[Editor's Note 2/3/2020: This post gets more traffic than any other asset location post on the blog. But it is far from my best post on the subject. In fact, I'd recommend you skip this one entirely and read either the follow-up post:
My Two Asset Location Pet Peeves
or the more comprehensive:
Six Principles of Asset Location, my best post on the subject. But if you want to read this one, knock yourself out.]
I've written about this subject before, but it has been over a year, and based on questions I'm seeing in the comments section and my email box, and recommendations I'm seeing in online forums, people aren't getting it. Many investing authorities over the years have recommended that if you have to use a taxable account, that you preferentially put very tax-efficient asset classes, such as stock index funds into it, while preferentially putting tax-inefficient asset classes, particularly bonds, into your tax-protected accounts, like 401Ks and Roth IRAs. This seems like a nice simple way to conceptualize the subject of asset location. However, as Einstein famously said, “Make things as simple as possible, but not simpler.” It turns out that basing your decision simply on the tax-efficiency of the asset class is making things “simpler.” You also have to take rate of return into consideration, and for bonds, that varies highly with changes in interest rates.
Remember The Value Of A Tax-Protected Account
To understand this, it is important to first be able to conceptualize the value of a tax-protected account. The value of a Roth (AKA after-tax, AKA tax-free) account is simply the value of the tax-free growth, i.e. how much more money you'd have in the Roth account than you would have if you had put the investment into a taxable account. Obviously, there are also some estate planning and asset protection benefits and perhaps some additional fees, but let's ignore all that for the moment. The value of a 401K (AKA tax-deferred) account is a little more complicated. First, you recognize that the money in a 401K is only partly yours. Part of it is the governments. If your marginal tax rate at contribution is 33%, only 2/3 of the 401K is yours. On that 2/3, the benefit is the same as the Roth. There is also the possibility of a tax arbitrage. If you contribute at a 33% marginal tax rate, but withdraw the money at a 17% effective tax rate, you get an additional significant benefit out of the 401K. Obviously, the arbitrage CAN be negative, but that's pretty unlikely for most doctors, even with a rise in marginal tax rates. More details on calculating the benefit of retirement accounts can be found in my recent post on the value of a 401K.
The important concept to understand here, of course, is that assets grow faster in the retirement account due to the lack of tax drag. So you will benefit from having a fast-growing asset (i.e. stocks) inside the retirement account. It expands your tax-protected space, increasing your tax benefit, estate planning benefit, and asset protection benefit as the years go by. You have to weigh that benefit against the additional tax drag of having a tax-inefficient asset class in the taxable account. At our current low interest rates, the first benefit dramatically outweighs the second. Let's look at two examples to understand this.
Why Bonds Go In Taxable
[Update: Post updated 2/11/2014 with some minor changes to the calculations to make them more accurate, based on criticism found in the comments section below. These changes made the numbers less impressive, but did not change the conclusions.]
Make a few reasonable assumptions, and it becomes easy to see why bonds belong in taxable. Let's assume a $200K portfolio, split half into a Roth IRA and half into a taxable account. Our physician investor has a 33% marginal tax rate, a 15% long-term capital gains/dividends rate, an 8% return for stocks of which 1.86% (the actual yield on the Vanguard Total Stock Market Fund on 2/11/2014) comes from qualified dividends/long-term capital gains, and a 2.16% return for municipal bonds and 2.69% for high quality taxable bonds (actual yields from appropriate Vanguard bond funds on 2/11/2014.) We'll also assume no rebalancing to keep things simple, but that won't distort the direction of results, only the magnitude. (Run the examples using just 1 year if you're not convinced of this.) Note that this physician will use muni bonds when bonds are in taxable, and taxable bonds when bonds are in the Roth since 2.69*(1-33%) < 2.16%.
If you put the bonds in the Roth, you get this:
Roth IRA
$100K Bonds grows at 2.69% for 30 years to $221,740 =FV(2.69%,30,,-100000,1)=221,740
Taxable
$100K Stocks grows at 8% -(15% * the 1.86% yield) = 7.72% to $930,873 over 30 years. =FV(7.72%,30,,-100000,1)
You don't pay capital gains on the original $100K, nor on the $183,177 in dividends received. So capital gains taxes on the $647,696 in gains are $97,154, leaving you with $930,873-$97,154 = $833,718.
Total = $1,055,459
Now, put the stocks in the Roth, and you'll get this.
Roth IRA
$100K Stocks grows at 8% for 30 years to $1,006,266
Taxable
$100K Municipal Bonds grows at 2.16% for 30 years to $189,857
Total = $1,196,123
You get $140,664 or 13% MORE by putting the stocks into the Roth.
Even Under Old Assumptions Bonds Should Be in Taxable
Now, what really surprised me, was when I ran the numbers using what someone could have assumed just a few years ago. It used to be that you could expect a return of about 8% from stocks, about 5% from taxable bonds, perhaps 4% from municipal bonds, and a 2% stock market yield. Capital gains and qualified dividends were taxed at 15%. Under those assumptions, you might think that it would be best to put taxable bonds into your retirement accounts, especially for a high earner with a 33% tax bracket. Surely at those higher bond rates of return and lower tax rate on stocks, we should put bonds into the Roth, no? Let's take a look.
First, bonds in the Roth
Roth IRA
$100K Taxable Bonds grows at 5% to $432,194 over 30 years
Taxable
$100K Stocks grows at 8% -(15% * the 2% yield) = 7.7% to $925,702 over 30 years. You then pay capital gains taxes on $629,367 ($94,405) and are left with a total of $831,297.
Total = $1,263,491
Then, we'll try bonds in taxable, (and given the 33% bracket, we'll use the muni bonds at 4%, although using taxable bonds with an after-tax return of 3.35% doesn't change the direction of results, only the magnitude)
Roth IRA
$100K Stocks grows at 8% to $1,006,266 over 30 years
Taxable
$100K Municipal Bonds grows at 4% to $324,340 over 30 years
Total = $1,330,606
Bonds in taxable STILL leaves you with $67,115, or 5% more. Now, I'm sure if we try hard enough we can come up with a set of assumptions that will favor putting bonds in tax-protected (it will likely involve a great deal of tax-loss harvesting and donation of shares or getting the step-up in basis at death), but under any reasonable assumptions in our current environment, it's pretty hard to justify that advice.
[Update 2/11/14: Using current yields, even if you assume that you DIE without ever selling any of the appreciated stock fund shares in the taxable account (meaning you get a complete step-up in basis at death), stocks in Roth still comes out ahead by $43,510 (4% more). If you use the “old assumptions,” and have a complete step-up in basis at death, stocks in taxable finally wins, but only by $27,290, or 2%. This demonstrates the importance of running these numbers yourself using current yields and other assumptions specific to your situation when making asset location decisions, rather than blindly following a rule of thumb, unless having 13% more money to spend doesn't matter to you.]
Keep in mind that this exercise is already slanted in favor of the bonds in the tax-protected account just by virtue of the fact that we're assuming we're pulling all that Roth money out exactly after we finish contributing it, when in reality, it will likely be withdrawn over the next 15-30 years by the retiree, and perhaps another 20 by the retiree's heir.
Too many investors, including many very knowledgeable investors and advisors, have been giving portfolio construction advice that is too simple, and costing those taking it real money. Even the Bogleheads just recently revised their Wiki on this subject to show that stocks in taxable isn't always right. Put your bonds in taxable and let proper asset location give your portfolio a boost.
What do you think? How did the dogma of “stocks in taxable” get started and why does it persist? Comment below!
Leave all your Roth investments in stocks as you probably will touch them last or probably pass them down
And open Roths for your kids to make them instant millionaires at retirement
Kind of defeats the purpose of tax diversification to never touch your Roths, no? Also, it’s possible your kids will be in a lower tax bracket than yours. Why not let them pay the tax on an tax-deferred inheritance? Don’t get me wrong, I’d rather inherit a Roth, but any inheritance is nice.
Would the conclusion change if you controlled for “Asset Allocation Drift”?
In your “Bonds in Roth” scenario, Bonds fall from 50% to 19% of the portfolio over the 30 years.
But in your “Stocks in Roth” scenario, Bonds fall from 50% to 16% of the portfolio over the 30 years.
We know that stocks beat bonds (given your assumptions), so it would follow that allocating more of the portfolio to stocks will win!
To what extent is an “Asset Allocation Effect” crowding out (or masking or skewing) an “Asset Location Effect”?
Hmmmm….an excellent point. I don’t think the 3% difference is enough to really make much of a difference though.
I looked at the asset allocation drift in your numbers again, but this time after taxes. “Bonds in Roth” ends up with 21% in bonds, while “Stocks in Roth” ends up with only 16% in bonds. This is a 5% difference in asset allocation (higher than 3% but still not massive).
Your overall point is excellent, which is that asset location is a function of BOTH expected returns and tax efficiency. That said, there may be bigger fish to fry in asset location than where to place bonds…
One wrinkle I’ve been researching is the FTC (foreign tax credit) and whether it is large enough to prioritize placing international index funds in taxable…because if you leave them in tax deferred you never get the credit! Turns out Emerging Markets are an interesting case because they are tax inefficient BUT offer a large FTC (available only in taxable).
For example, REIT’s are a total no brainer to go in tax deferred (high returns, low tax efficiency).
The approach I’ve read and like is to START with your no brainer funds – locate them first where they should go, and then “fill in” with the rest.
Excellent approach.
Based on this post and others, it seems like the outcome of fussing with asset location can be positive or negative, and depends a lot on assumptions about the future that are difficult or impossible to predict — future interest rates, future tax rates, future income, etc. Moreover, simple analyses can’t factor in possible mitigation strategies such as Roth conversion in a low-earning year and tax diversification in retirement. Trying to sort through the various scenarios makes my head hurt.
Thus, as a young investor starting my professional life (along with my wife who’s a new minted EM attending), I feel like the best course of action may just be to maintain virtually equivalent asset allocations across tax-deferred, Roth, and taxable accounts, with the exception of using muni bonds instead of treasuries in taxable and carving out tax-advantaged space for the slices of extremely inefficient assets (e.g., REITs).
This has the benefit of simplicity — especially in the accumulation phase where contributions are significant relative to total assets and the various accounts are growing at different rates — as well as flexibility in instances where alternative investment opportunities arise and there is a high premium on liquidity.
Thanks for the great blog! Cheers.
I think your conclusion is excellent – simple and practical.
Some of us are drawn to the asset location puzzle because we like puzzles, not because it ends up being a big driver of returns.
Emerging Markets (Vanugaurd’s VWO) is an interesting “puzzle” case. It is pretty tax inefficient (so put it in tax deferred) but it offers a large foreign tax credit (which is only available in taxable). So round and round we could go debating this an other fine points of asset location.
More broadly, if you are an index investor only, asset location is not a “big idea.” If you have a “vegas” account, on the other hand, obviously you would do that in tax deferred, and so on.
Excellent points and a reasonable solution that is highly likely to minimize your future regret.
Excellent post. I plan to do the same. Munis in taxable account if I have to.
For a taxable account, would it make more sense to use the Intermediate Term Tax Exempt Bond Fund (Expense ratio 0.20%) or the lower cost Total Bond Market Index Fund (Expense ratio 0.07%). Would the higher expense ratio of the tax exempt fund potentially outweigh any taxable advantage?
Probably not. Both of those fees are very low.
For your bond allocation, if you are using international bonds for diversification, consider putting the international bonds in taxable accounts first before US bonds – they typically pay less dividends than US bonds. This will make your bond allocation in taxable even more tax efficient.
I’m not sure that’s a better move for most high income professionals than putting muni bonds in taxable.
First I would like to thank you for your blog and the book. I had no idea about Backdoor Roth until I read your book. You have already saved me a ton of money! Thank you for doing a great service!
Question:
Vanguard Intermediate-Term Treasury Fund (with an SEC yield of 1.78%) should have a tax cost of approximately 0.45% per year (i.e., 25% of the 1.78 yield). But if you go to the fund’s website (https://personal.vanguard.com/us/funds/snapshot?FundId=0035&FundIntExt=INT#tab=1), or Morning Star (http://performance.morningstar.com/fund/tax-analysis.action?t=VFITX®ion=usa&culture=en-US), it appears that the Tax cost is much higher (Tax cost ~ 1%). Vanguard 500 Index Fund Investor Class
MUTF: VFINX has a tax cost ratio of only ~0.6%. Will that change your calculation? Should I just go by the Morning Star tax ratio and put higher cost funds in the tax deferred accounts until I run out of space, and but the remaining funds in the taxable account?
What do you think about Vanguard Balanced Index Fund Admiral Shares VBIAX in the taxable? It has a tax cost ratio of only ~0.8%.
I am in 33% tax bracket.
You can learn more about efficient tax location here:
https://www.whitecoatinvestor.com/my-two-asset-location-pet-peeves/
Bottom line: Tax location is not just about tax-efficiency, but also about the return of the investment. You’re only looking at tax-efficiency. But you’e right that the last tax efficient of those three funds is the bond fund, then the balanced fund, then the equity fund.
You can learn more about efficient tax location here:
https://www.whitecoatinvestor.com/my-two-asset-location-pet-peeves/
Bottom line: Tax location is not just about tax-efficiency, but also about the return of the investment. You’re only looking at tax-efficiency. But you’e right that the least tax efficient of those three funds is the bond fund, then the balanced fund, then the equity fund.
Thanks!. That was very helpful.
First, I love you post and the discussion it generated. I understand the simplicity of the using ROTH as a proxy for all tax-advantaged accounts. Your rule of thumb that a fraction (33%) of a tax-deferred account actually belongs to the government–and the rest is a ROTH–is a good one, but I think it misses a VERY important point here.
In the situation you outline of dramatically changing the traditional advice and putting stocks in tax-advantaged you will likely also dramatically increase the investor’s liability to a higher tax-bracket in the future. Tax-brackets in retirement are based on social security and any other income, but they are mostly–especially for the physician–based on the actual size of your tax-deferred account.
For example, if the investor wants to draw 100k a year, a dramatically larger amount of that would need to come from the 8% stocks vs the 2% bonds, wherever they are. If the stocks are in taxable the percentage the government actually ends up owning in the 401k is likely a very different number. Yes, you improve the total portfolio by 13%, but the government might easily own 30% of your 401k vs 15% had you gone with bonds (your tax bracket at distribution).
For every additional dollar the government gets, you get two. A larger tax-protected account is not a bad thing. It’s all about having the most money overall, so you have to run the numbers with your assumptions and see what drops out.
Sure, larger is always good but I think this analysis is ignoring a significant distinct disadvantage of having that large-ness be in a tax-advantaged account–that will give the government a higher share of your accounts as your tax-brackets are bumped up.
I think you’re underestimating the benefits of tax-deferral for 30-60 years and overestimating the effect you’re concerned about. Run the numbers out and you’ll see what I mean. The likelihood of doing this moving your marginal tax rate from 30% to 15% is pretty much nil and if all it is doing is dropping you from 28% to 25% or something similar isn’t enough to make up for the benefit.
Thanks for your thoughts on this. Just ran the numbers for someone with a social security payout of 34,688 (based on 150k salary) and your 100k stocks and bonds each (which accrues to about 1M) to cover the rest of their expenses with a Safe Withdrawal Rate of 3%. The example with stocks in deferred bumps up their taxable income from 36k a year to 60k (15% to 25% tax brackets) and total after-tax income in retirement is reduced from 65.1k to 61.3k when they follow the plan of bonds in taxable.
Maybe the physician saved more, so I also doubled that net worth to ~2M and re-ran (200k starting balances in each). In this situation taxable income was increased from 43k to 89k (and tax bracket again from 15% to 25%). Here total after tax income in retirement is reduced from 99k to 90.3k when they follow the plan of bonds in taxable.
I’m sure some sort of accurate retirement modelling software could improve my estimate. The takeway though seems to be that increasing taxable income by 24k or 55k respectively (a large increase) while at the same time increasing tax bracket from 15% to 25% take home pay is reduced.
Did you adjust for the fact that you have a less aggressive after tax asset allocation putting stocks in the tax protected account in the accumulation phase?
I’d highly recommend spending some time with James Lange’s Retire Secure book to really “get” this concept.
Looks like I misread the cap gains reduction line above so my math was off! According to my numbers stocks in taxable will beat bonds by only 1-3% not 5-10% like above (close to a wash)— Still seems like the progressive nature of the tax structure will negate the effect described above in some (most?) situations.
With some assumptions stocks in taxable comes out ahead. You really need to run the numbers for your situation.
Agreed. In these situations the deferred account is about 5x larger which is what creates the potentially less favorable tax bracket.
I considered it but decided to simplify. Stocks in taxable came out ahead though so reducing equities in the deferred account with stocks would throw the numbers further in the direction of stocks in taxable
Not sure you’re getting what I’m saying. If 25% of the tax-deferred account belongs to the government, then a 50/50 pre-tax portfolio is really a 43/57 or a 57/43 portfolio. You have to compare equally risky post-tax asset allocations to each other.
Yeah, agreed. Decreasing the taxable accounts to start at 75k to to make an equitable comparison with a 100k TDA would further advantage stocks to taxable in this instance
Is there a rule of thumb for when you wouldn’t want bonds in taxable?
The BogleHeads wiki on tax-efficient fund placement states that:
Today low yields are common, and a bond fund with an expected return of less than 1% can be more tax efficient than a stock fund with an expected return of 7% even though the bond fund’s return is taxed at a higher rate.
Bond yields are higher than 1%, so I’m not sure how accurate the wiki is. Thoughts?
I wish there were a rule of thumb. It is super complicated. I’ve spent a lot of time thinking about it and haven’t been able to come up with one. The good news is that with low interest rates it won’t hurt you much if you get it wrong.
great post. what are your thoughts on VWITX or VTEBX (or the ETF version VTEB) for taxable allocation? thanks!
My thoughts are first that it is rude to make me look up the fund names for the ticker symbols (I don’t know more than a half dozen from memory and those aren’t on the list).
My second thought, after looking up VWITX (Vanguard intermediate tax exempt bond fund) is it is a great holding for a taxable account, probably my first choice.
VTEBX is VAnguard tax-exempt bond index fund, also a fine holding. Very subtle differences between the two, which probably don’t matter. The holdings are almost identical. I guess maybe VTEBX would be a better choice since it has more securities in it.
What do you think about Vanguard California Intermediate-Term Tax-Exempt Fund (VCAIX) instead of VWITX (Vanguard intermediate tax exempt bond fund) for California residents?
I think it’s a good choice
Hi Dr. Dahle,
I’m trying to learn about asset allocation by reading old posts. Would you be so kind to explain a bit more simply the justification for the following: “Note that this physician will use muni bonds when bonds are in taxable, and taxable bonds when bonds are in the Roth since 2.69*(1-33%) < 2.16%." I had a hard time following the rationale for this.
Muni bonds are not taxed federally and sometimes by the state. So the yield on a taxable bond must be adjusted to make an apples to apples comparison.
For the typical physician, if they choose to hold bonds in taxable, they should use a municipal bond fund for its higher after-tax yield. If they choose to hold bonds in a tax protected account, they should use a taxable bond fund for its higher yield.
Hope that helps.
Thank you, I think it does. If a taxable fund is held in something like a Roth, that means that there is no actual tax generated from the fund, even though its named a “taxable” fund?
That’s correct.
Does this same thinking still apply to cash balance plans configured with a fixed target rate of return (say 4%)? In theory at least, they’re a bit of a different beast from 401(k)s and Roths, as follows.
If the rate of return is too low (eg if doing a Swenson-style allocation with 100% 10y bonds+TIPS that return ~3% these days), then I need to make up the difference by contributing more money. OTOH, my understanding is that this extra “penalty” money then becomes tax-deferred, effectively expanding my CBP contribution space for that year and meaning that my “mistake” would actually save me 32% in taxes or whatever my marginal bracket is. Since I want to FIRE and will make enough to have lots of otherwise-taxable money sitting around anyway, that’s not necessarily a bad thing.
OTOH, my actuary tells me that if the rate of return is too high (eg all stocks with 8% returns over a year or two), the IRS can levy a 50% excise tax on all CBP returns above 4%. Not ideal, but I wonder if this ever actually happens.
The math would seem to get a bit complex here. Can’t find any good info on CBP allocation on the Net that approaches this topic from both sides (undershooting vs overshooting).
As interest rates continue to increase will bonds once again need to be placed in sheltered accounts?
Higher interest rates favor bonds in tax-sheltered accounts, yes. Where that crossover point occurs is different for everyone and difficult to calculate given that so many of the variables are unknown and unknowable.
Can you please tell me how you arrived at $183,177 in dividends received. This is located in the “Why Bonds Go in Taxable Section”, First “Taxable” Section? Which Excel formula is used here? Thank you for your time and writing this excellent post!
I have the exact same question.
I found a spreadsheet at https://docs.google.com/spreadsheets/d/1gc2Br4ntVZqROGP0OCGuE7pp8eXtxFUDRh5p5RgrjUM/edit#gid=14
You will have to save the spreadsheet to your computer to run the numbers.
I entered the following numbers:
Step 1: Input expected initial investment amount: $100,000
Step 2: Input expected ongoing investment amount:
$0 per month
and/or… $0 per quarter
and/or… $0 per year
Step 3: Input expected annual stock price appreciation: 7.7% (Enter as a %) [I used 8% -(15% * the 1.86% yield) = 7.72%]
Step 3a: Input Beginning Price: $1.00 (Note that subsequent prices are NOT adjusted for possible splits.)
Step 4: Input initial dividend yield: 1.86 %
Step 5: Input expected annual dividend increase percentage: 7.72 % [same as stock price appreciation rate]
I got $ 289,443.45 as dividends received.
Looks like there might be a typo in WCI’s calculations.
He is probably using expected annual dividend increase percentage as 5.64%, rather than 7.72 %. Or in other words, a declining dividend rate (as percentage of stock price) with time
Also, this qualified dividend income is taxed at the capital gains tax rate anyway (you just pay it in taxes every year, rather than at the end of 30 years). So probably using it to change the basis is superfluous. I think we can just ignore the change in basis due to dividends and use the basis as the initial $100,000.
Or if you want to be more accurate, use Input initial dividend yield as 1.86 % -(15% * the 1.86% yield) = 1.581% in the above spreadsheet to account for 15% capital gain; and that will give you $232,548.76 in dividends which you can use to adjust the basis.
Hey John! Thanks for your response and the insanely comprehensive “Dividend Reinvestment and Growth Spreadsheet.” I agree with the majority of your calculations. I tweaked a few things, and this is what I came up with.
Step 1: Input expected initial investment amount: $100,000
Step 2: Input expected ongoing investment amount: $0 per month
and/or… $0 per quarter
and/or… $0 per year
Step 3: Input expected annual stock price appreciation: 7.72% (Enter as a %)
Step 3a: Input Beginning Price: $1.00 (Note that subsequent prices are NOT adjusted for possible splits.)
Step 4: Input initial dividend yield: 1.58 %
Step 5: Input expected annual dividend increase percentage: 6.63 %
Year 30 = $183,305.89 in Cumulative Dividends
As you stated in your post, I think we do need to account for the taxes that we pay on the dividends each year. They are a drag on our performance. I deducted for the taxes going forward and adjusted the Expected Annual Dividend Increase Percentage to 6.63%.
I’m hoping that WCI will respond so that we can check our answers!
Thanks again for that fantastic spreadsheet! Someone invested some time in that!
Guys, it’s been 5 years since I wrote this post. If I’m going to go back and redo the calculations, I’m going to redo the entire post.
The main thing that’s changed in the last five years is interest rates have gone up. The effect of that is that now the vast majority of people are going to find that bonds preferentially go in tax-protected whereas with the very low rates we had a few years ago it sometimes made sense to put bonds in taxable.
As of today
the 30 day SEC yields are
BND 3.21%, (Vanguard total domestic taxable bond)
VTI yield 2.13% (Vanguard Total Stock Market ETF)
Municipal bonds 2.37% (Vanguard Intermediate-Term Tax-Exempt Bond )
Assuming 2.87% annual rise in PRICE of the stocks, and 15% capital gains tax bracket, with dividend reinvestment (after 15% tax in taxable) for stocks I got the following numbers.
Bonds in Roth, Stocks in Taxable: Post tax Returns = 3.87%
Bonds in Taxable, Stocks in Roth: Post Tax Returns = 3.92%
So it appears that even with today’s interest rates, Bonds in taxable still appear to come out ahead as long as stock PRICE (excluding dividend) is going up by 3 % or more
That’s a pretty low assumption for stock market returns. Sounds like you’ve been reading Bernstein or something. 🙂
Quoting: “That’s a pretty low assumption for stock market returns.”
Exactly!
That’s why I think Bonds should still go in taxable. As even in one of the worst case scenario (only ~3% long term annual stock price rise), we will have better returns by putting bonds in taxable.
If I use 5.87% annual rise in PRICE of the stocks, the difference becomes much more stark.
Bonds in Roth, Stocks in Taxable: Post tax Returns = 5.70%
Bonds in Taxable, Stocks in Roth: Post Tax Returns = 6.14%
I hope that you write an updated post as I am very interested in your calculations as I have been unable to find a way to agree with your conclusion in the last post.
Quoting “The main thing that’s changed in the last five years is interest rates have gone up. The effect of that is that now the vast majority of people are going to find that bonds preferentially go in tax-protected…”
This post has already been updated once (or at least the subject has been covered on the blog since this post was written), as found here:
https://www.whitecoatinvestor.com/my-two-asset-location-pet-peeves/
There is a TON that goes into this and many of the assumptions are complete guesses- for example when you die and what ordinary and capital gain/dividend tax rates will be in a decade or two. Other things that affect it are tax loss harvesting, how much you give to charity etc.
It’s very hard to come to any sort of definitive conclusion that will apply to even a broad segment of my readers on this topic. Combine that with the fact that is makes less of a difference than so many other things in investing, it’s hard to get super excited about it. It’s really a pretty niche little segment of investing.
WCI,
I believe there is a miscommunication. John and I have the same question, “How did you arrive at $183,177 in dividends received in the Why Bonds Go in Taxable Section.” Things got off track a little bit because we have been trying to justify that number.
Interests rates have gone up over the years, and that’s fine. I’m not asking you to prove your claim with current rates. However, it would be nice for an updated post because Asset Location is vital to maximizing after-tax returns.
My only question is, “How did you come up with $183,177 in dividends using your data?” I’m referring to the 1.86% dividend on VTI, 1.58% Tax Adjusted.
You shared your Excel calculations for everything in the example except for the $183,177 in dividends. This is a complicated calculation because there are dividend growth rates etc. I was hoping that you could share the Excel calculation that you used or if you use an online calculator etc. to calculate the dividends.
Thanks!
You understand I have no memory of writing this post nor some excel spreadsheet hidden away somewhere right? It’s been 5 years. But it looks to me like the dividends come from the yield of 1.86% over 30 years. So in year 1, that’s $1,860. Each year after that it goes up. At the end of 30 years, it’s $183,177.
If you don’t think that’s accurate, what do you think the number should be?
1) Actually, looks like WCI did in fact consider the capital gain in his calculations.
In the formula $100K Stocks grows at 8% -(15% * the 1.86% yield) = 7.72% , he is reducing dividend yield by 15% annually
2) Also in the spreadsheet STEP 3 is the annual increase in PRICE (not growth, so subtract dividend from 8% growth)
Now the numbers become
Step 1: Input expected initial investment amount: $100,000
Step 2: Input expected ongoing investment amount:
$0 per month
and/or… $0 per quarter
and/or… $0 per year
Step 3: Input expected annual stock price appreciation: 6.14% (Enter as a %) [I used 8% – 1.86% yield = 6.14%]
Step 3a: Input Beginning Price: $1.00 (Note that subsequent prices are NOT adjusted for possible splits.)
Step 4: Input initial dividend yield: 1.58% ( 1.86% -(15% * the 1.86% yield))
Step 5: Input expected annual dividend increase percentage: 6.14 % [same as stock price appreciation rate]
I got $ 173,118.75 as dividends received.
Question.
Why did you adjust the Expected Annual Dividend Increase Percentage to 6.63% in your calculations?
Bingo John, I believe you nailed it on the inputs for the calculation. I like the spreadsheet that you are using because it shows the cash flows. I have tried other dividend reinvestment calculators and everyone one of them gives me a different result. Many variables could be different. Some might calculate dividends monthly and others annually. These differences get exacerbated when spread out over 30 years.
I used your inputs on this calculator https://buyupside.com/calculators/dividendreinvestmentdec07.htm
Initial Number of Shares: 1
Initial Stock Price per Share ($): 100000
Annual Dividend Paid ($): 1580
Dividend Annual Growth Rate (%): 6.14
Stock Price Annual Growth Rate (%): 6.14
Number of Years: 30
Results With Dividend Reinvestment
Total Value = $956,354.16
Number Shares = 1.60
Dividends Paid = $183,716.77
Annualized Return (%) = 7.82
This is very close to the result in the post. We know the truth is between $173,000 and $183,000. That would definitely not be satisfactory for an accountant, but it’s probably okay for running hypothetical scenarios 30 years out.
I haven’t run the calculations myself, but as it appears there has been a shift in the “bonds go in taxable” to “bonds go in tax deferred” within the last 5 years I assume there must not be a huge difference.
Right now I don’t have a taxable account (relatively new on the job), but I’ve been maxing out my Roth IRA and my fiance’s Roth IRA, my 401k, my HSA. I currently have each account set to my preferred asset allocation. Since I dump $ into the Roth IRAs all at once and the HSA/401k both have bi-weekly contributions that get automatically invested per my AA, I find this method to work well keeping my overall AA fairly consistent over the duration of the year. If I were to put ALL bonds in my 401k for instance, and put $12k into Roth IRA for the fiance and myself at the beginning of the year then my AA would be a little stock-heavy for my preference until my 401k is maxed at the end of the year.
I’ll probably start a taxable account this year however, and my contributions will likely not be a consistent amount every month. Do you think there is any downside to investing with the same AA in my taxable bucket? Or is it better to just put all bonds in the tax-deferred space.
Also, if I put all bonds in tax-deferred space, is it better to put that in my 401k as opposed to my Roth(s) and HSA since I’ll never have to pay taxes on those accounts again?
The main change is simply that rates went up in the last two years so bonds in tax-protected is now right for even more people than previously. But the difference is often not large.
I think this post will help and addresses all of your questions:
https://www.whitecoatinvestor.com/my-two-asset-location-pet-peeves/
You are merely demonstrating how much higher expected returns are if you increase your exposure to risky assets. A tax-advantaged account with identical assets to a taxable account will have a higher after-tax beta for an investor. In effect, you are demonstrating that levering up increases expected returns, and ignoring the fact that the riskiness of the returns is also magnified.
I decided to use your numbers and run some simulations (I did 1000 simulations on Excel). I assumed a 2.5% risk-free rate and a 5.5% equity risk premium (so that stocks return 8% in expectation). That results in an implied debt beta of 0.035 to get an expected 2.69% return for bonds that you used. I used a 20% annual standard deviation for the stock portfolio (the historical average for stocks). Then I computed the value of the two possible portfolios.
I followed the same procedures for the taxable accounts and assumed a 1.86% dividend taxed annually at 15% and the interest on the bonds was taxed at 33% annually. The results of 1000 tests were as follows: White Coat Method returned an effective 6.17% (versus 6.14% that you calculated in expectation). Traditional Method returned 5.6% (versus 5.7% that you calculated in expectation). In other words, my simulation demonstrated the EXPECTED returns you calculated are valid with the given parameters, and the White Coat Method indeed has a higher expected return.
HOWEVER, the riskiness of the two portfolios were also quite different. I took the standard deviation of the after-tax portfolio values and scaled them down by dividing by the initial investment and 100%. Then I took a ratio by dividing the expected return by this measure of risk. The results were approximately 1 unit of return per unit of risk in the traditional approach. However, the White Coat Method only got 0.84 units of return per unit of risk. In other words, the Sharpe ratio for an individual investor is vastly inferior when putting the riskier asset in the tax-advantaged account, despite the expected returns being half a percent higher.
Believe it or not, using the method you are suggesting, an investor is taking on about 25% more risk to achieve about 0.5% more yield. This is not a good deal. Investors are much better off putting bonds in tax-advantaged accounts and stocks in taxable accounts.
At a yield of 0%, that is not the case for anyone. At higher yields, eventually that will become true for every one. That’s the point of the article.
Thank you for your simple, and yet elegant reply. In three short sentences, you answered five paragraphs of rambling, and I think I do understand now what you are talking about. Your point about bond yields approaching 0% is well taken, and I hope you also appreciate the nuance of discount rate for the tax shield.
Here is the way I look at the decision:
Step 1 compute the future value of the tax shields:
Tax shield 1A)
Roth Bond FV – Taxable Bond FV = 221,740 – 189,857 = $31, 883
Tax shield 1B)
Roth Stock FV – Taxable Stock FV = 1,006,266 – 833,718 = $172,548
Step 2 discount the tax shields to the present value:
Present value 1A)
NPV(2.69%, 30 years, , $31,833, end) = $14,379
Present value 1B)
NPV(8%, 30 years, , $172,548, end) = $17,147
The difference FV 1B – FV 1A = $140,664 is not as helpful as PV 1B- PV 1A for making a decision because it does not reflect the difference in riskiness between tax shields A and B. The tax benefits are strongly correlated with the investment returns, so they should be discounted at the same rates.
The benefit of putting stock in the Roth is PV 1B – PV 1A = $2,769 in present value or $27,862 if you prefer to compound those savings at 8% for 30 years.
Side Note: We could use 7.72% for stocks and 2.16% for bonds if we assumed most other investors are using taxable accounts. The difference is pretty small, but I believe the before tax expected return is generally a more accurate discount rate.
In Example 2, Roth Bond FV – Taxable Bond FV = $107,854 and Roth Stock FV – Taxable Stock FV = $174,969
In this case, the present values are
Present value 2A)
NPV(5%, 30 years, , $107,854, end) = $24,955
Present value 2B)
NPV(8%, 30 years, , $174,969, end) = $17,388
So the conclusion is the opposite, and your portfolio is worth $7,567 more by putting bonds in the Roth.
I admire your do-the-math-yourself mindset, and I’m sure you inspire many people to do the same.
Thanks for sharing your preferred method.
It’s far more complicated than the math you’re using suggests. For example, we donate all of our appreciated shares to charity and tax loss harvest the losses. So stocks in taxable is even better for us than it is for most people. Now that rates are rising, I think it might even be time to do another asset location post on the blog. It’s a pretty small-fry subject in investing, but it can make a small difference and with rates having risen, bonds in taxable makes sense for even fewer people.
Great post, thank you. I’m creating a spreadsheet to try to help wrap my head around it all. “You don’t pay capital gains on the original $100K, nor on the $183,177 in dividends received. ” Could you tell me how the $183,177 was calculated?
It’s been years since this post was written. If the work isn’t there displayed in the post, I’d have to start over with the calculations. Comments like this are the reason I try to show my work as much as possible now as I write, because it’s too much effort to explain it months or years later.
To me the real lesson of this math is that you shouldn’t be investing in bonds at all if you’re looking at a 30-year time horizon and are appropriately diversified. If you put $100k in both Roth and regular accounts and did both in stocks you would come out ahead with either scenario, right? Of course the point of investing in bonds is for diversification in a down year, but on a 30-year time horizon that’s not a big concern.
Unless bonds outperform stocks over your investment career (could happen, even if less likely) or if you panic in a downturn due to the volatility of your all stock portfolio and end up selling low.
Did you figure out the excel formula for calculating the total dividends received?
Not sure what you’re referring to.
Here is a link to a website justifying bonds in taxable accounts and stocks in tax protected accounts.
https://www.biglawinvestor.com/bonds-in-a-taxable-account/
As interest rates climb, it makes less and less sense to put them in taxable. Of course, some of us don’t have much choice as our taxable:tax protected ratios are high and growing.
Hi in 2 years I’m estimating to have
125k in my taxable account
150k in my roth ira /pretax 401k/rollover ira accounts
275k total
If I decide to use a 90/10 risk portfolio I should put 27.5k in bonds in my taxable and keep my roth ira and rollover iras 100% stocks?
I’m a 35yo traveling physical therapist. Like a traveling nurse. I’m currently able to stay in the 12% tax and 0% capital gains bracket utilizing my pretax 401k and because of the fact that the majority of my pay is in tax free stipends.
Also do bonds in taxable matter if my capital gains tax is 0%? My goal is to semi retire early and stay in the inflation adjusted below 58,575 income (including single standard deduction and roth conversion ladders when I semi retire or when I completely stop working). Assuming tax laws don’t change.
I’m student loan and completrly debt free and have a current savings rate of 50k/year+.
Thank you!
It’s complicated. As interest rates rise, it makes more and more sense to have bonds in a tax protected location. This article was written at a time of pretty low interest rates.
At any rate, at 90/10, it doesn’t matter much, but most these days in your situation would probably put the bonds in your 401(k) or rollover IRA if there’s a reasonable good offering there.