Author Topic: How to find portfolios that better meet your needs  (Read 193062 times)

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
How to find portfolios that better meet your needs
« on: March 07, 2016, 05:09:39 PM »
So I've been playing around with asset data again, and 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.



This displays the Real CAGR and the single worst year since 1972 for every possible portfolio of up to five different assets that I have data for (more than 83,000 total portfolios).  The "TSM" marks the total stock market.  Believe it or not, there are about 40,000 portfolio options that returned at least as much as the stock market alone but with more manageable worst years.  Rather than immediately conceding that you should ramp up risk to improve returns, why not first explore how diversification can accomplish the same goal with less downside?

To help find the portfolios that work best for you, I've built a tool called the Portfolio Finder that navigates this cloud of options and sorts them by your own needs and preferences.  I recommend starting here for the detailed explanation.  While past performance cannot predict future returns, you can still learn a lot about the power of diversification in the process.

BTW, I'm sharing this not to try to talk people out of their investment plan, but simply to present an alternative way of looking at the problem.  Asset allocation is a lot more interesting and powerful than many people realize!
« Last Edit: March 08, 2016, 05:18:20 PM by Tyler »

Interest Compound

  • Pencil Stache
  • ****
  • Posts: 655
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #1 on: March 07, 2016, 06:23:25 PM »
Wow!

I am absolutely floored! Great work Tyler!

Putting on my skeptic hat, I don't see how this data can be used to formulate a portfolio. Here are the first things that come to mind:

Out of 83,000 total portfolios, about half of them did better than TSM in terms of risk/reward. So that means about half of them worse. Isn't that the point of just buying the whole stock market? I understand you're looking at many asset classes which aren't composed of stocks, but that's the most interesting part. Despite looking at so many different asset classes, that dynamic was still in play! Buy the whole stock market, and you'll get average returns.

After thinking about that, I started wondering how much of this is affected by Survivorship Bias. Completely out of your control of course, but:
  • Why is it that we only have data for these specific asset classes going back to 1972? Surely the 1972 investor had access to more. Is it because these asset classes did better than the rest?
  • If an asset class crashed and burned in 1980, would data still have been calculated on it for 36 more years, and been made available in a publicly available format so you could download it and include it in your charts? How many of these assets will be meaningless to us in 36 years? I can imagine the 1972 investor choosing their equivalent of Bitcoin...
  • Is this list full of assets which are here only because of their stellar risk/reward performance between 1972 and today? Looking over the asset list, this clearly doesn't apply to many of the options here, but maybe it's those options which are at the bottom of the chart. Gold, REITS, Emerging Markets, or Small Cap Value do seem to be a part of nearly every top performing asset here.

50/50 TSM/International stocks outperformed TSM during this period, but just barely. That might be a better comparison, but 5.96% compared to 5.87% isn't a big deal.

In the end, I'm still floored :) but the take-away seems to be, "If the 1972 investor tried to predict which combination of assets (which did well enough from 1972 to 2016 for the data to still exist) would've outperformed TSM on a risk/reward basis, they would've had about a 50/50 chance of getting it wrong. And to get in that top 50 you almost had to have picked one of the small handful of assets which really outperformed."

I just don't see how this information could be useful which choosing a portfolio, besides being used as evidence towards picking total market index funds.
« Last Edit: March 07, 2016, 06:28:43 PM by Interest Compound »

thd7t

  • Handlebar Stache
  • *****
  • Posts: 1348
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #2 on: March 07, 2016, 06:25:25 PM »
I am not terribly surprised to see this. It's very close to the average return. The risk for the worst year seems a bit high, but absent data about the best year, it's strange to show that data point.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #3 on: March 07, 2016, 08:28:19 PM »
Out of 83,000 total portfolios, about half of them did better than TSM in terms of risk/reward. So that means about half of them worse. Isn't that the point of just buying the whole stock market? I understand you're looking at many asset classes which aren't composed of stocks, but that's the most interesting part. Despite looking at so many different asset classes, that dynamic was still in play! Buy the whole stock market, and you'll get average returns.

That's one way to look at it. My interpretation is that buying the whole stock market does provide the average return but with above average risk.  But the good news is that you don't have to go to any extremes to generally improve the risk/reward tradeoff.  Just a little diversification goes a long way!  The other interesting takeaway is just how many portfolios had returns similar to the stock market but with low percentages of stocks and minimal down years.  I don't think many investors expect to see that.  So I do believe you can learn a lot by studying some of the clusters (both good and bad) for lessons in diversification strategy. 

The selection of assets is the widest variety I can find with good data going back as far as I can.  I understand your questions about certain assets and don't know all of the answers.  That's exactly why I included the option in the calculator to exclude ones you're not comfortable with. 
« Last Edit: March 08, 2016, 07:46:24 AM by Tyler »

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #4 on: March 07, 2016, 08:35:42 PM »
I am not terribly surprised to see this. It's very close to the average return. The risk for the worst year seems a bit high, but absent data about the best year, it's strange to show that data point.

For reference, the chart is roughly based on the concept of the "efficient frontier" that uses standard deviation as the proxy for portfolio "risk".  I personally find that measure to be not very intuitive or actionable in the real world, so I chose the worst year as something that investors can immediately relate to.  Nobody thinks "Oh crap, my standard deviation is too high right now!", especially if their portfolio is volatile to the upside.  But one bad year will spook a ton of investors out of the market altogether.

mrpercentage

  • Handlebar Stache
  • *****
  • Posts: 1235
  • Location: PHX, AZ
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #5 on: March 08, 2016, 12:25:49 AM »
Thank you for sharing

BFGirl

  • Pencil Stache
  • ****
  • Posts: 766
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #6 on: March 08, 2016, 06:36:52 AM »
Tyler,

I assume the portfolio charts site is yours?  I was on it last week and it made me feel much better about my more conservative investment strategy. 

I can't find that you mention much about municipal bonds or municipal funds and wondered if you had any information on those as part of a portfolio.

Thanks and great work!

Retire-Canada

  • Walrus Stache
  • *******
  • Posts: 8683
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #7 on: March 08, 2016, 08:30:28 AM »
You can be 100% invested in stocks and not just buy TSM. There are a number of stock markets around the world to diversify your investments.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #8 on: March 08, 2016, 09:22:11 AM »
Tyler,

I assume the portfolio charts site is yours?  I was on it last week and it made me feel much better about my more conservative investment strategy. 

I can't find that you mention much about municipal bonds or municipal funds and wondered if you had any information on those as part of a portfolio.

Thanks and great work!

Yep - I'm glad you find it helpful!

I don't currently have any municipal bond data, but may add some in the future. 
« Last Edit: March 08, 2016, 05:23:51 PM by Tyler »

thd7t

  • Handlebar Stache
  • *****
  • Posts: 1348
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #9 on: March 08, 2016, 10:13:37 AM »
I am not terribly surprised to see this. It's very close to the average return. The risk for the worst year seems a bit high, but absent data about the best year, it's strange to show that data point.

For reference, the chart is roughly based on the concept of the "efficient frontier" that uses standard deviation as the proxy for portfolio "risk".  I personally find that measure to be not very intuitive or actionable in the real world, so I chose the worst year as something that investors can immediately relate to.  Nobody thinks "Oh crap, my standard deviation is too high right now!", especially if their portfolio is volatile to the upside.  But one bad year will spook a ton of investors out of the market altogether.
It's funny, a few years ago, a 401k manager was really pushing the standard deviation to show how bad volatility could be (and to push people toward a hugely underperforming/high cost fund).

Matumba

  • 5 O'Clock Shadow
  • *
  • Posts: 81
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #10 on: March 08, 2016, 10:19:03 AM »
Following

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #11 on: March 08, 2016, 10:28:49 AM »
Can I assume the TSM is short for the total US stock market and includes no international?

Interest Compound

  • Pencil Stache
  • ****
  • Posts: 655
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #12 on: March 08, 2016, 10:34:23 AM »
Can I assume the TSM is short for the total US stock market and includes no international?

Correct. Here's what it looks like with the standard 3 fund portfolio:



I-TOT is Total International. The world cap-weighted stock market returned 5.96% compared to 5.87% with TSM alone.

brooklynguy

  • Handlebar Stache
  • *****
  • Posts: 2204
  • Age: 43
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #13 on: March 08, 2016, 11:07:28 AM »
Thanks for putting together another excellent tool, Tyler!

One important limitation that I think is worth highlighting is that all of this data (the entire cloud of data points in the chart which the calculator helps the user navigate) reflects a single cycle (from 1972 to the present), which seems obvious from your description but may not immediately register for those of us accustomed to thinking in terms of the multi-cycle outputs generated by cFIREsim and FIREcalc (and most of Portfolio Charts' other calculators).

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #14 on: March 08, 2016, 11:25:52 AM »
Thanks for putting together another excellent tool, Tyler!

One important limitation that I think is worth highlighting is that all of this data (the entire cloud of data points in the chart which the calculator helps the user navigate) reflects a single cycle (from 1972 to the present), which seems obvious from your description but may not immediately register for those of us accustomed to thinking in terms of the multi-cycle outputs generated by cFIREsim and FIREcalc (and most of Portfolio Charts' other calculators).

Yes -- that's true.  I hope to eventually incorporate that same multi-cycle methodology into this tool as well, but that will take a bit more development.  Just getting the results for one investment cycle down to a manageable file size was quite the effort!

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #15 on: March 08, 2016, 02:15:20 PM »
Tyler - it's great work and it's interesting however I don't use it and I'll state why:-

1. The limited data-set as listed above. This though isn't my biggest issue or concern.
2. It's backtesting asset allocations. My personal opinion is that getting the right asset allocation isn't a statistical science. Asset allocation in my opinion is extremely important however as you are using it to work out your current and future asset allocation. I don't think that a tool like that helps if you try and figure the optimal asset allocation for current/future needs.

So your admittedly mildly attention seeking thread doesn't really phase me. I don't believe in 100% stocks in your asset allocation and I believe in asset allocation however I think it's more of a personal preference and not just a returns based issue. I look more at principles and I think you need a decent amount of stocks in your portfolio. The decent amount is up to everyone's personal preference.


4tify

  • Bristles
  • ***
  • Posts: 347
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #16 on: March 08, 2016, 02:37:16 PM »
amazing stuff thank you!

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #17 on: March 08, 2016, 02:43:24 PM »
@Steveo -- I don't disagree with anything you said.  There's much more to picking a portfolio than simply relying on backtesting.  But I do believe that tools like this are valuable as one piece of the larger puzzle to validate portfolio theories and explore new ideas.  I recommend keeping an open mind while at the same time being smart about it. 
« Last Edit: March 08, 2016, 03:23:45 PM by Tyler »

pbkmaine

  • Walrus Stache
  • *******
  • Posts: 8927
  • Age: 67
  • Location: The Villages, Florida
The Case Against 100% Stocks (and what to do about it)
« Reply #18 on: March 08, 2016, 03:06:13 PM »
http://performance.morningstar.com/fund/ratings-risk.action?t=VFIAX&region=usa&culture=en_US

In terms of risk: the Vanguard 500 index fund has a lower standard deviation than its (large blend) category, a higher Sharpe ratio, greater upcapture and lower downcapture. These are all good things.
« Last Edit: March 08, 2016, 03:09:21 PM by pbkmaine »

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #19 on: March 08, 2016, 03:35:30 PM »
@Steveo -- I don't disagree with anything you said.  There's much more to picking a portfolio than simply relying on backtesting.  But I do believe that tools like this are valuable as one piece of the larger puzzle to validate portfolio theories and explore new ideas.  I recommend keeping an open mind while at the same time being smart about it.

I think what you've done and information is really good. I honestly think we are on the same page with investing - maybe not the same asset allocation but the same general principles.

The information that you've provided if used wisely can be a great help. If you are using it to nit pick tiny details in your asset allocation I think it's unlikely to work exactly that way in the future. If you use it to help work through a plan that you are comfortable with I think it's great.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #20 on: March 08, 2016, 05:24:29 PM »
You can be 100% invested in stocks and not just buy TSM. There are a number of stock markets around the world to diversify your investments.

Very true. I appreciate your point and updated the thread title to promote a more constructive conversation.

Interest Compound

  • Pencil Stache
  • ****
  • Posts: 655
Re: How to find portfolios that better meet your needs
« Reply #21 on: March 08, 2016, 08:52:38 PM »
You can be 100% invested in stocks and not just buy TSM. There are a number of stock markets around the world to diversify your investments.

Very true. I appreciate your point and updated the thread title to promote a more constructive conversation.

Now THIS is a conversation I can get behind :)

So if I do the math and calculate I only need a 5% real CAGR, and exclude everything that isn't either a total market index fund, or a treasury it looks like I have some nice options with much lower volatility:



#10 looks pretty interesting, let's plot that on a chart vs 50/50 USA/International stocks:




Ok, looks like I almost match the returns of 100% stocks, but with much lower volatility. Perfect! However, this is just a single deposit, sitting there for 44 years. Let's see what this would look like during the accumulation phase:

Initial Amount: $10,000
Contributing: $10,000 a year, inflation adjusted



Hey cool! I actually beat out 100% stocks by a little bit, and again with much lower volatility! Win! Let's see how this looks during the withdrawal phase, once I'm retired:

Initial Amount: $1,000,000
Withdrawal: $40,000 a year, inflation adjusted



Ouch! Sure I had less volatility, but I also had less money. I'd need a 50% gain just to match the 100% stock portfolio. Ok, ok, maybe this isn't a fair comparison. What would've happened if I contributed $10,000 a year until I reached $1,000,000, then withdrew an inflation adjusted $40,000 with the 4% rule?

100% stocks reached a million in 1985. Today the portfolio is worth $5,486,439.

33/33/34 USA/International/LTT reached a million in 1986. Today the portfolio is worth $5,766,920.

Seems like a wash, besides the extra year of retirement of course. I'm not really sure where I'm going with this. It seems the calculation is much more complex than it first seems. Despite the tool showing this portfolio was better, it still lost to 100% stocks in my opinion. This is assuming of course, that you're suited to handling the extra volatility, which I think is one of your main points in creating this. Most people aren't.

As you noted, "tools like this are valuable as one piece of the larger puzzle to validate portfolio theories and explore new ideas." I'd have to say I agree with that.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #22 on: March 08, 2016, 09:19:52 PM »
This is assuming of course, that you're suited to handling the extra volatility, which I think is one of your main points in creating this. Most people aren't.

Exactly.  A portfolio that one can't hold through the worst times is a poor choice in both accumulation and retirement -- even the best theoretical returns evaporate the moment you sell low.  Those blessed with both the willingness and the capacity to weather deep and prolonged losses without worry can invest however they please.  For others, exploring a more enjoyable road to the same goal is a worthwhile exercise.  There are many good roads!
« Last Edit: March 09, 2016, 09:22:32 AM by Tyler »

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Re: How to find portfolios that better meet your needs
« Reply #23 on: March 08, 2016, 09:25:17 PM »
It seems the calculation is much more complex than it first seems.

Exactly. You are also using backdated data. I can't see how you can justify an asset allocation based just on this or any available data-set.

As you noted, "tools like this are valuable as one piece of the larger puzzle to validate portfolio theories and explore new ideas." I'd have to say I agree with that.

It is a way to check if what you are stating is even workable but I think a simple 3 fund portfolio (domestic stocks/international stocks/domestic bonds) based on your own risk profile with low cost index options being chosen for each asset group has as much chance of doing well in the future than any other option. It's also so freaken easy to manage.
« Last Edit: March 08, 2016, 09:27:20 PM by steveo »

Seppia

  • Pencil Stache
  • ****
  • Posts: 616
  • Age: 43
  • Location: NYC
Re: How to find portfolios that better meet your needs
« Reply #24 on: March 08, 2016, 10:49:42 PM »
Thanks for your work Tyler, amazing as usual.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: How to find portfolios that better meet your needs
« Reply #25 on: April 03, 2016, 05:36:21 AM »
Great tool Tyler.  You continue to amaze.

Thanks for the insights, Interest Compound.  I enjoyed your perspective on it.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

wienerdog

  • Pencil Stache
  • ****
  • Posts: 587
Re: How to find portfolios that better meet your needs
« Reply #26 on: April 03, 2016, 07:58:11 AM »
Interest Compound what tool do you use to compare the two portfolios?

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: The Case Against 100% Stocks (and what to do about it)
« Reply #27 on: April 03, 2016, 09:45:35 AM »
@wienerdog: it's Portfolio Visualizer.  I love how there are several good tools out there for this kind of thing.   

You can be 100% invested in stocks and not just buy TSM. There are a number of stock markets around the world to diversify your investments.

After studying this a bit more, I found the distribution of all-stock portfolios interesting.  The blue dots are all of the portfolios that exclusively invest in some mix of US stocks, international stocks, and REITs.



Also, here's a chart of all of the Lazy Portfolios on the site.



BTW, since first posting this I added a feature to the Portfolio Finder to allow you to "require" assets.  So let's say you are interested in exploring new assets that might be good complements to the stock index fund you already own.  Now you can set your current asset as a requirement and easily surf portfolios that include it. 
« Last Edit: April 03, 2016, 09:58:26 AM by Tyler »

forummm

  • Walrus Stache
  • *******
  • Posts: 7374
  • Senior Mustachian
Re: How to find portfolios that better meet your needs
« Reply #28 on: April 03, 2016, 10:00:08 AM »
Tyler, I'm having trouble seeing how this analysis could possibly be accurate. Maybe I don't understand the methodology. But how could TSM have a lower CAGR than ~95% of all stocks? Perhaps the survivorship bias is at play here. Perhaps something else. TSM should be approximately in the middle of the equities dot unless you're doing something unexpected. What am I missing? How do you explain this seeming inconsistency?

forummm

  • Walrus Stache
  • *******
  • Posts: 7374
  • Senior Mustachian
Re: How to find portfolios that better meet your needs
« Reply #29 on: April 03, 2016, 10:07:49 AM »
This displays the Real CAGR and the single worst year since 1972 for every possible portfolio of up to five different assets that I have data for (more than 83,000 total portfolios).

Bolding is mine. Perhaps this is why your results are so unexpected. You probably don't have data for all of the stocks that existed during that time frame. You probably have data for ones that did relatively well compared to the market.

beltim

  • Magnum Stache
  • ******
  • Posts: 2957
Re: How to find portfolios that better meet your needs
« Reply #30 on: April 03, 2016, 10:13:38 AM »
Tyler, I'm having trouble seeing how this analysis could possibly be accurate. Maybe I don't understand the methodology. But how could TSM have a lower CAGR than ~95% of all stocks? Perhaps the survivorship bias is at play here. Perhaps something else. TSM should be approximately in the middle of the equities dot unless you're doing something unexpected. What am I missing? How do you explain this seeming inconsistency?

Tyler's not looking at individual stocks.  The different dots are different portfolios with varying proportions of stocks, including domestic and international, large and small cap, value and growth, as well as REITs.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #31 on: April 03, 2016, 10:29:01 AM »
Tyler, I'm having trouble seeing how this analysis could possibly be accurate. Maybe I don't understand the methodology. But how could TSM have a lower CAGR than ~95% of all stocks? Perhaps the survivorship bias is at play here. Perhaps something else. TSM should be approximately in the middle of the equities dot unless you're doing something unexpected. What am I missing? How do you explain this seeming inconsistency?

The total stock market in this case represents the cap-weighted market for US stocks only.  Practically speaking, the performance is indistinguishable from an S&P500 fund as the returns are overwhelmingly driven by a handful of US mega caps.  The All-Stock portfolios in the chart include combinations of all kinds of other stocks either not heavily represented in a TSM fund (like small caps and REITs) or not represented at all (like international stocks and emerging markets). 

So there are two ways of looking at this.  First, this could show that the period since 1972 was not the best for US large cap stocks relative to other options.  Second, this could be a visual example of how riskier stocks (small caps, emerging markets, etc) tend to provide higher returns in the long run provided you can handle the increased volatility. 

I'm working on studying multiple timeframes to try to answer this, but that will take time.  My instinct is that it's a little bit of both, but with option 2 being the dominant factor. 

Interest Compound

  • Pencil Stache
  • ****
  • Posts: 655
Re: How to find portfolios that better meet your needs
« Reply #32 on: April 03, 2016, 10:30:29 AM »
This displays the Real CAGR and the single worst year since 1972 for every possible portfolio of up to five different assets that I have data for (more than 83,000 total portfolios).

Bolding is mine. Perhaps this is why your results are so unexpected. You probably don't have data for all of the stocks that existed during that time frame. You probably have data for ones that did relatively well compared to the market.

Perhaps the survivorship bias is at play here.

This is exactly what I suspected in my first post (second post in the thread). These graphs (along with Tyler's incredible work in creating them) are amazing, as it highlights how easily Survivorship Bias slides it's way into our analysis. When everything is plotted on a graph like this it's clear to see there's something missing...but how many of us have ever seen a graph like this? How many of us have the technical prowess of Tyler to make a similar graph for the investments we're considering?

------------------------------------------
It is easy to fall victim to the Survivorship Bias.  After any process that leaves behind survivors, the non-survivors are often destroyed or muted or removed from your view. If failures becomes invisible, then naturally you will pay more attention to successes. Not only do you fail to recognize that what is missing might have held important information, you fail to recognize that there is missing information at all.
------------------------------------------

http://youarenotsosmart.com/2013/05/23/survivorship-bias/


Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #33 on: April 03, 2016, 10:45:14 AM »
This displays the Real CAGR and the single worst year since 1972 for every possible portfolio of up to five different assets that I have data for (more than 83,000 total portfolios).

Bolding is mine. Perhaps this is why your results are so unexpected. You probably don't have data for all of the stocks that existed during that time frame. You probably have data for ones that did relatively well compared to the market.

Not in this case.  Nine of the stock funds are subsets of the small/mid/large and value/blend/growth stocks already contained in the TSM fund.  The Portfolio Finder simply mixes them in different proportions to what you find in a typical market-cap weighted fund.  Same with the international options.  It's not survivorship bias, as the underlying companies are all the same. 

It's more an artifact of the natural risk/reward of certain assets.  People understand that higher volatility is compensated by higher rewards with stocks vs. bonds -- well, the same thing applies with some types of stocks over others. 

Another way to look at it is that one should expect an equal weight stock portfolio (where all stocks are purchased in equal proportions) to be in the middle of the blue blob.  A TSM fund is cap weighted, not equally weighted, and will naturally be biased towards one edge.
« Last Edit: April 03, 2016, 11:43:34 AM by Tyler »

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6633
Re: How to find portfolios that better meet your needs
« Reply #34 on: April 03, 2016, 10:50:56 AM »
Tyler - InterestCompound mentioned gold in an earlier post, but there was no follow up on that.  Do you include gold as one of the possible assets in the 1972-present modeling?

I agree with your FAQ (http://portfoliocharts.com/portfolio-finder-faq/) that the start date is important.  For gold, it skews things significantly owing to the 1973-1974 tripling of gold's price.  One approach would be to allow a trailing 40 year period (1976-2015?) as an alternative to the current 44 year time frame.

InterestCompound - Others should be warned that bonds did something unique from 1980 until now, where their rates dived a staggering amount.  So a back test would certainly show the capital gains from long-term treasuries, but we shouldn't expect the same thing in the upcoming decades.  A period of 10% interest rates is very different from our current situation of 2-3% rates.  More for other's benefit than yours, but thought it should be mentioned so people don't automatically invest in a great looking back test.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #35 on: April 03, 2016, 11:03:09 AM »
Yes, gold is one of the options.  You can exclude or require any asset you choose.

Also from the FAQ:

Quote
Won’t the recommended portfolios change if you study different timeframes?

They will.  Unlike most calculators on the site, the Portfolio Finder only studies one investing period — the longest one I have data for from 1972 to the present.  If one runs the same calculations starting in a different year, the final returns may be very different.  But importantly, some portfolios are more start date sensitive than others.

The calculator is designed specifically to seek out the most stable portfolios that have weathered anything the markets have thrown at them over the last four and half decades with minimal losses.  While a single average return does not tell the whole story, many of the most appealing low loss options are naturally the least start date sensitive and are good examples for further study.

I absolutely recommend studying all portfolios from a broader perspective that accounts for varying timeframes.  The Heat Map and Portfolio Growth calculators are particularly helpful for visualizing returns uncertainty and identifying historical trends.  Consider the Portfolio Finder a starting point in your research, not the final stop.

The site offers a wide assortment of calculators to help you study portfolios over all possible timeframes with any combinations of assets that you like.  This includes times when gold and bonds did very poorly, and also other times when stocks lost money for more than a decade.  Use them to study Portfolio Finder results, and the start date sensitivity is very clear. For example, compare this to this
« Last Edit: April 03, 2016, 11:47:12 AM by Tyler »

dandypandys

  • Pencil Stache
  • ****
  • Posts: 545
  • Age: 47
  • Location: USA
Re: How to find portfolios that better meet your needs
« Reply #36 on: April 03, 2016, 11:23:21 AM »
This might help me soon, although at the moment, it is all a bit confusing.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #37 on: April 03, 2016, 11:26:55 AM »
This is exactly what I suspected in my first post (second post in the thread). These graphs (along with Tyler's incredible work in creating them) are amazing, as it highlights how easily Survivorship Bias slides it's way into our analysis. When everything is plotted on a graph like this it's clear to see there's something missing...

Sometimes what's missing isn't alternative data that we want to believe exists to prove our present beliefs, but an important unseen bridge in our own incomplete understanding of how things work.

I'm not an economist and I don't claim to have all of this figured out.  I enjoy making charts and tools like these because it helps me learn about my own investments, and I like sharing them with others as I do.  It's amazing to think about what I believed as an article of investing faith just a few years ago before exploring the topic for myself.  So don't make the same mistake by taking my data at face value and running with it without question, but also don't dismiss it just because it challenges your assumptions.  I like to think we can all learn new things together. 
« Last Edit: April 03, 2016, 05:38:25 PM by Tyler »

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #38 on: April 03, 2016, 11:33:50 AM »
This might help me soon, although at the moment, it is all a bit confusing.

Yeah, this discussion is admittedly getting into some advanced investing topics.  Don't worry -- it doesn't have to be as complicated as it sounds.  For starters, I recommend just browsing the many lazy portfolios on the site to see how asset allocation can affect investment performance.  Pick the simplest one that makes sense to you, and you'll be on the right track. 

dandypandys

  • Pencil Stache
  • ****
  • Posts: 545
  • Age: 47
  • Location: USA
Re: How to find portfolios that better meet your needs
« Reply #39 on: April 03, 2016, 11:47:16 AM »
:)
Love this page you linked, so pretty! This is more my level, thanks for your hard work.

wapiti

  • 5 O'Clock Shadow
  • *
  • Posts: 36
Re: How to find portfolios that better meet your needs
« Reply #40 on: April 03, 2016, 12:11:49 PM »
Thanks for this data.Quite impressive.

To improve the simulation, it would be good to test portfolio with unequal weight of each asset.

It would be also good to compare with some MSCI indice.
For exemple, the max CAGR (what ever the drowdown) is SCV, I-EM, I-SML with 10.33.
If we check the MSCI WORLD and MSCI WORLD VALUE gross returns since 31 Dec 1974 (yes three years are missing compare to the data in portfolio finder) https://www.msci.com/resources/factsheets/index_fact_sheet/msci-world-value-index.pdf, the GAGR is 10.60 and 11.61.
« Last Edit: April 03, 2016, 12:18:42 PM by titi22 »

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #41 on: April 03, 2016, 12:53:26 PM »
Studying equal weight portfolios was the approach that bounded the problem just enough to be technically doable.  That should get you in the ballpark.  For refining that further with unequal weights, I recommend playing with the site calculators

Regarding the MSCI numbers, please note that the ones in your link are nominal and the ones at PortfolioCharts are real (adjusted for inflation).  But it's true that different sources may have different values for past returns. 
« Last Edit: April 03, 2016, 12:58:50 PM by Tyler »

wapiti

  • 5 O'Clock Shadow
  • *
  • Posts: 36
Re: How to find portfolios that better meet your needs
« Reply #42 on: April 03, 2016, 01:06:07 PM »
I will check the calculator.
However, unegal weight complexify a lot the calculation, but could be interesting.

I have really missed that your numbers were adjusted to the inflation.
I will try to find or create the real number of MSCI indices.

YoungInvestor

  • Bristles
  • ***
  • Posts: 409
Re: How to find portfolios that better meet your needs
« Reply #43 on: April 03, 2016, 04:23:55 PM »
Very interesting. I am in complete agreement with your thought that more diversification could lead to better results with less risk.

A good idea to see if these patterns are consistent would be to run the same test for 1972-1993 and 1994-2015 and see if portfolios end up in the same "quadrant" both times ( better risk-reward).

i.e.: how many of these 40k portfolios were better over the two 22-year periods? If it is significantly more than 20k, you may be on to something really interesting.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #44 on: April 03, 2016, 06:12:13 PM »
A good idea to see if these patterns are consistent would be to run the same test for 1972-1993 and 1994-2015 and see if portfolios end up in the same "quadrant" both times ( better risk-reward).

I agree.  The Portfolio Finder currently defines "risk" as the worst historical year, but others may define it as start date dependency.  I'm working on a way to quantify that (and other metrics) in an efficient and repeatable way so that I can apply it directly to the Portfolio Finder.  It just takes time. 

In the meantime, hopefully that gives people a reason to experiment with the other calculators to explore that for themselves.  ;)

missmoneymachine

  • 5 O'Clock Shadow
  • *
  • Posts: 39
Re: How to find portfolios that better meet your needs
« Reply #45 on: April 03, 2016, 08:17:47 PM »
Thought this was an interesting thread considering I'm in the process of researching various asset allocations for my investment plan.  Portfolio visualizer and portfolio charts are great tools.

I think this shows how important diversity can be (#2 is a rough draft of a portfolio I'm considering). The start year is 1972.  Sorry if the image is too large


AdrianC

  • Handlebar Stache
  • *****
  • Posts: 1211
  • Location: Cincinnati
Re: How to find portfolios that better meet your needs
« Reply #46 on: April 04, 2016, 08:51:38 AM »
The site offers a wide assortment of calculators to help you study portfolios over all possible timeframes with any combinations of assets that you like.  This includes times when gold and bonds did very poorly, and also other times when stocks lost money for more than a decade.  Use them to study Portfolio Finder results, and the start date sensitivity is very clear. For example, compare this to this.

Would it be useful (and possible) to have an input for date range?

As has been pointed out several times, portfolios with lots of gold are going to look really good when starting in 1972.

I used the Bogleheads back-test spreadsheet to compare a few portfolios:

1972-2015 (44 years)
TSM real CAGR 5.87%
Permanent Portfolio real CAGR 4.35% with rebalancing, 3.95% without rebalancing.

1985-2015 (31 years)
TSM real CAGR 7.9%
Permanent Portfolio real CAGR 4.61% with rebalancing, 5.25% without rebalancing.

Interesting results.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1198
Re: How to find portfolios that better meet your needs
« Reply #47 on: April 04, 2016, 10:03:03 AM »
Because there are so many portfolios involved, I have to do all of the processing for the Portfolio Finder on my own computer and create a single output file small enough to work on the site.  Right now a user-selected date range isn't practical.  But as I mentioned before, I'm working on different metrics for the tool including one that measures start date dependency.  Keep an eye out.  For now, the Portfolio Finder allows you to exclude gold or any other asset for any reason you wish. 

FWIW, the Heat Map calculator does the exact same multi-timeframe calculation you did, but instead of picking just two it does all possible 990 at the same time.  Use the calculator on the site, and you can hover your mouse over each cell to see the specific CAGR.



When you look at the big picture, the gold explosion in the first two years really didn't skew the results outside of the historic norm for that particular portfolio.  Low volatility in all kinds of extreme economic environments is a hallmark of the Permanent Portfolio, and is the reason it's designed the way it is.   Most "traditional" portfolios (especially the TSM) are far more start date dependent.  That said, I really don't want to make this discussion about debating specific portfolios or assets as there are plenty of other threads for that.  I just want to point out that I discuss the issue of not looking at a single average when comparing portfolios more than anyone else I'm aware of. 

The Portfolio Finder is simply a tool to help you explore new ideas you've perhaps never considered.  Don't take the quoted return as gospel.  When you find a portfolio that seems interesting, always check the uncertainty and trends with the other tools to make sure it meets your needs. 
« Last Edit: April 04, 2016, 01:49:03 PM by Tyler »

FrugalFan

  • Pencil Stache
  • ****
  • Posts: 895
Re: How to find portfolios that better meet your needs
« Reply #48 on: April 04, 2016, 10:16:41 AM »
Keep up the great work, Tyler. I'm still learning and don't understand everything yet, but this definitely helps.
« Last Edit: April 04, 2016, 10:18:25 AM by Travelling Biologist »

AdrianC

  • Handlebar Stache
  • *****
  • Posts: 1211
  • Location: Cincinnati
Re: How to find portfolios that better meet your needs
« Reply #49 on: April 05, 2016, 05:49:50 AM »
FWIW, the Heat Map calculator does the exact same multi-timeframe calculation you did, but instead of picking just two it does all possible 990 at the same time.  Use the calculator on the site, and you can hover your mouse over each cell to see the specific CAGR.


Thanks, Tyler. I wasn't aware of the mouse hover feature (it doesn't work on the tablet). It works great on the PC and is just what I was looking for. Nice job.