Gyroscopic Investing

This whole document put out by Credit Suisse is pretty incredible:
http://www.investmenteurope.net/digital ... al_web.pdf

However, I would like to draw your attention to the country breakdown that starts at page 35. It is fascinating how many individual countries have had terrible performance in terms of financial market performance whether it be equities, bills, or bonds. What is equally fascinating is how risky bills and bonds appear to be when extended to a long time horizon.

Perhaps I am being seduced by a rear view mirror look, but a global cap weighted VP looks very compelling.

melveyr wrote: This whole document put out by Credit Suisse is pretty incredible:
http://www.investmenteurope.net/digital ... al_web.pdf

However, I would like to draw your attention to the country breakdown that starts at page 35. It is fascinating how many individual countries have had terrible performance in terms of financial market performance whether it be equities, bills, or bonds. What is equally fascinating is how risky bills and bonds appear to be when extended to a long time horizon.

Nice find!

The charts you point out show the impacts of wars, revolutions, etc. in the changing world. Markets are not stable. Even more, they can become unstable very quickly and leave investors with little time to react. They also can remain flat, or declining, for very long times.

When you live in a country that has been stable for your lifetime, it's easy to think it must always be that way. But history has shown repeatedly that volatility and uncertainty are the norm, not stability. You've got to diversify.

Perhaps I am being seduced by a rear view mirror look, but a global cap weighted VP looks very compelling.

No I think you are right. I think cap weighted indices are the best way to invest because they rely on the wisdom of the markets to allocate resources where they should be on a global scale. You take the gambling element out of it. Just because an investor thinks Venzuela is the next hot market, does not mean the rest of the planet thinks that way. In many cases, the rest of the planet is probably right.

I would like to do more eventually with a framework for a global permanent portfolio. I think that a cap weighted world index fund like Vanguard Total would actually work quite fine for the stock portion of the PP even today. The gold could also remain. But the cash and bonds are a problem. An investor's life is under total influence of the central bank where they reside. So diversifying out of your home currency into the bonds/cash of other regions could cause just as much harm as benefit. It's an interesting thing to think about.

Thanks for your post...

craigr wrote:
I would like to do more eventually with a framework for a global permanent portfolio. I think that a cap weighted world index fund like Vanguard Total would actually work quite fine for the stock portion of the PP even today. The gold could also remain. But the cash and bonds are a problem. An investor's life is under total influence of the central bank where they reside. So diversifying out of your home currency into the bonds/cash of other regions could cause just as much harm as benefit. It's an interesting thing to think about.

Thanks for your post...

Craig,

I have also been interested in the idea of a global PP. Betting solely on your own country reminds me of strictly buying the stock of your employer for your retirement account.

One global PP I have thought about is...

25% VT (Global stock index)
25% BWX (International intermediate sovereign debt)
25% TLH (Domestic intermediate sov debt)
25% GLD (Gold)

I subbed intermediate bonds for the barbell just for simplicity.

On a short term basis the allocation appears riskier than a domestic PP when you translate the return into dollars because of the foreign currency exposure. However, it would insulate you from a local currency crisis more so than a vanilla PP.

Additionally, if you changed the bond weightings to be based off of equity weightings (so the US is 40% of the global equity market so US bonds would be 40% of the PP's bonds) than the allocation plan would naturally adjust over time as the power structure of the world changed. It is certainly a lot to think about.

The sticking point really is going to be when you buy foreign debt, you are at their mercy. If everyone engages in a race to the bottom, you'll get trampled. It has been the case in the past with countries having debt problems that the first thing they do is backstab the foreign debt owners (who can't vote and have no political power).

Really I can only think of a handful of countries on the entire planet that I'd want to own in terms of foreign debt. Even then, not being aware of the full societal political situations in those places, it's still a risk. Off the top of my head in terms of countries who have bonds that are worth considering that are non-U.S.:

- Germany
- Switzerland
- Norway/Sweden/Finland/Denmark
- New Zealand/Australia
- Japan
- Netherlands
- U.K.
- Canada

They are *not* in order, again just thinking out loud on the list and then I'd need to whittle it down further. This list basically reflects my view on what non-U.S. countries are least likely to screw you with their debt.

Again this is off the top of my head without having looked at the financial condition of these places. It's more of my theory of only dealing with non-corrupt countries for investing and leaving the emerging markets to wallow in their dysfunction without my money being on the table. You'll also notice I left off the list many developed countries in Europe based on their history of mis-management (Spain, Italy, etc.)

It seems to me the problem with a comparison between "owning your country" vs. "owning your employer" is that it is orders of magnitude easier to change your employer than to change your country, and employers are typically rather limited in the control they can exercise.

(All this is from the perspective of a U.S. resident.  Living in another country, especially a less developed nation or a nation without a well developed rule of law or financial infrastructure imposes an entirely different perspective and circumstances such that my perspective is of limited applicability.)

Perhaps the recommendation to store a portion of gold outside your country is sufficient protection when weighed against the dilution of protection when you diversify other assets outside your own country.

Of course, if you can see the end coming, all bets are off.  But if you can do that, wouldn't it be better to simply leave?  And how many ever correctly see the end coming?  Even in hyperinflationary scenarios in the 20th century when changing countries was relatively easy compared to historical norms, it seems very few people left.

It seems that the natural reaction when times get bad is to hunker down and hope for the best - that the bad times pass quickly.  In that scenario, wouldn't it be better to have more of your assets available in a form usable where you are rather than denominated in a country which may have just imposed sanctions on you?  And when it does get bad enough so that significant numbers of people flee, assets that you can carry and assets that are already elsewhere are good.  But the more immediate problem is how to survive.

Tangent...

I wonder if the mid- to late-20th century wasn't a peak in the ease of changing countries.  It used to be very difficult to surmount the cost and time of travel.  By the mid-20th those were significantly reduced and now the challenge is becoming more about how to get past the barriers to entry.  Looking at the current and near future, barriers to exit are starting to appear.  Perhaps those barriers are the first sign of the end...

BWX top ten holdings:

Japan
U. K.
France
Belgium
Netherlands
Spain
Germany
Italy
Canada
Austria

Duration: 7.28 yrs.
Which is more similar to the duration of IEF than TLH.
Why melveyr do you prefer the latter?

And as I know that you know how to do it, Could you please check what is the BWX correlation with the other components?

escafandro wrote: BWX top ten holdings:

Japan
U. K.
France
Belgium
Netherlands
Spain
Germany
Italy
Canada
Austria

Duration: 7.28 yrs.
Which is more similar to the duration of IEF that TLH.
Why melveyr do you prefer the latter?

And as I know that you know how to do it, Could you please check what is the BWX correlation with the other components?

Great question.

If I had my way I would prefer if BWX had longer duration  I suppose you could use IEF and BWX together if you wanted, but it might make sense to boost the weightings to give them a little more power.

If you want to read more about my thoughts on intermediate bonds in the PP you could start here:
http://gyroscopicinvesting.com/forum/ht ... 408#p53408

Punchline: You can use shorter duration bonds in a PP, but it makes sense to hold more of them. Longer duration bonds give you the luxury of being able to put more in other asset classes effectively offering a form of leverage.

Correlations are easily found using the backtest feature of etfreplay.com

AgAuMoney wrote:
Perhaps the recommendation to store a portion of gold outside your country is sufficient protection when weighed against the dilution of protection when you diversify other assets outside your own country.

Of course, if you can see the end coming, all bets are off.  But if you can do that, wouldn't it be better to simply leave?  And how many ever correctly see the end coming?  Even in hyperinflationary scenarios in the 20th century when changing countries was relatively easy compared to historical norms, it seems very few people left.

I guess what really struck me was that for long periods of time some countries had negative real returns for stocks, bonds, and bills. I don't think I could count on gold to always be going up in those scenarios because it might be going down in real terms due to factors in other countries. I see gold as a global asset, and so combining it with just local assets doesn't seem as robust to me as putting it in a global portfolio. 

Thanks for the etreplay.com tip.

melveyr wrote: Additionally, if you changed the bond weightings to be based off of equity weightings (so the US is 40% of the global equity market so US bonds would be 40% of the PP's bonds) than the allocation plan would naturally adjust over time as the power structure of the world changed. It is certainly a lot to think about.

I like the idea, but, Here we should add then emerging country debt...

melveyr wrote: Perhaps I am being seduced by a rear view mirror look, but a global cap weighted VP looks very compelling.

Global cap weighted is stupid.  It throws good money after bad, i.e. Apple.  Country value-weighted is where you want to go.  Markets are not "efficient"; they are prone to neglect, ignorance, stupidity, fads and keeping up with the Joneses mentality.  This will cost you several percentage points per year.

I have half of my PP equity allocation to non-U.S. companies.

MachineGhost wrote:

melveyr wrote: Perhaps I am being seduced by a rear view mirror look, but a global cap weighted VP looks very compelling.

Global cap weighted is stupid.  It throws good money after bad, i.e. Apple.  Country value-weighted is where you want to go.  Markets are not "efficient"; they are prone to neglect, ignorance, stupidity, fads and keeping up with the Joneses mentality.  This will cost you several percentage points per year.

Are there any funds that fit the bill, or is assembling your own individual stock positions really the only way to go for this?

MachineGhost wrote:

melveyr wrote: Perhaps I am being seduced by a rear view mirror look, but a global cap weighted VP looks very compelling.

Global cap weighted is stupid.  It throws good money after bad, i.e. Apple.  Country value-weighted is where you want to go.  Markets are not "efficient"; they are prone to neglect, ignorance, stupidity, fads and keeping up with the Joneses mentality.  This will cost you several percentage points per year.

I have half of my PP equity allocation to non-U.S. companies.

Calling it stupid seems unreasonable to me. Cap weighting ensures that you always have higher than average net investment returns because you are earning the average investment return without any of the expenses. Sounds pretty smart to me.

Interesting report. Thanks for posting!

melveyr wrote:Cap weighting ensures that you always have higher than average net investment returns

Uh, no.  In the real world returns are primarily determined by how much you allocate to the future best performing asset compared to the other assets.  Cap weighting is an explicit assumption that the past best will be the future best.  Sometimes it is sometimes it isn't.

It seems the only possible way your claim could be true is in the universe where "average net investment returns" is defined to make it true.

So how do you define average to make it true?

If you define average as the return of the same portfolio cap equal weight then the real world counter example is RSP vs SPY.

In this world market caps change every day and your cap weighted portfolio must be reconstituted to match on some sort of schedule.  That incurs at least the same transaction costs as maintaining an equal weighted portfolio on the same schedule.

Or perhaps you are comparing to some alternate way of position sizing entirely unrelated to market cap?  What kind of rational and at least somewhat used position sizing is going to have more expenses than the same portfolio cap-weighted?

One of the others I see talked a lot about in the discussions I frequent (re. growing dividends) is position sizing by income generated.  Not really appropriate to most emerging market equities.

AgAu,

The cap weighted index replicates the stock market (which is composed of active managers).  It is literally impossible for active managers aggregated as a group to outperform themselves. My statement wasn't empirical (although there are tons of studies that prove it) but definitional.

Before expenses, active manager returns aggregated as a group perfectly match the benchmarks return. After expenses, they always lag it. If you index by cap weighting you are always better than the aggregated group of active managers after expenses.

melveyr wrote: The cap weighted index replicates the stock market (which is composed of active managers).  It is literally impossible for active managers aggregated as a group to outperform themselves. My statement wasn't empirical (although there are tons of studies that prove it) but definitional.

Before expenses, active manager returns aggregated as a group perfectly match the benchmarks return. After expenses, they always lag it. If you index by cap weighting you are always better than the aggregated group of active managers after expenses.

That's what I thought you meant.  You are wrong and you entirely ignored my counter example: SPY vs RSP.

"Definitions" can be wrong, because they only apply in their scope.  Investing is a much broader scope than your definition, and in that broader scope your definition is wrong.

melveyr wrote: If you index by cap weighting you are always better than the aggregated group of active managers after expenses.

And you know what, on thinking about it, even in scope your definition is wrong.

"The aggregated group of active managers" that is the market experiences in the aggregate the return we know of as the market return.

That market return is measured by what we call an index.  The index never has friction.  But investors in that index always have friction.  So until you find an index fund with a negative expense ratio, if you invest in the index you will always get LESS than that market return because you have more expenses than the benchmark index.

AgAuMoney wrote:

melveyr wrote: The cap weighted index replicates the stock market (which is composed of active managers).  It is literally impossible for active managers aggregated as a group to outperform themselves. My statement wasn't empirical (although there are tons of studies that prove it) but definitional.

Before expenses, active manager returns aggregated as a group perfectly match the benchmarks return. After expenses, they always lag it. If you index by cap weighting you are always better than the aggregated group of active managers after expenses.

That's what I thought you meant.  You are wrong and you entirely ignored my counter example: SPY vs RSP.

"Definitions" can be wrong, because they only apply in their scope.  Investing is a much broader scope than your definition, and in that broader scope your definition is wrong.

AgAu,

When I talk about average I am talking about the average investment dollar. When you start doing something like RSP as your benchmark that is fine but you are no longer tracking the investment universe, but an arbitrary selection of securities. I am concerned with how an investor fares versus other investors, not an arbitrary selection of securities.

People don't use cap weighting as the primary benchmark because it is easy or simple, they do it because it represents the amount of wealth gained or lost by all investors aggregated together. Some people do better than others, but on a dollar basis of outperformance it all nets out to zero before fees are taken into account.

Finally, nothing about this argument is my creation. This all very widely understood knowledge that most professional active managers understand. You can still make a case for active management even if you agree with everything I just said. You just have to argue that you are so much better than average that you can make up for the greater expenses associated with seeking outperformance.

Edit: If you find an active fund/strategy with less friction than a TSM index than please let me know.

melveyr wrote:

AgAuMoney wrote: you entirely ignored my counter example: SPY vs RSP.

"Definitions" can be wrong, because they only apply in their scope.  Investing is a much broader scope than your definition, and in that broader scope your definition is wrong.

When I talk about average I am talking about the average investment dollar. When you start doing something like RSP as your benchmark that is fine but you are no longer tracking the investment
universe, but an arbitrary selection of securities. I am concerned with how an investor fares versus other investors, not an arbitrary selection of securities.

The market return is that average.

The S&P 500 is the most widely recognized benchmark and is most commonly what is meant when experts and lay people say "the market did (whatever)."  The DJIA is a distant second, but usually referred to as "the DOW" instead of "the market."

RSP isn't normally considered a benchmark, nor do I consider it so.

RSP is a fund which holds all the companies of the S&P 500.  It is not "an arbitrary selection of securities" any more than the market is an arbitrary selection of securities.

RSP is much less "an arbitrary selection of securities" than the typical TSM fund, because RSP actually holds every company in its index, while no TSM fund actually holds the total stock market.

People don't use cap weighting as the primary benchmark because it is easy or simple, they do it because it represents the amount of wealth gained or lost by all investors aggregated together.

Correct.  And meaningless when it comes to answering the original question or making a decision.

Some people do better than others, but on a dollar basis of outperformance it all nets out to zero before fees are taken into account.

I love that fees disclaimer.

Finally, nothing about this argument is my creation. This all very widely understood knowledge that most professional active managers understand.

Wrong.

"this argument" IS your creation.  You misrepresent the "widely understood knowledge" when you claim an index fund will outperform the market average after fees.  Either because you do not understand the principles you are trying to argue, or you do not know how to apply them.  There is no way to outperform the market by investing in a market index fund if that fund truly tracks the market.  Well, no way unless the fund has a negative expense ratio.

I don't expect there is an index fund with a negative expense ratio.  But if you happen to find such an fund that can then generate your cited performance in excess of the market, I'm not the only one who'd like to know.  I'm not the least bit interested in a TSM fund or an active fund, so I won't be researching that for you.


But the original question was whether to hold a portfolio cap weighted or not?

I showed how two funds which hold exactly the same companies have radically different performance simply by changing the weighting.  That is, in the race between the cap weighted SPY which is as near to perfectly tracking the market as you are going to find, and the exact same companies held in the equal weight RSP, RSP is the obvious winner over its life.  (There are two co-existing explanations in the "widely understood knowledge" and academic research, but this missive is entirely too long already.)


Equal weight by market value is a very easy management approach because you can tell at a glance when something is off.  Cap weighting as you mentioned is much harder.  So hard it's arguably best left to a fund manager if that approach is really what you want.

Another weighting approach is equal weighting by income produced.  I tried this for nearly two years. Didn't like it.

Another weighting approach is equal risk weighting (where risk is usually equated to volatility with beta as a proxy for volatility, following classic MPT practice).  This is, if you believe in MPT, on the efficient frontier.

Some modify equal risk weighting to lower overall portfolio risk by overweighting the lower risk positions.  This happens enough the approach has a name that I don't recall right now.  This is how MPT tunes your portfolio along the efficient frontier to match your risk tolerance or demanded return.


With my own core holdings I go equal market value weighting.  I'm not hard-nosed about everything being exactly the same amount, or even doing any scheduled rebalance adjustment.  I only adjust if something gets significantly out of whack.  Sometimes I sell the high if it is really high, but everything is always moving so normally I can make the adjustment by turning off dividend reinvestment on the higher one(s) and dedicating the income (and incoming contributions if any) to the lower ones.  Much cheaper that way than selling in order to buy elsewhere.


Bottom line is if you are managing your portfolio yourself, you might consider simplicity important.  Equal risk or equal market value are probably the easiest, and equal market value seems way easier to me.  Plus I like the results I see in the widely understood knowledge and research, and over the past 10+ years I like the results I've experienced far more than I like the market results.

AgAu,

"this argument" IS your creation.  You misrepresent the "widely understood knowledge" when you claim an index fund will outperform the market average after fees.  Either because you do not understand the principles you are trying to argue, or you do not know how to apply them.  There is no way to outperform the market by investing in a market index fund if that fund truly tracks the market.  Well, no way unless the fund has a negative expense ratio.

If by "market average" one means "the average earned by all participants in the market" rather than "the S&P 500" (or for that matter "the TSM" ) the a passive index fund investor CAN outperform after fees. The index investor will earn the market's return minus the index fund's tiny fee and all the other market participants (whether investing on their own by picking stocks, investing through an index fund, or investing through an actively managed fund) will on average earn whatever the total market returns (it IS true that some will do better and some will do worse than the total market but they will all together on average do exactly the same as the market; the returns will just be unevenly distributed between the investors). But because at least SOME fraction of the "other" market participants will be investing through funds that charge a higher expense ratio than the index fund then the index investor will do better after fees than the overall average of all the "other" investors.

By way of example:

TSM is, let's say, $20,000,000,000,000 in market cap. Index investor has $20,000 to invest. "Everyone else" (which includes some other index investors as well as active investors) combined has $19,999,999,980,000 to invest. Index investor buys an index fund (let's say it has a 0.1% ER) and everyone else buys whatever sections of the rest of the market they want (which on average is done through vehicles with an ER of, let's say,  0.75%) but their investment adds up to $19,999,999,980,000. After a year the market is up 10% so the total market cap is now $22,000,000,000,000. Index investor now has (after applying the ER) $21,978. Everyone else now has (split up various ways between them since some did better overall than the total market and some did worse) 21,999,999,978,000 BUT after applying the ER of 0.75% they actually only have $21,834,999,978,165. The index investor earned roughly 9.9% (9.89%) and what everyone else (who were as an aggregate invested in the total market) earned on average overall was roughly 9.25% after the ER (actually 9.18% since in this example I charged the ER on ending asset amount and didn't spread it evenly over the year but to be fair I did the same with the index investor's ER too).

Oh, and while I'm on the subject, their are no TSM or S&P 500 funds with a true negative ER (at least that I am aware of...and by "negative ER" I mean an actual NEGATIVE amount charged on assets under management) but you CAN get a "quasi-negative ER" by investing in VIIIX or VITNX (respectively the high-end, very high minimum investment "institutional plus" versions of VFINX and VTSMX). These funds have such low ERs that the amount they earn from securities lending outstrips what they earn from ERs on AUM. VIIIX, for instance, would have turned $10,000 invested on 1-1-2003 into $19,922.87 as of 12-31-2012 whereas the same $10,000 invested into a hypothetical no-expense ratio S&P 500 total return (with dividends reinvested) tracker would have only left you with $19,858.

Not that you, I, or anyone else on this forum would likely be directly investing in such funds...the minimum initial investment on each is $200 million.

Pointedstick wrote: Are there any funds that fit the bill, or is assembling your own individual stock positions really the only way to go for this?

Not yet, but one is in the works.  In the meantime, you could just buy the 10 cheapest country ETF's based on Shiller P/E.

The irony here is the country ETF's are cap-weighted.

D1984 wrote: The index investor will earn the market's return minus the index fund's tiny fee and all the other market participants (whether investing on their own by picking stocks, investing through an index fund, or investing through an actively managed fund) will on average earn whatever the total market returns (it IS true that some will do better and some will do worse than the total market but they will all together on average do exactly the same as the market; the returns will just be unevenly distributed between the investors). But because at least SOME fraction of the "other" market participants will be investing through funds that charge a higher expense ratio than the index fund then the index investor will do better after fees than the overall average of all the "other" investors.

The short form of this argument is that the fees paid by the index fund investor are lower than the average of all the fees paid by all the other investors.

The argument is probably true if you average by the beneficial owners.  It's probably not true if you average by the number of retail investors.  Those counts are unknown and confusing because nobody quite knows if I am or any investor is one investor or 10.  So nobody knows for sure.

Research into costs was done and reported on in the context of eliminating high-frequency trading.  They reported that the average fee per dollar invested is approximately the SEC fee.  (0.{lots of zeros}{some number}% added to the SEC fee)  So the argument is false if you average costs by dollars, which is typically how investors pay costs.

Way many dollars are invested by very large institutions, including financial institutions who run the index fund, and these large investors pay on average almost nothing beyond the SEC fee.

So who knows.

But the widely known knowledge is that by investing in an index fund you can come as close as you can guarantee to receiving the market return. But expect just a bit less.  Or choose a more appropriate target.

melveyr wrote: This whole document put out by Credit Suisse is pretty incredible:
http://www.investmenteurope.net/digital ... al_web.pdf

How do you convert it to a single page PDF? Might be interesting stuff, but totally unreadable this way.

Why the global portfolio is placed in the Variable Portfolio sub-folder? It just makes sense for a permanent portfolio for a non-US person, doesn't it?

Anyways, what we need here is multivariate testing: http://gyroscopicinvesting.com/forum/ht ... 535#p69535