Author Topic: How to find portfolios that better meet your needs  (Read 19744 times)

dandypandys

  • Bristles
  • ***
  • Posts: 469
  • Age: 43
  • Location: USA
Re: How to find portfolios that better meet your needs
« Reply #50 on: April 05, 2016, 06:14:23 AM »
very neat!

ender

  • Walrus Stache
  • *******
  • Posts: 5111
Re: How to find portfolios that better meet your needs
« Reply #51 on: April 05, 2016, 07:27:26 AM »
These charts would be a lot more useful if they included actual axis information to indicate what the dots were.

And frankly were not started in 1972. That seems like an awfully weird date to randomly pick for this sort of analysis. How consistent is your plot if you vary the start date in increments of 3 years? Do the 1969, 1966, 1963 and then also the 1975, 1978, 1981... 1995 charts all show the same trends?

Are you including dividend reinvestment in your TSM return? Or is that just share appreciation? Presumably your "real" return is post-inflation? What are some examples of the 83,000 portfolios that you picked that were better than the TSM? Where did the data come from, particularly for returns? What were the asset classes?

Any post showing a chart like yours and stating:

Quote
in the process I uncovered a nice visual demonstration of why just putting all of your money in a total stock market fund may not be the optimal choice from a risk management perspective.

Leaves me, as an engineering/data analytics type of person, quite skeptical.

Honestly, more than anything else that chart actually convinces me how unbelievably good of an investment market index funds are, especially for working Mustachians. When over a 44 year period an index fund is still that competitive across a huge range of portfolios when it took over 10 years to get any real return from the start date, that makes it is still a very competitive option to me (the post-inflation SP500 dividend reinvested fund took over 10 years to recover).

That sounds really, really damn nice and really a good investment choice.

It seems somewhat intellectually dishonest to build this sort of framework around a 1972 retirement start date and make the claims you are stating about risk analysis. Every single one of your statements needs qualification of "if you started in 1972" until you can show that the same trends hold true regardless of time period start date. Especially considering you are using gold/TSM in your calculations. Heavy gold portfolios will get hammered hard in this analysis if you move the start date back even only a few years ($263/ounce in Jan 1972 vs $362 in Jan 1973 vs nearly $750 in Jan 1974).

If many of those "better" portfolios are including very large gold holdings then they will be very sensitive to your start date because you have arbitrarily picked a year where Gold gained nearly all of its inflation-adjusted return in the past 45 years. etc.

This is just one of the reasons why it is incredibly important to not cherry pick data and make a more wholistic approach when doing this sort of analysis.


Interest Compound

  • Pencil Stache
  • ****
  • Posts: 658
Re: How to find portfolios that better meet your needs
« Reply #52 on: April 05, 2016, 08:05:22 AM »
These charts would be a lot more useful if they included actual axis information to indicate what the dots were.

And frankly were not started in 1972. That seems like an awfully weird date to randomly pick for this sort of analysis. How consistent is your plot if you vary the start date in increments of 3 years? Do the 1969, 1966, 1963 and then also the 1975, 1978, 1981... 1995 charts all show the same trends?

Are you including dividend reinvestment in your TSM return? Or is that just share appreciation? Presumably your "real" return is post-inflation? What are some examples of the 83,000 portfolios that you picked that were better than the TSM? Where did the data come from, particularly for returns? What were the asset classes?

Any post showing a chart like yours and stating:

Quote
in the process I uncovered a nice visual demonstration of why just putting all of your money in a total stock market fund may not be the optimal choice from a risk management perspective.

Leaves me, as an engineering/data analytics type of person, quite skeptical.

Honestly, more than anything else that chart actually convinces me how unbelievably good of an investment market index funds are, especially for working Mustachians. When over a 44 year period an index fund is still that competitive across a huge range of portfolios when it took over 10 years to get any real return from the start date, that makes it is still a very competitive option to me (the post-inflation SP500 dividend reinvested fund took over 10 years to recover).

That sounds really, really damn nice and really a good investment choice.

It seems somewhat intellectually dishonest to build this sort of framework around a 1972 retirement start date and make the claims you are stating about risk analysis. Every single one of your statements needs qualification of "if you started in 1972" until you can show that the same trends hold true regardless of time period start date. Especially considering you are using gold/TSM in your calculations. Heavy gold portfolios will get hammered hard in this analysis if you move the start date back even only a few years ($263/ounce in Jan 1972 vs $362 in Jan 1973 vs nearly $750 in Jan 1974).

If many of those "better" portfolios are including very large gold holdings then they will be very sensitive to your start date because you have arbitrarily picked a year where Gold gained nearly all of its inflation-adjusted return in the past 45 years. etc.

This is just one of the reasons why it is incredibly important to not cherry pick data and make a more wholistic approach when doing this sort of analysis.

For what it's worth, the year 1972 is the furthest back we have reliable information for most of these asset classes. That's the same reason why https://www.portfoliovisualizer.com/ only goes back to 1972, while Cfiresim goes back to 1871. Cfiresim doesn't include those asset classes.

While I agree it's a bit dangerous to weigh this information too strongly when choosing a portfolio, I believe Tyler would agree with that statement as well. The real flaw here, as you mentioned, is the lack of context for those who might be new to investing. Even if he put "if you started in 1972" next to every chart, people would still interpret that information inaccurately. I'd like to see each chart be a Cfiresim style chart that runs through each available starting point from 1972 onward.

I'd also like to see a comparison of the portfolios listed on his site, with other portfolios which ended up failing, but seemed equally reasonable. In other words, if someone did this same analysis in 1972, and choose a similar slice-n-dice portfolio that looked good over the previous 40 years (1932-1972), what kind of portfolio would they have likely ended up with? How would that have done over the next 40 years? The 1972 investor had access to millions of different portfolios, what are the chances they would've been successful if they employed this same logic?

-----------------------------------
"There may be better investment strategies than owning just three broad-based index funds, but the number of strategies that are worse is infinite." ~John Bogle, Vanguard founder
-----------------------------------

From all the evidence I've seen, it's impossible to predict ahead of time, so the only logical portfolio is a total market portfolio.
« Last Edit: April 05, 2016, 08:12:24 AM by Interest Compound »

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #53 on: April 05, 2016, 08:29:07 AM »
It looks like I touched a nerve.

@Ender -- Some things are best presented as a central resource rather than turning every message board post into an annotated scientific paper.  That's why I started a website, which was linked in the OP and repeatedly throughout the thread.  If you take the time to read it, you'll find all of the answers you're looking for. 

Post explaining the chart in detail: http://portfoliocharts.com/2016/03/07/the-ultimate-portfolio-guide-for-all-types-of-investors/

Tool to help you browse all of the portfolios in the chart: http://portfoliocharts.com/portfolio/portfolio-finder/

Calculation Methodology: http://portfoliocharts.com/methodology/

Asset information for everything used in the tool, including where the data comes from and why it starts in 1972: http://portfoliocharts.com/assets/

Frequently Asked Questions: http://portfoliocharts.com/portfolio-finder-faq/

We've also had a nice discussion in this thread about things like gold and start date dependency, and there are examples of multiple ways to deal with that.  That includes using other tools that I built specifically for studying this issue (http://portfoliocharts.com/calculators/), and using the Portfolio Finder "exclude" feature to avoid certain assets entirely.  I've also talked about how I'm looking to expand the original analysis to do this directly. 

I hope that helps.  If after reading the provided resources you still have questions, feel free to PM me.  And if the information makes you even more confident in your current investing strategy, that's great! 
« Last Edit: April 05, 2016, 09:45:46 AM by Tyler »

aFrugalFather

  • Stubble
  • **
  • Posts: 219
    • Life/Finance Blog
Re: How to find portfolios that better meet your needs
« Reply #54 on: April 05, 2016, 09:12:55 AM »
Interesting analysis, thank you for going through the work.  I am letting it all settle in and then will have to follow up when I am awake  :)

ender

  • Walrus Stache
  • *******
  • Posts: 5111
Re: How to find portfolios that better meet your needs
« Reply #55 on: April 05, 2016, 11:42:53 AM »
It looks like I touched a nerve

I suppose I really don't like people cherry picking data and burying all the fine print/assumptions in offsite resources while simultaneously making fairly global claims/suggestions about the cherry picked data they presented being applicable in a fairly general perspective.

/shrug


arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28260
  • Age: -999
  • Location: Seattle, WA
Re: How to find portfolios that better meet your needs
« Reply #56 on: April 05, 2016, 12:09:55 PM »
That's a little harsh.
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with three kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #57 on: April 05, 2016, 04:08:05 PM »
I suppose I really don't like people cherry picking data and burying all the fine print/assumptions in offsite resources while simultaneously making fairly global claims/suggestions about the cherry picked data they presented being applicable in a fairly general perspective.

I talk about the dangers of cherry picking data more than anyone I know, so I appreciate your concern.  The Portfolio Finder methodology is completely above board for those who take the time to look deeper before passing judgment, and I'm always happy to answer questions.  I've also freely acknowledged the shortcoming you're concerned about, and have offered specialized tools designed to study multiple timeframes in detail while I continue to make improvements.  One is free to disagree and do their own thing, but I'll just say that I believe the intellectual dishonesty accusation is unfair.  :/
« Last Edit: April 05, 2016, 05:33:37 PM by Tyler »

wienerdog

  • Pencil Stache
  • ****
  • Posts: 506
Re: How to find portfolios that better meet your needs
« Reply #58 on: April 05, 2016, 05:53:49 PM »
Tyler quick question.  In the heat map when you do hover over each block that is the CAGR for that year or the running average for all the blocks to the left of it?  Thanks again for all the great tools on your site!

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #59 on: April 05, 2016, 07:10:38 PM »
Tyler quick question.  In the heat map when you do hover over each block that is the CAGR for that year or the running average for all the blocks to the left of it?  Thanks again for all the great tools on your site!

Each cell in the heat map represents the inflation-adjusted CAGR for the timeframe starting in the year listed on the left and with the duration stated at the top.  For example, if you want to know the 10-year annualized return starting in 1980 you find the row that corresponds to 1980 and the column that corresponds to year 10.  If you want to see the individual annual returns, just read down the first column.  Make sense?
« Last Edit: April 05, 2016, 07:13:38 PM by Tyler »

wienerdog

  • Pencil Stache
  • ****
  • Posts: 506
Re: How to find portfolios that better meet your needs
« Reply #60 on: April 05, 2016, 07:27:38 PM »
Make sense?

Yup! That is what I thought and meant, you just explain it so much better.  Thanks again!

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #61 on: April 06, 2016, 08:42:43 PM »
I thought a lot about this thread yesterday, which led to a unique level of focus.  Late last night I had a technical breakthrough, and today has been a whirlwind of work!

As promised, I've updated the Portfolio Finder to now take a start-date-independent approach to returns calculations.  Take a look!  I bet the results may surprise you even more than before. 

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28260
  • Age: -999
  • Location: Seattle, WA
Re: How to find portfolios that better meet your needs
« Reply #62 on: April 06, 2016, 11:54:58 PM »
That damn golden butterfly, always making me doubt my AA.
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with three kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

MustacheAndaHalf

  • Handlebar Stache
  • *****
  • Posts: 1964
Re: How to find portfolios that better meet your needs
« Reply #63 on: April 07, 2016, 01:00:15 AM »
Including "gold" and years 1973-1974 is cherry picking the data.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28260
  • Age: -999
  • Location: Seattle, WA
Re: How to find portfolios that better meet your needs
« Reply #64 on: April 07, 2016, 01:08:59 AM »
Including "gold" and years 1973-1974 is cherry picking the data.

...did you even read the latest article, linked two posts above yours?

I don't know why I'm asking, clearly not.

Go read it.  :)
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with three kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

ender

  • Walrus Stache
  • *******
  • Posts: 5111
Re: How to find portfolios that better meet your needs
« Reply #65 on: April 07, 2016, 06:03:42 AM »
I thought a lot about this thread yesterday, which led to a unique level of focus.  Late last night I had a technical breakthrough, and today has been a whirlwind of work!

As promised, I've updated the Portfolio Finder to now take a start-date-independent approach to returns calculations.  Take a look!  I bet the results may surprise you even more than before.

This is a really interesting chart:



I would be curious what the minimum CAGR return plotted onto the "since 1972 CAGR" chart would look like (ie instead of using the worst year as X axis using the "since 1972 CAGR" return). Though I'm not sure how well this would work but it might be interesting.

 It seems a good comparison for risk analysis would be looking at what portfolios do better overall but have a worse minimum 15-year CAGR vs those that are less great overall but a much better 15-year minimum.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #66 on: April 07, 2016, 08:53:14 AM »
I would be curious what the minimum CAGR return plotted onto the "since 1972 CAGR" chart would look like (ie instead of using the worst year as X axis using the "since 1972 CAGR" return). Though I'm not sure how well this would work but it might be interesting.

That's an interesting idea.  I need a break from the data crunching, but I'll look into it. 

It seems a good comparison for risk analysis would be looking at what portfolios do better overall but have a worse minimum 15-year CAGR vs those that are less great overall but a much better 15-year minimum.

The best tool I have for studying this for individual portfolios is the Long Term Returns calculator.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #67 on: April 08, 2016, 11:03:58 PM »
I would be curious what the minimum CAGR return plotted onto the "since 1972 CAGR" chart would look like (ie instead of using the worst year as X axis using the "since 1972 CAGR" return). Though I'm not sure how well this would work but it might be interesting.

Here you go -- just be careful interpreting this one.  As you've pointed out, it's possible that portfolios may move along the X-axis a bit depending on the long-term start year.  ;)  But based on the methodology, they can't move up the Y.  Clearly some portfolios are more consistent than others


« Last Edit: April 08, 2016, 11:56:56 PM by Tyler »

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28260
  • Age: -999
  • Location: Seattle, WA
Re: How to find portfolios that better meet your needs
« Reply #68 on: April 09, 2016, 12:09:15 AM »
SON OF A

That damn golden butterfly, always making me doubt my AA.
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with three kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

dandypandys

  • Bristles
  • ***
  • Posts: 469
  • Age: 43
  • Location: USA
Re: How to find portfolios that better meet your needs
« Reply #69 on: April 09, 2016, 07:45:58 AM »
Hello,
 I am investigating the site a bunch, it is fun, I'm still learning though only a couple months into it all. I am trying to see if my asset allocation I have currently, is a close match to any of the well known portfolios.
Anyway, because my 403b does not have total bond, i was forced to choose something else, and was advised over on bogleheads by a nice advisor called Taylor to choose https://www.standard.com/eforms/16276.pdf   Apparently this stable asset fund  as a good option from what i had available to replace the total bond index.

my question is, is this type of fund on this list here under a different name or not included for some reason? https://portfoliocharts.com/assets/
Thanks!

beltim

  • Magnum Stache
  • ******
  • Posts: 2863
Re: How to find portfolios that better meet your needs
« Reply #70 on: April 09, 2016, 08:11:41 AM »
I would be curious what the minimum CAGR return plotted onto the "since 1972 CAGR" chart would look like (ie instead of using the worst year as X axis using the "since 1972 CAGR" return). Though I'm not sure how well this would work but it might be interesting.

Here you go -- just be careful interpreting this one.  As you've pointed out, it's possible that portfolios may move along the X-axis a bit depending on the long-term start year.  ;)  But based on the methodology, they can't move up the Y.  Clearly some portfolios are more consistent than others



That's pretty cool.  Any chance we could get you to identify a few of the portfolios that have CAGRs over 9%, or minimum 15 year returns over 6%?  It's be interesting to look at those.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #71 on: April 09, 2016, 10:26:58 AM »
Hello,
 I am investigating the site a bunch, it is fun, I'm still learning though only a couple months into it all. I am trying to see if my asset allocation I have currently, is a close match to any of the well known portfolios.
Anyway, because my 403b does not have total bond, i was forced to choose something else, and was advised over on bogleheads by a nice advisor called Taylor to choose https://www.standard.com/eforms/16276.pdf   Apparently this stable asset fund  as a good option from what i had available to replace the total bond index.

my question is, is this type of fund on this list here under a different name or not included for some reason? https://portfoliocharts.com/assets/
Thanks!

It's true that many accounts don't have all of the options you'd like, but that does seem like a pretty decent intermediate fixed income option. 

I don't have data on stable value funds like that mostly because they're a mixed financial product rather than a true index.  And a secondary issue is getting good data all the way back to 1972.  The total bond market is probably the closest comparison on the site, but consider the numbers simply directional rather than an exact representation of that fund. 

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1162
Re: How to find portfolios that better meet your needs
« Reply #72 on: April 09, 2016, 10:39:33 AM »
That's pretty cool.  Any chance we could get you to identify a few of the portfolios that have CAGRs over 9%, or minimum 15 year returns over 6%?  It's be interesting to look at those.

Ignore the rankings -- I just manually typed these in. 



Basically, they all have very heavy small cap bias.  The tradeoff is that the worst individual years are particularly painful.  Those last few including LTTs are pretty interesting, though.  They're the least painful of the bunch, and if you plug them into a heat map that's about as blue as you can get.

I'm not sure I'd recommend that anyone invest this way, but it's certainly an interesting data point.

SON OF A

That damn golden butterfly, always making me doubt my AA.

:)  That made me LOL.
« Last Edit: April 09, 2016, 10:49:46 AM by Tyler »

dandypandys

  • Bristles
  • ***
  • Posts: 469
  • Age: 43
  • Location: USA
Re: How to find portfolios that better meet your needs
« Reply #73 on: April 09, 2016, 01:14:05 PM »
Hello,
 I am investigating the site a bunch, it is fun, I'm still learning though only a couple months into it all. I am trying to see if my asset allocation I have currently, is a close match to any of the well known portfolios.
Anyway, because my 403b does not have total bond, i was forced to choose something else, and was advised over on bogleheads by a nice advisor called Taylor to choose https://www.standard.com/eforms/16276.pdf   Apparently this stable asset fund  as a good option from what i had available to replace the total bond index.

my question is, is this type of fund on this list here under a different name or not included for some reason? https://portfoliocharts.com/assets/
Thanks!

It's true that many accounts don't have all of the options you'd like, but that does seem like a pretty decent intermediate fixed income option. 

I don't have data on stable value funds like that mostly because they're a mixed financial product rather than a true index.  And a secondary issue is getting good data all the way back to 1972.  The total bond market is probably the closest comparison on the site, but consider the numbers simply directional rather than an exact representation of that fund.

ok good to know.. thanks!
I listened to the podcast with you being interviewed on The Voluntary Life, it helped.
Just needed the youtube screen capture to go with it :)

AdrianC

  • Handlebar Stache
  • *****
  • Posts: 1035
  • Location: Cincinnati
Re: How to find portfolios that better meet your needs
« Reply #74 on: April 09, 2016, 01:30:48 PM »
SON OF A

That damn golden butterfly, always making me doubt my AA.

The Swenson looks pretty good too.

I'm reading Swenson's book right now. He's making a lot of sense.

Monkey Uncle

  • Handlebar Stache
  • *****
  • Posts: 1530
  • Location: West-by-god-Virginia
Re: How to find portfolios that better meet your needs
« Reply #75 on: April 10, 2016, 05:03:25 AM »
SON OF A

That damn golden butterfly, always making me doubt my AA.

The Swenson looks pretty good too.

I'm reading Swenson's book right now. He's making a lot of sense.

It seems like there is a cluster of portfolios that were pretty stable while providing total returns that matched or exceeded TSM (golden butterfly, coffeehouse, ivy, swensen, and merriman).  I suspect the actual positioning of these relative to each other would be different depending on the historical time period selected.  What's important to note is that all of them combine a diversity of partially correlated, volatile but potentially high return assets (more so than the S&P 500) with a substantial component of bonds for ballast.  So the take-away message for me isn't so much that one should try to match one of these portfolios precisely.  Rather, there are a few principles that seem to be emerging:

1. Focus the core of the portfolio on high-return, high-risk asset classes.
2. Those high-risk, high-return assets should be un-correlated as much as possible to balance out the ups and downs (i.e., don't just divide everything among various categories of u.s. small caps, for example). 
4. Some exposure to real estate and commodities seems to be a recurring theme.
3. Keep about a third of the portfolio in low-risk bonds to further cushion the downs during major melt downs when panic-selling impacts almost all equity-type asset classes (e.g., 2008).