The Money Mustache Community

Learning, Sharing, and Teaching => Investor Alley => Topic started by: germandude on June 09, 2013, 11:14:14 AM

Title: Building your own index fund?
Post by: germandude on June 09, 2013, 11:14:14 AM
Hi,

In Europe we can`t buy Vanguard index funds and the cheapest index fund for the MSCI World has a TER of 0,45%. And it has additional risk because it is not 100% physical replicated and stocks can be borrowed, so there is additional counter-party risk.
Because of all this hassle, i came to the conclusion that it could be not so wrong to build my own index fund and simply buy the stocks that represent around 80% of the marketcap of the MSCI World or S&P500. I want to equal weight the stocks in the beginning and cap at 5% weight of the complete portfolio to minimize trading costs. 80-100 Stocks should be enough to reach that goal and with around 2000-3000€ per position the commission for buying the portfolio should be around 0,5%. After that the TER is 0% and i have the option to hold all stocks in my name. :)

Any thoughts?
Title: Re: Building your own index fund?
Post by: kyleaaa on June 09, 2013, 11:56:37 AM
Vanguard does have some Ireland-domiciled index funds and ETFs with expense ratios significantly below 0,45%. I know of a few people living all over Europe who successfully own them.

I also know of some people who do what you plan to do. There are problems with the approach, but it's workable and possibly more tax-efficient in the long run.
Title: Re: Building your own index fund?
Post by: germandude on June 09, 2013, 12:31:38 PM
Thanks i didnt know about the vanguard funds from ireland. I looked them up and found the S&P500 ETF with a TER of 0,09% and the world ETF with 0,25%.
But the problem i have with them is that they are currently only traded in Berlin and they have only assets under 100 million $ in that fund. That is too risky for me to invest 6 digit sums into it, i won`t feel that good with that way. Hopefully they grow larger in europe the next years to become a real alternative to ishares.

Can you tell me more about the problems that arise with my approach?
Title: Re: Building your own index fund?
Post by: Reepekg on June 09, 2013, 02:29:35 PM
Can you tell me more about the problems that arise with my approach?

Most people don't have enough money to cost-effectively own equities in an entire index. As an example, buying the S&P 500 at $10 a trade would cost $5000 to set up which is a 5% expense ratio on a $100k portfolio. Every time you rebalance, you are facing another $5000 in expenses.

Basically, I see one problem being that unless you have free trades, you can't cost effectively follow the index. If you save by rebalancing less frequently, you aren't tracking the index properly. The second problem is that it is a pain to constantly track the performance of all the equities, including which are added and dropped from the index. The point of the index fund is this is all automated and trading fees are a tiny percentage of a billion dollars.
Title: Re: Building your own index fund?
Post by: aclarridge on June 09, 2013, 04:00:11 PM
Can you tell me more about the problems that arise with my approach?

Most people don't have enough money to cost-effectively own equities in an entire index. As an example, buying the S&P 500 at $10 a trade would cost $5000 to set up which is a 5% expense ratio on a $100k portfolio. Every time you rebalance, you are facing another $5000 in expenses.

Basically, I see one problem being that unless you have free trades, you can't cost effectively follow the index. If you save by rebalancing less frequently, you aren't tracking the index properly. The second problem is that it is a pain to constantly track the performance of all the equities, including which are added and dropped from the index. The point of the index fund is this is all automated and trading fees are a tiny percentage of a billion dollars.

Yeah even if you could get the transaction costs down such that rebalancing/adding funds isn't prohibitively expensive, actually doing the rebalancing with 80-100 stocks would be an extremely tedious task unless you had a sophisticated trading platform that was either made for that purpose or was extremely easy and fast to use. I know that with my discount brokerage, it would take me several hours to do it, during which time all the stocks would be moving around in value screwing up my calculations - minor problem but still.

But yeah the transaction costs would have to be near-zero...typically you would be adding funds in smallish chunks, maybe 5-20k. So that's like $50-$200 per stock. Even at, say, $1/trade, this seems wayyyy too expensive - and it doesn't even take into account the bid-ask spread costs!
Title: Re: Building your own index fund?
Post by: germandude on June 09, 2013, 10:06:46 PM
I dont want to buy the complete index and i only want to cap at 5%, so transaction costs after setup are minimal. My approch is not to match the index exactly, why should i?
Going with the pareto approach 80% of the marketcap of the index should be enough. The DOW JONES only has 30 stocks and gives nearly the same performance as the S&P 500, so for the world 80-100 stocks should be enough.
Title: Re: Building your own index fund?
Post by: Khan on June 10, 2013, 03:34:24 AM
And 80-100 stocks, at 10$ a trade is 800-1000$ just on that alone. It's just not worth the effort.

Also, the S&P 500 is capitalization weighted, meaning that you're still going to have to go through the trouble of figuring out total capitalization of the stocks you're interested in, and then % allocation to each stock.

Edit: S&P 500 is float weighted, which is just a further branch of capitalization weighting AFAICT.
Title: Re: Building your own index fund?
Post by: germandude on June 10, 2013, 04:32:59 AM
If i buy an index fund for 300.000€ i pay 1350€ every year through the hidden TER. And buying the index fund will cost also. At my broker this will result to around 300-400€ or more because he will split this amount into more orders.
And then there is the counterparty risk. If we talk about holding the assets forever, i want the least risk of losing the assets as possible. When holding an index fund there are at minimum three parties involved which can go broke or simply steal your assets. Read about MF Global.
Perhaps i am a bit paranoid about that risk, but we are talking about lifetime savings here.
Title: Re: Building your own index fund?
Post by: Christof on June 10, 2013, 05:31:30 AM
ETFs are Sondervermögen in Germany, so there's protection against insolvency of either the broker or the depot managing bank. It doesn't protect against illegal activities such as selling all shares of an index fund and then filing insolvency. I don't see, however, how holding shares would be different. You still have a bank that manages your depot and a second that manages shares with potential to sell your shares unbeknownst to you.

I also don't understand why managing short of 100 million Euros is not sufficient? What is the risk that you see here?
Title: Re: Building your own index fund?
Post by: germandude on June 10, 2013, 08:57:14 AM
You still have a bank that manages your depot and a second that manages shares with potential to sell your shares unbeknownst to you.

I also don't understand why managing short of 100 million Euros is not sufficient? What is the risk that you see here?

I can diversify that risk with using 3 or more different banks and/or i can choose to hold the stocks in my name and not the name of the bank. This costs extra money, so for now i live with this risk to loose 1/3 of the assets. I don`t want to invest in funds that are not liquid, because they can easier be closed and shut down and it is harder to sell them at the stock exchange.

And we haven`t talked about the risk of losing assets because the fund borrows them for money. Most of the profit of borrowing goes directly to the fund company, so they have a strong argument to do it. But this adds the greatest risk of all, because you cannot be sure if your assets are even there. When living in US i can imagine to buy and trust Vanguard`s funds, but it is solely on trust. But i can`t really trust a postbox in ireland.
Title: Re: Building your own index fund?
Post by: Christof on June 10, 2013, 10:02:55 AM
Do you insist on the MSCI World index, or would another index or a combination do, such as MSCI Europe/USA/Emerging Markets? That is offered by ETFlab, a subsidiary of the Sparkassen, at 0.3% TER as a fully replicated fund (no borrowed assets). Or the iShares MSCI World Minimum Volatility index that is offered by iShares, a UK-based company for 0.3% TER.

Getting both via different brokers would minimize your risk similar to replicating a partial index yourself. Holding stocks in your name and physical deliver them is more expensive then holding a number of index funds via different brokers at 0-5-0.8% TER.
Title: Re: Building your own index fund?
Post by: germandude on June 10, 2013, 10:36:40 AM
I want to diversify at minimum over usa/japan/europe. I know the ETFs of ETFLab, they are under the best ETFs available for a european citizen. But even they have borrowing agreements in every ETF they offer, even the physical replicated ones. In my calcuation it is much cheaper to buy 100 stocks and hold them for 20 years (around 0,5% onetime), then to buy index funds (for 0,25% commission) and pay every year 0,3%+its safer. But this is only true for big portfolios and when you have no problem if the performance can differ from the index. (sometimes more than the index sometimes less, because i have less smallcap stocks)
Title: Re: Building your own index fund?
Post by: kyleaaa on June 11, 2013, 09:16:58 AM
Just be aware that by doing it this way, you will effectively be creating a value index that will behave differently than the S&P 500, MSCI Global, etc.
Title: Re: Building your own index fund?
Post by: Gremlin on June 12, 2013, 04:41:48 AM
This is effectively my investment strategy.  I buy packets of $10,000 shares (Australian).  Each time my investment "fund" hits $10k, I buy a parcel of the individual stock that best rebalances my portfolio towards the ASX200 index.  My transaction costs are lower than a Vanguard fund and it seems to work well enough for me.
Title: Re: Building your own index fund?
Post by: pom on June 13, 2013, 08:03:53 AM
I thought of doing the same thing. It is definitely doable and it is true that with a large portfolio these 0.3% end up costing you quite a bit.

As long as you are comfortable that while you have the same expected return, day-to-day and year-to-year it will differ from that of the index, I don't see the problem.

One point that I want to make is that taking the highest market cap might not be the best thing to do. That is not how the DJIA is built, they build it by looking at the main industries in the country and chose representative stocks of that industry.

Maybe a way to do it is through a random sample of large caps, middle caps and small caps and adjust the weights to more or less replicate the respective weights of the category in your total index. Then every year or two you see if your weights became unbalanced and re-adjust.

Anyway, I think your idea makes perfect sense, It adds a little bit of risk while increasing return a little bit. It will required discipline though.
Title: Re: Building your own index fund?
Post by: Gremlin on June 13, 2013, 10:08:58 PM
I thought of doing the same thing. It is definitely doable and it is true that with a large portfolio these 0.3% end up costing you quite a bit.

As long as you are comfortable that while you have the same expected return, day-to-day and year-to-year it will differ from that of the index, I don't see the problem.

One point that I want to make is that taking the highest market cap might not be the best thing to do. That is not how the DJIA is built, they build it by looking at the main industries in the country and chose representative stocks of that industry.

Maybe a way to do it is through a random sample of large caps, middle caps and small caps and adjust the weights to more or less replicate the respective weights of the category in your total index. Then every year or two you see if your weights became unbalanced and re-adjust.

Anyway, I think your idea makes perfect sense, It adds a little bit of risk while increasing return a little bit. It will required discipline though.

Yes, that's pretty much what I do.  I look at the mix by market cap and industry in my portfolio relative to the ASX200.  If I'm underweight mid-cap and underweight Materials, it'll prioritise a mid-cap Materials stock.  I don't automatically reinvest dividends in the underlying stocks so I'm getting a regular stream of income to choose where to reinvest, which helps the rebalancing.

Sigh... I'm a nerd.

EDIT:  just to add that it also gives me flexibility to manage my after tax returns better.  If selling A and buying B leaves me with essentially the same portfolio mix relative to the index but a better tax position, then net of tax I'm better off than in an index fund.
Title: Re: Building your own index fund?
Post by: pom on June 15, 2013, 10:30:51 AM
I got curious to see when it is better to just create your own index vs a pro one so I did some simulation for the fun of it using the S&P500 and assuming:
Index fund fees of 0,05%
$8 a trade (I use e-trade)
50 trades a year would be needed after the initial 500 trades
0,05% bid-ask spread

I assumed that you would need 50 trades a year due to changes to the index and reinvestment of dividends. You could argue that you need 500 trades to be perctecly in-sync with the index but I think that as long as you are over 95% matched, it is ok.

My conclusion is that if you have above $700 000 and an horizon of 30 years, it is better to buy all of the stocks in the index.

Using interactive broker and its $1 trades, you would be better off with creating your own index fund even with only 100 000.

Admitedly I did not take into account the time it takes to manage that. Probably a good day or two to set it up and  then half day every year.
Title: Re: Building your own index fund?
Post by: germandude on June 15, 2013, 01:49:21 PM
Nice and that with 0,05% TER? With 0,3% TER it then should be better above 120.000$. Thank you for your calculation.
Title: Re: Building your own index fund?
Post by: grantmeaname on June 15, 2013, 02:40:04 PM
Do you guys think index fund fees will stay static at .05% and trades will stay that cheap over the lifetime you did the comparison?
Title: Re: Building your own index fund?
Post by: Reepekg on June 15, 2013, 10:31:43 PM
Do you guys think index fund fees will stay static at .05% and trades will stay that cheap over the lifetime you did the comparison?

This is actually a really interesting question. In my very short investing experience, I've only seen a downward trend in both ERs and cost per trade. I would guess it would continue to get cheaper through competition and tech improvements. Is there a more expensive future scenario to worry about?
Title: Re: Building your own index fund?
Post by: grantmeaname on June 16, 2013, 07:39:09 AM
I'm wondering if there's a minimum price for either such that one will plateau and the other will continue to decline, not necessarily an increase.
Title: Re: Building your own index fund?
Post by: pom on June 17, 2013, 02:31:02 AM
Hi Grantmeaname,

That is true that I should have considered the trend, which I didn't. If both continue to decrease at the same rate (which would be a decent guess), most likely the breakeven amount will be higher due to the initial 500x8 investment taking more time to amortize. I don't think that it would go to up all that much unless we assume a very quick decrease.

I am really mulling over doing it (by first switching to $1 trades at interactive brokers). I have about 300k invested in the S&P500 index and while 0.05% doesn't sound like much, it is still $150 a year. I need to think about it a little more but saving half of that amount would be badass.
Title: Re: Building your own index fund?
Post by: Khan on June 17, 2013, 05:27:03 AM
Except pom, then you wouldn't be tracking the index, and you would have to pay more attention to it.

How many hours are you willing to spend doing it? How much is that time worth to you? If you choose wrong with your investments, and some of them go sour, are you willing to take an entire-portfolio view of your investments and hold the course vs. holding the index? Can you sleep well at night with those choices?

The moment you step away from an index you -have- to ask yourself these questions, and .05% is NOT worth making these considerations for. I don't primarily index invest because I enjoy doing what I do. Even then, I'm starting to look into value and dividend funds, but those do have higher ER's then .05%, but they would probably provide me as much piece of mind and I will be putting much of my further funds towards such ends.
Title: Re: Building your own index fund?
Post by: pom on June 17, 2013, 07:09:33 AM
I can accept some tracking error with my replicated index (even Vanguard has a tracking error though very small).

That would be as long as:
- The error is very small. For exemple, that I track 95% of the index return.
- The error is unbiaised. In that same exemple, the 5% remaining is "random". By that I mean that there are equal chancess that it will above the index than there are that it will be under.

So if the index does +10%, I would be happy with a portfolio that generates a mean return of between 9.5% and 10.5%.

I am pretty disciplined (I assume most people here are) so I am not worried about having a portfolio view. Also with 300k, the largest positions would be 9k in AAPL and XOM ... I don't think that I would panic even if one decreases 50%.

Your other question is the one that makes me hesitate: time.  I would guess that I will have to spend half to a full day of "work" every year to compare my holding to the index and rebalance the parts that need rebalancing. It might be fun and keep my mind busy though.
Title: Re: Building your own index fund?
Post by: grantmeaname on June 17, 2013, 07:14:58 AM
Hi Grantmeaname,

That is true that I should have considered the trend, which I didn't. If both continue to decrease at the same rate (which would be a decent guess), most likely the breakeven amount will be higher due to the initial 500x8 investment taking more time to amortize. I don't think that it would go to up all that much unless we assume a very quick decrease.

I am really mulling over doing it (by first switching to $1 trades at interactive brokers). I have about 300k invested in the S&P500 index and while 0.05% doesn't sound like much, it is still $150 a year. I need to think about it a little more but saving half of that amount would be badass.
I wasn't saying it was a critical flaw, just that it was an interesting question that probably matters a lot over a time period as long as thirty years.
Title: Re: Building your own index fund?
Post by: pom on June 17, 2013, 08:35:37 AM
Grantmeaname,

It's cool, I actually appreciate that you took the time to challenge my model. Sometimes I get caught up in the calculation process and forget that my assumptions might be flawed.
Title: Re: Building your own index fund?
Post by: grantmeaname on June 17, 2013, 08:39:37 AM
What would be difficult, but very interesting, would be to go dig up old Vanguard and e-Trade press releases about expense ratio/trade price reductions, then plot both over time and see which is decreasing faster and whether either has leveled off.
Title: Re: Building your own index fund?
Post by: kyleaaa on June 20, 2013, 12:59:04 PM
Do you guys think index fund fees will stay static at .05% and trades will stay that cheap over the lifetime you did the comparison?

It's pretty inevitable that index funds will eventually have ERs of 0.00%, at least in the US and for popular core asset classes. Management companies can earn a decent profit off securities lending alone once the asset base grows large enough. I've heard of some management companies already experimenting with this.
Title: Re: Building your own index fund?
Post by: dmn on June 27, 2013, 08:17:39 AM
In Europe we can`t buy Vanguard index funds and the cheapest index fund for the MSCI World has a TER of 0,45%. And it has additional risk because it is not 100% physical replicated and stocks can be borrowed, so there is additional counter-party risk.
Because of all this hassle, i came to the conclusion that it could be not so wrong to build my own index fund and simply buy the stocks that represent around 80% of the marketcap of the MSCI World or S&P500. I want to equal weight the stocks in the beginning and cap at 5% weight of the complete portfolio to minimize trading costs. 80-100 Stocks should be enough to reach that goal and with around 2000-3000€ per position the commission for buying the portfolio should be around 0,5%. After that the TER is 0% and i have the option to hold all stocks in my name. :)

I like this idea a lot. Apart from the lower TER, it may also be more tax-efficient in the long run, because you are not forced to realize capital gains when e.g. your ETF is closed down. (Capital gains taxes in Germany are 25% and there is talk about increasing them to 32%.) You can also reduce transaction costs by not selling and buying stocks whenever the composition of an index changes.

I am not sure about how much capital this approach requires initially. Is there any problem with doing this during the accumulation phase, if I can buy a new stock for 2000-3000$ every two months? Sure, my portfolio will be more volatile in the beginning, but will the fluctuations not mostly cancel out anyway? After two years I would have 12 stocks, surely at least some of them would do reasonably well. I haven't done the math though, so maybe I underestimate the volatility of 12-stock-portfolios.


It's pretty inevitable that index funds will eventually have ERs of 0.00%, at least in the US and for popular core asset classes. Management companies can earn a decent profit off securities lending [...].

Compared to the index, you still have a lower expected return due to the expected losses from counterparty risk. (This risk may or may not materialize during your lifetime, but your *expected* return is impacted a bit.)
Title: Re: Building your own index fund?
Post by: pom on June 27, 2013, 08:32:07 AM
I suspect that with 12 share you will have quite a bit of non-systemic volatility but if you can stomach the added risks and work ... I think this may be a good way to go.
Title: Re: Building your own index fund?
Post by: Christof on June 27, 2013, 11:56:19 PM
Capital Gain taxes in Germany are subject to Solidarity Surcharge tax and Church tax, so the total is closer to 28%.
Title: Re: Building your own index fund?
Post by: kyleaaa on June 30, 2013, 05:58:30 PM
Compared to the index, you still have a lower expected return due to the expected losses from counterparty risk. (This risk may or may not materialize during your lifetime, but your *expected* return is impacted a bit.)

Before I call this completely false, could you please describe the mechanism whereby this would theoretically take place? Certainly nothing like this occurs with standard mutual funds or even ETFs. Perhaps this is something that's only an issue in Germany/Switzerland/wherever due to peculiarities of the legal structure of investment companies over there. In the US, if my mutual fund company goes bankrupt, I'm not even slightly impacted. There is no risk I would lose any money at all.
Title: Re: Building your own index fund?
Post by: Christof on July 01, 2013, 12:12:40 AM
To reduce TER funds lend shares to other market participants. Those parties sell shares and later buy them back hoping for a lower price before returning them to the fund. If the third party goes bankrupt during the lending phase, the shares are lost.
Title: Re: Building your own index fund?
Post by: dmn on July 01, 2013, 02:11:01 AM
If the third party goes bankrupt during the lending phase, the shares are lost.

As a safety mechanism, the third party gives some collateral to the fund during the lending phase. When the third party goes bankrupt, the fund can sell the collateral and rebuy the shares.

Losses can only be incurred if the third party goes bankrupt and
(a) the borrowed shares' price goes up, or
(b) the collateral's price goes down
so that the collateral is worth less than the shares that were lost.

Due to the collateral, people are usually not worried that ETFs will sustain huge losses from lending shares, but some losses may be incurred during serious market turmoil, during which sudden bankruptcies are more likely and the supposedly safe collateral may become more volatile.