginners or non-math oriented readers can use and still get close to the right answer.
No real preference regarding the addition of an "Eligibility" section and the edits therein.
Someone added some language in there during my LOA, and after reading it it seemed like a good addition. If we're going to discuss eligibility it should be segregated from the tax-based TvsR issues, in my opinion.
Line 31: Don't see any difference between the versions, so...?
Yea, not sure what happened there.
- The change from the simple(st) case involving the commutative property, to what happens when the pre-tax amount one wants to use is greater than the contribution limit, is common enough that it should be toward the top of the article.
I would agree with a mention/link, or possibly a brief summary, but are you saying you want to move the whole "maxing out retirement accounts" section more toward the top of the article? I don't think that's a good idea - from what I've seen in the forum, many people have enough trouble understanding the reasoning behind the simplest case. Moving that into this section would separate the marginal rate discussion with the two key sections on how to calculate current and future marginal rates.
Leaving "Comparing marginal rates between contribution (or conversion now) and withdrawal in the future is the most direct way to achieve this goal" as a separate paragraph seems better, as it makes the point more visible.
Yeah, but it's directly tied to the prior two sentences - the "this" as the second to last word would become loose. Maybe there's another good way to emphasize? Bold/italics/underline?
The older version directs the reader to "see the forest" instead of going "tree by tree", and seeing the bigger picture seems preferable.
Sorry, not sure what you mean. I thought it would be helpful to list the bulletized methods. You don't like this change?
Need to distinguish between pensions that start immediately vs. those that can be deferred in return for higher payments
Sure, but in either case the pension is going to start some time, and when it does the income needs to be accounted for. Royalty checks may get cut off at a certain point, but I'm not sure getting into the time phasing of these incomes is something we can do for readers here, they'll need to figure that out for themselves. Maybe we say, "Estimate any guaranteed retirement income, including pensions, royalties, and rental property and investment income, and the time-phasing of when this income will start and stop." ?
Due to the unequal downsides of getting t vs. R "wrong" - having too much in traditional is a "first world" problem, while too much in Roth due to unexpected income shortfall can be more serious - advising people to use "conservative" return estimates is preferable to advising "realistic" return estimates.
I've thought a lot about this, and I'm still not convinced Traditional provides much benefit in terms of safety. If we were recommending to readers how much
of their income to save for a retirement goal, then yes, conservative estimates for investment return are clearly appropriate. And it's definitely true that in a shortfall scenario, a declining tax rate will help partially offset the shortfall, but it's a pretty small effect. I just ran some numbers, and I assumed a single retiree hoped to have $2.5M saved and fell short by 20% ($2.0M). His AVERAGE tax rate on a 4% withdrawal ($100,000) would have been 14.85%, but it's now only 13.01% on an $80,000 withdrawal, so his after-tax availability of his traditional account increased from 85.15% to 86.99%... only a 2% increase. So, he falls short on savings by 20% and only gets 2% of that back due to tax savings by going traditional... that seems like a small benefit to me. Future tax law changes could also easily wipe out this benefit.
There's an offsetting risk of Traditional, and that is if you are contributing anything like a fixed dollar amount, either to get a matching or maxing out, you have more spending money today due to the tax savings, and are thus actually saving less than with a Roth. If you don't have the discipline to save extra money, in or outside of your retirement accounts, then you'll have a shortfall, whereas if you go Roth, at least you "guarantee" the tax portion of your retirement savings, even if the investment side remains uncertain. Small issue admittedly, especially for our readers, but seems on par with the other one.
After thinking about this more, I think this "benefit" is small and tenuous enough to not get mentioned. What do you think? Am I missing something?
The timing of SS benefits is important.
Yes, definitely, although I think Roth conversions are better accounted for in the traditional line item than SS. How about adding something like, "...and the decision of when to begin receiving Social Security benefits." ?
Just do it every year.
Eh, call me lazy, but in my situation, living in a very high tax state and planning to retire to a tax-free state, if not move there much sooner, it's not worth crunching extra numbers every year. The answer is going to come out the same. Ditto for all those who are way ahead or behind on savings. I'm not hard-over on this though, we can say every year and those who are lazy like me can pare it down.
The usefulness of cautioning people about self-defeating predictions is discussed in another post.
I don't think "self-defeating" is the right word. The more you contribute to Traditional, the lower the marginal value of those contributions will be, because of rising future tax rates. Once future rates equal current rates, switch to Roth for further contributions. It's just a mathematical balance, but it's an overstatement to say that traditional contributions are self-defeating. In any case, this is still under the umbrella of the "simple case" where we assume constant marginal rates, so I don't think this effect should be emphasized here. The right place to discuss it is further down, in the "straddling brackets" section.
Using "no further traditional contributions" as a basis for estimating future tax rates is discussed in another post.
Yeah, and I don't think this is the correct strategy. The problem with assuming no future contributions is that if you check the math and it shows a lower future marginal rate, you're left wondering if you contributed all traditional whether your future rate would be equal or higher than today. But if you check assuming full planned future contributions and traditional comes out better, then you're confident that's the right answer. I suppose if your future rate were higher
than current rate with no future contributions, then you'd be confident that Roth was the right answer, but that's a unlikely circumstance in a year with normal income. Far more likely, it seems, is someone who's behind on savings and/or retiring to a tax-free state, which is why I wrote the example out the way I did.
- Using "no further traditional contributions" as a basis for estimating future tax rates is discussed in another post.
- We could simply delete the "qualitative considerations" section - as noted in the "Line 2" comments above, it's not practical to list all the permutations - but if we want to retain them, it seems better to put them towards the end.
See my comments above. Also, the example before I edited it involved partial Social Security taxation and had a marginal rate of 22.2% - I think that's too complex for the simplest example. I chose numbers that would guarantee that all SS income would be taxed at 85% to remove that from the equation. There's a big SS section further down that deals with that issue.
- Maybe include the "risk of shortfall" wording where we make conservative assumptions about investment returns?
See above analysis; I'm now in favor of removing this section completely.
No problem with including that link.
At some point we might want to take some ideas from the Investment Order
post and revise the Prioritizing investments
article, but that's for another day.
Agreed that's a good reference, although I'm not sure I agree 100%. For instance, it has readers never pay off debts with less than 3% IR above the 10-year T-note. That seems really high to me; I'd definitely recommend readers pay off debts with lower IR than that. MAYBE hold onto a 1-2% nominal debt, but even that I would pay off myself for simplicity and cash flow. As you say, a discussing for another time.
Line 253, 308, 318, 345, 371, 420
- All changes look good. Especially nice on the math work!
- When the personal finance world stops making the mistake of "saving marginal but paying effective" (e.g., see Roth versus Tax-Deferred: The Critical Concept of Filling the Brackets - The White Coat Investor - Investing & Personal Finance for Doctors from a month ago), then we can dip our toes into the "instead of marginal" waters.
Until then, we should stay consistent with our use of "The main reason to prefer one type of account over the other is the comparison of marginal tax rates."
Over glasses of our favorite beverages, and just among us posting here, we likely all agree on the distinction between "instantaneous marginal" vs. "average marginal" (insert effective or cumulative, etc., for average if desired). But to those thinking the IRS is giving a sweet deal on traditional withdrawals, a more forceful distinction is warranted.
Reasons why this is good the way it is:
- This is the wording we agreed to in a previous post.
- The "marginal" and "cumulative" terms exactly match what's in the personal finance toolbox.
- I scrupulously avoided the use of the term "average" to avoid confusion.
- We mention that exact issue in the "common misconceptions" section.
- WCI doesn't use the term "cumulative" in his post.
- By the time someone is this far down in a long and technical article, the chances of them being confused by this phrasing are even lower.
- If we deviate from this phrasing, then it raises the question of what call the "actual" marginal rate if we are using "marginal" for cumulative. I've never gotten a good answer to this question.
We definitely should try to be clear to our readers, but changing definitions around from what are most commonly accepted is a bad way to do this - it creates more confusion than it fixes. The way it's written now is clear and consistent, both internally and with external sources.