Expanding on The Rewards Of Multiple Asset Class Investing

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Image

This is one of my all time favorite diagrams from all of my investment reading. It’s from Gibson’s Asset Allocation: Balancing Financial Risk. The specific asset classes aren’t what’s significant. What’s significant is the increased likelihood of getting close to the northwest corner with less than perfectly correlated sources of return. How much more could one improve his odds of success by diversifying even more? In an earlier post today I talk about intuitive belief and conviction. The intuitive belief is not just in the individual factors or sources of return. It’s also in how they combine in a portfolio. When I start talking about adding size, value, Momentum, Alt Lending, Reinsurance, or whatever potential source of uncorrelated return to a portfolio, this is the diagram I have in the back of my mind. The more esoteric one gets, the more expensive. But he’s also moving towards the north west corner. What a potential new addition to a portfolio contributes depends on expected return, correlations (and when correlations change), volatility, and of course costs.

Dave
ThriftyPhD
Posts: 870
Joined: Mon Jul 31, 2017 10:43 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by ThriftyPhD »

I wonder how much the plot changes as you adjust the years. If you changed 2009 to 2007, or 1972 to 1976, would you reach the same conclusion?
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Taylor Larimore »

Dave:

Sorry to say but Gibson's chart is very misleading as a guide for investors. Let me explain.

The chart circles four top performing portfolios. The one thing these four portfolios have in common is their allocation to commodities which did very well during the 1972-2009 period. Unfortunately, commodities (as a group) have lost money ever since. Meanwhile, Vanguard's Total Stock Market Index Fund has more than doubled.

This is a Morningstar link to the commodities index:

http://performance.morningstar.com/fund ... ture=en_US

Investors who followed the winning "Multiple Asset Class Portfolios" would have done much better with a simple Three-fund Portfolio of total market index funds.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

I think this diagram is very useful. The issue is not the time period nor the asset classes. It is much more general than that. It is how less than perfectly correlated, weakly correlated, uncorrelated, negatively correlated components can mix in a portfolio. Returns and correlations will change over time, but the conceptual and practical benefit of combining different sources of return within a portfolio should not change.

Dave
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by SimpleGift »

The rational appeal of multi-asset class portfolios is undeniable. As the old Harry Markowitz quote goes, diversification "is the only free lunch in finance." And Roger Gibson's writings have been some of the best at explaining these benefits to investors in easy-to-read, layman terms. His early book, Asset Allocation, was my primer when I first started investing in the 1990s.

But it should be point out that, on the emotional front, equity diversification is far from a free lunch. The performance of multi-asset equity portfolios will inevitably diverge from the broad market indexes, sometimes significantly and often for long periods of time. To reap the rewards of diversification, one needs to be willing to feel like a fool for extended periods — and more importantly, have the patience and faith that the returns of various asset classes and factors will eventually mean-revert in time.

In short, without the stomach to stick with parts of your portfolio that perform horribly over long periods, one is probably better off with a simple 2 or 3 fund portfolio, as Taylor suggests.
Last edited by SimpleGift on Wed Nov 01, 2017 3:42 pm, edited 1 time in total.
MIretired
Posts: 1059
Joined: Fri Sep 06, 2013 12:35 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by MIretired »

I imagine the chart changes yearly just as the efficient frontier graph does. As Taylor said, maybe more-so using many volatile asset classes. I just imagine it would.
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Taylor Larimore »

Random Walker wrote: Wed Nov 01, 2017 3:35 pm Returns and correlations will change over time, but the conceptual and practical benefit of combining different sources of return within a portfolio should not change.
\Dave
Dave:

We agree. It is the primary reason for favoring the total market portfolio which has the "conceptual and practical benefit of combining different sources of return within a portfolio."

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Taylor,
We almost agree. When one creates a portfolio of different assets, the expected return is the weighted average of the expected return of the individual assets. But because of the less than perfect correlations between portfolio components, the expected standard deviation of the portfolio is less than the weighted average of the standard deviations. I would venture to say that is why you recommend the three fund portfolio over a two fund portfolio comprised of just TSM and TBM. TSM and Total Int should have pretty much same expected return over time (neglecting current valuation differences), but the less than perfect correlation improves portfolio efficiency. When one starts diversifying beyond just geography, into factors or alternative sources of return, this benefit can be markedly further increased. Of course, at increased cost too.

Dave
Snarfanio
Posts: 11
Joined: Tue Oct 03, 2017 2:10 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Snarfanio »

Random Walker stated: "When one creates a portfolio of different assets, the expected return is the weighted average of the expected return of the individual assets."

I know that even some of the best minds (including William Bernstein's) have made that mistake in the past, but it's simply not true. The return of a portfolio of different asset classes depends on the correlation between such asset classes. A simple example: when the allocation between large company stocks and intermediate-term government bonds in a portfolio is varied, the result is a curve, not a straight line.

Regards, Sergio.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Sergio,
Perhaps that’s the difference between expected returns and actual returns?

Dave
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by patrick013 »

Aren't there 10 sectors, or asset classes, or at least industries. Commodities
have value but not much earning power so not really investment grade IMO.
Seasonality and high inflation were relevant to their past successes.
age in bonds, buy-and-hold, 10 year business cycle
Snarfanio
Posts: 11
Joined: Tue Oct 03, 2017 2:10 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Snarfanio »

Random Walker wrote: Wed Nov 01, 2017 7:50 pm Sergio,
Perhaps that’s the difference between expected returns and actual returns?

Dave
Dave,
No, it's not that at all. It's Investing 101, risk-return curve basics. See the third graph here: http://www.swapmeetdave.com/Bible/USForeign.htm
All portfolios with asset classes that don't have a perfect +1 correlation, have a risk-return curve, not a straight line.

Regards, Sergio.
JBTX
Posts: 11205
Joined: Wed Jul 26, 2017 12:46 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by JBTX »

Random Walker wrote: Wed Nov 01, 2017 2:02 pm Image

This is one of my all time favorite diagrams from all of my investment reading. It’s from Gibson’s Asset Allocation: Balancing Financial Risk. The specific asset classes aren’t what’s significant. What’s significant is the increased likelihood of getting close to the northwest corner with less than perfectly correlated sources of return. How much more could one improve his odds of success by diversifying even more? In an earlier post today I talk about intuitive belief and conviction. The intuitive belief is not just in the individual factors or sources of return. It’s also in how they combine in a portfolio. When I start talking about adding size, value, Momentum, Alt Lending, Reinsurance, or whatever potential source of uncorrelated return to a portfolio, this is the diagram I have in the back of my mind. The more esoteric one gets, the more expensive. But he’s also moving towards the north west corner. What a potential new addition to a portfolio contributes depends on expected return, correlations (and when correlations change), volatility, and of course costs.

Dave
Interesting. Is the increased return due to the buy low sell high impact of period rebalancing? I always knew that diversifying decrease risk for a given return. I did not know (or likely forgot) that diversification can actually increase return.

Of course, rebalancing in itself may be a chore, in that I doubt you can find a fund that balances all these classes, so you will have to do it yourself, and it will take fortitude to keep selling the winners and keep buying more of the dogs as you rebalance, sometimes for decades. I wonder how often these portfolios are rebalanced?

Finally, on the commodities, is that the actual return of an investable fund/etf, or the return of just an index? Its my understanding that with some commodities - specifically perishable and consumable ones, like crops and oil, you can't actually buy the index, as you would have to hold it and take delivery - but either have to try to replicate via futures, where traders will eat your lunch when you liquidate to avoid taking delivery, or else find a financial institution that will synthetically recreate the index, for a fee, and some risk (risk that the fin institution goes bust and can't back up the etf/fund)

It points out one of the things I probably haven't done well is rebalancing. I do have stocks, bonds, commodities, and REITS, but I don't really religiously rebalance them - other than stocks vs bonds to a degree. Over the long term that could cost me some return.


Taylor Larimore wrote: Wed Nov 01, 2017 3:18 pm Dave:

Sorry to say but Gibson's chart is very misleading as a guide for investors. Let me explain.

The chart circles four top performing portfolios. The one thing these four portfolios have in common is their allocation to commodities which did very well during the 1972-2009 period. Unfortunately, commodities (as a group) have lost money ever since. Meanwhile, Vanguard's Total Stock Market Index Fund has more than doubled.

This is a Morningstar link to the commodities index:

http://performance.morningstar.com/fund ... ture=en_US

Investors who followed the winning "Multiple Asset Class Portfolios" would have done much better with a simple Three-fund Portfolio of total market index funds.

Best wishes.
Taylor
It would be interesting to see it extended through 2017. Your point is valid in the time periods selected somewhat cherry pick the beginning and end of two different strong periods for commodities. Nonetheless, the concept still holds. Of course, during an up market like 2009-2017 stocks are going to look better than a multi asset portfolio. I imagine through 2017, or just about any oher 40 year period, the combined risk/return combination will look much better for the combined asset class portfolio

Simplegift wrote: Wed Nov 01, 2017 3:35 pm

But it should be point out that, on the emotional front, equity diversification is far from a free lunch. The performance of multi-asset equity portfolios will inevitably diverge from the broad market indexes, sometimes significantly and often for long periods of time. To reap the rewards of diversification, one needs to be willing to feel like a fool for extended periods — and more importantly, have the patience and faith that the returns of various asset classes and factors will eventually mean-revert in time.
This is really true, and practically next to impossible. Significantly lagging the most successful broad index can cause massive fund outflows. Look at Grantham for instance. He is lauded for eventually calling a bubble (often way too early) and then criticized when his mean reversion portfolio lags the broad indexes in extended bull markets. I doubt over the long term he beats the markets, but I'll bet he comes close and his risk/return relationship looks a lot better.
ryman554
Posts: 1634
Joined: Sun Jan 12, 2014 8:44 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by ryman554 »

Simplegift wrote: Wed Nov 01, 2017 3:35 pm To reap the rewards of diversification, one needs to be willing to feel like a fool for extended periods — and more importantly, have the patience and faith that the returns of various asset classes and factors will eventually mean-revert in time.
I would go one step further -- you will feel like (or at least look like) a fool *all of the time*. But that's because you're never invested totally in "what's hot".

Of course, then you remember that, to reap the rewards on investing in the latest craze, you have to have been invested in them from the beginning. When you would look like a fool for doing so.
rkhusky
Posts: 17652
Joined: Thu Aug 18, 2011 8:09 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by rkhusky »

Random Walker wrote: Wed Nov 01, 2017 2:02 pm When I start talking about adding size, value, Momentum, Alt Lending, Reinsurance, or whatever potential source of uncorrelated return to a portfolio, this is the diagram I have in the back of my mind. T
Size, value and momentum are not like the others. In the plot are 4 different assets, which are 4 different sets of companies. Alt lending and reinsurance are also baskets of different companies. With the factors, one does not have separate groups of companies (i.e. disjoint groups of small companies, high value companies, and high momentum companies), but instead one has one set of companies that all have the same size, value and momentum characteristics.

On the other hand, one could do a similar analysis in the factor world, with the baskets (with 2 factors) being Small Value, Small Growth, Large Value, and Large Growth. I wonder what a comparable chart with these 4 baskets would look like? Or perhaps going long with Small Value and Large Value and shorting Small Growth and Large Growth?
User avatar
Ethelred
Posts: 384
Joined: Sun Oct 30, 2016 9:38 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Ethelred »

Taylor Larimore wrote: Wed Nov 01, 2017 3:18 pm Dave:

Sorry to say but Gibson's chart is very misleading as a guide for investors. Let me explain.

The chart circles four top performing portfolios. The one thing these four portfolios have in common is their allocation to commodities which did very well during the 1972-2009 period. Unfortunately, commodities (as a group) have lost money ever since. Meanwhile, Vanguard's Total Stock Market Index Fund has more than doubled.

This is a Morningstar link to the commodities index:

http://performance.morningstar.com/fund ... ture=en_US

Investors who followed the winning "Multiple Asset Class Portfolios" would have done much better with a simple Three-fund Portfolio of total market index funds.

Best wishes.
Taylor
Taylor,

I just looked at the chart and read your post, and I'm afraid it's not quite right. The one thing that the four portfolios have in common is their allocation to both commodities and real estate. Commodities (asset D on the chart) actually performed worst (lowest CAGR), and also with the highest standard deviation. The diversification (lack of correlation) between real estate and commodities is what's driving the allocations in the best four portfolios. Conversely, the high correlation between US and non-US stocks explains why the "AB" portfolio performs worst of all the portfolios with more than one asset class.

I'm not writing that to claim that the three-fund portfolio is in any way wrong though. The diversification effect is instead provided by bonds, which do not feature on this chart.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

It seems to me that many of the comments are ignoring the forest from the trees. The specific asset classes are irrelevant. The time frame is irrelevant. What is relevant is the big picture. The big picture is that combining sources of return with less than perfectly correlated behavior improves portfolio efficiency. By dampening portfolio volatility and taking advantage of rebalancing, one is most likely to maximize return per unit portfolio volatility if they diversify across sources of return. One can extrapolate the basic results from this diagram way beyond the 4 asset classes it uses. One can look at anything: asset classes, factors, styles, alternatives. The question is whether adding a new potential investment to a portfolio increases the likelihood of landing in the upper left corner. When one decides whether to add something new, he needs to consider how it mixes with the rest of the portfolio. Characteristics to consider include expected return, cost, volatility, correlations, when correlations rise or fall.

Dave
staythecourse
Posts: 6993
Joined: Mon Jan 03, 2011 8:40 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by staythecourse »

Taylor Larimore wrote: Wed Nov 01, 2017 3:18 pm Dave:

Sorry to say but Gibson's chart is very misleading as a guide for investors. Let me explain.

The chart circles four top performing portfolios. The one thing these four portfolios have in common is their allocation to commodities which did very well during the 1972-2009 period. Unfortunately, commodities (as a group) have lost money ever since. Meanwhile, Vanguard's Total Stock Market Index Fund has more than doubled.

This is a Morningstar link to the commodities index:

http://performance.morningstar.com/fund ... ture=en_US

Investors who followed the winning "Multiple Asset Class Portfolios" would have done much better with a simple Three-fund Portfolio of total market index funds.

Best wishes.
Taylor
No that is not a misleading chart. It fits in with concept of diversification. The idea of the chart is to show how adding different asset classes together with different return, risk, and correlation coefficient alters the return of the portfolio. That's it. It is not meant as a, "Hey add commodities to your portfolio because of this chart". I have read Mr. Gibson's book and to be honest it is still one of the best books on the science of investing I have read. Just because it is not written by a bogleheads does not mean its message does not have merit.

If you love the three fund that is great. No bid deal. I don't think it is right to degrade another author's views just because they don't align with yours.

Good luck.
"The stock market [fluctuation], therefore, is noise. A giant distraction from the business of investing.” | -Jack Bogle
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by SimpleGift »

Simplegift wrote: Wed Nov 01, 2017 3:35 pmIn short, without the stomach to stick with parts of your portfolio that perform horribly over long periods, one is probably better off with a simple 2 or 3 fund portfolio, as Taylor suggests.
Just to add another caution (though I'm a fan of multi-asset investing and have benefited from the "free lunch" over the years):

In taxable accounts, if one starts out with a diverse collection of equity asset classes, eventually over decades each of these funds becomes highly-appreciated with lots of unrealized capital gains (which is good!). But later in life, if one ever wants to simplify their equity portfolio, it becomes very difficult to sell or weed out this large collection of funds, due to the capital gains tax liability. In other words, one becomes stuck for life with these highly-appreciated funds.

As an example, when starting out in the early 1990s, I invested in Vanguard's Energy Fund (VGENX) for commodities exposure — but in later years realized that it was very highly correlated with Vanguard's Emerging Markets Index, so was not adding much diversification. But I couldn't reasonably sell it, due to the capital gains taxes. My best option was ultimately to gift these highly-appreciated shares to Vanguard's donor-advised fund, then make charitable contributions from there, in place of direct cash donations.

In short, without the stomach to be stuck with and manage many, small, highly-appreciated equity holdings in one's later retirement years, one is likely better off starting with a simpler 2 or 3 fund portfolio.
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Taylor Larimore »

staythecourse wrote: I don't think it is right to degrade another author's views just because they don't align with yours.
Staythecourse:

I agree with you that it is improper to "degrade another author's views just because they don't align with yours." I would never intentionally do that.

I own and treasure Roger Gibson's Third Edition of "Asset Allocation" published in 2000. I consider it one of the best books on my bookshelf. I included Mr. Gibson's book in my Investment Gems. Mr. Gibson knows much more than I do.

I could be wrong, but I believe the particular chart in the Opening Post is misleading and I explained why. I hope you will understand.

Asset Allocation by Roger Gibson -- A Gem

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by patrick013 »

Random Walker wrote: Thu Nov 02, 2017 11:59 am It seems to me that many of the comments are ignoring the forest from the trees. The specific asset classes are irrelevant. The time frame is irrelevant. What is relevant is the big picture. The big picture is that combining sources of return with less than perfectly correlated behavior improves portfolio efficiency.
Well let me say this. When I relate International to my hypothetical portfolio my beta metric
is awful. I'm better off with Utilities for lessening price fluctuations. Correlations can rise
and fall but more importantly is geographic separation. A global recession such as 2008 may
happen but geographic separation can help. A recession in Mexico but not in China, in USA but
not in South America, in Europe but not in Indonesia, etc..

I fail to see the usefulness of correlations when old and new data don't lead to fundamental
improvements. For domestics use beta, for International you have geographic separation,
low returns, and that's about it. I honestly don't see total returns jumping off the charts for
2 or 3 reasons. Even international PE's are holding steady at long term averages, nothing
optimistic or indicating super growth ahead. The data for commodities is so old, what does
newer data inform ? For hedgers mostly.

Hate to be pessimistic but having more assets classes than fewer, having many stocks instead
of few is not really being sold with that chart. So as usual I'm weak on international and adverse
to commodities. A chart with technology and health care wouldn't be broad based either. The
Dow Theory is even more broad based.
age in bonds, buy-and-hold, 10 year business cycle
JBTX
Posts: 11205
Joined: Wed Jul 26, 2017 12:46 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by JBTX »

rkhusky wrote: Thu Nov 02, 2017 10:26 am
Random Walker wrote: Wed Nov 01, 2017 2:02 pm When I start talking about adding size, value, Momentum, Alt Lending, Reinsurance, or whatever potential source of uncorrelated return to a portfolio, this is the diagram I have in the back of my mind. T
Size, value and momentum are not like the others. In the plot are 4 different assets, which are 4 different sets of companies. Alt lending and reinsurance are also baskets of different companies. With the factors, one does not have separate groups of companies (i.e. disjoint groups of small companies, high value companies, and high momentum companies), but instead one has one set of companies that all have the same size, value and momentum characteristics.

On the other hand, one could do a similar analysis in the factor world, with the baskets (with 2 factors) being Small Value, Small Growth, Large Value, and Large Growth. I wonder what a comparable chart with these 4 baskets would look like? Or perhaps going long with Small Value and Large Value and shorting Small Growth and Large Growth?
Technically, aren't REITs and Reinsurance part of the broader markets? Although they may be small enough such that their unique uncorrelated nature doesn't give a lot of diversification to the total ball of wax?

Is there a way one can break up the total market (either domestic, and/or international) into sectors that have the lowest correlation to the broad markets, and you overweight those particular sectors? REITS and Reinsurance may be examples. Perhaps others (mining companies or timber, for instance). Would that result in a higher potentially lower standard deviation and maybe higher long term returns?

We always go with equal market weighting because that is the easiest to track and to measure against. But in theory, could one build a portfolio of all, or most stocks, but weight them based upon their optimized impacts on standard deviation and return, which i would assume would mean companies or sectors with low correlations to the broad market get higher weightings.

Is that possible? Any merit to that? Perhaps just a ignorant ranting from a BH noobie. :D
staythecourse
Posts: 6993
Joined: Mon Jan 03, 2011 8:40 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by staythecourse »

Ethelred wrote: Thu Nov 02, 2017 10:58 am
Taylor Larimore wrote: Wed Nov 01, 2017 3:18 pm Dave:

Sorry to say but Gibson's chart is very misleading as a guide for investors. Let me explain.

The chart circles four top performing portfolios. The one thing these four portfolios have in common is their allocation to commodities which did very well during the 1972-2009 period. Unfortunately, commodities (as a group) have lost money ever since. Meanwhile, Vanguard's Total Stock Market Index Fund has more than doubled.

This is a Morningstar link to the commodities index:

http://performance.morningstar.com/fund ... ture=en_US

Investors who followed the winning "Multiple Asset Class Portfolios" would have done much better with a simple Three-fund Portfolio of total market index funds.

Best wishes.
Taylor
Taylor,

I just looked at the chart and read your post, and I'm afraid it's not quite right. The one thing that the four portfolios have in common is their allocation to both commodities and real estate. Commodities (asset D on the chart) actually performed worst (lowest CAGR), and also with the highest standard deviation. The diversification (lack of correlation) between real estate and commodities is what's driving the allocations in the best four portfolios. Conversely, the high correlation between US and non-US stocks explains why the "AB" portfolio performs worst of all the portfolios with more than one asset class.

I'm not writing that to claim that the three-fund portfolio is in any way wrong though. The diversification effect is instead provided by bonds, which do not feature on this chart.
Exactly. It is an example of adding a dog asset class standing alone to a more efficient portfolio when added to other asset classes. It is an example of MPT. Now if that happens in real life is an argument, but a great illustration nevertheless.

The 3 fund is a great portfolio, but the one question I have had is wouldn't it be better to substitute Total international for international small value? You get better currency diversification, lower correlations with US TSM, attempt to exploit some value and small premiums which have low correlations with beta. Seems like it would be a good addition for those looking for one addition to a US TSM/ US bond combo.

Good luck.
"The stock market [fluctuation], therefore, is noise. A giant distraction from the business of investing.” | -Jack Bogle
User avatar
Ethelred
Posts: 384
Joined: Sun Oct 30, 2016 9:38 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Ethelred »

staythecourse wrote: Thu Nov 02, 2017 2:25 pm The 3 fund is a great portfolio, but the one question I have had is wouldn't it be better to substitute Total international for international small value? You get better currency diversification, lower correlations with US TSM, attempt to exploit some value and small premiums which have low correlations with beta. Seems like it would be a good addition for those looking for one addition to a US TSM/ US bond combo.

Good luck.
I would be wary about asking questions like that.

For some, the Bogleheads philosophy is about investing in index funds with low expenses, and that can include the application of Modern Portfolio Theory, factor investing, and whatever other ideas that are of value. For others, the Bogleheads philosophy specifically includes the three-fund portfolio, invested in index funds with low expenses. Diversification means different but related things to these two groups: To the first group, diversification is about maximizing risk-adjusted return on the efficient frontier; to the second, diversification is about "owning the market". Theory and back-testing versus simplicity and philosophy.
Last edited by Ethelred on Thu Nov 02, 2017 2:47 pm, edited 1 time in total.
rkhusky
Posts: 17652
Joined: Thu Aug 18, 2011 8:09 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by rkhusky »

JBTX wrote: Thu Nov 02, 2017 2:05 pm We always go with equal market weighting because that is the easiest to track and to measure against. But in theory, could one build a portfolio of all, or most stocks, but weight them based upon their optimized impacts on standard deviation and return, which i would assume would mean companies or sectors with low correlations to the broad market get higher weightings.
Stock market correlations are not static, but change over time. I imagine that correlations at the individual stock level are very noisy.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Stay the course,
I think you are thinking correctly. ISV would be a better diversifier for the reasons you mention. Size and value are not correlated with market, and International small is more affected by local economies. I think the goal is maximize the likelihood of being in the northwest on that chart. And many diversifier can push the portfolio in that direction: size, value, TS Momentum, CS Momentum, reinsurance, Variance Risk Premium, are examples I know of.

Dave
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by patrick013 »

JBTX wrote: Thu Nov 02, 2017 2:05 pm
Is there a way one can break up the total market (either domestic, and/or international) into sectors that have the lowest correlation to the broad markets, and you overweight those particular sectors?
Easier said than done, although beta is a little different than correlation of course.
Low beta stocks just keep doing whatever they are doing at a slower pace.

https://myphotos.mypclinuxos.com/images ... table2.png
age in bonds, buy-and-hold, 10 year business cycle
JBTX
Posts: 11205
Joined: Wed Jul 26, 2017 12:46 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by JBTX »

rkhusky wrote: Thu Nov 02, 2017 2:47 pm
JBTX wrote: Thu Nov 02, 2017 2:05 pm We always go with equal market weighting because that is the easiest to track and to measure against. But in theory, could one build a portfolio of all, or most stocks, but weight them based upon their optimized impacts on standard deviation and return, which i would assume would mean companies or sectors with low correlations to the broad market get higher weightings.
Stock market correlations are not static, but change over time. I imagine that correlations at the individual stock level are very noisy.
True. It’s mostly just a thought exercise not something easy to act upon and implement. Instead of stocks, maybe sectors? To the extent an indexer buys a REIT index on the side that is somewhat of an example.
R2D2
Posts: 393
Joined: Tue Aug 19, 2014 4:37 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by R2D2 »

Snarfanio wrote: Wed Nov 01, 2017 7:26 pm Random Walker stated: "When one creates a portfolio of different assets, the expected return is the weighted average of the expected return of the individual assets."

I know that even some of the best minds (including William Bernstein's) have made that mistake in the past, but it's simply not true. The return of a portfolio of different asset classes depends on the correlation between such asset classes. A simple example: when the allocation between large company stocks and intermediate-term government bonds in a portfolio is varied, the result is a curve, not a straight line.

Regards, Sergio.
No, I think Random Walker (and Bernstein) got it right the first time. The expectation of a sum is the sum of the expectations. Risk is a different story, but RW didn't mention risk.

Sorry if I missed something.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

I’m no math or stats guy, but I’m pretty sure I’m right as well. The straight line of the expected return seems intuitive. The curved line of actual returns in a rebalanced portfolio comprised of two uncorrelated assets (and correlations change over time too) seems very difficult to predict.

Dave
User avatar
triceratop
Posts: 5838
Joined: Tue Aug 04, 2015 8:20 pm
Location: la la land

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by triceratop »

R2D2 wrote: Thu Nov 02, 2017 5:50 pm
Snarfanio wrote: Wed Nov 01, 2017 7:26 pm Random Walker stated: "When one creates a portfolio of different assets, the expected return is the weighted average of the expected return of the individual assets."

I know that even some of the best minds (including William Bernstein's) have made that mistake in the past, but it's simply not true. The return of a portfolio of different asset classes depends on the correlation between such asset classes. A simple example: when the allocation between large company stocks and intermediate-term government bonds in a portfolio is varied, the result is a curve, not a straight line.

Regards, Sergio.
No, I think Random Walker (and Bernstein) got it right the first time. The expectation of a sum is the sum of the expectations. Risk is a different story, but RW didn't mention risk.

Sorry if I missed something.
(I am a math/stats person, but definitely not a finance person)

Snarfanio is confusing expectation with realizations. The actual return of a portfolio depending on realized returns and realized correlations has nothing to do with the ex ante expected returns.
"To play the stock market is to play musical chairs under the chord progression of a bid-ask spread."
User avatar
Johnnie
Posts: 587
Joined: Sat May 28, 2016 3:18 pm
Location: Michigan

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Johnnie »

Simplegift wrote: Wed Nov 01, 2017 3:35 pm ...But it should be pointed out that, on the emotional front, equity diversification is far from a free lunch. The performance of multi-asset equity portfolios will inevitably diverge from the broad market indexes, sometimes significantly and often for long periods of time. To reap the rewards of diversification, one needs to be willing to feel like a fool for extended periods — and more importantly, have the patience and faith that the returns of various asset classes and factors will eventually mean-revert in time.
Merriman's "Ultimate Buy and Hold Equity Portfolio" contains equal shares of 10 asset classes, which ideally means one might only have to feel like a 10 percent fool at any given time. :) In reality that's unlikely, and one can still expect to feel like a 50 percent fool for long periods, and perhaps more than 50 percent at times.

Specifically, half the portfolio is internationals, including EM, so if the dollar or US equities are on a tear and the rest of the world is dogging it that could be painful. It's also 40 percent value - half US and half international - another source of feeling like a fool for an extended time.

Still, his "fine tuning table" shows the portfolio providing an 11.4 percent compound annual return since 1970 with a 14.7% SD, vs. a 10.3% annualized return on the S&P 500 with a 15.2% SD. You can compare the year-to-year returns there too.

So it looks like more evidence supporting Random Walker's OP.

The asset classes are SP 500, LCV, SC, SCV, US Reits, International LC Blend, ILCV, ISC, ISCV, EM.
"I know nothing."
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

I’ve posted before on the concept of volatility drag. If two portfolios have the same weighted average annual mean return, but different volatilities, the geometric return of the more volatile portfolio will be less than the geometric return of the less volatile portfolio. Geo Mean = Simple Mean - 1/2 SD^2. It is the geometric return (compounded return) of the portfolio that matters to us. Volatility is a real cost to a portfolio. If someone can construct a portfolio with an SD of 14 with the same expected return as another portfolio with an SD of 18, he’s Going to gain about 0.6% annually from a more efficient portfolio.
Volatility drag = 1/2 (SD1^2-SD2^2) = 1/2 (.18^2-.14^2) = 0.6%.
This is achievable by creating a portfolio with the same expected return as a TSM type portfolio, but increasing tilts to small, value, emerging, decreasing overall equity allocation, increasing overall bond allocation. As Larry Swedroe points out, this lower beta / higher tilt portfolio will have smaller SD and smaller left and right tails.

Dave
Last edited by Random Walker on Thu Nov 02, 2017 11:40 pm, edited 1 time in total.
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

1-yr, 3-yrs 5-yrs and 10-year results for 10 Professionally Designed Portfolios

Post by Taylor Larimore »

Johnnie wrote: Thu Nov 02, 2017 7:51 pm
Merriman's "Ultimate Buy and Hold Equity Portfolio" contains equal shares of 10 asset classes, which ideally means one might only have to feel like a 10 percent fool at any given time. :) In reality that's unlikely, and one can still expect to feel like a 50 percent fool for long periods, and perhaps more than 50 percent at times.

Specifically, half the portfolio is internationals, including EM, so if the dollar or US equities are on a tear and the rest of the world is dogging it that could be painful. It's also 40 percent value - half US and half international - another source of feeling like a fool for an extended time.

Still, his "fine tuning table" shows the portfolio providing an 11.4 percent compound annual return since 1970 with a 14.7% SD, vs. a 10.3% annualized return on the S&P 500 with a 15.2% SD. You can compare the year-to-year returns there too.
Johnnie:

You can also compare 10 professionally designed portfolios on MarketWatch. Mr. Merriman's Ultimate Buy and Hold Portfolio is last for 1-yr, 3-yrs, 5-yrs and 10-yr returns.

The Second Graders Three-Fund Portfolio is tied for first place:

Past performance does not forecast future performance.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
User avatar
Johnnie
Posts: 587
Joined: Sat May 28, 2016 3:18 pm
Location: Michigan

Re: 1-yr, 3-yrs 5-yrs and 10-year results for 10 Professionally Designed Portfolios

Post by Johnnie »

Taylor Larimore wrote: Thu Nov 02, 2017 9:04 pm Johnnie:

You can also compare 10 professionally designed portfolios on MarketWatch. Mr. Merriman's Ultimate Buy and Hold Portfolio is last for 1-yr, 3-yrs, 5-yrs and 10-yr returns.

The Second Graders Three-Fund Portfolio is tied for first place:

Past performance does not forecast future performance.

Best wishes.
Taylor
I thought I had partially preempted this by explicitly comparing the 100% percent equity versions of these portfolios. I did so because the "lazy portfolios" comparison Taylor often cites to make this same point in more recent threads are are not apple-to-apple comparisons, because they all contain different proportions of fixed income. I can't tell if that applies to this cite but I suspect it may.
"I know nothing."
Jiu Jitsu Fighter
Posts: 179
Joined: Mon Nov 21, 2016 9:22 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Jiu Jitsu Fighter »

Great post. The 3-fund portfolio looks great after a huge bull market. Owning a decent portion of value stocks during the 2000-2002 bear market made it way more tolerable. I thought that period was more psychologically distressful than the 2008 plummet even though my investible assets were much more going into '08. Apple at $900 billion market cap??? U.S. large-growth has been dominating this bull. Does anyone remember AOL's $224 billion market cap? Gold has still outperformed since the turn of the century. You don't have to make things super complicated to obtain different sources of risk/return.
User avatar
Portfolio7
Posts: 1128
Joined: Tue Aug 02, 2016 3:53 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Portfolio7 »

I admire Gibson's observation, but you can't just slap 10 asset classes together and hope it all works out. There are a lot of portfolios where the parts don't play so well together, no potential advantage over a 3 fund portfolio, plenty of potential disadvantage. I think to take advantage of this idea, you need sustainably high return asset classes, where volatility tends to cancel out somewhat - then you may outperform a 3-fund over time (past return data seems to suggest one could have done so pretty consistently with certain portfolios over the last 30 years, which is probably just enough time to feel meaningful yet be misleading :? . I would like at least a full interest rate cycle. Of course that leads to the efficient frontier discussion, which is a valid consideration.) As Taylor noted w/r/t commodities, there are flash in the pan asset classes and investing trends. Despite the hype and some very impressive research, Factor funds and Alt investments may turn out to be not very useful over the long term. As much as I like min volatility funds, for example, I have growing questions about their proper long term use within portfolios.

Anyways, I figure using a high number of asset classes involves higher risk outside of mere volatility. For me, that risk starts with 'sustainably' high returns. Are higher EM or SCV returns sustainable? I believe there is a certain amount of incremental faith required for a slice and dice approach that's not necessary in a 3-fund, simply because there is no guarantee of continued outperformance - whether that's faith in factors, China's GDP growth, or whatever it is you believe in. In any of those cases, you have to be able to weather more frequent and maybe bigger disappointments, which bring me around to the Munger/Buffet tidbit that successful investing is mostly about avoiding the big mistake. I believe that.
"An investment in knowledge pays the best interest" - Benjamin Franklin
rkhusky
Posts: 17652
Joined: Thu Aug 18, 2011 8:09 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by rkhusky »

triceratop wrote: Thu Nov 02, 2017 6:59 pm (I am a math/stats person, but definitely not a finance person)

Snarfanio is confusing expectation with realizations. The actual return of a portfolio depending on realized returns and realized correlations has nothing to do with the ex ante expected returns.
In other words, the expectations involve only average returns and do not include any time variation, where correlation could enter the analysis.
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by patrick013 »

https://myphotos.mypclinuxos.com/images ... 7beta2.png

Here's some figures I threw together. Beta's are 5 year, guesstimates
are based on 10 year returns give or take based on my viewpoint. So,
a multiple fund portfolio is displayed. They all hover around 6% today.
age in bonds, buy-and-hold, 10 year business cycle
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Just thought I’d bump this post. I think the initial diagram at the top of the thread is invaluable. Diversification across less than perfectly correlated sources of return works.

Dave
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Thought I’d bump this again. Recent talk of alternatives got me envisioning this diagram again.

Dave
User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by in_reality »

1972 -2007 CAGR (per Portfolio visualizer - rebalance annually [monthly or quarterly don't seem to change much])

100% US stocks

8.31% CAGR
15.11% StDev
-50.89% Max drawdown

------------------

33% US stocks
33% Int stocks
34% REIT


5.79% CAGR
17.85% StDev
-57.97% Max drawdown

-------------------

50% US stocks
50% Int stocks


5.78% CAGR
16.37% StDev
-54.79% Max drawdown

----------------------

100% Int stocks

3.04% CAGR
18.63% stDev
-58.50% Max Drawdown

-----------------------

25% US stocks
25% Int stocks
25% REIT
25% Commodities


2.72% CAGR
16.85% StDev
-57.46% Max drawdown

-------------------

I'm not sure simply adding additional asset classes really helps. Maybe a different weighting has some effect.
staythecourse
Posts: 6993
Joined: Mon Jan 03, 2011 8:40 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by staythecourse »

Portfolio7 wrote: Thu Nov 02, 2017 11:06 pm I admire Gibson's observation, but you can't just slap 10 asset classes together and hope it all works out. There are a lot of portfolios where the parts don't play so well together, no potential advantage over a 3 fund portfolio, plenty of potential disadvantage. I think to take advantage of this idea, you need sustainably high return asset classes, where volatility tends to cancel out somewhat - then you may outperform a 3-fund over time (past return data seems to suggest one could have done so pretty consistently with certain portfolios over the last 30 years, which is probably just enough time to feel meaningful yet be misleading :? . I would like at least a full interest rate cycle. Of course that leads to the efficient frontier discussion, which is a valid consideration.) As Taylor noted w/r/t commodities, there are flash in the pan asset classes and investing trends. Despite the hype and some very impressive research, Factor funds and Alt investments may turn out to be not very useful over the long term. As much as I like min volatility funds, for example, I have growing questions about their proper long term use within portfolios.

Anyways, I figure using a high number of asset classes involves higher risk outside of mere volatility. For me, that risk starts with 'sustainably' high returns. Are higher EM or SCV returns sustainable? I believe there is a certain amount of incremental faith required for a slice and dice approach that's not necessary in a 3-fund, simply because there is no guarantee of continued outperformance - whether that's faith in factors, China's GDP growth, or whatever it is you believe in. In any of those cases, you have to be able to weather more frequent and maybe bigger disappointments, which bring me around to the Munger/Buffet tidbit that successful investing is mostly about avoiding the big mistake. I believe that.
I am not sure why this is so difficult to figure out. The equation for portfolio return is simply the summation of the weighted averages of the individual asset classes. So plainly speaking 3 fund will outperform if the weighted averages of TSM, TISM, and TBM do better in the time period of analysis vs. weighted averages for a four fund diversified (TSM, US SCV, Total international large, and total international small value) (for example). It is that simple.

It is up to you to decide which asset classes to add, but there is NO PREORDAINED reason one portfolio HAS to do better then another unless you know which asset classes will outperform going forward. But then again if you did why diversify at all when you can just put 100% into the best performing asset class. That is why we diversify since we don't know IN ADVANCE which will do the best, no?

Good luck.
"The stock market [fluctuation], therefore, is noise. A giant distraction from the business of investing.” | -Jack Bogle
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Thought I’d give this post a bump. The diagram at the start really is the foundation of my investment philosophy and it’s in the back of my mind whenever we talk factors and alternatives.

Dave
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Thought I’d give this post another bump in light of some recent threads, especially the one on Larry’s article : Looking beyond 60/40 Portfolio.

Dave
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Here are two more references for those interested in the issue of bringing compounded return closer to simple mean return.

Expected Returns by Antti Ilmanen pp 485-486.
Successful Investing Is A Process by Jacques Lussier pp. 54-63

Dave
Fclevz
Posts: 651
Joined: Fri Mar 30, 2007 11:28 am

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Fclevz »

in_reality wrote: Mon Jan 29, 2018 10:20 pm 1972 -2007 CAGR (per Portfolio visualizer - rebalance annually [monthly or quarterly don't seem to change much])

100% US stocks

8.31% CAGR
15.11% StDev
-50.89% Max drawdown

------------------

33% US stocks
33% Int stocks
34% REIT


5.79% CAGR
17.85% StDev
-57.97% Max drawdown

-------------------

50% US stocks
50% Int stocks


5.78% CAGR
16.37% StDev
-54.79% Max drawdown

----------------------

100% Int stocks

3.04% CAGR
18.63% stDev
-58.50% Max Drawdown

-----------------------

25% US stocks
25% Int stocks
25% REIT
25% Commodities


2.72% CAGR
16.85% StDev
-57.46% Max drawdown

-------------------

I'm not sure simply adding additional asset classes really helps. Maybe a different weighting has some effect.
Whoa, check your dates. Portfolio Visualizer only has data for global ex-US back to 1986, REIT to 1994, and commodities to 2007, so those comparisons are all covering totally different time periods.
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Found an article updating Gibson’s original graph. As I said at the start of the thread, I don’t think the specific asset classes are significant. What I do think is important is an appreciation for the benefits of diversifying across uncorrelated sources of return.

https://www.forbes.com/sites/greggfishe ... 98f6912231

Dave
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Just thought I’d give my old thread a bump. As stated at the top, the specific asset classes are almost irrelevant. What is relevant is how less than perfectly correlated assets mix in a portfolio. Whether we are talking asset classes, styles, alternatives, geography, they all have the potential to move a portfolio towards the northwest corner, more efficient. Lately there have been discussions regarding the presence or absence of the small cap premium, current valuations, whether REITs are a separate asset class. It’s all relevant within this framework.

Dave
Topic Author
Random Walker
Posts: 5561
Joined: Fri Feb 23, 2007 7:21 pm

Re: Expanding on The Rewards Of Multiple Asset Class Investing

Post by Random Walker »

Thought I’d give this thread another bump in light of recent threads on diversifying across factors.

Dave
Post Reply