Net total return vs. gross total return - a fallacy?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Net total return vs. gross total return - a fallacy?

Postby Charybdis » Mon Jun 11, 2012 5:26 am

Have you ever wondered what is the difference between net total return, and gross total return? Index providers like MSCI usually use three index types:

Prices indexes don't count dividends, only the price of the stocks.
Gross total return indexes reinvest as much as possible of a company’s dividend distributions. The reinvested amount is equal to the total dividend amount distributed to persons residing in the country of the dividend-paying company1. Gross total return indexes do not, however, include any tax credits.
Net total return indexes reinvest dividends after the deduction of withholding taxes, using (for international indexes) a tax rate applicable to non-resident institutional investors who do not benefit from double taxation treaties.

But those ETFs/index funds who invest in foreign stocks (foreign = relative to the domicile of the fund) usually track the net total return indexes.

Isn't it a fallacy?

1) Bogleheads' goal is to capture the whole market return, after all taxes and expenses. The whole market return is the Gross total return index. So I think ETFs/index funds should compare themselves to gross total return indexes, not to net total return indexes!

2) The net total return index assumes the worst case scenario, when a fund pays the full dividend withholding tax after foreign stocks. But only the dumbest fund with the dumbest manager pays the full dividend withholding tax after foreign stocks.

3) The performance of a fund is depends on the domicile of the fund. For example, Ireland/Luxembourg domiciled funds don't pay the dividend withholding tax after European stocks, but pay the withholding tax after US/emerging market stocks. US domiciled funds on the other hand don't pay the US dividend withholding tax, but pay the tax after European stocks etc.

4) Security lending can increase the income of a fund, and also decrease the amount of dividend withholding tax payable. For example, if a fund lends a dividend paying stock to someone who doesn't pay the dividend withholding tax after that stock, then the fund avoids paying the dividend withholding tax in the end.

5) Tax treaties between countries could be more favorable than the dividend withholding tax level is assumed by the index provider, because the index provider assumes the worst case scenario.

6) The replication method affect the dividend withholding tax payable a great deal. An ETF can use not just full replication, but also optimized replication and swap-based synthetic replication. The optimized replication could avoid those stocks where the amount of dividend withholding tax payable is too high. And the swap-based ETFs just "outsource" the index tracking to a counterparty. How this counterparty manages to track the index is unclear. ETFs don't disclose this, as far as I know. So the counterparty could use other clever methods to avoid the dividend withholding taxes we don't even know about.


I always wondered how can the index funds/ETFs produce so low, even positive tracking error, when tracking an international index. I think it is because they use the wrong index! They should track the gross total return indexes, not the net total return indexes.

Because as an investor, I care about the gross total return indexes. This is the market return.

In short, ETFs pay much lower dividend withholding taxes than the level of taxes assumed by the net total return index provider, therefore net total return indexes are easy to beat.

Isn't it a fallacy that international ETFs/index funds track the net total return indexes? Net total return makes the ETFs/index funds look much better than they really are.

And those studies which examined the diversification benefit of investing in foreign stocks, did they count the foreign dividend withholding taxes? So is it worth diversifying internationally, even after paying all the dividend withholding tax?

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Tue Jun 12, 2012 10:10 am

Does anyone know, how much dividend withholding tax does MSCI assume, when calculating the net total return indexes?

For example, here are the withholding tax levels used by STOXX for their total return indexes: http://www.stoxx.com/indices/taxes.html

Of course, no ETF pays that much withholding tax like the figures in the linked spreadsheet, due to tax treaties and withholding tax optimization via security lending.

That is why you must compare international ETFs to the gross total return indexes, which is actually the market return. The net total return index is not the market return. When investing books talk about long term equity returns, they use the gross total return index, but the international ETFs compare themselves to the net total return indexes.

HongKonger
Posts: 851
Joined: Tue Jun 21, 2011 10:35 am
Location: Deep in the Balkans

Re: Net total return vs. gross total return - a fallacy?

Postby HongKonger » Tue Jun 12, 2012 10:30 am

Charybdis wrote:Of course, no ETF pays that much withholding tax like the figures in the linked spreadsheet, due to tax treaties and withholding tax optimization via security lending.

.


Well I pay 30% witholding on all dividends paid by US originated ETFs because there is no tax treaty ... and I know that in HK its the same for several individual stocks such as Prada (Italy) which IPO'd here a year or so ago. Not everywhere has tax treaties!

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Tue Jun 12, 2012 10:41 am

HongKonger wrote:
Charybdis wrote:Of course, no ETF pays that much withholding tax like the figures in the linked spreadsheet, due to tax treaties and withholding tax optimization via security lending.

.


Well I pay 30% witholding on all dividends paid by US originated ETFs because there is no tax treaty ... and I know that in HK its the same for several individual stocks such as Prada (Italy) which IPO'd here a year or so ago. Not everywhere has tax treaties!


It doesn't matter how much do you pay as an individual.

I am talking about the internal rate of return of the ETFs. ETFs are domiciled in those countries where the tax environment is the most investor friendly. US ETFs of course don't pay the 30% US withholding tax, which is assumed by the index provider when calculating the net total return. For example, a US ETF could track the MSCI ACWI net total return index. MSCI assumes that this ETF will pay 30% US withholding tax, which is of course not the case. So it is easy to produce so low, or even positive tracking error, when an ETF tracks the net total return index, instead of the gross total return index.

And the European ETFs are domiciled in either Ireland or Luxembourg. Surprise, these countries don't pay that much dividend withholding tax like the assumed numbers by index providers. And there is also the security lending which lowers the dividend withholding tax.

BTW why don't you invest in for example db X-trackers accumulating ETFs? They don't pay out dividends (they reinvest the received dividends back into the ETF right away), and they don't pay 30% US withholding tax, only 15% minus security lending.

HongKonger
Posts: 851
Joined: Tue Jun 21, 2011 10:35 am
Location: Deep in the Balkans

Re: Net total return vs. gross total return - a fallacy?

Postby HongKonger » Tue Jun 12, 2012 6:12 pm

Charybdis wrote:
HongKonger wrote:
Charybdis wrote:Of course, no ETF pays that much withholding tax like the figures in the linked spreadsheet, due to tax treaties and withholding tax optimization via security lending.

.


Well I pay 30% witholding on all dividends paid by US originated ETFs because there is no tax treaty ... and I know that in HK its the same for several individual stocks such as Prada (Italy) which IPO'd here a year or so ago. Not everywhere has tax treaties!


It doesn't matter how much do you pay as an individual.

I am talking about the internal rate of return of the ETFs. ETFs are domiciled in those countries where the tax environment is the most investor friendly. US ETFs of course don't pay the 30% US withholding tax, which is assumed by the index provider when calculating the net total return. For example, a US ETF could track the MSCI ACWI net total return index. MSCI assumes that this ETF will pay 30% US withholding tax, which is of course not the case. So it is easy to produce so low, or even positive tracking error, when an ETF tracks the net total return index, instead of the gross total return index.

And the European ETFs are domiciled in either Ireland or Luxembourg. Surprise, these countries don't pay that much dividend withholding tax like the assumed numbers by index providers. And there is also the security lending which lowers the dividend withholding tax.

BTW why don't you invest in for example db X-trackers accumulating ETFs? They don't pay out dividends (they reinvest the received dividends back into the ETF right away), and they don't pay 30% US withholding tax, only 15% minus security lending.


Because I don't use synthetic ETFs.

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Wed Jun 20, 2012 5:44 pm

Here is a detailed article describing this problem: http://www.indexuniverse.eu/europe/publ ... =1&start=5

In summary, when you buy an international index fund or ETF, there is a hidden cost, which is about 0.2-0.6% / year :moneybag .
This hidden cost is that the funds track the net total return index, and compare themselves to the net total return index (which is not the market return), but they should compare themselves to the gross total return index (which is the market return). The net total return index lags behind the gross total index by about 0.2-0.6% / year, the exact amount of difference depends on the current dividend yield.

Because you want to earn the market return, right?

Metallising
Posts: 11
Joined: Sun Jun 24, 2012 4:14 pm
Location: Switzerland

Re: Net total return vs. gross total return - a fallacy?

Postby Metallising » Fri Jul 06, 2012 7:59 pm

Interesting topic, I had never thought about that.

I understand your point and based on my knowledge about the subject I would say you are right. It would be much more clear and logical if the ETFs tracked the gross total return index.

If an EU investor wanted to buy Ishares SP500, do you think he/she should buy the fund domiciled in US or EU? As I see it:
1 - If the investor bought the fund domiciled in US, then a witholding tax of 15% (double taxation treaty applied) will be applied on dividends. However the fund itself wouldn't pay any witholding taxes on dividends since the companies that constitute the SP500 are US based, like the domicile of the fund.
2 - If the investor bought the fund domiciled in EU, then he/she wouldn't pay any witholding tax on dividends distributed by fund. However the fund itself would pay witholding taxes on dividends since it held foreign companies, probably a tax of 15% would be applied or maybe even less, due lending securities. In the end is not clear which witholding taxes rates are applied to the fund.

The conclusion I take from this is that it's the same investing in EU or US domiciled funds, however there is a chance that the EU domiciled fund reduce the 15% witholding tax a bit by lending securities. Even so it might be better to stay with the US domiciled fund since it has smaller spreads and less fees. What do you think?

In point 2 I said that a EU investor wouldn't pay any witholding tax on EU domiciled funds, but I'm not totally sure about this. Can you confirm this please?

8-)

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Sat Jul 07, 2012 2:01 am

Metallising wrote:Interesting topic, I had never thought about that.

I understand your point and based on my knowledge about the subject I would say you are right. It would be much more clear and logical if the ETFs tracked the gross total return index.

If an EU investor wanted to buy Ishares SP500, do you think he/she should buy the fund domiciled in US or EU? As I see it:
1 - If the investor bought the fund domiciled in US, then a witholding tax of 15% (double taxation treaty applied) will be applied on dividends. However the fund itself wouldn't pay any witholding taxes on dividends since the companies that constitute the SP500 are US based, like the domicile of the fund.
2 - If the investor bought the fund domiciled in EU, then he/she wouldn't pay any witholding tax on dividends distributed by fund. However the fund itself would pay witholding taxes on dividends since it held foreign companies, probably a tax of 15% would be applied or maybe even less, due lending securities. In the end is not clear which witholding taxes rates are applied to the fund.

The conclusion I take from this is that it's the same investing in EU or US domiciled funds, however there is a chance that the EU domiciled fund reduce the 15% witholding tax a bit by lending securities. Even so it might be better to stay with the US domiciled fund since it has smaller spreads and less fees. What do you think?

In point 2 I said that a EU investor wouldn't pay any witholding tax on EU domiciled funds, but I'm not totally sure about this. Can you confirm this please?

8-)


That's right. You cannot escape the 15% withholding tax when investing in US equities. Either you or the ETF pays that.
You cannot fully escape any other withholding taxes either.

I wouldn't buy any US domiciled funds. Those aren't designed to EU investors. Buy only the Ireland/Luxembourg domiciled funds.

The majority of the EU domiciled funds don't distribute dividends, they just reinvest the net dividends, after paying the withholding tax.
US funds must always distribute income, unlike EU domiciled ETFs.

When you invest in US funds, the US government withholds 30% dividend tax. Then you must claim back 15% actively, because according to the tax treaty they should have withheld only 15%. But when you buy a US fund anonymously on the stock exchange, they don't know whether your country has a tax treaty with the US or not. That is why they withhold the maximum 30%.

It is complicated paperwork to claim back the 15% withholding tax, so that you pay only 30%-15% = 15%. And there is a delay, you don't receive it back immediately. During that delay your rate of return on the dividend is 0%.

And after you received the dividend, you must reinvest it back into the ETF. You pay brokerage commission and spread. If you cannot reinvest the dividend immediately (because for example your are on holiday), then there is a delay yet again, while you don't earn any money.

That is why I think a capitalizing/accumulating EU domiciled ETF is always better, which doesn't distribute dividends, so that you don't have to deal with reinvesting the dividends and pay extra cost when doing so.

For example, the db X-trackers MSCI USA TRN INDEX ETF outperformed the tracked index since inception: index return: 5.77%; ETF return: 5.90%. Note: it is a synthetic ETF.

Don't just look at the TER, look at the actual tracking error between the index and the ETF, that is what matters, not the TER.

But this outperformance is a fallacy, because that db X-trackers MSCI USA TRN INDEX ETF tracks the MSCI USA Total Return Net Index. The MSCI USA Total Return Net Index assumes 30% dividend withholding tax, but the ETF must pay only 15%.

It is easy to outperform the benchmark when you choose the wrong benchmark.

Another thing to consider: are you from Switzerland? Do you invest in CHF or in EUR? Do you want to spend your earnings in the same currency you invested in?

Then US domiciled funds have another disadvantage: you must exchange your EUR or CHF for USD, and you must pay the bid/ask spread for this currency exchange.

But you can buy EU domiciled ETFs directly in EUR or in CHF. So you avoid the currency exchange cost.

There is currency risk either way, between the USD and EUR, or USD and CHF. So you should invest directly either in CHF or in EUR, to avoid the currency exchange cost and simplify things. You can find ETFs in the CHF trading currency at SIX Swiss Exchange: http://www.six-swiss-exchange.com/funds ... ew_en.html So even if the base currency of an ETF is USD, you can buy it in CHF, so you avoid the currency exchange. But of course then there is a currency risk between the USD and CHF. But if you exchange your CHF for USD, invest in USD, then when you want to spend your earnings and exchange the USD for CHF, then you accomplished the same thing but at a higher cost, as if you invested directly in CHF.

Only invest in the USD currency if you want to spend USD later.

Metallising
Posts: 11
Joined: Sun Jun 24, 2012 4:14 pm
Location: Switzerland

Re: Net total return vs. gross total return - a fallacy?

Postby Metallising » Sat Jul 07, 2012 10:04 am

When you invest in US funds, the US government withholds 30% dividend tax. Then you must claim back 15% actively, because according to the tax treaty they should have withheld only 15%. But when you buy a US fund anonymously on the stock exchange, they don't know whether your country has a tax treaty with the US or not. That is why they withhold the maximum 30%.
It is complicated paperwork to claim back the 15% withholding tax, so that you pay only 30%-15% = 15%. And there is a delay, you don't receive it back immediately. During that delay your rate of return on the dividend is 0%.


Depends. If you use Interactive Brokers (probably many others US big brokerage firms do the same) as your intermediary they will apply the tax treaty immediately when receiving dividends. You don't have to claim nothing back because you are going to receive the right amount.

Don't just look at the TER, look at the actual tracking error between the index and the ETF, that is what matters, not the TER.


True, I noticed that sometimes higher TER ETFs have better performance than lower ones. The problem is that the great majority of EU ETFs have a lack of history, I wouldn't consider 2 years of history enough to analyse the tracking error of a fund.

For example, the db X-trackers MSCI USA TRN INDEX ETF outperformed the tracked index since inception: index return: 5.77%; ETF return: 5.90%. Note: it is a synthetic ETF.


I have been avoiding other than full replication method ETFs. Mainly because I don't understand the synthetic ETFs, in fact, I think it's not supposed to understand since the prospectus don't explain it clearly to the investors.

The idea of capitalizing dividends instead of distributing is great and at this moment we have at least two reliable ETF providers doing this with physical and full replicated products. Credit Suisse and Ishares. If an EU investor build a portfolio with capitalizing ETFs from those two I think there isn't much to worry about. I don't see disadvantage by owning capitalizing ETFs instead of distributing ones.
Amundi, Deutsche Bank, Lyxor are also options, but I would avoid them because of the replication method.

About the currency, I really don't care about it. Currency risk and volatility is a benefit for an international diversified portfolio. The currency in which a ETF is quoted isn't a problem in my view, since the currency exchange cost from your home currency to the foreign one can be as low as 1/2 pips. For this, you can use a broker like Interactive Brokers, which allow you to hold a multi-currency account and charges 1/2 pips in the major pairs. So, if you have CHF and want to buy a ETF quoted in EUR the broker wouldn't do the conversion for you, you must buy by yourself EUR and hold it in your account in order to buy it.
Also, I'm not a Swiss citizen, I don't know where I will be living in 3, 5, 10 years, times to times I move to a new country and I plan to keep doing this in the future. Most likely I will be spending in CHF, EUR and USD in the future, but I'm not sure.

Only invest in the USD currency if you want to spend USD later.


Not true. It's a complex subject, but a few studies have been done and the conclusion is that currency volatility and risk in the end add 0% to a international diversified portfolio. If you hedge, add 0% minus hedging costs.

Edit: Forgot to ask. Do you know if EU investors pay witholding taxes on distributions by Irish domiciled funds?

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Sat Jul 07, 2012 10:32 am

Well, let's see. The US domiciled Vanguard Total Stock Market ETF has an expense ratio of just 0.06%. It distributes dividends quarterly. Question: are you able to invest new money each quarter? Because if you invest new money quarterly, then you could buy that Vanguard Total Stock Market ETF, because then you can reinvest the distributed dividends along with your new capital.

So maybe it could be a good idea in your case to buy that Vanguard Total Stock Market ETF.

However, I wouldn't buy the Vanguard Total International Stock ETF, because that ETF pays dividend withholding taxes to foreign (non-US) governments, then it distributes dividends to you, and you pay 15% dividend tax yet again.

But an important question: when you receive dividend from an ETF, do you pay dividend tax in Switzerland? If you pay that dividend tax, then forget all the US domiciled ETFs and EU domiciled distributing ETFs, and buy ONLY EU domiciled capitalizing ETFs.

ETFs are still very new in the EU so we are guinea pigs. You can't do much about it.

It is true that synthetic ETFs suck. However, I bought one: the db X-trackers FTSE EPRA/NAREIT GLOBAL REAL ESTATE ETF. Because it is the only global REIT ETF which doesn't distribute dividends. I know how they work and understand the risk. But

When I wrote that "Only invest in the USD currency if you want to spend USD later", I meant that if you have EUR or CHF, then I think you should buy the chosen ETF in the EUR or CHF trading currency (if available). You cannot avoid the currency risk either way.

I heard that Ireland doesn't apply dividend withholding tax, but it can change in the future.

Metallising
Posts: 11
Joined: Sun Jun 24, 2012 4:14 pm
Location: Switzerland

Re: Net total return vs. gross total return - a fallacy?

Postby Metallising » Sat Jul 07, 2012 12:31 pm

I heard that Ireland doesn't apply dividend withholding tax, but it can change in the future.


Recently I received dividends from an Ishares ETF domiciled in Ireland and paid 0% in witholding taxes. However, I'm not sure why. My broker for those assets is EU-based but I would like to know what would happen if I used a US-based broker.

Well, let's see. The US domiciled Vanguard Total Stock Market ETF has an expense ratio of just 0.06%. It distributes dividends quarterly. Question: are you able to invest new money each quarter? Because if you invest new money quarterly, then you could buy that Vanguard Total Stock Market ETF, because then you can reinvest the distributed dividends along with your new capital.


In my view there isn't much difference owning US stocks via a fund domiciled in Ireland or US. It's you (US domiciled) or the fund (Irish domiciled) who will pay the witholding taxes. However the US domiciled have lower TER and lower spreads and if you use a broker like Interactive Brokers (IB), you will pay only 1$ for transaction in US. And yes, I invest new money every months.

But an important question: when you receive dividend from an ETF, do you pay dividend tax in Switzerland? If you pay that dividend tax, then forget all the US domiciled ETFs and EU domiciled distributing ETFs, and buy ONLY EU domiciled capitalizing ETFs.


I think every country consider dividends as income and it must be declared. In Switzerland when I declare dividends, the witholding taxes that I paid to foreign countries are credited back to me in form of a check. However, later I will receive a bill from Swiss tax authorities to pay a tax on those dividends, the tax is based on my total annual income, usually around 9%.
Conclusion: By filling the tax form, I can have the witholding tax back, it's a very simple process in Switzerland. Of course, then, dividends are going to be summed on my total annual income which is subject to taxes.

When I wrote that "Only invest in the USD currency if you want to spend USD later", I meant that if you have EUR or CHF, then I think you should buy the chosen ETF in the EUR or CHF trading currency (if available). You cannot avoid the currency risk either way.


In my opinion there isn't any good reason to avoid US ETFs. First, currency exchange costs can be insignificant, you can make the currency conversion yourself via a Forex position, luckily you can take some profit from it. Second, ETF traded in US stock exchange have lower TER and lower spreads. Third, if you use a broker such IB you can trade ETF in US market paying a 1$ comission, or even better, if you use TDameritrade you can trade up to 100 ETF comission-free.

If you use a EU broker and trade on EU stock markets it's hard pay less than 5€ for transaction. The best I know is 4€ to German markets and 5€ to Euronext (Portugal, France, Netherlands and Belgium) markets.

It is true that synthetic ETFs suck. However, I bought one: the db X-trackers FTSE EPRA/NAREIT GLOBAL REAL ESTATE ETF. Because it is the only global REIT ETF which doesn't distribute dividends. I know how they work and understand the risk.


I think you did well. Honestly I don't see any better solution if you really want to hold REITs.

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Sat Jul 07, 2012 12:50 pm

I don't quite understand your tax situation regarding dividends, but when you buy a capitalizing EU ETF, you don't receive dividends, so no tax is payable. So you must carefully calculate which is better in your situation.

Also when you receive dividend from a US domiciled ETF which invests in non-US equities, then you are double taxed: first, when the ETF itself pays the foreign dividend withholding tax, second when you pay the US dividend withholding tax.

this can be avoided by buying an EU domiciled capitalizing ETF. So again, carefully calculate this in a spreadsheet.

One time trading costs and bid/ask spreads aren't significant on the long haul. Your biggest costs will be the TER, the actual tracking error and dividend withholding taxes, and your taxes you pay in your home county (if any).

All in all, your total costs can easily exceed 0.8-1% per year. Given that the equity premium is about 3% if we are lucky, then costs eat 33% of your equity premium.

I am not saying that then you shouldn't invest in equity ETF, but this is something you must be aware of. I also own equity ETFs of course.

Also Interactive Brokers charges an inactivity fee.

Metallising
Posts: 11
Joined: Sun Jun 24, 2012 4:14 pm
Location: Switzerland

Re: Net total return vs. gross total return - a fallacy?

Postby Metallising » Sat Jul 07, 2012 1:46 pm

I don't quite understand your tax situation regarding dividends, but when you buy a capitalizing EU ETF, you don't receive dividends, so no tax is payable. So you must carefully calculate which is better in your situation.


It may be my fault as I didn't explain clearly. To resume, if I receive dividends, no matter if they come from US or Irish domiciled funds, those dividends are always going to be consider income by Swiss tax authorities and subject to tax in Switzerland. Witholding foreign taxes are always credited back to me by Swiss National Bank.
So yes, capitalizing ETFs still a better option since I don't have to declare dividends which would be considered income subject to taxes.

One time trading costs and bid/ask spreads aren't significant on the long haul. Your biggest costs will be the TER, the actual tracking error and dividend withholding taxes, and your taxes you pay in your home county (if any).


True, in the long term taxes paid on dividends and the TER is what really matters.

Also Interactive Brokers charges an inactivity fee.


Yes, you have to spent at least 10$ per month, if you don't, it will be debited from your account. Let's say you spend only 5$, then, IB will debit the remaining 5$ from your account.
I consider it a good deal since I buy shares on a monthly basis.
I put it this way: I pay 120$ in commissions per year and re-balance and add new money in a monthly basis.

Even if you only add new money and re-balance yearly, assuming that you own more than 5 ETFs and use a EU-based broker with average prices, you won't pay much less than 120$ in commissions.

Other good thing about this broker is that they transfer funds back to you free of charge.

User avatar
Charybdis
Posts: 242
Joined: Sun Apr 25, 2010 4:38 am
Location: Budapest, Hungary

Re: Net total return vs. gross total return - a fallacy?

Postby Charybdis » Sat Jul 07, 2012 5:56 pm

I think you shouldn't add new money each month and rebalance each month, you will pay too much spread and broker commission.

Add money for example quarterly, and rebalance yearly.

Don't invest in too many ETFs, 3-4 ETFs max. should do it.

Also don't forget that you should pay for real time market data, because you should trade ETFs using limit orders. But for limit orders you should know the real time prices. And the market makers of certain EU domiciled ETFs increase the spread dramatically at random times and without notice. This is why you should carefully monitor the spread in real time, and buy only when the spread is tight. If you invest new money each month, then you must pay for the real time market data each month. If you invest quarterly, you only pay 4 months. I suggest not to buy ETFs without real time market data; what you win by not paying for real time data, you might lose it at unexpectedly high spreads.

US domicled funds have dramatically smaller spreads, though.

Metallising
Posts: 11
Joined: Sun Jun 24, 2012 4:14 pm
Location: Switzerland

Re: Net total return vs. gross total return - a fallacy?

Postby Metallising » Sat Jul 07, 2012 10:59 pm

I think you shouldn't add new money each month and rebalance each month, you will pay too much spread and broker commission.


I wouldn't sell any position, just add new money on losing assets. You pay the same spread no matter if you buy 10 shares monthly or 100 shares yearly. Since I'm planning to open an account with IB, I have to assume that I will pay 120$ yearly in commissions, so I think it worth to add new money on a monthly basis. I think this is a common investment strategy, dollar cost averaging as "they" call it.
How much do you spend yearly in commissions?

Don't invest in too many ETFs, 3-4 ETFs max. should do it.


Why? In fact I own 22 ETFs. I slice every part of my portfolio. For example, for international exposure I have different ETFs to Emerging Markets, Pacific and US. And I go even further, for example for Pacific, I use two, one Pacific ex-Japan, and another just to Japan. Since Japan weights 60% in MSCI Pacific index I prefer to separate it and rebalance it independently. Do you remember the tsunami and nuclear disaster in Japan? By owning a ETF specifically for Japan I had the opportunity to take advantage during the chaos and buy more shares during the panic in the markets.
I just consider two important things to have in mind before slice and dice a portfolio: 1- Since a large number of funds will be traded, commissions should be as low as possible. 2- Avoid exotic ETFs like small caps in Pakistan or something like that in order to avoid very high spreads and liquidity problems.
Slice and dice is a very common strategy and very discussed in this forum. I don't see any good reason to don't slice and dice. It might be a little more complex but it worth for better risk management and little extra return.

Also don't forget that you should pay for real time market data, because you should trade ETFs using limit orders. But for limit orders you should know the real time prices. And the market makers of certain EU domiciled ETFs increase the spread dramatically at random times and without notice. This is why you should carefully monitor the spread in real time, and buy only when the spread is tight. If you invest new money each month, then you must pay for the real time market data each month. If you invest quarterly, you only pay 4 months. I suggest not to buy ETFs without real time market data; what you win by not paying for real time data, you might lose it at unexpectedly high spreads.


As a rule of thumb I never buy ETFs with a market cap inferior to 200 millions, this way I avoid being ripped of by market makers once usually funds with this dimension trade frequently.

tuscan girl
Posts: 1
Joined: Sat Nov 10, 2012 6:11 am

Re: Net total return vs. gross total return - a fallacy?

Postby tuscan girl » Sat Nov 10, 2012 6:51 am

Hello all,
I’d like to ask a closely-related accumulating ETF question (that may show my basic misunderstanding – sorry!) that is superficially a cross-border tax treatment question, but is (I think) really a question about how accumulating etfs employ their accumulated dividends from underlying instruments. U.S. friends may not be able to answer the first half of my question, but may be able to answer the second part which is, I think, generic – so please read the end!

In this case, the two tax domains are Italy (where etfs are held) and the UK where I live at the moment. Unfortunately the EU hasn’t sensibly integrated member states’ taxation and it’s hell!

Consider one share of a simple index tracking ETF. In Italy at sale there are two tax components:

(i) ‘income from capital’ = (NAV at sale - NAV at purchase). That’s at 20%.
(ii) ‘capital gain/loss’ = [(Sale Price – Purchase Price) – income from capital (above) ] also at 20%

If the etf price deviates from NAV (the same as tracking error, I think) during taxation period, this can produce odd results.

Now, with double taxation, in the UK, as it’s an accumulating etf, the distributed income is zero so the tax is yearly on ‘reported income’. So my question is whether in an accumulating ETF's ‘reported income’ is the same as it's ‘change in NAV’ (i.e. (i) above). Unless they can be somewhat equated (obviously over the ownership period..) I can’t offset Italian tax against UK tax and will get double taxed!

Generic bit: Underneath all of this is my lack of clarity over how an accumulating ETF that tracks an index employs the underlying dividends it earns (assuming they are physical) whilst continuing to track the index? It would seem it should outperform if it bought more underlying instruments with the dividend funds. From what you all say, I think the message is that it should come close to tracking a 'gross return index' rather than a 'net return index'. However, an accumulating etf in question - Credit Suisse CSPX.L (CSSPX.IM in Milan) tracks ‘S&P 500 Composite (NR)’ which, I assume, means ‘Net Return’ although the fund sheet does not explicitly state this as far as I can see.

If this is true, then the CS, tracking, accumulating, etf tracks a net return index – so I’m even more confused as to how the accumulated underlying dividends ulimately benefit me, the shareholder compared with a distributing SP500 tracker?

I’m sorry: it’s probably clear, I’ve just missed the point! Thanks......

Atomium
Posts: 13
Joined: Mon Jul 06, 2015 3:15 pm
Location: Belgium

Re: Net total return vs. gross total return - a fallacy?

Postby Atomium » Fri Aug 28, 2015 3:03 pm

But those ETFs/index funds who invest in foreign stocks (foreign = relative to the domicile of the fund) usually track the net total return indexes.

Isn't it a fallacy?

1) Bogleheads' goal is to capture the whole market return, after all taxes and expenses. The whole market return is the Gross total return index. So I think ETFs/index funds should compare themselves to gross total return indexes, not to net total return indexes!

2) The net total return index assumes the worst case scenario, when a fund pays the full dividend withholding tax after foreign stocks. But only the dumbest fund with the dumbest manager pays the full dividend withholding tax after foreign stocks.

I fully agree with the opening post and all following posts of Charybdis. For example there are synthetic trackers, that follow the Stoxx Europe 600 Net Return Index
The net return index assumes that dividends minus withholding tax is reinvested. You can find the rates here: https://www.stoxx.com/withholding-taxes
F.e. for the USA they assume 30% withholding tax, while this can be reduced to 15%. So the synthetic tracker behaves as if they/you paid 30% withholding tax, while it's only 15% in reality. This 15% tax advantage is extra profit for the ETF issuer.
Huge fan of iShares ETFs. | I put a lot of effort & research in my posts, if you believe what I'm saying is incorrect please contact me!

User avatar
galeno
Posts: 1003
Joined: Fri Dec 21, 2007 12:06 pm
Location: Alajuela, Costa Rica

Re: Net total return vs. gross total return - a fallacy?

Postby galeno » Fri Aug 28, 2015 5:08 pm

We are USA-NRA investors in a country w/o a tax treaty with the USA. We hold a 2 ETF portfolio: 60% VWRD + 35% IUAG + 5% CASH.

The interest income from IUAG has a 0% withholding tax. Our USA equities have a 15% dividend withholding tax. Our non-USA equities have a 6% dividend withholding tax.

E.g. VWRD costs us 0.45% (ER = 0.25%. TR = 0.20%). VUSA would cost us 0.37% (ER=0.07%. TR=0.30%). World wide equity diversification costs us an additional 0.08%. It's a bargain.
AA = 40/55/5. Expected CAGR = 3.8%. GSD (5y) = 6.2%. USD inflation (10 y) = 1.8%. AWR = 2.8%. TER = 0.5%. Port Yield = 2.0%. Term = 35 yr. FI Duration = 6.2 yr. Portfolio survival probability = 100%.

Micks
Posts: 312
Joined: Tue Feb 25, 2014 12:38 pm
Location: the Netherlands
Contact:

Re: Net total return vs. gross total return - a fallacy?

Postby Micks » Fri Aug 28, 2015 5:28 pm

Atomium wrote:F.e. for the USA they assume 30% withholding tax, while this can be reduced to 15%. So the synthetic tracker behaves as if they/you paid 30% withholding tax, while it's only 15% in reality. This 15% tax advantage is extra profit for the ETF issuer.


The ETF issuer itself does not receive extra profit if the fund outperforms its index, likewise if it underperforms (the issuer's compensation is included in the total expense ratio).

I agree that comparing to the worst-case scenario is weird, but is something that both synthetic and physical ETFs do.
Kind regards, Mick

Clive
Posts: 1951
Joined: Sat Jun 13, 2009 5:49 am

Re: Net total return vs. gross total return - a fallacy?

Postby Clive » Fri Aug 28, 2015 7:18 pm

As a UK investor, if I buy into Vanguard UK's S&P500 ETF then the factsheet indicates that it benchmarks to the SPTR500N index which is the total return less 30% standard US withholding tax. The fund is domiciled in Ireland which sees that withholding tax reduced to 15%. As the fund has a 0.07% fee then generally the fund outperforms the benchmark, but lags the S&P500 gross total return by around 0.4%.

For example the factsheet reported figures versus the S&P indexes gross total return figure for the reported July based yearly figures

To July Gross BM Fund
2014 16.94 16.21 16.48
2015 11.21 10.53 10.8

Vanguard is one of the more transparent providers and even then things aren't clear at all IMO. Other for profit fund managers can exploit costs much more.

I agree in that I would also prefer funds benchmarked to the gross index. That way, as a example, for 2015 July you'd see the 11.21 gross index figure, and perhaps be aware of around a 2.2% dividend, 15% withholding tax and 0.07% fund expense fee and would figure a 2.2% x 0.15 dividend withheld = 0.33%, combined with 0.07% expense = 0.4% and deducting that from the 11.21% gross index figure leaves 10.81% and see that the 10.8% fund return aligned to that and be more aware of the 0.4% gross index lag factor. If/when dividend yields spike to perhaps 5%, 15% withholding tax = 0.75%, 0.07% fee, 0.82% lag expected.

Another factor that UK investors have to be wary of is if a foreign fund isn't UK reporting registered then HMRC will tax capital gains derived from such funds as income. You also wont be able to tax harvest capital losses as offsets against capital gains elsewhere in such cases. Even if a fund is UK reporting registered if at ANY time when you held that fund the fund wasn't UK reporting registered, even just very briefly, then the whole purchase to sale date capital gains will be charged as income, not capital gain.

For me that make synthetics the more appealing. Where the counter-party offers a swap for the gross total return, ideally where the swap is collateralised and I prefer leveraged versions myself such that a third in 3x, two thirds retained in bonds, no dividends from the swap and only 33% at risk between rebalances. If/when bonds earn more than what the fund/counterparty charge for providing leverage (borrow), then that closes down or even reverses rewards (outperform the gross index).

Personally I dislike stocks paying dividends (income) as I use stocks for growth and bonds for income (and seek to minimise taxes on bond interest so that bonds better keep up with inflation (future purchase power)). So positions somewhat along the lines of https://www.portfoliovisualizer.com/bac ... ount=10000 where dividends are small (0.15% in that particular case). QLD is a 2x QQQ with a low dividend yield (less than 0.5%), BRK-B pays no dividends and no fund fees. 33% in 2x, 33% in BRK-B combines to make the equivalent of 100% long stock. Gold pays no dividends. For that example since 2007 = 14.4% annualised compared to SPY total return of 6.9% annualised.

Atomium
Posts: 13
Joined: Mon Jul 06, 2015 3:15 pm
Location: Belgium

Re: Net total return vs. gross total return - a fallacy?

Postby Atomium » Sat Aug 29, 2015 12:45 pm

Micks wrote:The ETF issuer itself does not receive extra profit if the fund outperforms its index, likewise if it underperforms (the issuer's compensation is included in the total expense ratio).
You are right, my statement about them earning extra profit is incorrect. It remains strange though that a synthetic index follows the Net Return version of some index: the gross return of that Index minus (fictional) maximum withholding taxes.
They do what they promise, but it's a strange promise for what I'm concerned.
Last edited by Atomium on Sat Aug 29, 2015 12:54 pm, edited 2 times in total.
Huge fan of iShares ETFs. | I put a lot of effort & research in my posts, if you believe what I'm saying is incorrect please contact me!

Federiko25
Posts: 5
Joined: Sat Nov 24, 2012 4:54 pm

Re: Net total return vs. gross total return - a fallacy?

Postby Federiko25 » Thu Sep 24, 2015 7:47 pm

galeno wrote:We are USA-NRA investors in a country w/o a tax treaty with the USA. We hold a 2 ETF portfolio: 60% VWRD + 35% IUAG + 5% CASH.

The interest income from IUAG has a 0% withholding tax. Our USA equities have a 15% dividend withholding tax. Our non-USA equities have a 6% dividend withholding tax.

E.g. VWRD costs us 0.45% (ER = 0.25%. TR = 0.20%). VUSA would cost us 0.37% (ER=0.07%. TR=0.30%). World wide equity diversification costs us an additional 0.08%. It's a bargain.


Galeno,
Shouldn't a NRA still pay income tax on the semi-annual IUAG distrubution?

Clive
Posts: 1951
Joined: Sat Jun 13, 2009 5:49 am

Re: Net total return vs. gross total return - a fallacy?

Postby Clive » Fri Sep 25, 2015 1:31 pm

tuscan girl wrote:...Now, with double taxation, in the UK, as it’s an accumulating etf, the distributed income is zero so the tax is yearly on ‘reported income’. So my question is whether in an accumulating ETF's ‘reported income’ is the same as it's ‘change in NAV’ (i.e. (i) above). Unless they can be somewhat equated (obviously over the ownership period..) I can’t offset Italian tax against UK tax and will get double taxed!

From a UK investors perspective even though a fund reinvests dividends there will still be reportable income (the amount of dividends reinvested). After paying tax on that reportable income to HMRC the cost of stock needs to be revised (upwards) so that you don't later pay more capital gains to HMRC than you should when you sell the stock. If the fund isn't UK reporting registered HMRC will consider all income/gains to be income (and you can't use capital losses to offset capital gains elsewhere), taxed at your marginal tax rate. Provided the offshore fund is UK reporting registered then capital/losses gains will be considered as capital gains/losses by HMRC.

I use stocks for growth, bonds for income/drawdown (2 basket) and avoid stock dividends. For bonds I just hold domestic. Which greatly simplifies matters. I anticipate net total stock reward that compares to the gross total index reward.

Micks
Posts: 312
Joined: Tue Feb 25, 2014 12:38 pm
Location: the Netherlands
Contact:

Re: Net total return vs. gross total return - a fallacy?

Postby Micks » Fri Sep 25, 2015 2:09 pm

Federiko25 wrote:Galeno,
Shouldn't a NRA still pay income tax on the semi-annual IUAG distrubution?


Federiko, the IUAG ETF is domiciled in Ireland, which means its distributions are not taxed by the Irish government (see also the wiki for this). The wikipage in brackets also shows you this fund does not have tax leakage coming from the interest it receives from its bonds. This is all assuming you are not an Irish investor, in that case Irish income taxes etcetera may apply.

You may of course be taxed by your own government on distributions/cap gains/reinvested dividends.
Last edited by Micks on Fri Sep 25, 2015 11:37 pm, edited 1 time in total.
Kind regards, Mick

User avatar
galeno
Posts: 1003
Joined: Fri Dec 21, 2007 12:06 pm
Location: Alajuela, Costa Rica

Re: Net total return vs. gross total return - a fallacy?

Postby galeno » Fri Sep 25, 2015 4:28 pm

For some reason IUAG's interest income comes thru without any taxes. Compared to a USA-domiciled investor using AGG (AGG=IUAG), the IUAG holder gets AGG's SEC yield minus the difference in ER (0.25% - 0.08% = 0.17%). So the TER of IUAG will always be 0.25% no matter what interest rates go.

Assuming a SEC rate = 2.00%, if we held AGG instead of IUAG, it would cost us 0.08% + .60% = 0.68%. With a SEC rate = 3.00% the cost goes to 0.98%.

Federiko25 wrote:Galeno, Shouldn't a NRA still pay income tax on the semi-annual IUAG distrubution?
AA = 40/55/5. Expected CAGR = 3.8%. GSD (5y) = 6.2%. USD inflation (10 y) = 1.8%. AWR = 2.8%. TER = 0.5%. Port Yield = 2.0%. Term = 35 yr. FI Duration = 6.2 yr. Portfolio survival probability = 100%.


Return to “Investing - Theory, News & General”

Who is online

Users browsing this forum: bayview, Big Dog, Dead Man Walking, Jacotus, jocdoc, Oblivious, Random Walker and 76 guests