2017 Relative Tax Efficiency

[triceratop]

My birthday gift to the Bogleheads:

This is Pt. III of (2015) Relative tax efficiency including Foreign Tax Credit (VSS, VWO, VBR, etc.) and 2016 Relative Tax Efficiency, with numbers for the 2017 tax year. The google spreadsheet link is here.

Enjoy!

[livesoft]

Happy Birthday! Thx.

[triceratop]

Happy Birthday! Thx.

I didn't state it was my birthday (I also did not state it was not). It is, however, the Bogleheads forum's birthday.

[triceratop]

Thanks to Longtermgrowth for pointing out that DES underwent a 3:1 split in November. The figures have been adjusted to correctly account for this.

[spacecadet610]

This is great. thanks for doing this.

Wish all funds could be compared.

I am in a high tax bracket and it helps to know which funds to buy for tax efficiency.

My VSS in my taxable account has a tax efficiency of 0.94! Had no idea it was so tax inefficient. I may have to think about putting it in my tax deferred accounts or find a different fund.

[triceratop]

This is great. thanks for doing this.

Wish all funds could be compared.

I am in a high tax bracket and it helps to know which funds to buy for tax efficiency.

My VSS in my taxable account has a tax efficiency of 0.94! Had no idea it was so tax inefficient. I may have to think about putting it in my tax deferred accounts or find a different fund.

When you compare to other funds, do make sure you are comparing apples to apples. VSS is an interesting fund because it is a true "Total" International Small Cap ETF -- it has exposure to EM small caps. Many other funds do not, so it may be deceptive to conclude they are truly more tax efficient, since to compare on an equal basis you would need to complement those other funds with a distinct EM small cap fund.

[Ethelred]

Thanks for posting this, it's always helpful.

[Longtermgrowth]

This has become my favorite part of the New Year. Ball drop? Fireworks? Tax Efficiency thread? I'll gladly skip the first two, if I had to choose between them, and wait a bit for the Relative Tax Efficiency spreadsheet! Very informative and helpful

[fctu]

Does past tax efficiency predict future tax efficiency?

[triceratop]

Does past tax efficiency predict future tax efficiency?

I cannot prove that it does.

It is well discussed that as International stock yields fluctuate the relative efficiency to US stocks varies.

However, it was easy to see that IXUS was more efficient than VXUS in 2016. That saved me money on my 2017 taxes since I was able to switch funds.

[sal paradise]

triceratop - thanks for posting this info. can you enlighten me on focusing on the tax efficeincy over total cost? shouldn't i just look at total cost? for instance, i've been debating whether to keep vxus (vanguard total int) or voo (vanguard s and p) in taxable. in my case, vxus has better "tax efficienty" (0.03 vs 0.06) but a greater "total cost" (0.14 vs 0.10). minimal difference, but still a difference).

[triceratop]

triceratop - thanks for posting this info. can you enlighten me on focusing on the tax efficeincy over total cost? shouldn't i just look at total cost? for instance, i've been debating whether to keep vxus (vanguard total int) or voo (vanguard s and p) in taxable. in my case, vxus has better "tax efficienty" (0.03 vs 0.06) but a greater "total cost" (0.14 vs 0.10). minimal difference, but still a difference).

For evaluating fund placement between funds you know you want to hold: use the tax efficiency number not the total cost. The total cost is a number which I find interesting to know but should not be used to make decisions on placement between tax-advantaged and taxable.

[ruralavalon]

My birthday gift to the Bogleheads:

This is Pt. III of (2015) Relative tax efficiency including Foreign Tax Credit (VSS, VWO, VBR, etc.) and 2016 Relative Tax Efficiency, with numbers for the 2017 tax year. The google spreadsheet link is here.

Enjoy!

Thanks for posting.

This is always interesting.

[historyforsale]

Thanks for the info!

I think you made a mistake listing the expense ratio for FTIPX as 0.035. As far as I can tell, it should be 0.1.

[triceratop]

Thanks for the info!

I think you made a mistake listing the expense ratio for FTIPX as 0.035. As far as I can tell, it should be 0.1.

Corrected, thank you.

[mgensler]

Thank you for posting this. I am trying to compare AGG and FSITX. I can't seem to locate the QDI Ratio for either. Perhaps there isn't one? Without using QDI my calculations show AGG to be slightly more tax efficient in my tax bracket.

[triceratop]

Bond income is not qualified.

[jackal]

Any idea if the tax strategies of dimensional for funds such as dtmvx or dwusx add value above and beyond say ijs or vss? Looking at it, the long term tax cost impact of dwusx seems to be lower but not sure if I’m misreading it. I looked at the prospectus on the public portal and am guessing that included the benefits of foreign tax credits.

[aristotelian]

Is there a typo on IXUS? (.02)

[livesoft]

Is there a typo on IXUS? (.02)

The beauty of publishing the data and the spreadsheet for all to see and use is that it can be peer-reviewed, so that others can double-check it and point out any errors --- and their corrections.

[triceratop]

For my situation, IXUS actually had negative tax cost in 2016.

[triceratop]

Any idea if the tax strategies of dimensional for funds such as dtmvx or dwusx add value above and beyond say ijs or vss? Looking at it, the long term tax cost impact of dwusx seems to be lower but not sure if I’m misreading it. I looked at the prospectus on the public portal and am guessing that included the benefits of foreign tax credits.

Interesting. What data did you use and did you use this spreadsheet? Can you please share the data for dtmvx and dwusx with the rest of the class? Thanks!

[jackal]

Any idea if the tax strategies of dimensional for funds such as dtmvx or dwusx add value above and beyond say ijs or vss? Looking at it, the long term tax cost impact of dwusx seems to be lower but not sure if I’m misreading it. I looked at the prospectus on the public portal and am guessing that included the benefits of foreign tax credits.

Interesting. What data did you use and did you use this spreadsheet? Can you please share the data for dtmvx and dwusx with the rest of the class? Thanks!

Sure!
Looked up the prospectus for dwusx on https://us.dimensional.com.
From inception on 11/1/12: total return was 11.4. After taxes on distributions but before selling shares it was 10.71.

For Vss, from vanguard portal, for 5 years, the returns were 8.56 and 7.66.
So approximate tax cost was 0.7 vs 0.9.
Not sure if both included foreign tax credits or if they calculated tax impact differently.
Was this what you were asking?

[ivk5]

The google spreadsheet link is here.

@triceratop

Thank you so much for this- a great resource for the community. I've cloned and played with adding a few others for comparative purposes.

One erratum: looks to me like the 2017 version is still pulling 12/30/2016 and 12/31/2016 prices in the "# shares / $10,000" formulas...

[triceratop]

The google spreadsheet link is here.

@triceratop

Thank you so much for this- a great resource for the community. I've cloned and played with adding a few others for comparative purposes.

One erratum: looks to me like the 2017 version is still pulling 12/30/2016 and 12/31/2016 prices in the "# shares / $10,000" formulas...

I should hope so? Why would 12/31/2017 prices tell me anything about the cost to hold the security in 2017? No, not an erratum

[ofckrupke]

[I should hope so? Why would 12/31/2017 prices tell me anything about the cost to hold the security in 2017? No, not an erratum

Sorry, that's a little too glib. The movie is feasibly/tolerably viewable in reverse, there are no obviously irreversible processes.

Since we don't know until the end of the year how much was distributed during the year in what form including foreign taxes paid (and with Vanguard funds, you don't even know the ER until afterward), using BOY/EOPY share prices is just a convention that we forward projectors find more appealing.

Of course I too used this convention for similar calculations, before your spreadsheet became available.

[triceratop]

[I should hope so? Why would 12/31/2017 prices tell me anything about the cost to hold the security in 2017? No, not an erratum

Sorry, that's a little too glib. The movie is not really any different watched in reverse, there are no obviously irreversible processes.

Since we don't know until the end of the year how much was distributed during the year in what form including foreign taxes paid (and with Vanguard funds, you don't even know the ER until afterward), using BOY/EOPY share prices is just a convention that we forward projectors find more appealing.

Of course I was using the same convention as you for similar calculations, before your spreadsheet became available.

No, not too glib. The spreadsheet measures the tax costs of holding a security for a calendar year. In order to be exposed to the tax consequences for 2017, one must hold those assets during the year. If the price of an asset doubles on 12/31/2017 from 12/30/2017 then did my tax liability halve for year 2017? No.

[ofckrupke]

If the price of an asset doubles on 12/31/2017 from 12/30/2017 then did my tax liability halve for year 2017? No.

The tax liability doesn't depend on the asset price at either end of the year. Only this measure of the year's tax liability per dollar holding. Since the asset price varies throughout the year, some method must be chosen to determine a denominator. It doesn't have to be one end or the other; it could as well be some well behaved function of the asset price throughout the year.

[triceratop]

If the price of an asset doubles on 12/31/2017 from 12/30/2017 then did my tax liability halve for year 2017? No.

The tax liability doesn't depend on the asset price at either end of the year. Only this measure of the year's tax liability per dollar holding. Since the asset price varies throughout the year, some method must be chosen to determine a denominator. It doesn't have to be one end or the other; it could as well be some well behaved function of the asset price throughout the year.

Alright, you make a good point. I'll be happy to add flexibility for this very legitimate case when time travel is widespread / time reversibility is observed.

For now, I think it would be confusing to be anything but 12/31/2016 for year 2017. Yes, I apologize for my spreadsheet's limitations.

[ivk5]

Interesting discussion but I agree I was thinking about this the wrong way and stand corrected.

[livesoft]

I think that if I bought the shares on the last day of the year, then I would probably have no tax liability that year created by those shares.

[ofckrupke]

For now, I think it would be confusing to be anything but 12/31/2016 for year 2017. Yes, I apologize for my spreadsheet's limitations.

I forgot to point out that the marginal brackets aren't settled until the year is over too.

Of course I too prefer the end-previous-year pricing convention....because it's more in the spirit of making projections about an uncertain future.
Unless an alternate convention would produce significantly different rank orderings for like-type assets with some sort of predictability, I don't see a case being made for rejection of the EOPY convention.

But no change needs to be made to the spreadsheet: anyone can already install whatever prices they want - end year, mid-year, average-of-the-two, geometric mean, etc. Or to use acquisition price for a mid-year purchase. And if someone wants to separately tally a unit tax cost at each distribution using its reinvestment price as a denominator, they can easily implement that too...it's just arithmetic. So...about the spreadsheet apology: "nothing is forgiven".

Anyway, wrt time travel, if you knew what even one fund was going to close at on 12/31/18 you wouldn't really care about the tax drag cost, right? You'd make as many bets as you could find, against anyone willing to bet it would be anything else, higher or lower.

[triceratop]

I think that if I bought the shares on the last day of the year, then I would probably have no tax liability that year created by those shares.

You might very well think that, but what if time reversed after you bought the shares? Your thinking is stuck in the past where time only moves forward. Very simplistic thinking, livesoft, I am disappointed.

[livesoft]

Well, I have ridden my bike faster than the speed of light: I started at the top of a steep hill and just took off. I pedaled faster and faster in the top gear, but the real trick was I reached over to my left wrist where my analog watch was and pulled the stem out. Then I turned the minute hand backwards as fast as I could. So I got to the bottom of the hill before I left the top of the hill with hours to spare ... at least according to my watch.

[johan851]

Thanks, this sheet has been really helpful for planning recently, as I'm starting to invest significantly more in taxable this year.

So far I've been going off the "general advice", including the wiki, which indicates that large cap or total market should be the most efficient, followed by mid or small cap funds. A lot of opinions I've read on the forum point out the foreign tax credit being an advantage for putting international in taxable, with some even recommending prioritizing foreign stocks in taxable to take advantage of it.

I'm a bit surprised to find that in my tax bracket the situation is somewhat reversed. See my copy of the sheet below, where I've added small and mid blend ETFs from Vanguard and iShares.

https://docs.google.com/spreadsheets/d/ ... sp=sharing

First, foreign equities were less efficient than domestic for me in 2017. As mentioned earlier in this thread, IXUS was much more efficient than VXUS, despite virtually identical performance. But both were less efficient than domestic total market or large caps - by 10bp (IXUS) and 17bp (VXUS).

Also a surprise, VB, VO, and VXF (Vanguard's small, mid-cap, and extended market ETFs, respectively) were all more efficient than the total market (VTI) and large cap (VV, VOO) ETFs.

IJH and IJR, the iShares small and mid-cap ETFs, were the most efficient of the lot. Note that these use S&P indexes and can't be compared directly to their similarly-named Vanguard counterparts. I tilt my portfolio a bit towards small - prior to plugging the numbers in I would've avoided a small cap index in taxable, but using IJR in my Fidelity taxable brokerage account would have been a very efficient choice.

The resulting stack rank for me (and, I think, for many) is, from most to least tax efficient:

• Domestic small/mid cap
• Domestic large cap / total market
• International total market
• International small

Question 1
Did I make a mistake? I'm surprised by the results. 20% + 3.8 NIIT results in a 23.8% QDI tax rate, correct?

Question 2
Assuming I didn't screw it up, smaller cap funds were generally more tax efficient than large cap and total market funds. This seems to be the case so long as the QDI tax rate is 15%+. Why is that? Do these indexes tend to avoid the large companies that pay dividends? And if small cap funds are generally more efficient then why is the reverse often recommended?

Question 3
IJH and VB are both mid/small blends that have performed very similarly for about 20 years. There's clearly considerable overlap. But IJH was more efficient. Is it the index (S&P 400 vs CRSP Small) or is Blackrock doing something else to make this happen?

[triceratop]

Johan,

I have not checked your numbers. However, they do not seem unreasonable.

So far I've been going off the "general advice", including the wiki, which indicates that large cap or total market should be the most efficient, followed by mid or small cap funds. A lot of opinions I've read on the forum point out the foreign tax credit being an advantage for putting international in taxable, with some even recommending prioritizing foreign stocks in taxable to take advantage of it.

Indeed. As was noticed when this series of topics was started in 2015, common wisdom is often not so wise. Opinions are rarely informed by the data in this sheet, which I don't mind so much because the ranked ordering of the list may be volatile anyway and the principles in the advice given on the forum are generally correct, even if the result is not optimal.

I will note that for me, foreign is more tax efficient than US, but I recognize that that can change due to personal circumstance or the particular way capital is returned to investors. It is highly dependent on your personal situation, which is why I like this sheet so much.

As for your other questions, I don't have too much to say. The answer to (1) is "That is the correct rate", (2-3) "You should be able to see the source of the higher efficiency whether that be lower dividend yield, higher percentage of qualified dividend.

[johan851]

I've seen comments about mid cap indexes being inefficient due to companies' valuations dropping into small cap space or rising into large cap space. Is that sort of "index churn" reflected in the short or long term gains? Are the generic "distributions" only due to dividend payouts?

I like this sheet too. I'll take data over speculation anytime.

[triceratop]

I've seen comments about mid cap indexes being inefficient due to companies' valuations dropping into small cap space or rising into large cap space. Is that sort of "index churn" reflected in the short or long term gains? Are the generic "distributions" only due to dividend payouts?

I like this sheet too. I'll take data over speculation anytime.

Vanguard ETF (and mutual fund for that matter, read on for why) capital gains payouts are exceedingly rare. Any distribution marked as "dividend" is exclusively income and not related to the need to sell securities when they move out of the index. Vanguard can avoid most capital gains distributions using its tax management techniques (and especially thanks to the dual ETF share classes that many of their funds have).

[StrangePenguin]

This is a really great tool. I'm really intrigued to see the tax efficiency number for Small Cap Value (VBR) being so good compared to e.g. S&P 500 VOO.

Is it possible that the numbers for VBR are not updated for 2017? I do not have the expertise to check the QDI number but the dividend / share numbers look like maybe they are not updated for 2017, checking against: https://personal.vanguard.com/us/funds/ ... true#tab=4

The net result doesn't change much though, assuming QDI ratio remains the same.

[triceratop]

This is a really great tool. I'm really intrigued to see the tax efficiency number for Small Cap Value (VBR) being so good compared to e.g. S&P 500 VOO.

Is it possible that the numbers for VBR are not updated for 2017? I do not have the expertise to check the QDI number but the dividend / share numbers look like maybe they are not updated for 2017, checking against: https://personal.vanguard.com/us/funds/ ... true#tab=4

The net result doesn't change much though, assuming QDI ratio remains the same.

Looks fine to me now.

That is, I just fixed it

[JustinR]

My birthday gift to the Bogleheads:

This is Pt. III of (2015) Relative tax efficiency including Foreign Tax Credit (VSS, VWO, VBR, etc.) and 2016 Relative Tax Efficiency, with numbers for the 2017 tax year. The google spreadsheet link is here.

Enjoy!

Thank you triceratop.

How do you calculate your QDI Tax Rate? I'm in California. I'm seeing a lot of 18.8%, should I just go with that?.

And just so, I'm sure:
- Federal Marginal Tax Rate is your highest tax bracket right? 12%, 22%, 24%, 32%, etc.
- State Tax Rate is your highest tax bracket for your state right? So in CA it would be 8%, 9.3%, etc.

[triceratop]

QDI tax rate is typically a federal thing, at least as far as the capabilities of my spreadsheet. There are, to my knowledge, a handful of states which give this treatment but I don’t handle that. California is not one such state.

Marginal tax rate may or may not correspond to one’s marginal tax bracket, due to the way income stacking works. The marginal tax rate is the tax owed on $1 of extra income/distributed gains of the fund.

Marginal tax rate is specific to an individual and is not the maximum tax bracket of the state they live in. It may oftentimes be the top marginal bracket one finds themselves in.

[ivk5]

I'm seeing a lot of 18.8%, should I just go with that?.

Federal QDI rate depends on your taxable income. It's either 0%, 15% or 20%. You can Google the income ranges. If you are subject to the 3.8% Net Investment Income Tax (NII), you'll need to factor that in too, so 23.8% is effectively the top Federal QDI rate (assuming no unusual secondary effects from phaseouts/etc).

[JustinR]

I'm seeing a lot of 18.8%, should I just go with that?.

Federal QDI rate depends on your taxable income. It's either 0%, 15% or 20%. You can Google the income ranges. If you are subject to the 3.8% Net Investment Income Tax (NII), you'll need to factor that in too, so 23.8% is effectively the top Federal QDI rate (assuming no unusual secondary effects from phaseouts/etc).

Thank you!

[JustinR]

VWO (emerging markets) is really tax efficient, any ideas why?

Also interesting that VSS is way more inefficient than several bond funds, including Total Bond.

[JustinR]

My birthday gift to the Bogleheads:

This is Pt. III of (2015) Relative tax efficiency including Foreign Tax Credit (VSS, VWO, VBR, etc.) and 2016 Relative Tax Efficiency, with numbers for the 2017 tax year. The google spreadsheet link is here.

Enjoy!

If you're interested in adding any more, could you consider doing VNQ (REIT)?

I assume it would be best in a tax-advantaged account but it'd be interesting to see the actual numbers of how tax-inefficient it is.

[grabiner]

VWO (emerging markets) is really tax efficient, any ideas why?

VWO is tax-efficient because of the foreign tax credit. Emerging markets tend to withhold more than developed markets in foreign taxes, and you lose this credit if you hold the fund in an IRA.

Also interesting that VSS is way more inefficient than several bond funds, including Total Bond.

The high tax cost is the result of a high yield (which I do not believe is normal for this fund) and a low fraction of qualified dividends (which is normal). However, you are looking at the wrong line; VSS is more tax-efficient (0.28% with the spreadsheet's 12% federal and 6% stae tax) than BND (0.41%).

[JustinR]

Also interesting that VSS is way more inefficient than several bond funds, including Total Bond.

The high tax cost is the result of a high yield (which I do not believe is normal for this fund) and a low fraction of qualified dividends (which is normal). However, you are looking at the wrong line; VSS is more tax-efficient (0.28% with the spreadsheet's 12% federal and 6% stae tax) than BND (0.41%).

Try it with these tax rates:

Federal Marginal Tax Rate: 0.24
QDI Tax Rate: 0.15
State Tax Rate: 0.093

VSS gives me 0.91
BND gives me 0.78

[grabiner]

Also interesting that VSS is way more inefficient than several bond funds, including Total Bond.

The high tax cost is the result of a high yield (which I do not believe is normal for this fund) and a low fraction of qualified dividends (which is normal). However, you are looking at the wrong line; VSS is more tax-efficient (0.28% with the spreadsheet's 12% federal and 6% stae tax) than BND (0.41%).

Try it with these tax rates:

Federal Marginal Tax Rate: 0.24
QDI Tax Rate: 0.15
State Tax Rate: 0.093

VSS gives me 0.91
BND gives me 0.78

With tax rates on non-qualified and qualified dividends so close, this tax situation (CA resident, presumably) favors the lower-yielding fund.

However, the 3.59% yield for 2017 was abnormal; Vanguard FTSE All-World Ex-US Small-Cap Index Tax Distributions on the wiki says that the normal yield is about 2.5%.

[JustinR]

My birthday gift to the Bogleheads:

This is Pt. III of (2015) Relative tax efficiency including Foreign Tax Credit (VSS, VWO, VBR, etc.) and 2016 Relative Tax Efficiency, with numbers for the 2017 tax year. The google spreadsheet link is here.

Enjoy!

If you're interested in adding any more, could you consider doing VNQ (REIT)?

I assume it would be best in a tax-advantaged account but it'd be interesting to see the actual numbers of how tax-inefficient it is.

So I tried putting in VNQ myself and I got 1.42. Can someone check my work?

Expense ratio: 0.12
Dividend distribution: =(0.00572+0.17301+0.41627+0.0077+0.23291+0.56039+0.0082+0.24832+0.59747+0.01214+0.36742+0.88404)
Next four rows as: 0

It's a little complicated because the distribution page has, among dividends, "return of capital" and "unrecap sec 1250 gains" which I'm unsure of what to categorize it as.