Direct indexing for income

[stevesemire]

As a notoriously bad stock picker,I am very drawn to the idea of direct indexing and looked into several indices worth replicating. For a diversified income strategy,direct indexing of the "msci world high dividend yield" appears to be ideal. The index selects 300 global companies, screens for quality and yield,weighs according to market cap and offers a healthy 3% yield. It has a long track record and has outperformed its parent index,the Msci World, over the last 25 years. Unlike all of these garbage dividend etfs that focus exclusively on high yield,it seems to be a well constructed dividend index. You can find the constituents and their weighings at msci.com/constituents. I could buy the 300 companies at my cheap discount broker that offers 1€ or even free trades. I would rebalance once a year with saved up cash and avoid the 0.4% fee the iShares etf has for this index. Am I stupid? Missing something?

[Chris]

Record keeping and tax reporting would be more burdensome, for one. And if you want to track the index as closely as the ETF does, you'll be making a lot of trades per year.

[stevesemire]

@Chris. Thank you for your response. Please excuse my newbie question: Sure,tracking a market weighted index closely would be quite tricky but,according to msci,the index just rebalances twice per year. Therefore I assumed that no further changes are made.

[Dream of Freedom]

I've heard talk of the idea, but never heard from someone who actually did something like it. You would save $400 for every hundred thousand invested. If you go through with it I would be interested in seeing an update about your experience.

How do you plan to deal with dividends? If one of the companies gets bought out and you get cash how do you reinvest it quickly? Do you need a broker that will handle fractional shares? I imagine it is hard to get a very precise percentage of a stock in your portfolio if that stock is high priced. Luckily Amazon shouldn't be in a dividend index. It's $3,443. It would increment like $0, $3443, $6886, and so on.

[jacob]

There are some thread(s) around here with manual implementations of indexing. I think someone (@slowtraveler?) even made one eventually. See e.g. viewtopic.php?t=8282

[Dream of Freedom]

Mr. slowtraveler's last message in that thread was just saying he would do it. A message from another thread yields

My experiment is far more volatile than the vanilla index fund. This is expected from Siegel's paper but interesting to experience. I'm up 3% relative to the index now. The reverse was true last week. Got surprised by my cost basis changing for Abbvie upon a merger with AGN. Debating on whether to direct index the proceeds or dump them into the 2 new S&P 500 companies that both seem richly valued.

That was months ago. It was only one small post. Most of the details and even the morals of the story are left blank.

[jacob]

I'm pretty sure there was another thread or post on it and how it was implemented in detail (the hundreds of trades). I can't find it though.

[Lucky C]

The MSCI World High Dividend Yield index is naturally more concentrated than MSCI World index. The top 10 holdings are 26% of the index. It's likely that if you only held the top 30 in the index rather than shooting for all 300, you would be >90% of the way to replicating the index with only 10% of the trouble. It would also be much easier to replicate with whole shares rather than trying to add the 300th stock at less than 0.1% of your portfolio (if you are doing cap weighted).

Some observations on the history of the index vs. its parent MSCI World Index:
- MSCI World High Dividend Yield (WHDY) data starts in 1995 (from what I can see), and total return pretty much matches the MSCI world from 1995 to the 2000 tech bubble peak. WHDY outperformed World for most of the late 90's but then the tech bubble caught up such that they were even again in 2000.
- WHDY easily beat World during the downturn from 2000 to 2003, just as any fund that didn't hold glamorous tech stocks would have done. E.g. the DJIA, or any value investor, could easily beat the NASDAQ or even S&P 500 from 2000 to 2003.
- WHDY returns from 2003 through 2012 (last year in the Excel data set I was able to find) match MSCI World returns very closely. They look like nearly identical indices for that time period.
- WHDY returns from 2013 through 2017 still closely match MSCI World returns based on the MSCI fact sheet.
- WHDY has trailed MSCI World since the early 2018 melt-up peak, since generally growth/momentum stocks have been beating value/dividend stocks since then. See for example the close tracking of VOO (Vanguard S&P500) vs. VTV (Vanguard Value) up until 2018 when they also started to diverge.

Also, the MSCI World High Dividend Yield index has only been "live" tracked by MSCI since June 2006. Data before that (starting 1995) has been back-calculated. If you were to ignore back-calculated data and only look at performance since 2006, you'll see (based on the PDF linked above) that its total return has trailed the MSCI World index. The entirety of its "25 years" of outperformance can be attributed to the 1995-2003 (back-calculated) run where it beat the World index that had been skewed by the tech bubble and its crash. Again, any fund that didn't hold the popular dot-com stocks over that period should have also done well vs. the MSCI World.

So my conclusions are:
- This index should beat the World index when value outperforms growth, e.g. when a tech bubble bursts, or generally when you start in a period when growth/momentum is much more popular than value, followed by a period when value becomes more popular relative to growth/momentum.
- This index should perform about the same as any type of value-oriented global portfolio that doesn't hold extremely popular and overvalued tech stocks. I think that is more important than the amount of dividend being paid or the quality factor, at least if we are really in another tech bubble that will burst in the coming years.
- Assuming you're in the US, it would be easy to do individual stocks for US companies but more limited for international stocks. You could instead have a portfolio of US stocks plus international funds and overall still have very low expenses. If you wanted to match total world stock indices you would have something like 60% US value or dividend stocks, 40% in an EAFE value/dividend fund, and 10% in an EM value/dividend fund. For the US stocks you could just do the top 20 or 30 within the MSCI WHDY index, and then just pick ETFs for international and not worry about it. I think the total return of such a portfolio would match the MSCI WHDY index very well, and the tracking error would probably be just as likely to go in your favor than against you, based on the historical data I've seen.

[stevesemire]

Thank you all for your very informative feedback. Thank you Jacob for slowtraveler's link to the s&p500 experiment. I agree with your views on the index' past outperformance. Still I found it a remarkably well constructed global dividend index: given that it almost excludes IT and just holds 1/6th (300) of the holdings of its partent index (1700) it has a good track record that matches the Msci World. That fact that even Vanguard uses it in their research (See their paper/PDF: "Analysis of dividend oriented equity strategies") makes me at least a little more confident that it is a dividend index worth considering.

After having read slowtraveler's direct indexing experiment I see why it makes sense to create an equally weighted replication of the index. It is easier to set up and given that small-cap is expected to outperform large-cap (fama-french) equal weight is a way to maybe capture beta+ performance. Also you will not crash into each and every bubble as with market-cap weight.

Avoiding fees is not the only reason for my direct indexing idea. I am from the EU and thus regulations,taxes and costs around funds have been all over the place and change constantly. I guess holding the stocks directly,without the wrapper,gives me more freedom. Also,I do not want to be dealing with total return and then swr for my country. In Germany the swr is below 1% because we suffered from hyperfinflation in the 20s and 30s (See Wade Pfau's research on global swr). This data is included and thus calculated as swr. I'd much rather avoid swr discussions,focus on income and geo arbitrage when the shit hits the fan.

Thank you again for all your time and feedback. Much appreciated!

[IlliniDave]

@Chris. Thank you for your response. Please excuse my newbie question: Sure,tracking a market weighted index closely would be quite tricky but,according to msci,the index just rebalances twice per year. Therefore I assumed that no further changes are made.

Actually, it is the opposite. If you track a market-weighted index by using the same weighting as the index, your portfolio will be self adjusting relative to the index with no action required. Until they redefine the index (change constituents or change weighting strategy). That's why when you look at index finds they typically have much lower turnover than other funds and lower cost.

I'm not real familiar with the indices discussed. Sounds like they are more complicated to track, which is probably why they cost more than the typical 0.1% or less for a simpler index. In my 401k we have a MSCI ACWI ex-US fund whose fee is 0.0455%. Is there an similar index fund that would give you the ex-US exposure you want that could be combined with a US total market fund where you could rebalance between the two yourself (if even desired)? Seems like you could lose 3/4 of the 0.4% fee that way and get yourself at or below 0.1% using funds.

But that's probably less fun.

If I ever tried to build my own index I would equal-weight, taking on the hassle with the hope of beating the index.

[stevesemire]

Actually, I wonder why the cost for this fund is that high. It only covers developed markets and only buys highly liquid stocks (T,PG,KO etc). I have the impression that in this low interest environment,"dividend","income" and "yield" have become nice marketing labels for investment advisors to justify the price tag. There is a cheaper fund option here that tracks the same index but its tracking difference is horrible (I guess I could do better;-)) and its aum is so low that I would not want to have the risk of them shutting down the entire fund because it is not profitable.

The whole reason for this experiment might not seem rational to many of you but somehow, I just do not buy into the whole (almost mainstream) advice of just buy vt/bnd and take out 4%. No offense,bogleheads, I am really open for anyone's input but thinking that ETFs as a vehicle are "forever",as many suggest,seems absurd to me in such a rapidly changing market. I want the freedom,control and flexibility and I guess that is why I want to hold the assets as directly as possible.

Anyway,I am getting off track here. I am open for your feedback,criticism and highly appreciate the time,effort and work many put into this wonderful forum. As suggested,I might start buying the top stocks of the index first and work my way through its holdings. If you are interested,I'll keep you all updated. These zero cost brokers nowadays are definitely fun if you like a little creativity :p

[classical_Liberal]

,,,

[jacob]

Actually, I wonder why the cost for this fund is that high. It only covers developed markets and only buys highly liquid stocks (T,PG,KO etc).

Foreign investments has overhead than a domestic ditto. What is liquid to a one-person retail investor buying 500 shares is not liquid to a fund that needs to buy 400% of the daily volume, say. This requires specialized block traders. Typically when you're moving that much money, you need an account either at a very big bank or you need to buy your own seat at a given exchange. All this + the added accounting and taxwork for investing internationally adds up. This is why international funds cost rather more than domestic equivalents. These funds are basically paying for services that retail investors have come to expect to get for free---and when going international this cost accumulates for each country they enter. For an active fund that needs a new team of analysts for each new country, it gets even worse.

[IlliniDave]

Foreign investments has overhead than a domestic ditto. What is liquid to a one-person retail investor buying 500 shares is not liquid to a fund that needs to buy 400% of the daily volume, say. This requires specialized block traders. Typically when you're moving that much money, you need an account either at a very big bank or you need to buy your own seat at a given exchange. All this + the added accounting and taxwork for investing internationally adds up. This is why international funds cost rather more than domestic equivalents. These funds are basically paying for services that retail investors have come to expect to get for free---and when going international this cost accumulates for each country they enter. For an active fund that needs a new team of analysts for each new country, it gets even worse.

Makes and I guess that's why the .0455% we get charged is about twice what they charge for an equivalent US market fund. Those are more-or-less institutional rates (Megacorp has clout), I should add.

[IlliniDave]

The whole reason for this experiment might not seem rational to many of you but somehow, I just do not buy into the whole (almost mainstream) advice of just buy vt/bnd and take out 4%. No offense,bogleheads, I am really open for anyone's input but thinking that ETFs as a vehicle are "forever",as many suggest,seems absurd to me in such a rapidly changing market. I want the freedom,control and flexibility and I guess that is why I want to hold the assets as directly as possible.

Okay, my bad. You started off by saying you felt you were a lousy stock picker and proceeded to list some hefty fees you seemed to want to avoid in an etf that interested you so I pointed out an avenue for lower-cost, non-picking alternatives. Etfs are no more or less forever than stocks. Dunno where you heard that, but if it was at bogleheads.org either someone misspoke or something got lost in translation. There's nothing magic about index funds. Their secret sauce is that they are cheap and relatively tax efficient outside retirement accounts. The fundamental premise is to get as much of the full return for the risk you take with your money with the least getting siphoned off by the middle men.

That said, one of the most important boglehead principles is SWAN (sleep well at night). Any strategy that helps avoid behavioral errors is better than one that makes you nervous. If you can avoid the errors, DIY should be about as cheap as it comes.

I share some of classical_Liberal's thoughts regarding non-correlation. I would be wealthier if I stuck with the total stock/total bond bare bones "3-fund" strategy, but within the bounds of what I can readily invest in, I do some contrarian maneuvering (i.e., buy what's headed down and sell what's headed up). Hurts the bottom line but does seem to dampen volatility.

Good luck.