The Money Mustache Community

Learning, Sharing, and Teaching => Investor Alley => Topic started by: Radagast on October 05, 2019, 08:08:16 PM

Title: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 05, 2019, 08:08:16 PM
For a long time I have noticed that naively filling every asset class (or every second, third, or fourth asset class) often gave better results than most "expert" portfolios using both https://portfoliocharts.com/ and https://www.portfoliovisualizer.com/. Here is my formal investigation into it.

Rules: I equally weighted every or every 2nd, 3rd, or 4th asset class using both withdrawal types (safe and perpetual) at https://portfoliocharts.com/portfolio/portfolio-matrix/. Spare change went to cash, or was taken away from cash, to keep all other assets equally weighted. This was a form of sandbagging my hypothesis, because Cash is Trash as Portfolio Charts confirms.

Note 1: I like to use withdrawal rates best of all backtesting objectives. Are long term returns more important, or is stability more important? With safe withdrawal rate there is no bullshitting about one or the other is more important: if it was more important why didn't it result in a higher safe withdrawal rate?

Note 2: There are really only let's generously say 3 data points affecting the results here. All of the portfolios except Japan had their worst case in 1970-1974. Some of the countries give slightly different results so lets call that a second effective datapoint, as we can throw cold water on some of the weakest contenders. Pay special attention to Japan, as that is the only true out of sample datapoint for most "expert" portfolios.

Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant. I did on odd occasions throw out TSM and International TSM for simplicity.

Note 4: SWR is determined based on the single worst year in the record for 30-year rolling returns. Average returns would make a very different ranking.

Here are the results. Naive portfolios are white, "experts" get colors. The withdrawal rate is listed for each, could be some typos but the rankings should be correct. I am surprised to note that even with cash sandbagging, the "experts" mostly demonstrated no skill and underperformed naive asset class selection. Only a few expert portfolios appear to show skill, including the Ivy, 7Twelve, and Pinwheel. Even the 7Twelve and Pinwheel portfolios seem to be more about extreme diversification than skill, with 12 and 8 slices. Perhaps a future data point will show that the Ivy Portfolio was just lucky? The average "expert" portfolio looks to have underperformed the average naive portfolio.
(https://i.imgur.com/YVTlwn4.jpg)

And here is what went into the naive portfolios. TMI I know.
(https://i.imgur.com/z9w7nNh.jpg)

If I ever make my own portfolio I will call it the "I'm an Idiot Portfolio" :D

Thanks mjr! Had to run just as I posted, not sure what went wrong.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 07, 2019, 11:32:56 AM
A post like this deserves way more attention.  Nice work!

Note 2: There are really only let's generously say 3 data points affecting the results here. All of the portfolios except Japan had their worst case in 1970-1974. Some of the countries give slightly different results so lets call that a second effective datapoint, as we can throw cold water on some of the weakest contenders. Pay special attention to Japan, as that is the only true out of sample datapoint for most "expert" portfolios.

Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant. I did on odd occasions throw out TSM and International TSM for simplicity.

I think these are very important points.  You may call your selection methodology "naive", but I'd argue that the built-in diversification with an eye on Japan is a very intelligent approach.  My big takeaway is not that the various gurus are unskilled, but that perhaps they should pay more attention to home country bias and the benefits of true asset diversification rather than simple stock tilts.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: kenmoremmm on October 07, 2019, 12:58:06 PM
where do i send the check for all of this work?

after compiling these data, did it change your investment strategy?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 07, 2019, 01:42:55 PM
Finally PTF. For the first time ever.

Quite intrigued.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 07, 2019, 07:38:52 PM
A post like this deserves way more attention.  Nice work!

Note 2: There are really only let's generously say 3 data points affecting the results here. All of the portfolios except Japan had their worst case in 1970-1974. Some of the countries give slightly different results so lets call that a second effective datapoint, as we can throw cold water on some of the weakest contenders. Pay special attention to Japan, as that is the only true out of sample datapoint for most "expert" portfolios.

Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant. I did on odd occasions throw out TSM and International TSM for simplicity.

I think these are very important points.  You may call your selection methodology "naive", but I'd argue that the built-in diversification with an eye on Japan is a very intelligent approach.  My big takeaway is not that the various gurus are unskilled, but that perhaps they should pay more attention to home country bias and the benefits of true asset diversification rather than simple stock tilts.
Thank you for bumping! I was hoping for discussion but it's bad form to bump my own post, I was getting scared the whole thing would be a one post wonder.

I agree that 1/n weighting in general is more sophisticated than naive. Both for individual stocks and for portfolio components I have read research shoing that 1/n is surprisingly effective, for example people in retirement plans who simply apply equal weight to every available option. Provided the expense ratios are low, that is actually a very good strategy.

You are also probably thinking that including a "real" asset is also something many "gurus" miss, and I am coming around to that line of thought, especially as it did not seem to hurt the Japan case, provided they are used in moderation. In fact I think the Portfolio Charts weighting which is about 40% home country stock*, 36% foreign stock, 16% bonds cash, 8% commodities is probably a pretty ideal asset allocation. *Depending how you count REITs. Take the Merriman Ultimate Buy and Hold: if it had diverted a percentage of its bonds to commodities it would have been one of the best portfolios out there (at the cost of even more slices).

One thing I notice is that guru portfolios do best in Japan and Germany, which I expect is a result of heavy bond allocations. Relative to the English speaking countries, I believe bond yields on those two started higher, and are now much lower. This created a huge tailwind for gurus who favored huge helpings of  long term and intermediate bonds, but did no favors to Bernstein who strongly prefers large amounts of short term bonds. I expect this to reverse at some point, and for that reason I think gurus who have outsized allocations to either extreme are poorly diversified.

But I am not convinced now that any guru is better than randomly selecting a variety of funds with relatively low expense ratios.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 07, 2019, 07:54:57 PM
where do i send the check for all of this work?

after compiling these data, did it change your investment strategy?
No check, like most posts this was for self confirmation, vindication, and gratification! Often people post a list of funds and ask what to invest in, and forumers would say things like "76.2% S&P500, 14.2% midcap, 9.6% small cap will give you the equivalent of total market which is what you want" while I would be like "cut the precision, you have large, mid, small, bond, and international funds with low expense ratios, just equally weight those and call it a day, it's not like anyone has any idea anyhow." So I am basically just trying to prove myself right.

No changes, I put these together in a few hours. My strategy don't turn on no dime. It did influence me in two ways though. 1st, I have always been compelled by the zen of just putting it all in total world stock market and saying "ooohhhmmmm." However I notice that allocations with fewer than four slices often make trips to the bottom, and in some cases live there permanently. All of the naive allocations had at least 7 slices plus cash and did very well. I have increasingly been thinking that 7-15 slices might be better although many people would call that crazy, especially if spread around nine accounts. 2nd, while there is only one solid data point for gold/commodities, it did not really have a significant downside anywhere (in moderation) and the Ivy Portfolio did great for Japan in the out-of-sample point. I am warming toward a 8%-16%ish allocation to those.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 07, 2019, 07:58:20 PM
Also big thanks to mjr. For some reason my original links did not embed the images, and I had to leave for a few hours just as I posted. mjr made working image links and then deleted the post.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: effigy98 on October 07, 2019, 08:25:51 PM
Interesting. Max diversification looks like the most consistent winner.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: sailinlight on October 07, 2019, 09:30:28 PM
Is there a dummy version of the results of this? It seems that the takeaway is that for a 30 year retirement, Golden Butterfly affords the largest withdrawal rate (or the more useful converse 'greatest chance of not running out of money').
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: vand on October 08, 2019, 02:24:03 AM
Diversification is the only free lunch in investing.

All this shows that there are many successful ways to build a portfolio. It may not be an overexaggeration to suggest that the "naive" approach is to think that a larg concentration in stocks is the One True Way to wealth accumulation.

Title: Re: Portfolio Design: Idiots v. Gurus
Post by: effigy98 on October 08, 2019, 04:35:55 PM
What exactly happened in Japan to the Golden Butterfly? Is there not enough deflation protection?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Kierun on October 08, 2019, 05:25:17 PM
whut?  can someone dumb this down for me as to what it I'm looking at here? looks interesting but I'm an idiot and don't understand, all I get is...colors.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 08, 2019, 07:23:39 PM
@Kierun, I think it's working like this:

Bottom section (the part with all the little black blocks):
Each sector under a yellow label, such "0th" (zero-th) or "5th" (fifth) shows the allocation of investments in one of the "idiot" or "naive" portfolios that the OP made. Each little block shows an investment type, and the number inside the block is the % of the portfolio that went into that investment type. So the first or "0th" example, in the Domestic or USA row, the left-most investment block is TSM (which stands for Total Stock Market); TSM had 4% of the portfolio invested in it. You can go through all the rows in the portfolio, and each portfolio's numbers should add up to 100%. The abbreviations (TSM for Total Stock Market, LCB for Large Cap Blend or something like that, etc) are explained in the website that the OP used to make all these calculations (portfoliocharts.com), which has a ton of calculators based on historical investment data.

Top section:
The investment results of the idiot portfolios are shown, along with the investment results of some other, more carefully crafted portfolios recommended by investment "experts" such as the authors of various investment books. The color coded boxes show the results of the experts' portfolios, referred to as gurus in the thread title; the white boxes show the results of the "idiot" aka "naive" portfolios, where the OP basically picked investments using various random methods (his choices were detailed in the bottom section of the colorful post). Each colored box in the top section gives the common name of the expert portfolio that it describes. Each white box refers to bottom section - "1st naive" in the white box means the portfolio labeled "1st" in the bottom section; the portfolio header "1st" is yellow in the bottom section.

The plain red boxes in the top section refer to portfolios that were one of the worst two portfolios in some country (any country). So the red portfolios were historically risky for some investors; implicitly, perhaps a red flag. The other colors show expert portfolios that escaped the scary cellar of these league standings, so to speak. The expert portfolios are each described in detail on portfoliocharts.com, which Tyler the Mustachian created as his gift to the world for free portfolio calculations based on historical data.

Each box in the top section ends in a number giving the highest withdrawal rate that worked in all investment years for that portfolio during the period where data could be found (I think 1970 to present) - in other words, if you started by investing in the worst possible year, the box shows the investment rate you could have used. I think that this required 30 year periods though, so years that started less than 30 years ago may not be included. Perhaps investors who started in 2000, for example, who suffered from the tech bust of 2001-2003 and then the 2008-09 Great Financial Crisis, will have lower withdrawal rates than shown.

I think the "Safe" columns meant "didn't run out of money for 30 years at that rate". You can see that for each country, the portfolios are in order from highest withdrawal rate to lowest, starting at the top. The "Perpetual" columns show the rate where the portfolio ended the test period with as much money as it started; the perpetual rate is how much you could withdraw per year historically and still end up with the original investment value. You can see that's a lower rate than the so-called safe rate, but it would let your investments last longer, so it's good to know for people who have longer investment timeframes, such as people retiring at 35 and living to 90. For them, the perpetual rate might actually be safer!

Clear as mud, eh??
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 08, 2019, 07:46:22 PM
Is there a dummy version of the results of this? It seems that the takeaway is that for a 30 year retirement, Golden Butterfly affords the largest withdrawal rate (or the more useful converse 'greatest chance of not running out of money').
Not really :-| in general I am saying that explanations of why past stock, bond, and "real assets" did what they did might not be correct so much as they are good stories. As Vand says, diversification is the only free lunch.

My asset allocation advice (which I am tweaking slightly as a result of this) is pretty simple:
At least 50% of total assets in stocks
At least 25% but not more than 50% of total allocation in your home country/currency stocks (or other business)
Meaningful allocation to international stocks (unhedged)
Include 15-40% bonds
Include a "real asset," and yes your home or rental real estate count.
In fact investment-grade real estate would be fine for 80% of this
Plus lots of caveats!

Like I said in OP, Japan was the only true test of the Golden Butterfly. Performance in the US was dominated by a single year, 1973, which doesn't give much information going forward. My explanation of why a single data point is not very useful.here (https://forum.mrmoneymustache.com/investor-alley/4-swr-(or-3-x)-has-never-been-modeled-with-current-interest-rates/msg2459046/#msg2459046). Also gold behaved oddly in 1970-75 because over that period the government started off with price controls (at a very low price btw), and then gradually released it, so it isn't a particularly useful data point. You would have to experiment with substituting it with REITs or commodities or even cash to get a non-interventionist picture.
What exactly happened in Japan to the Golden Butterfly? Is there not enough deflation protection?
Both the largest stocks in Japan and the smaller value companies did poorly in that period. Gold was in the dead center of a two decade bear market when the Japanese needed it most in 1990. Cash did its usual thing and went nowhere. That left the 20% allocation to long term bonds to carry the entire thing. You might say it did not have enough deflation protection, but I say it did not have enough diversification ;). A 6th equally weighted slice to Total International Stock Market would have brought the Japan case way up, and if the 6th slice was International Small Cap Value instead it would have backtest as the best portfolio in the English speaking world in addition to being near the top in Japan and Germany.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 08, 2019, 07:54:02 PM
I think it's working like this:

Bottom section (the part with all the little black blocks): Each sector under a yellow label, such "0th" (zero-th) or "5th" (fifth) shows the allocation of investments in one of the "idiot" or "naive" portfolios that the OP made. Each little block shows an investment type, and the number inside the block is the % of the portfolio that went into that investment type. So the first or "0th" example, in the Domestic or USA row, the left-most investment block is TSM (which stands for Total Stock Market); TSM had 4% of the portfolio invested in it. You can go through all the rows in the portfolio, and each portfolio's numbers should add up to 100%. The abbreviations (TSM for Total Stock Market, LCB for Large Cap Bonds or something like that, etc) are explained in the website that the OP used to make all these calculations (portfoliocharts.com), which has a ton of calculators based on historical investment data.

Top section: The investment results of the idiot portfolios are shown, along with the investment results of some other, more carefully crafted portfolios recommended by investment "experts" such as the authors of various investment books. The color coded boxes show the results of the experts' portfolios, referred to as gurus in the thread title; the white boxes show the results of the "idiot" aka "naive" portfolios, where the OP basically picked investments using various random methods (his choices were detailed in the bottom section of the colorful post). Each colored box in the top section gives the common name of the expert portfolio that it describes. Each white box refers to bottom section - "1st naive" in the white box means the portfolio labeled "1st" in the bottom section; the portfolio header "1st" is yellow in the bottom section. I think that all the plain red boxes refer to portfolios viewed as generic, such as all stock or 60% stock 40% bonds, while other colors refer to specific expert portfolios. The expert portfolios are each described in detail on portfoliocharts.com, which Tyler the Mustachian created as his gift to the world for free portfolio calculations based on historical data.

Each box in the top section ends in a number giving the highest withdrawal rate that worked in all investment years for that portfolio - in other words, if you started by investing in the worst possible year, the box shows the investment rate you could have used. I think the "Safe" columns meant "didn't run out of money for 30 years at that rate". You can see that for each country, the portfolios are in order from highest withdrawal rate to lowest, starting at the top. The "Perpetual" columns show the rate where the portfolio ended the test period with as much money as it started; the perpetual rate is how much you could withdraw per year historically and still end up with the original investment value. You can see that's a lower rate than the safe rate, but it would let your investments last longer, so it's good to know for people who have longer investment timeframes, such as people retiring at 35 and living to 90.
Thanks BicycleB that is completely correct! Well except one minor change: I used red for any portfolio which made it to the bottom 2 rows for any country, which is why the "Larry Portfolio" is red, and also the Bill Bernstein "No-Brainer."
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 08, 2019, 07:56:09 PM
As a side note there might be other ways to accomplish a higher safe withdrawal rate beside a static asset allocation which are commonly discussed around here. For example, the reverse glide path using a 5-year CD ladder, or a variable withdrawal rate.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 08, 2019, 07:59:07 PM
@Radagast, thanks for answering.

I have a lot of boxes that appear on my computer as red. In the USA Safe column, is the red for Larry Portfolio supposed to be the same as the color for the bottom three fields in that column?

ETA: Never mind, I now realize what you said. Read means bottom of ANY column - as in, historically risky somewhere! Very thoughtful. Thx, I can edit my explanation post now.


ETA: Ok, main explanation post edited.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 08, 2019, 08:09:31 PM
@Radagast, thanks for answering.

I have a lot of boxes that appear on my computer as red. In the USA Safe column, is the red for Larry Portfolio supposed to be the same as the color for the bottom three fields in that column?

ETA: Never mind, I now realize what you said. Read means bottom of ANY column - as in, historically risky somewhere! Very thoughtful. Thx, I can edit my explanation post now.
Yep, to my mind picking future winners is hard, but avoiding the worst outcomes is probably possible. To bad about the "No Brainer" though, if it had used intermediate term bonds and small value stocks instead of just small it would have been notably higher. Even the "Larry Portfolio" would probably have been better in real life in the UK, as it dedicated half its 70% bond allocation to inflation protected bonds.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 08, 2019, 10:48:29 PM
Each box in the top section ends in a number giving the highest withdrawal rate that worked in all investment years for that portfolio during the period where data could be found (I think 1970 to present) - in other words, if you started by investing in the worst possible year, the box shows the investment rate you could have used. I think that this required 30 year periods though, so years that started less than 30 years ago may not be included. Perhaps investors who started in 2000, for example, who suffered from the tech bust of 2001-2003 and then the 2008-09 Great Financial Crisis, will have lower withdrawal rates than shown.

Normally you're correct about 30-year SWRs for periods that started less than 30 years ago, but I actually figured out a way to account for that without trying to predict the future. You can read about it here: https://portfoliocharts.com/2017/03/21/how-to-predict-withdrawal-rates-without-a-crystal-ball/
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: vand on October 09, 2019, 03:57:54 AM
As Vand says, diversification is the only free lunch.


Haha, you probably know that they're not my words... I'm just requoting Harry Markowitz, whose work I believe puts him firmly in any list of the top 10 most influential investors of the 20th century alongside the likes of Buffett and Bogle.  If today's major trend is low cost passive indexing, I think the next big trend in financial markets may well be the rise of multi-asset diversification.  This may well come about when the current paradigm eventually morphs into something more hostile for traditional portfolios.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: effigy98 on October 09, 2019, 09:13:47 AM
What exactly happened in Japan to the Golden Butterfly? Is there not enough deflation protection?
Both the largest stocks in Japan and the smaller value companies did poorly in that period. Gold was in the dead center of a two decade bear market when the Japanese needed it most in 1990. Cash did its usual thing and went nowhere. That left the 20% allocation to long term bonds to carry the entire thing. You might say it did not have enough deflation protection, but I say it did not have enough diversification ;). A 6th equally weighted slice to Total International Stock Market would have brought the Japan case way up, and if the 6th slice was International Small Cap Value instead it would have backtest as the best portfolio in the English speaking world in addition to being near the top in Japan and Germany.

Ahhh yes, that would have helped a lot. Based on the Dalios drum beat for emerging, probably the best place vs too much developed to have less correlation with US market.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Kierun on October 09, 2019, 10:04:05 AM
@BicycleB Many mahalos!  It all makes so much more sense now.  And thanks @Radagast for all that info. 
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 09, 2019, 10:18:56 AM
I wish someone could analyze Nassim Taleb’s black swan portfolio, which consists of a huge allocation to treasuries plus a small allocation to long call options on the stock market. I wouldn’t expect it to do amazingly well, but it would definitely avoid major damage during corrections and apply upside-only leverage in recoveries. It’s too bad we have no easy way to measure it because historical options pricing is so hard to find or calculate.

What I really wish someone could analyze is a protected put or collared portfolio, which is another way to avoid big drawdowns. I suppose you’d have to build a spreadsheet options pricing model and feed it with historical volatility data, interest rates, etc. The slightest inaccuracy, such as misestimating theta or gamma, would throw it completely off.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ardrum on October 09, 2019, 11:01:48 AM
I don't have it open right now, but doesn't portfoliocharts show the highest SWRs with something like a 2:1 ratio of US small cap value to international small cap/small cap value with a smidge (5-10%) of gold?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 09, 2019, 11:27:02 AM
^Even if the data show that for the period described, how likely is it that such a precise portfolio will continue its outperformance?

AFAIK, there's no logic-based reason why portfoliocharts.com limits its data to 1970 or so to the present. I think it's just that that's the era Tyler was able to find data for. Is that long enough of a period to distinguish between strategies that have enduring value, vs strategies that had a hot streak during that time?

Questions like these are why I find Radagast's results so intriguing. Maybe they suggest that diversification as a general principle, rather than any particular allocation, is what has enduring value.

Still pondering how to sort these things out.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 09, 2019, 12:11:53 PM
AFAIK, there's no logic-based reason why portfoliocharts.com limits its data to 1970 or so to the present. I think it's just that that's the era Tyler was able to find data for.

Correct.  One could argue gold data under the Bretton Woods system is irrelevant today or that very old stock data is sketchy due to company count in the sample, but neither drove the 1970 date on Portfolio Charts.  It's just the limits of what I can find for so many assets, and I'll happily expand that if/when I find reliable new sources. 

I will note that Eugene Fama and Ken French have done tons of research on the small and value premiums going back to 1927 with plenty of data and reasoning to back it up.  Larry Swedroe has also written prolifically on the topic.  I encourage anyone who would like to know more to think beyond the PC data and read about it for yourself before either dismissing it entirely or betting your retirement on it.  My numbers are just one voice, not the final word. 

I also generally discourage people from getting too crazy with seeking out the highest SWR for a concentrated portfolio of any one asset.  Be smart about it. All numbers aside, diversification is also your friend when it comes to risk management. 
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ardrum on October 09, 2019, 01:28:15 PM
Great discussion and analysis.  It's hard to disagree with an approach grounded in broad diversification at minimal tax/cost.

I always find it fun to look at past data even though it's definitely no crystal ball.  I'd bet sticking to whatever allocation one goes with is a good way to stay out of the bottom of the list vs frequently jumping into and out of different allocations chasing results.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 09, 2019, 01:54:03 PM
Diversification is the only free lunch in investing.

Yep. Just as interesting as the variety of portfolios at the top of Radagast’s list is the remarkable consistency at the bottom.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: kenmoremmm on October 09, 2019, 02:25:38 PM
is there a way to cleanly rank the portfolios from an overall perspective? visually, golden butterfly, ivy, or 7twelve or one of the naives looks to be best. can you capture the average SWR's  as measured across all countries on the list?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 09, 2019, 03:41:51 PM
Diversification is the only free lunch in investing.

Yep. Just as interesting as the variety of portfolios at the top of Radagast’s list is the remarkable consistency at the bottom.

Here is the kicker. Will those portfolios on the bottom outperform over the next 20-50 years. I have real doubts being heavily invested in gold and bonds will perform well over that time period and those portfolios at the top appear to be heavy on gold and bonds. I have some bonds for safety but I'm not purchasing any gold. I also wonder if the world economy will continue to perform so poorly. I'm more than comfortable with having 50% of my portfolio in a world index tracker. In fact I think that this is hugely important based on being as diversified as possible within the equity market.

Diversification works but it's not just across markets but within markets.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 09, 2019, 08:25:06 PM
As a follow up, anyone care to guess how the naive portfolios did in terms of average returns? 'Cause I had some guesses and they totally fell short.

More notes:
Along with withdrawal rates, I also like to look at average returns. For one thing, the safe withdrawal rate is a single data point, meaning we can't draw very strong conclusions from it. Average still has the problem of being backward looking over a specific time period, but at least every year contributed equally to it. Also, it is good to get a perspective on how you will do in the 90% of the time when things are not going terribly.

These are arithmetic average returns, if I am reading this correctly. That would make them a little on the optimistic side for volatile portfolios because an 80% decrease has a far greater impact on your money than an 80% increase, which would not be reflected in the arithmetic average.

I never said it, but for the US "world stock" is 50% US/50% ex-US, which may give a rebalancing bonus. The other countries have a straight up global allocation.

The Pinwheel Portfolio and Ivy continue to shine, while Swensen, Merriman, and coffeehouse continue along the path of mediocrity which is probably good in terms of long term survival.

The Larry Portfolio has the distinction of being the only portfolio to be colored red for both safe withdrawal rate and average return. Whatever else you do, don't invest like this. 95% of the time this portfolio will have no upside, and 5% it has significant downside. I have always been mystified why this one exists because even a casual review of history shows government bonds have very little return and very little risk, until the issuing country gets into trouble and jacks up inflation while capping interest rates and the previously "safe" bonds switch to pure risk. The US did it in the 1940's, and other countries in this list did so to a much greater extreme.

So about those naive portfolios: don't forget they all have cash allocations. Cash is not a huge drawback for safe withdrawal rates, but it is a big bag of bowling balls for average returns.

Aaaannd I think it is pretty obvious that experts have no talent. They were already looking "meh" with the whole worst case withdrawal rates thing, but on average they are terrible.
(https://i.imgur.com/4HsoYE0.jpg)
Man I wish I could be this big an idiot. Is it still being naive if you try to act as if you were naive, or does the act of trying to act naive no longer make you truly naive?

Edit: fixed typo in image
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 09, 2019, 08:30:59 PM
What exactly happened in Japan to the Golden Butterfly? Is there not enough deflation protection?
Both the largest stocks in Japan and the smaller value companies did poorly in that period. Gold was in the dead center of a two decade bear market when the Japanese needed it most in 1990. Cash did its usual thing and went nowhere. That left the 20% allocation to long term bonds to carry the entire thing. You might say it did not have enough deflation protection, but I say it did not have enough diversification ;). A 6th equally weighted slice to Total International Stock Market would have brought the Japan case way up, and if the 6th slice was International Small Cap Value instead it would have backtest as the best portfolio in the English speaking world in addition to being near the top in Japan and Germany.

Ahhh yes, that would have helped a lot. Based on the Dalios drum beat for emerging, probably the best place vs too much developed to have less correlation with US market.
I recall that international small cap value generally backtests better than emerging, but I agree that going forward there is no reason to expect that. When in doubt, do both?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 09, 2019, 08:35:09 PM
I wish someone could analyze Nassim Taleb’s black swan portfolio, which consists of a huge allocation to treasuries plus a small allocation to long call options on the stock market. I wouldn’t expect it to do amazingly well, but it would definitely avoid major damage during corrections and apply upside-only leverage in recoveries. It’s too bad we have no easy way to measure it because historical options pricing is so hard to find or calculate.

What I really wish someone could analyze is a protected put or collared portfolio, which is another way to avoid big drawdowns. I suppose you’d have to build a spreadsheet options pricing model and feed it with historical volatility data, interest rates, etc. The slightest inaccuracy, such as misestimating theta or gamma, would throw it completely off.
I am skeptical of Taleb's barbell strategies. How do you know which asset will turn out to be riskless and which will turn out to have the highest return? If you are wrong you could end up with either all downside or no upside. The Larry portfolio attempts this (without options) and look how it ended up. How does he know that walking and weight lifting are all you need, and running and biking are pointless? I prefer the smorgasbord approach.

That said I wish I knew more about options, but rounding to the nearest 1/10th of a percent, my knowledge rounds to 0.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 09, 2019, 08:39:50 PM
I forgot my most important thank you to @Tyler ! I have done a lot of backtesting but am a lightweight. Tyler put together the entire Portfolio Charts website with a gajillion backtesting methods and personally found a huge amount of data which would otherwise not be available at all. Portfolio Charts fills a huge gap between Portfolio Visualizer which has a million assets and includes actual funds but scarcely goes back to the year 2000, and CfireSim which goes to 1872 but with a narrow range of assets and only for the US.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 09, 2019, 10:08:05 PM
It's interesting but it looks like some of the portfolios with a higher SWR have lower returns and vice versa. This is why I'm skeptical of the portfolios that are heavy in gold and cash type assets (bonds, treasuries, etc). Sure we might end up with poor returns in equities markets that result in those portfolios doing well that have gold and cash type assets but there is no guarantee that this will occur.

I like simple and I suppose naive portfolios foot the bill there.

It will be interesting how this changes over the next 50 odd years.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: SotI on October 10, 2019, 12:50:11 AM
Just posting to say "thanks"! This is quite some food for thought.
I find it also interesting how relatively poorly the "simple" Total Stock or Classic 60:40 have been performing.

My portfolio is regionally diversified, but so far I have not included my home property or REITs in it.

Diversification is the only free lunch in investing.

Yep. Just as interesting as the variety of portfolios at the top of Radagast’s list is the remarkable consistency at the bottom.

Here is the kicker. Will those portfolios on the bottom outperform over the next 20-50 years. I have real doubts being heavily invested in gold and bonds will perform well over that time period and those portfolios at the top appear to be heavy on gold and bonds. ...
I am wondering, too, considering FED/ECB policy over the last decade (from a European pov, I am particularly looking at Germany, UK and even Japan here for comparison).
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: vand on October 10, 2019, 04:09:37 AM
You can learn so much about investing from the myriad of good blogs (and books) available. Tyler's Portfoliocharts is just one example, but there are lots out there.

While I fully respect and enjoy MMM but it's not really an Investing blog, per se. Your universe becomes overly narrow if all you read is FIRE blogs which multiply the same lines of thinking when it comes to personal finances and how to invest.

For example I'm currently working my way through https://ofdollarsanddata.com/

Heer is an excellent post he wrote on the subject of AA and portfolio efficiency:

https://ofdollarsanddata.com/where-to-invest-when-youre-investing/

Again, none of this should come as a surprise if you have already really studied in detail the concepts of portfolio theory, but few FI blogs tackle the topic of investing with this level of throughness (and of course, it could be argued, why should they).

None of this contradicts what the FIRE community already preach (he is highly scathing of the Active/Hedge Fund industry, for example), but its about expanding your knowledge, digging beneath simplistic headline numbers - ie acknowledging that "distribution of outcomes matters greatly" - and tailoring your investing accordingly.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 10, 2019, 07:55:14 AM
It's interesting but it looks like some of the portfolios with a higher SWR have lower returns and vice versa. This is why I'm skeptical of the portfolios that are heavy in gold and cash type assets (bonds, treasuries, etc). Sure we might end up with poor returns in equities markets that result in those portfolios doing well that have gold and cash type assets but there is no guarantee that this will occur.

I like simple and I suppose naive portfolios foot the bill there.

It will be interesting how this changes over the next 50 odd years.

True. So much of this is history-specific. Since 1970, the US went off the gold standard, had a tech boom in equities, inflated real estate prices with government subsidized financing, and from the early 1980s to the present has had a bond rally of massive proportions. Meanwhile investors endured stagflation in the 70s, a tech bubble that burst in 2000, a real estate bubble that burst in 2008, and the worst recession since the 1930s. Will any of that happen in the next 50 years?

Selecting the optimal AA now would require knowledge of the future, not the past. Will there be a biotech boom driven by gene editing? Will bonds revert to their norms or go negative like in Europe? Will Fannie and Freddie Mac be broken up and sold off? Will the tech-driven attention economy hit a natural growth limit or see a backlash as their products become recognized as addictive? Will average savings rates continue to decline or are millennials naturally more frugal? Will there be a major war? Will blockchain disrupt banking and tank the dollar? How will demographic graying affect aggregate demand? Will the US remain a democracy? Will self-driving cars that we don’t own become the new transportation economy? Will parts of New Orleans, Miami, Charleston, Tampa, NYC, and Houston be abandoned and who pays for that? Will China continue growing, stagnate at some point, or collapse? What will be the average inflation rate for the next 3 decades?

If one knows any of this with certainty, portfolio selection becomes easy.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 11, 2019, 02:08:11 AM
It's interesting but it looks like some of the portfolios with a higher SWR have lower returns and vice versa. This is why I'm skeptical of the portfolios that are heavy in gold and cash type assets (bonds, treasuries, etc). Sure we might end up with poor returns in equities markets that result in those portfolios doing well that have gold and cash type assets but there is no guarantee that this will occur.

I like simple and I suppose naive portfolios foot the bill there.

It will be interesting how this changes over the next 50 odd years.

True. So much of this is history-specific. Since 1970, the US went off the gold standard, had a tech boom in equities, inflated real estate prices with government subsidized financing, and from the early 1980s to the present has had a bond rally of massive proportions. Meanwhile investors endured stagflation in the 70s, a tech bubble that burst in 2000, a real estate bubble that burst in 2008, and the worst recession since the 1930s. Will any of that happen in the next 50 years?

Selecting the optimal AA now would require knowledge of the future, not the past. Will there be a biotech boom driven by gene editing? Will bonds revert to their norms or go negative like in Europe? Will Fannie and Freddie Mac be broken up and sold off? Will the tech-driven attention economy hit a natural growth limit or see a backlash as their products become recognized as addictive? Will average savings rates continue to decline or are millennials naturally more frugal? Will there be a major war? Will blockchain disrupt banking and tank the dollar? How will demographic graying affect aggregate demand? Will the US remain a democracy? Will self-driving cars that we don’t own become the new transportation economy? Will parts of New Orleans, Miami, Charleston, Tampa, NYC, and Houston be abandoned and who pays for that? Will China continue growing, stagnate at some point, or collapse? What will be the average inflation rate for the next 3 decades?

If one knows any of this with certainty, portfolio selection becomes easy.

This is exactly how I feel about portfolio theory. It's a guide. You can't predict the portfolios that will be successful in the future because you don't know what will happen in the future.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 13, 2019, 02:12:57 AM
Each box in the top section ends in a number giving the highest withdrawal rate that worked in all investment years for that portfolio during the period where data could be found (I think 1970 to present) - in other words, if you started by investing in the worst possible year, the box shows the investment rate you could have used. I think that this required 30 year periods though, so years that started less than 30 years ago may not be included. Perhaps investors who started in 2000, for example, who suffered from the tech bust of 2001-2003 and then the 2008-09 Great Financial Crisis, will have lower withdrawal rates than shown.

Normally you're correct about 30-year SWRs for periods that started less than 30 years ago, but I actually figured out a way to account for that without trying to predict the future. You can read about it here: https://portfoliocharts.com/2017/03/21/how-to-predict-withdrawal-rates-without-a-crystal-ball/

I had wondered about that and now I find out you wrote about it over two years ago! 
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: vand on October 13, 2019, 07:47:22 AM
It's interesting but it looks like some of the portfolios with a higher SWR have lower returns and vice versa. This is why I'm skeptical of the portfolios that are heavy in gold and cash type assets (bonds, treasuries, etc). Sure we might end up with poor returns in equities markets that result in those portfolios doing well that have gold and cash type assets but there is no guarantee that this will occur.

I like simple and I suppose naive portfolios foot the bill there.

It will be interesting how this changes over the next 50 odd years.

True. So much of this is history-specific. Since 1970, the US went off the gold standard, had a tech boom in equities, inflated real estate prices with government subsidized financing, and from the early 1980s to the present has had a bond rally of massive proportions. Meanwhile investors endured stagflation in the 70s, a tech bubble that burst in 2000, a real estate bubble that burst in 2008, and the worst recession since the 1930s. Will any of that happen in the next 50 years?

Selecting the optimal AA now would require knowledge of the future, not the past. Will there be a biotech boom driven by gene editing? Will bonds revert to their norms or go negative like in Europe? Will Fannie and Freddie Mac be broken up and sold off? Will the tech-driven attention economy hit a natural growth limit or see a backlash as their products become recognized as addictive? Will average savings rates continue to decline or are millennials naturally more frugal? Will there be a major war? Will blockchain disrupt banking and tank the dollar? How will demographic graying affect aggregate demand? Will the US remain a democracy? Will self-driving cars that we don’t own become the new transportation economy? Will parts of New Orleans, Miami, Charleston, Tampa, NYC, and Houston be abandoned and who pays for that? Will China continue growing, stagnate at some point, or collapse? What will be the average inflation rate for the next 3 decades?

If one knows any of this with certainty, portfolio selection becomes easy.

This is exactly how I feel about portfolio theory. It's a guide. You can't predict the portfolios that will be successful in the future because you don't know what will happen in the future.

The purpose of MPT is not to try to predict which assets will be the ones to hold going forward. The point is to help determine how to trade return for risk in the most efficient manner possible.

People who say that they don't see how a multi-asset portfolio can hold up going forward "given the current price of X/Y/Z" should study the components of the Permanent Portfolio:

- A quarter is made up of gold, which got crushed for 20 years between 1980-2000
- another quarter from cash which has been dismal for almost just as long
- another quarter from stocks which had a lost decade from 2010
- another quarter in long term bonds, which people have been trying to call the bottom on for most of the last 20 years

and yet the portfolio's 40 year record still holds up and the strategy continues to perform well in the current FY to date. You don't need everything in a portfolio going up in order for the strategy to be effective.. in fact, the strategy relies on different parts of the portfolio not doing well for periods of time. If every part of your portfolio is doing well,...it's probably not diversified enough.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 13, 2019, 07:43:17 PM
Before I lay down a six-figure sum buying an "investment" from someone else (who has reasons they would rather have the cash), I would like to at least have a rationale for why it will earn money or appreciate over time. For example:

"This diversified stock portfolio is expected to return 6% over a decade because earnings will grow with inflation and the economy, because of stock buybacks, because the companies can make leveraged investments and earn a higher ROI than their interest, and because this has been the pattern for decades."

"This rental property is expected to yield an increasing return of 6% to 12% over a decade because [pulls out spreadsheet] expenses will be this much, rents will be this much, and appreciation will be this much assuming a vacancy rate of 15%, rent inflation of 3%, and property and property tax appreciation of 2%."

"This bond portfolio is expected to have a nominal YTM of 4% assuming a coupon of 5% and defaults/write-offs of 1% per year."


What exactly am I supposed to say about gold, or potentially worse, cash?

The best I can think of is:

"I will start by selling monthly at-the-money puts for GLD shares worth half my allocation until I am assigned. Upon assignment of half my allocation, I will sell at-the-money puts for the other half of my allocation, while writing at-the-money covered calls for the shares that have been assigned to me. If my 2nd set of puts are assigned, I will write far-ITM calls for half my allocation until assigned, and then resume the strategy of writing cash-secured puts for half my allocation and covered calls for the other half. If my covered call is assigned, I will write far-ITM puts for half my allocation until assigned, and then resume the strategy. I will make money while doing this because premium received each month will partially offset losses from large moves and because historically the options market's implied volatility has been slightly higher than actual volatility."

...which is pure gambling, but has a stronger rationale than the gambling strategy of buying and holding commodities in the hopes of them appreciating faster than treasuries.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: MustacheAndaHalf on October 13, 2019, 08:00:12 PM
"Note 4: SWR is determined based on the single worst year in the record for 30-year rolling returns. Average returns would make a very different ranking."

I think average returns are more common, so I'm glad you posted data for that in the replies.

One thing I find confusing: the naive 1st and naive 5th both tilt towards value.  They do quite well.  Larry Swedroe advocates both a value tilt and a small-cap tilt, but didn't do that well.

Maybe it's the small cap tilt, since none of the naive portfolios focus on small cap.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 13, 2019, 11:03:01 PM
Before I lay down a six-figure sum buying an "investment" from someone else (who has reasons they would rather have the cash), I would like to at least have a rationale for why it will earn money or appreciate over time. For example:

"This diversified stock portfolio is expected to return 6% over a decade because earnings will grow with inflation and the economy, because of stock buybacks, because the companies can make leveraged investments and earn a higher ROI than their interest, and because this has been the pattern for decades."

"This rental property is expected to yield an increasing return of 6% to 12% over a decade because [pulls out spreadsheet] expenses will be this much, rents will be this much, and appreciation will be this much assuming a vacancy rate of 15%, rent inflation of 3%, and property and property tax appreciation of 2%."

"This bond portfolio is expected to have a nominal YTM of 4% assuming a coupon of 5% and defaults/write-offs of 1% per year."


What exactly am I supposed to say about gold, or potentially worse, cash?

The best I can think of is:

"I will start by selling monthly at-the-money puts for GLD shares worth half my allocation until I am assigned. Upon assignment of half my allocation, I will sell at-the-money puts for the other half of my allocation, while writing at-the-money covered calls for the shares that have been assigned to me. If my 2nd set of puts are assigned, I will write far-ITM calls for half my allocation until assigned, and then resume the strategy of writing cash-secured puts for half my allocation and covered calls for the other half. If my covered call is assigned, I will write far-ITM puts for half my allocation until assigned, and then resume the strategy. I will make money while doing this because premium received each month will partially offset losses from large moves and because historically the options market's implied volatility has been slightly higher than actual volatility."

...which is pure gambling, but has a stronger rationale than the gambling strategy of buying and holding commodities in the hopes of them appreciating faster than treasuries.

I don't understand either the rationale of why the naive portfolios would often have cash and gold, yet do so well. But I would like my investing rationale to adequately explain the past before I would expect it to predict the future. Even in the average returns table, it looks like the naive portfolios as a group beat the expert ones.

In the period covered by the data, was the value of the uncorrelated variances so large that it produced more return than was lost by the underlying disadvantage of the low-return asset classes?

If so, is it reasonable to think that the value of the uncorrelated variances would continue? That's an investing rationale in itself, right?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 13, 2019, 11:22:45 PM
Man,what a great thread. Thanks for doing all that work and sharing with us, @Radagast !

I was a TSM guy for a long long time, but reading the Permanent Portfolio blog posts and playing with the data eventually changed my mind (it still took years of doing that). Now I happily slice and dice a bit, confident my portfolio is all the better for it.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: AdrianC on October 14, 2019, 06:04:09 AM
People who say that they don't see how a multi-asset portfolio can hold up going forward "given the current price of X/Y/Z" should study the components of the Permanent Portfolio:

- A quarter is made up of gold, which got crushed for 20 years between 1980-2000
- another quarter from cash which has been dismal for almost just as long
- another quarter from stocks which had a lost decade from 2010
- another quarter in long term bonds, which people have been trying to call the bottom on for most of the last 20 years

and yet the portfolio's 40 year record still holds up...

1980-2019, $10K initial, annual rebalance.
Final Balance:
Permanent Portfolio $172,324
60/40 (Intermediate Treasuries) $456,621

https://www.portfoliovisualizer.com/backtest-asset-class-allocation

Does anyone remember that old Little Caesar's commercial "One banana or two bananas... he seems to prefer two"?
https://www.youtube.com/watch?v=J47o4YZ_yrM&list=PLXUmNlnWp46X18xtT4wjIXjcfTfYuLTb-&index=35

(Note that I'm the orangutan here).
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: bacchi on October 14, 2019, 01:21:19 PM
0th Naive in the USA-Perpetual column isn't ordered correctly.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: bacchi on October 14, 2019, 02:19:32 PM
There's a lot to unpack here.

If you weren't in the US, buying "international" in the 1970-90s meant buying US. That could explain the good showing of the Ivy portfolio for non-US based portfolios, which has 20% in international stocks and 20% in international bonds (i.e., "international" is majority US for many of the time periods).

F&F might explain why some of the value portfolios do well in other countries. The 3rd does ok to great in most countries, where "international" means US for most of the time periods.

Great chart, Radagast!
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 07:33:40 PM
It's interesting but it looks like some of the portfolios with a higher SWR have lower returns and vice versa. This is why I'm skeptical of the portfolios that are heavy in gold and cash type assets (bonds, treasuries, etc). Sure we might end up with poor returns in equities markets that result in those portfolios doing well that have gold and cash type assets but there is no guarantee that this will occur.

I like simple and I suppose naive portfolios foot the bill there.

It will be interesting how this changes over the next 50 odd years.

True. So much of this is history-specific. Since 1970, the US went off the gold standard, had a tech boom in equities, inflated real estate prices with government subsidized financing, and from the early 1980s to the present has had a bond rally of massive proportions. Meanwhile investors endured stagflation in the 70s, a tech bubble that burst in 2000, a real estate bubble that burst in 2008, and the worst recession since the 1930s. Will any of that happen in the next 50 years?

Selecting the optimal AA now would require knowledge of the future, not the past. Will there be a biotech boom driven by gene editing? Will bonds revert to their norms or go negative like in Europe? Will Fannie and Freddie Mac be broken up and sold off? Will the tech-driven attention economy hit a natural growth limit or see a backlash as their products become recognized as addictive? Will average savings rates continue to decline or are millennials naturally more frugal? Will there be a major war? Will blockchain disrupt banking and tank the dollar? How will demographic graying affect aggregate demand? Will the US remain a democracy? Will self-driving cars that we don’t own become the new transportation economy? Will parts of New Orleans, Miami, Charleston, Tampa, NYC, and Houston be abandoned and who pays for that? Will China continue growing, stagnate at some point, or collapse? What will be the average inflation rate for the next 3 decades?

If one knows any of this with certainty, portfolio selection becomes easy.

This is exactly how I feel about portfolio theory. It's a guide. You can't predict the portfolios that will be successful in the future because you don't know what will happen in the future.

The purpose of MPT is not to try to predict which assets will be the ones to hold going forward. The point is to help determine how to trade return for risk in the most efficient manner possible.

People who say that they don't see how a multi-asset portfolio can hold up going forward "given the current price of X/Y/Z" should study the components of the Permanent Portfolio:

- A quarter is made up of gold, which got crushed for 20 years between 1980-2000
- another quarter from cash which has been dismal for almost just as long
- another quarter from stocks which had a lost decade from 2010
- another quarter in long term bonds, which people have been trying to call the bottom on for most of the last 20 years

and yet the portfolio's 40 year record still holds up and the strategy continues to perform well in the current FY to date. You don't need everything in a portfolio going up in order for the strategy to be effective.. in fact, the strategy relies on different parts of the portfolio not doing well for periods of time. If every part of your portfolio is doing well,...it's probably not diversified enough.
First, it figures you would say that, as in the past the Permanent Portfolio did better in the UK than anywhere else. You are probably a little biased though ;)
Second, I can't recommend the portfolios that ended up in red with white text for either of the two charts I posted. Well, I could recommend the No-Brainer, with tweaks.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 07:44:24 PM
"Note 4: SWR is determined based on the single worst year in the record for 30-year rolling returns. Average returns would make a very different ranking."

I think average returns are more common, so I'm glad you posted data for that in the replies.

One thing I find confusing: the naive 1st and naive 5th both tilt towards value.  They do quite well.  Larry Swedroe advocates both a value tilt and a small-cap tilt, but didn't do that well.

Maybe it's the small cap tilt, since none of the naive portfolios focus on small cap.
Of course they are more common :D. One thing is that the "average returns" favor portfolios with greater volatility because they use the arithmetic mean. Take a portfolio which returned 20% per year then lost 80% in the tenth year. It is 1.2*1.2*1.2*1.2*1.2*1.2*1.2*1.2*1.2*0.2 versus (0.2+0.2+0.2+0.2+0.2+0.2+0.2+0.2+0.2-0.8)/10 which give very different results. However if you are making monthly contributions and are less than half way to your FIRE goal, then the arithmetic average is exactly what you should be using (yay me!) A person looking at average safe withdrawal rates would be looking somewhere between the two charts posted here.

I think all the portfolios focus on small cap. "Large Blend" and the other large slices are just a little bit larger than total market, but all the portfolios have more mid and small cap slices than large cap. Thus, every portfolio has a small cap tilt, it is a large cap tilt which is missing (because Portfolio Charts does not feature a megacap selection). It is growth and value which do not have a bias one way or another with my/Portfolio Charts/conventional acedemic method, although you can find either in practice.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 07:54:19 PM
Before I lay down a six-figure sum buying an "investment" from someone else (who has reasons they would rather have the cash), I would like to at least have a rationale for why it will earn money or appreciate over time. For example:

"This diversified stock portfolio is expected to return 6% over a decade because earnings will grow with inflation and the economy, because of stock buybacks, because the companies can make leveraged investments and earn a higher ROI than their interest, and because this has been the pattern for decades."

"This rental property is expected to yield an increasing return of 6% to 12% over a decade because [pulls out spreadsheet] expenses will be this much, rents will be this much, and appreciation will be this much assuming a vacancy rate of 15%, rent inflation of 3%, and property and property tax appreciation of 2%."

"This bond portfolio is expected to have a nominal YTM of 4% assuming a coupon of 5% and defaults/write-offs of 1% per year."


What exactly am I supposed to say about gold, or potentially worse, cash?

The best I can think of is:

"I will start by selling monthly at-the-money puts for GLD shares worth half my allocation until I am assigned. Upon assignment of half my allocation, I will sell at-the-money puts for the other half of my allocation, while writing at-the-money covered calls for the shares that have been assigned to me. If my 2nd set of puts are assigned, I will write far-ITM calls for half my allocation until assigned, and then resume the strategy of writing cash-secured puts for half my allocation and covered calls for the other half. If my covered call is assigned, I will write far-ITM puts for half my allocation until assigned, and then resume the strategy. I will make money while doing this because premium received each month will partially offset losses from large moves and because historically the options market's implied volatility has been slightly higher than actual volatility."

...which is pure gambling, but has a stronger rationale than the gambling strategy of buying and holding commodities in the hopes of them appreciating faster than treasuries.

I don't understand either the rationale of why the naive portfolios would often have cash and gold, yet do so well. But I would like my investing rationale to adequately explain the past before I would expect it to predict the future. Even in the average returns table, it looks like the naive portfolios as a group beat the expert ones.

In the period covered by the data, was the value of the uncorrelated variances so large that it produced more return than was lost by the underlying disadvantage of the low-return asset classes?

If so, is it reasonable to think that the value of the uncorrelated variances would continue? That's an investing rationale in itself, right?
Yup, that is the entire point, to demonstrate that knowledge may not is probably not beneficial to the investing game. The real big points seem to be
1) Buy and hold
2) Be broadly diversified
3) Keep costs as low as reasonably possibly given 2)
And those all seem equally important. My take is, if you have some other theory it will almost certainly be wrong at some point, so try to avoid having as many as you can. Even the "total market is the most efficient" theory does not look to be correct, at least not using data during the period when experts made that hypothesis, which seems to give it bad odds for the future as well. Or at least it is not correct with respect to outcomes, I understand it is the minimum effort portfolio no matter what.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 07:56:26 PM
Man,what a great thread. Thanks for doing all that work and sharing with us, @Radagast !

I was a TSM guy for a long long time, but reading the Permanent Portfolio blog posts and playing with the data eventually changed my mind (it still took years of doing that). Now I happily slice and dice a bit, confident my portfolio is all the better for it.
Thanks ARS! I know that a few long time posters, including yourself, intuited these results a long time ago and are already using portfolios that resemble the naive portfolios.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 08:02:33 PM
People who say that they don't see how a multi-asset portfolio can hold up going forward "given the current price of X/Y/Z" should study the components of the Permanent Portfolio:

- A quarter is made up of gold, which got crushed for 20 years between 1980-2000
- another quarter from cash which has been dismal for almost just as long
- another quarter from stocks which had a lost decade from 2010
- another quarter in long term bonds, which people have been trying to call the bottom on for most of the last 20 years

and yet the portfolio's 40 year record still holds up...

1980-2019, $10K initial, annual rebalance.
Final Balance:
Permanent Portfolio $172,324
60/40 (Intermediate Treasuries) $456,621

https://www.portfoliovisualizer.com/backtest-asset-class-allocation

Does anyone remember that old Little Caesar's commercial "One banana or two bananas... he seems to prefer two"?
https://www.youtube.com/watch?v=J47o4YZ_yrM&list=PLXUmNlnWp46X18xtT4wjIXjcfTfYuLTb-&index=35

(Note that I'm the orangutan here).
Although I agree with disagreeing with Vand on this, I try to use years that end in "5" as starting points for backtesting, as for some reason those seem to give the least biased results. 1970 nearly worst case for stocks and bonds and best case for gold, 1980 inverse of 1970, 1990 peak of Japan bubble, 2000 peak of tech bubble, 2010 bottom of financial crisis... 5's are the way to go IMO.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 08:06:13 PM
Just posting to say "thanks"! This is quite some food for thought.
I find it also interesting how relatively poorly the "simple" Total Stock or Classic 60:40 have been performing.

My portfolio is regionally diversified, but so far I have not included my home property or REITs in it.
I should say that is optional. Or at least, if you already have home equity or real estate then there may be no compelling reason for additional real assets. However, if you do not have a home or other direct real estate, then adding some type of real assets seems unlikely to hurt and may very likely improve your odds, based on 1970-2018.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 08:07:08 PM
0th Naive in the USA-Perpetual column isn't ordered correctly.
Thanks, fixed. I caught so many before posting I expected to see another.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 08:12:52 PM
There's a lot to unpack here.

If you weren't in the US, buying "international" in the 1970-90s meant buying US. That could explain the good showing of the Ivy portfolio for non-US based portfolios, which has 20% in international stocks and 20% in international bonds (i.e., "international" is majority US for many of the time periods).

F&F might explain why some of the value portfolios do well in other countries. The 3rd does ok to great in most countries, where "international" means US for most of the time periods.

Great chart, Radagast!
Yup, it is really hard to untangle. I guess the point is more "everybody is clueless" and not so much "do this." I do think Fama-French would explain a lot of this, although I am not sure if it is for the reasons they give. In other words, "naive diversification is the best diversification" is something I am leaning towards. Although I do like the general weightings of 36:36:16:12 that result from this experiment, plus or minus about 10 in any direction.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 14, 2019, 08:56:48 PM
if you already have home equity or real estate then there may be no compelling reason for additional real assets.

I donno. I hold plenty of real estate (so don't use REITS, but also wouldn't for other reasons), but still like holding gold in my portfolio.

Much easier to rebalance into/out of IAU than a rental property.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 14, 2019, 10:42:37 PM
if you already have home equity or real estate then there may be no compelling reason for additional real assets.

I donno. I hold plenty of real estate (so don't use REITS, but also wouldn't for other reasons), but still like holding gold in my portfolio.

Much easier to rebalance into/out of IAU than a rental property.
Yeah but for you it is less necessary. You would have plenty of options for selling a bit of the real estate empire, raising rent, or others in the event of one of the bigger financial events. You might argue IAU is easier but even without it you would have options that somebody with neither real estate nor IAU would have. So if someone has a house and doesn’t want gold I’d say great! Personally we only buy houses we can live in and rent out other parts easily. That gives lots of options, live in the big part rent out the small, live in the small rent out the big part, live in a room rent out the house, cash out refi, lump sum mortgage payoff, there are lots of options there to respond to changing circumstances. What matters is knowing we don’t know the future and giving ourselves options.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 14, 2019, 10:58:35 PM

What exactly am I supposed to say about gold, or potentially worse, cash?


Gold has historically been volatile and relatively uncorrelated to negatively correlated with stocks, thereby allowing a portfolio incorporating gold to have a higher SWR despite gold having a lower total return than stocks (same can be said for any number of non-stock assets).  Isn't it on you to make a story about why that behavior would change in the future?  Will investors suddenly treat gold differently?  If so, it's a valid reason to reconsider.

Cash can be used to buy goods and services when I need them, and I expect that to be true in the future.  Good enough for me (it's my least favorite portfolio asset)
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 15, 2019, 11:26:48 AM

What exactly am I supposed to say about gold, or potentially worse, cash?


Gold has historically been volatile and relatively uncorrelated to negatively correlated with stocks, thereby allowing a portfolio incorporating gold to have a higher SWR despite gold having a lower total return than stocks (same can be said for any number of non-stock assets).  Isn't it on you to make a story about why that behavior would change in the future?  Will investors suddenly treat gold differently?  If so, it's a valid reason to reconsider.

Cash can be used to buy goods and services when I need them, and I expect that to be true in the future.  Good enough for me (it's my least favorite portfolio asset)

I suppose the rationale for holding a weakly-correlated, low or zero return asset is something like this:

I plan for my cash/gold allocation to dramatically underperform stocks, bonds, and RE in the long term (expected real ROI = 0%), but function as an emergency fund during bear markets, allowing me to avoid selling productive assets low. By not selling productive assets or by rebalancing to purchase more productive assets,I plan for these nonproductive assets to allow me to reduce portfolio beta. Assuming a significant but typical bear market (say a 40% drop over 5 years) occurs during my retirement, I will pivot out of my cash/gold allocation through rebalancing toward a more aggressive AA and taking withdrawals from the cash/gold position. When the market again reaches its previous high, I will return to my previous allocation of cash/gold. The underlying assumption is that such bear markets could be common enough in the future, or returns on productive assets low enough in the future, that the net benefit of reallocating during these occasional events will exceed the foregone gains one could have received by investing in productive assets.   

I set up a spreadsheet model to understand the dynamics of that last assumption. For simplicity, I started with $1M and a flat $40k annual withdrawal for 40 years. The return on nonproductive assets is 0% and the return on productive assets is 6%, except in a “crash year” when the return on productive assets is -30%. Two portfolios were set up: a 100% productive asset portfolio and a 70/30 productive to nonproductive portfolio. Allocations were assumed constant (I.e. continually rebalanced) I tried different sequences of flipping various years to being a “crash year”.

The 70/30 outperforms in some sequences and underperforms in others. If the crashes occur in years one and ten, for example, the 100% portfolio is exhausted in year 26 while the 70/30 lasts partway into year 27. If crashes occur in years 5 and 15, however, the 100% portfolio lasts 36 years and the 70/30 dies 5 years earlier.

TLDR: Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

Interesting note: The expected real returns of nonproductive assets have always been zero, but the collapse in bond yields and stock E/P ratios means the spread between asset types is smaller than ever. Thus, the expected “cost” of sitting in cash, for example, is lower than ever before. One’s mattress cash is only missing out on that fat 1.8% ten year treasury yield (just a few years back that would have been 4-5%). In other words, having a conservative/hedged portfolio has never been cheaper, so maybe in these anomalous times the gamble on low-correlation assets makes more sense than ever.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Buffalo Chip on October 15, 2019, 01:40:34 PM
Interesting. Max diversification looks like the most consistent winner.

“diversifying well is the most important thing you need to do in order to invest well.”

Ray Dalio
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Buffalo Chip on October 15, 2019, 02:10:01 PM
I don't have it open right now, but doesn't portfoliocharts show the highest SWRs with something like a 2:1 ratio of US small cap value to international small cap/small cap value with a smidge (5-10%) of gold?

That was about the same conclusion I came to when I ran the numbers a few months ago. With a big, huge, monstrous caveat: these are historical numbers. Future returns are not necessarily predicted by historical results. I’ve seen some folks postulate that the small cap advantage has been arbitraged away. Personally I don’t think it has, but it’s definitely possible.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 15, 2019, 05:43:17 PM
TLDR: Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

I'd take the opposite conclusion: I don't know when the market will crash, only that it will at some point (this everyone agrees on). So I hold a diversified portfolio consistently that I can stick with. I expect it to perform better in the long run in terms of both SWR and CAGR.

I am not holding this due to expecting an impending crash, or changing my AA based on what the current market is doing.

In fact, I'd flip your statement I quoted above:
Spoiler: show
TLDR: Not holding a diversified asset allocation is a bet on a lack of correction occurring sooner rather than later as well as a very specific bet on the particular asset class you're holding performing well in the short and medium term. In other words, it’s market timing.


Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Buffalo Chip on October 15, 2019, 06:57:31 PM

TLDR: Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

Interesting note: The expected real returns of nonproductive assets have always been zero, but the collapse in bond yields and stock E/P ratios means the spread between asset types is smaller than ever. Thus, the expected “cost” of sitting in cash, for example, is lower than ever before. One’s mattress cash is only missing out on that fat 1.8% ten year treasury yield (just a few years back that would have been 4-5%). In other words, having a conservative/hedged portfolio has never been cheaper, so maybe in these anomalous times the gamble on low-correlation assets makes more sense than ever.

That’s an interesting way of looking at it. So given that bonds are not really negatively correlated with equities (and offer crap returns in the long run anyways), what else is there in the way or negatively correlated assets other than cash and some durable commodities like gold? 
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 15, 2019, 09:22:35 PM
TLDR: Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

I'd take the opposite conclusion: I don't know when the market will crash, only that it will at some point (this everyone agrees on). So I hold a diversified portfolio consistently that I can stick with. I expect it to perform better in the long run in terms of both SWR and CAGR.

I am not holding this due to expecting an impending crash, or changing my AA based on what the current market is doing.

In fact, I'd flip your statement I quoted above:
Spoiler: show
TLDR: Not holding a diversified asset allocation is a bet on a lack of correction occurring sooner rather than later as well as a very specific bet on the particular asset class you're holding performing well in the short and medium term. In other words, it’s market timing.


Diversification among productive assets is a very good thing, but adding in double-digit allocations of nonproductive assets is, IMO, pretty darn close to market timing.

I'm defining productive assets as things that have earnings or yield something, such as equities, bonds, real estate, mortgage payoff, etc. I'm defining nonproductive assets as things that earn/yield nothing or have negative expected returns, such as cash, precious metals, collectibles, lotto tickets, and cryptocurrencies.

In my simple spreadsheet described above, I assumed a 6% expected ROI on the productive assets in a portfolio and 0% on the nonproductive. The 100% productive portfolio earns 6% per year except when markets crash, and the 70/30 productive/nonproductive portfolio earns 0.7(.06)+0.3(0)=4.2% per year, except when markets crash. For the 70/30 portfolio to outlast the 100% productive portfolio, a crash must occur soon enough that the outperformance of the 100% portfolio in normal years does not exceed the difference in performance between the two portfolios in the crash year.

Example: If one year productive assets drop 30% and nonproductive assets drop 0%, the all-productive portfolio loses 30% and the 70/30 portfolio only loses (.7*.3=) 21%. However, for each year up until this market crash, the all-productive portfolio has been earning (6-4.2=) 1.8% more than the 70/30 portfolio. If the all-productive portfolio has been out-earning the 70/30 portfolio by 1.8% for 5 years (1.8*5=9), the market crash only sets it back to where the 70/30 portfolio already is. So one's selection of portfolios in this simplified example would be a bet on whether one foresees a 30% market crash occurring in less than 5 years, or more than 5 years. If less than 5 years, the 70/30 outperforms, if more than 5 years, the all-productive outperforms.

The market timer who says "I read in the financial press that a crash is coming next year so I'm moving 30% of my portfolio into cash and gold", and does this every year, is making the exact same decision as someone who calls it their long term AA.
Spoiler: show
Their rationales do not affect their returns. Only their AA matters.


I don't agree the reverse statement is also market timing, when we are talking about productive vs. nonproductive assets. Market timing implies a temporary move out of or into risky assets on the hopes that one can profitably reverse the move later. This differs from setting an AA and rebalancing, where the goal is to maintain the percentage of a portfolio in each asset class for the long term.

An investment in nonproductive assets is a bet on negative returns for productive assets, which is a relatively rare thing. One would have to get their "timing" just right to make such a strategy work, whereas productive assets are all but certain to outperform nonproductive assets in the long term so no timing is required, just consistency.


Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 15, 2019, 10:14:53 PM
You:

Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

Also you:

Market timing implies a temporary move out of or into risky assets on the hopes that one can profitably reverse the move later. This differs from setting an AA and rebalancing, where the goal is to maintain the percentage of a portfolio in each asset class for the long term.


So by your second statement, a consistent AA with “nonproductive assets” is NOT market timing.  This is what we are discussing here.  Nobody is saying anything like “cash out, buy gold for the next crash and then jump back in”

It’s nice that you can change the definition of your own words when they are used against you
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 15, 2019, 10:57:32 PM

An investment in nonproductive assets is a bet on negative returns for productive assets, which is a relatively rare thing.

No, it’s a bet (in the broadest use of the term) on underperformance by productive assets. A gold allocation can be net benefit if it outperforms stocks on the right timeframe.  Stocks do not have to go negative.

Life insurance may also be a “bet” that you will die before you save enough money to support your heirs, but that doesn’t mean it’s a bad choice due to “death timing” (it May be a bad choice For other reasons)

Btw, I’m using your term “nonproductive asset” for convenience, but that doesn’t mean I agree it’s apt.  Cash in particular is rarely nonproductive (does anyone really have a significant allocation to mattress stash?).  Short term treasuries have historically done surprisingly well vs inflation and there are even better options available (ibonds)
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 16, 2019, 12:09:32 AM

In my simple spreadsheet described above, I assumed a 6% expected ROI on the productive assets in a portfolio and 0% on the nonproductive. The 100% productive portfolio earns 6% per year except when markets crash, and the 70/30 productive/nonproductive portfolio earns 0.7(.06)+0.3(0)=4.2% per year, except when markets crash. For the 70/30 portfolio to outlast the 100% productive portfolio, a crash must occur soon enough that the outperformance of the 100% portfolio in normal years does not exceed the difference in performance between the two portfolios in the crash year.

Example: If one year productive assets drop 30% and nonproductive assets drop 0%, the all-productive portfolio loses 30% and the 70/30 portfolio only loses (.7*.3=) 21%. However, for each year up until this market crash, the all-productive portfolio has been earning (6-4.2=) 1.8% more than the 70/30 portfolio. If the all-productive portfolio has been out-earning the 70/30 portfolio by 1.8% for 5 years (1.8*5=9), the market crash only sets it back to where the 70/30 portfolio already is. So one's selection of portfolios in this simplified example would be a bet on whether one foresees a 30% market crash occurring in less than 5 years, or more than 5 years. If less than 5 years, the 70/30 outperforms, if more than 5 years, the all-productive outperforms.


This model might have a key assumption incorrect: namely, that the assets you are calling "nonproductive" are essentially fixed. If the nonproductive assets fluctuate, they could produce results different from the fixed assumption.

For example, we can stipulate that gold will have a long term return of zero, but still agree that it could fluctuate quite a bit year to year. Suppose that the unproductive asset in the example is gold, which in this case drops 4%/year for five years, then rises 25% during the stock crash to end at its original value. Then initial return on 70/30 drops to 3%/year (is reduced by .04 x 30% = 1.2% reduction), so after 5 years, about 15% gain before compounding...spitballing, maybe 16.2%. The crash produces 21% stock drop (.3 x .7) but 7.5% gold gain (.25 x .3), so 13.5% drop from a base of 116.2; I guess a drop of 15.1 or so, final result maybe 101.1% of original investment, a slight gain.

If gold never fluctuated during these 6 years, the quoted model would be accurate, but would produce a loss. The 4.2% return would 21% before compounding, maybe 23.2 after, so peak value 123.2; loss in crash 21% x 123.2 or roughly 25.9, final result about 97.3, several percent lower than the fluctuation case.

IRL, gold doesn't fluctuate exactly opposite to stock. Maybe that's why these assets are described sometimes as "uncorrelated" - their fluctuations don't exactly match (or linearly diverge from) stocks. The difference could provide gains during ordinary years sometimes, not just crash years.  The fluctuations in the "nonproductive" assets can actually be productive. I think the results of the naive portfolios are showing that such variances can produce higher gains than we'd expect.

Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 16, 2019, 07:33:44 AM
You:

Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

Also you:

Market timing implies a temporary move out of or into risky assets on the hopes that one can profitably reverse the move later. This differs from setting an AA and rebalancing, where the goal is to maintain the percentage of a portfolio in each asset class for the long term.


So by your second statement, a consistent AA with “nonproductive assets” is NOT market timing.  This is what we are discussing here.  Nobody is saying anything like “cash out, buy gold for the next crash and then jump back in”

It’s nice that you can change the definition of your own words when they are used against you

Frequent changes to an “asset allocation” are generally market timing, regardless of what the person calls it. Can we agree there?

My next assumption is anyone putting a big chunk of their wealth into an asset that has no real earnings (cash, PMs, cryptocurrency, Beanie Babies) has a plan to trade out of those assets when their exchange rates are favorable (I.e. they predict a time in the future when prices will be favorable). Because nonproductive assets will never grow in size or increase earnings/dividends/buybacks, this is a bet on market sentiment, also known as greater fool theory. Can we agree on this?

If we agree on #1 and #2, then maybe the question is how long out can a market timer make predictions and still remain a market timer?

One day trade - market timer for sure
One month trade - market timer for sure
One year trade - probably market timing?
Five year trade - market timer or long term investor?

IMO, any plan to trade assets for a profit based on market predictions / greater fool theory is market timing, regardless of the timeframe. Investments in productive assets (businesses, bonds, real estate) have a different rationale- the value will go up because earnings are this or the yield is that. It is the reasoning process that makes a market timer or a long term investor.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 16, 2019, 09:40:28 AM
Investments in productive assets (businesses, bonds, real estate) have a different rationale- the value will go up because earnings are this or the yield is that.

Once you realize that cash in a portfolio is a type of bond you may think of it differently (https://portfoliocharts.com/2017/05/12/understanding-cash-will-make-you-a-better-and-happier-investor/).  It's how interest rates work, and it's why cash really is a productive asset.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 09:44:07 AM
You:

Big nonproductive asset allocations are a bet on a big correction occurring sooner rather than later. In other words, it’s market timing.

Also you:

Market timing implies a temporary move out of or into risky assets on the hopes that one can profitably reverse the move later. This differs from setting an AA and rebalancing, where the goal is to maintain the percentage of a portfolio in each asset class for the long term.


So by your second statement, a consistent AA with “nonproductive assets” is NOT market timing.  This is what we are discussing here.  Nobody is saying anything like “cash out, buy gold for the next crash and then jump back in”

It’s nice that you can change the definition of your own words when they are used against you

Frequent changes to an “asset allocation” are generally market timing, regardless of what the person calls it. Can we agree there?

My next assumption is anyone putting a big chunk of their wealth into an asset that has no real earnings (cash, PMs, cryptocurrency, Beanie Babies) has a plan to trade out of those assets when their exchange rates are favorable (I.e. they predict a time in the future when prices will be favorable). Because nonproductive assets will never grow in size or increase earnings/dividends/buybacks, this is a bet on market sentiment, also known as greater fool theory. Can we agree on this?

If we agree on #1 and #2, then maybe the question is how long out can a market timer make predictions and still remain a market timer?

One day trade - market timer for sure
One month trade - market timer for sure
One year trade - probably market timing?
Five year trade - market timer or long term investor?

IMO, any plan to trade assets for a profit based on market predictions / greater fool theory is market timing, regardless of the timeframe. Investments in productive assets (businesses, bonds, real estate) have a different rationale- the value will go up because earnings are this or the yield is that. It is the reasoning process that makes a market timer or a long term investor.
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

It is also not about gold, the methodology randomly assigns gold allocations to some portfolios for an average of 3-4% allocation, but that is a pretty minor point. If it makes you feel better just ignore the naive portfolios which have gold, the results are essentially the same (although you would only have 6 naive portfolios to consider rather than 10).
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: AdrianC on October 16, 2019, 10:06:26 AM
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 16, 2019, 10:36:03 AM

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?

Yes, all portfolios are rebalanced annually at the end of the year.  There's nothing special about that date except for the fact that it's when the annual returns data is reported.  Rebalancing on another date might change the numbers a bit, but relatively speaking I would not expect it to make a huge difference in the order of either of Radagast's lists.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 16, 2019, 11:58:17 AM
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?

Somewhere I read an article where, supposedly, someone studied the question of "some other reasonable time period" by backtesting many instances of rebalancing on the basis of timeframes of different length. Their conclusion was that the best results from rebalancing occurred between a quarterly basis and an annual basis. In other words, if you balanced more often than quarterly, it was too often and you lost money, relatively. Same thing if you rebalanced less than once year - you left money on the table. The authors felt either quarterly or annually would produce maximum results. I suppose a period between them, such as every six months, would be similarly effective but the article didn't specify that. Sorry I don't know the actual study.

Googling, the closest I found was the article below, in which some person did a similar study, but not on a formal basis. This author concluded the same thing.
https://seekingalpha.com/article/4077399-often-rebalance-portfolio
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 16, 2019, 01:06:59 PM
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?

Somewhere I read an article where, supposedly, someone studied the question of "some other reasonable time period" by backtesting many instances of rebalancing on the basis of timeframes of different length. Their conclusion was that the best results from rebalancing occurred between a quarterly basis and an annual basis. In other words, if you balanced more often than quarterly, it was too often and you lost money, relatively. Same thing if you rebalanced less than once year - you left money on the table. The authors felt either quarterly or annually would produce maximum results. I suppose a period between them, such as every six months, would be similarly effective but the article didn't specify that. Sorry I don't know the actual study.

Googling, the closest I found was the article below, in which some person did a similar study, but not on a formal basis. This author concluded the same thing.
https://seekingalpha.com/article/4077399-often-rebalance-portfolio

There are two goals of rebalancing.... one is to obtain a “bonus” return and the other is to keep your risk in line with your original AA.  As far as the bonus goes, any rebalancing is a “bet” that volatility will be high on the timescale of your rebalancing period.  You can of course data mine a historically “best” rebalancing period for and specific collection of assets, but then you are betting that the past will look like the future.  So much betting in this thread!  For some collections of assets, the highest return comes from NEVER rebalancing.  It’s been a free lunch historically but could hurt you in the future

I’m sure you guys all know this but seemed relevant to mention.

Also don’t forget “rules based” rebalancing where you only rebalance when some assets get too high or low relative to your original AA.  You can also data mine the best rebalancing bands for your assets and a specific time period, but again maybe it will be different in the future
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: bacchi on October 16, 2019, 01:08:01 PM
Somewhere I read an article where, supposedly, someone studied the question of "some other reasonable time period" by backtesting many instances of rebalancing on the basis of timeframes of different length. Their conclusion was that the best results from rebalancing occurred between a quarterly basis and an annual basis. In other words, if you balanced more often than quarterly, it was too often and you lost money, relatively. Same thing if you rebalanced less than once year - you left money on the table. The authors felt either quarterly or annually would produce maximum results. I suppose a period between them, such as every six months, would be similarly effective but the article didn't specify that. Sorry I don't know the actual study.

Googling, the closest I found was the article below, in which some person did a similar study, but not on a formal basis. This author concluded the same thing.
https://seekingalpha.com/article/4077399-often-rebalance-portfolio

Vanguard? https://www.vanguard.com/pdf/ISGPORE.pdf

I remember reading a study where it claimed that rebalancing less frequently (~5-7 years) was better than annually in order to allow maximum growth. I'll have to search for that paper.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: talltexan on October 16, 2019, 02:14:56 PM
If you're young, your contributions are so large relative to existing investments that rebalancing wouldn't really change much.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 16, 2019, 02:22:49 PM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Buffalo Chip on October 16, 2019, 04:58:38 PM
Investments in productive assets (businesses, bonds, real estate) have a different rationale- the value will go up because earnings are this or the yield is that.

Once you realize that cash in a portfolio is a type of bond you may think of it differently (https://portfoliocharts.com/2017/05/12/understanding-cash-will-make-you-a-better-and-happier-investor/).  It's how interest rates work, and it's why cash really is a productive asset.

I agree. Unless cash is buried in a jar in the backyard, it is like a short term bond with a relatively low return.

A completely unrelated question for you. For the TSP, how would one back test the “S” fund on portfoliocharts? Seems like it would be a combination of US SCB and MCB. But what proportion? Something else completely?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 16, 2019, 04:59:06 PM
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?

I don't know if it is a concern but it masks the result of different withdrawal strategies. If you read McClungs living off your money his idea is to withdraw from your bond portfolio unless the market is up 20% since your FIRE date. So it's not a straight rebalancing. Bonds/Cash are used to cater for SORR or just a down market.

I like this approach a lot more than having gold in my portfolio for instance. I figure stocks have to be the driver of your long term portfolio longevity and I just need to be diversified enough to be able to handle market down turns.

If I become really rich I might consider gold and commodities in my portfolio but I think I am still more comfortable with a 100% stock portfolio and a low WR than bothering with having gold and/or commodities.

I also own my home which to me means I have more than enough in real estate.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 16, 2019, 07:06:30 PM
A completely unrelated question for you. For the TSP, how would one back test the “S” fund on portfoliocharts? Seems like it would be a combination of US SCB and MCB. But what proportion? Something else completely?

While the percentages may change a little over time, as of December 2018 the S fund covered the smallest 13% of the market.  For reference, SCB on Portfolio Charts covers the bottom 15%.  That's a very good match well within the normal margin of error for how funds are managed, so I would use SCB.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 16, 2019, 07:31:08 PM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)

The point of the spreadsheet being, had the sequence of returns been very slightly different or had the starting point been different, the list of best performing funds would have been different too. For example, a study starting in 1982 would have ranked anything with gold in it near the bottom.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 07:52:17 PM
This thread is not about market timing, and none of the portfolios makes any market timing assumptions. Withdrawal rates are for 30 year holding periods or extrapolated 30 year holding periods, while average returns were averages of every year of 1970-2018.

Quibble: all portfolios are rebalanced annually at years end, are they not?

Why is that the correct time to rebalance?
What happens if  they're rebalanced on some other date, or some other reasonable time period?
Should it be a concern?
I'd say no. I have seen a crazy number of rebalancing interval ideas including continuously, daily, weekly, monthly, quarterly, twice annually, annually, the day before the two year US election, only through new money and withdrawals for spending, and letting your winners run and never rebalancing. Plus dozens of "bands" strategies. They pretty much all had very intelligent people behind them and great rationales and usually backtesting, and given variability in backtesting there is often not a consistent winner especially depending on your reason. Ultimately my opinion is that any rebalancing back to a consistent predetermined allocation is never market timing, and neither is never rebalancing.

It seems like there is a lot of evidence that 1-3 years is a good frequency. However, since more naive is apparently more better, I endorse "annually on your birthday" as the official strategy of naive or would-be-naive investors everywhere. Which I have no reason to believe is better or worse than annually at year end.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 07:55:01 PM
If you're young, your contributions are so large relative to existing investments that rebalancing wouldn't really change much.
#mustachianproblems
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 08:04:30 PM
Btw, I’m using your term “nonproductive asset” for convenience, but that doesn’t mean I agree it’s apt.  Cash in particular is rarely nonproductive (does anyone really have a significant allocation to mattress stash?).  Short term treasuries have historically done surprisingly well vs inflation and there are even better options available (ibonds)
Savings bonds started as a way to pay the government to bomb the shit out of evil foreign countries. Waaaayyyyy more productive (and socially and environmentally conscious!) than layabout gold.

But more seriously, in someway or another ibonds are part of my best-practice bond allocation ever since I knew about them. Equal split ZROZ, LTPZ, ibonds FTW!
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 08:13:14 PM
I figure stocks have to be the driver of your long term portfolio longevity and I just need to be diversified enough to be able to handle market down turns.
I agree completely with that, and add "or income real estate, or your own business" to the list. However I can't agree with 100% stocks (well, I can if you are still at the start of your earning/saving journey). In the short term markets can do very crazy things like lose 27% in a single day, or 90% in just a few years. And that is in normal times. So you need something to get you through those trough times, even with VPW.

20% S&P500
20% US Small Value
20% International
20% Bonds (see above)
20% Your House
Rebalance on your birthday

Seems like a naively winning strategy to me!
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 16, 2019, 08:14:17 PM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)

The point of the spreadsheet being, had the sequence of returns been very slightly different or had the starting point been different, the list of best performing funds would have been different too. For example, a study starting in 1982 would have ranked anything with gold in it near the bottom.
Yup, that's why you should look carefully at Japan, the only truly independent data point.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 16, 2019, 08:57:52 PM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)

The point of the spreadsheet being, had the sequence of returns been very slightly different or had the starting point been different, the list of best performing funds would have been different too. For example, a study starting in 1982 would have ranked anything with gold in it near the bottom.

But what if we could test many different starting points?
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 16, 2019, 10:19:49 PM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)

The point of the spreadsheet being, had the sequence of returns been very slightly different or had the starting point been different, the list of best performing funds would have been different too. For example, a study starting in 1982 would have ranked anything with gold in it near the bottom.

But what if we could test many different starting points?

Would be useless without some kind of visualization to go with it
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 17, 2019, 12:07:59 AM
In my simple spreadsheet described above

If only there were some site that had historical data that we could model a portfolio with a mix of "productive" and "nonproductive" assets versus one of just "productive" assets... maybe even a random mix of the former, we could call "Naive" and specific mixes of the latter we could call "Guru" and then post a thread to discuss them.

Alas, we'll probably have to go with the overly generic spreadsheet that returns 6% every year.

;)

The point of the spreadsheet being, had the sequence of returns been very slightly different or had the starting point been different, the list of best performing funds would have been different too. For example, a study starting in 1982 would have ranked anything with gold in it near the bottom.

But what if we could test many different starting points?

Would be useless without some kind of visualization to go with it

Hmm, good point. Or maybe Charts, instead.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 17, 2019, 12:10:18 AM
What I'm getting at, ChpBstrd, is that holding diversified assets (even "nonproductive" ones, for the purposes of non-correlated rebalancing for lower volatility and thus higher SWR and sometimes even higher CAGR) isn't "market timing," it's more the similar idea to most investing: that the future won't look substantially different than the past.

I think Portfolio Charts has shown via its blog posts and available data (such as was demonstrated in the OP) that these type of portfolios, held consistently and rebalanced into (e.g. not "I'm going to go into X asset now, and out of it later" as some of your posts implied via the market timing comments) will result in improved performance over many standard portfolios like TSM or 60/40.

A spreadsheet with overly simple assumptions seems very lacking compared to what you can actually see in the data, especially with regards to SWR/PWR, IMO.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 17, 2019, 12:52:21 AM
I figure stocks have to be the driver of your long term portfolio longevity and I just need to be diversified enough to be able to handle market down turns.
I agree completely with that, and add "or income real estate, or your own business" to the list. However I can't agree with 100% stocks (well, I can if you are still at the start of your earning/saving journey). In the short term markets can do very crazy things like lose 27% in a single day, or 90% in just a few years. And that is in normal times. So you need something to get you through those trough times, even with VPW.

20% S&P500
20% US Small Value
20% International
20% Bonds (see above)
20% Your House
Rebalance on your birthday

Seems like a naively winning strategy to me!

I get that stocks can crash. I expect to see a couple of 50% drops in my post FIRE life. I think I'd personally be okay with it if I was at a really low WR. That is though a personal preference. I'd love to have 100% TSM  in one fund just for simplicity.

I bet that portfolio you've listed would do extremely well over time.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: vand on October 17, 2019, 04:17:11 AM
Just to throw something else into the discussion...

https://www.marketwatch.com/story/bank-of-america-declares-the-end-of-the-60-40-standard-portfolio-2019-10-15?mod=MW_story_top_stories

I have been aware that Bonds and Stocks have not always been historically negatively correlated, but I was rather stunned at how the relationship has changed over time:

(https://ei.marketwatch.com/Multimedia/2019/10/15/Photos/NS/MW-HT335_bondco_20191015123101_NS.jpg?uuid=2cecd798-ef69-11e9-b7c9-9c8e992d421e)

A return to pre-1995 sort of relationship would change the current paradigm of mainstream investing.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: talltexan on October 17, 2019, 08:58:32 AM
The hidden variable that you don't see on that graph: inflation!

I'd be more interested in a 12-month correlation than in a 2-year correlation.

Despite my complaints, I still think it's a really interesting graph.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 17, 2019, 09:48:38 AM
What I'm getting at, ChpBstrd, is that holding diversified assets (even "nonproductive" ones, for the purposes of non-correlated rebalancing for lower volatility and thus higher SWR and sometimes even higher CAGR) isn't "market timing," it's more the similar idea to most investing: that the future won't look substantially different than the past.

I think Portfolio Charts has shown via its blog posts and available data (such as was demonstrated in the OP) that these type of portfolios, held consistently and rebalanced into (e.g. not "I'm going to go into X asset now, and out of it later" as some of your posts implied via the market timing comments) will result in improved performance over many standard portfolios like TSM or 60/40.

A spreadsheet with overly simple assumptions seems very lacking compared to what you can actually see in the data, especially with regards to SWR/PWR, IMO.

Not to discount the hard work that went into understanding the last 50 years of investing, but I think it is important to keep in mind that we are studying history. This is not a scientific endeavor to uncover the physical laws of the universe that will apply everywhere and always in the future. If the future does not exactly resemble the past, then the precise mix of assets that did best in the past will not do best in the future. Some drivers of asset performance over the last 50 years can never happen again at this point. The US can never go off the gold standard again and treasury yields (presumably) cannot fall by 10% from here over the next 40 years. Why pick the portfolio that did best under those specific conditions? Similarity, the next five decades will be full of events that never happened before.

Second, there are a near-infinite number of nonproductive assets for sale, and if we adopt the mantra that more diversification = always good, we will end up buying things like vacant land, futures contracts on foreign currencies, art, carbon credits, patents, URLs, antiques, lotto tickets, mineral rights, futures on silver, platinum, cobalt, copper, steel, pork bellies, cotton, corn, timber, wheat, and several grades of crude and refined oils, and of course the old standby ammo and canned goods. I’m sure some of these gambles will pay off big, but the losses from others will likely offset the winners. In a nutshell, the expected return of this whole nonproductive pile is between zero and negative. Buying and selling them is generally a zero sum game. Perhaps in 30 years people will look back with the benefit of hindsight and realize that a portfolio with 64% emerging markets stocks, 12% lithium futures, 2% the cryptocurrency that comes out next year, 2% URLs, and the rest in condo timeshares was the ideal mixture, conveniently forgetting the rest of the nonproductive choices a person faced back in 2019, such as the 5,000 cryptocurrencies that would go defunct.

Last, if we are using nonproductive assets to hedge against SORR events to protect our portfolios, why buy assets like gold whose weak counter-correlation with stocks varies from year to year and is subject to influences like jewelry demand in India or semiconductor production? Why not hedge with put options, which have a precisely known and a mathematically certain payout in the event of a correction or lack thereof, require a much smaller allocation, and if LEAPS are bought on a low volatility day they can suffer less annual decay than the amount of underperformance one is likely to suffer with a large allocation to gold (avg 5% worse than stocks over the past 45 years* but that’s history of course :).

* https://seekingalpha.com/article/4296091-gold-vs-stocks?_gl=1*ar01en*_ga*YW1wLTVOM2FUVm9ZeHQ3a3R4ZE5DSFhhVXc.



Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 17, 2019, 10:26:39 AM
Not to discount the hard work that went into understanding the last 50 years of investing, but I think it is important to keep in mind that we are studying history. This is not a scientific endeavor to uncover the physical laws of the universe that will apply everywhere and always in the future. If the future does not exactly resemble the past, then the precise mix of assets that did best in the past will not do best in the future.

That's true.  But while the future is never perfectly predictable, not all portfolios are equally unpredictable.  By studying the past beyond simple averages -- the best times, the worst times, and everything in between -- we can learn a lot about portfolio consistency even in wildly changing markets.  Thanks to diversification, some portfolios are more trustworthy than others and can help us plan for an uncertain future with confidence.

That's why I like Radagast's use of SWRs in his analysis.  Because they're defined by the worst-case scenarios (https://portfoliocharts.com/2015/11/17/how-safe-withdrawal-rates-work/), the portfolios with high SWRs are usually impressively consistent in all economic conditions.  I'd actually expect the spread of SWRs to get even wider on the list over time, as by definition SWRs can only go down with more data and never go up.  And the volatile portfolios at the bottom of the list are more likely to set new lows than the consistent ones at the top.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 17, 2019, 12:29:17 PM
Woah, was that list supposed to all be nonproductive assets?  If so, I think you’re going off the deep end there.

I do think if there was a cheap way to hold everything you listed, store it, and rebalance, the result might be surprisingly good.  Gold is often chosen specifically because it’s relatively easy to hold and transact.  Of course by holding stocks you probably get a good proxy ownership of all the things you listed.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 17, 2019, 01:15:38 PM
Second, there are a near-infinite number of nonproductive assets for sale, and if we adopt the mantra that more diversification = always good, we will end up buying things like ... futures on silver, platinum, cobalt, copper, steel, pork bellies, cotton, corn, timber, wheat, and several grades of crude and refined oils

Luckily all of these futures contracts are already included in the commodities data, allowing us to compare portfolios that include such naive "nonproductive" diversification to ones that don’t.  Radagast's analysis shows that several portfolios including commodities did quite well.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 17, 2019, 02:47:28 PM
I would like to acknowledge the @ChpBstrd has put his finger on a really important question. How do you know which unloved items for sale are assets are worth investing in, and which unloved items for sale are unloved because they truly are useless piles of junk?

Some of the things we see on portfoliocharts.com were probably very small scale options or relatively new in 1970. Buying them then wasn't quite like buying them now. How do we know what today's equivalents are? Is there a selection bias in being on portfoliocharts.com that makes us miss something important? (No offense re portfoliocharts.com... I use it here as a best-in-class example of conventional wisdom).

Somewhere I saw a wonderful set of graphs that compared stocks, bonds and real estate in France, Germany, and the US (I think) from something like 1917 to 2016. If the data underlying the graphs was accurate, real estate in France right after WWI was super cheap. But "real estate in France after WWI" probably included a lot of land with burned out buildings, and a huge % of men who might buy property had been killed, and prices for what you could produce from land (food, rent) were probably quite low, so low prices were rational. Also, if you lived near a large country that had invaded you twice in living memory, the fact that a house is not mobile could be viewed as a significant drawback. Which  it indeed turned out to be just 20 years later!

That was only an example. These three major asset classes apparently performed quite differently from country to country, and differently from each other within a country. It seemed to make a good argument for diversification for sure. I just wonder how to know what the right categories are to diversify in.

Part of me worries about this thing I read once about British bonds. Some were issued in the 1700s and paid until 2016 or so, when they were retired. As a class, British bonds apparently paid about 2%/year after tax from 1700 or so until 1930s IIRC. But then they lost a mind boggling 80% of value. How do we tell which beacon of eternal strength is set for a permanent fall?

How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio? Crypto seems like a bad "investment" to me, but... well, where is the line?  "The answer is stocks" seems too glib. Even "the ones Tyler wisely chose because of a bunch of data" just seems...well, probably good, but I feel like we're missing something. Even if the categories on portfoliocharts.com are the "right" ones historically, it seems likely to me that they probably include something (stocks?) that are ready for a permanent fall, and exclude something with a shorter history or other exclusion reason that will perform much better. Is diversifying amoung Things That Did Well Up To Now really our best option?

Sorry if I'm rambling. This part seems really confusing to me.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 17, 2019, 03:22:28 PM
Is there a selection bias in being on portfoliocharts.com that makes us miss something important?

That's a fair question.  Without sidetracking this discussion, the short story is that one of my goals is to fight the data availability bias very common in investing analysis by offering a wide variety of modern asset options both at home and abroad.  But it's an ongoing project, and the current list of assets is definitely not all-inclusive of everything I think might be useful.  For example, I think TIPS are nice products and would love to include them on the site if I can find an accurate method to simulate their performance prior to their introduction in 1997. It's a difficult problem to solve, and I'm still researching models behind the scenes. So I encourage people to use the data to expand their thinking rather than artificially limit it.

How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio?

Since there's no single portfolio suitable for all people, to a large extent I think the answer is personal.  If you really hate a specific asset like gold, then choose a portfolio that doesn't have it!  There are lots of good options.  My one overarching piece of advice, however, is to try to stop tasting each portfolio ingredient in isolation and instead think about its important contribution to the overall recipe.  Nobody thinks baking soda tastes good, but there's a reason it's in the cookies.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 17, 2019, 08:09:44 PM
^ Good answer, thanks. Still pondering, but good answer.

Part of me instantly goes "Ah, crypto = baking soda". To me it tastes terrible and, by itself, seems likely to fall flat. But since crypto is highly variable, with possibly a tendency to gain value when ordinary currencies and investments tremble, it's conceivable that crypto could be part of a tasty cookie.

I mean, it scares me too much to actually do it. Just thinking out loud.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 18, 2019, 02:34:17 AM
Is there a selection bias in being on portfoliocharts.com that makes us miss something important?

That's a fair question.  Without sidetracking this discussion, the short story is that one of my goals is to fight the data availability bias very common in investing analysis by offering a wide variety of modern asset options both at home and abroad.  But it's an ongoing project, and the current list of assets is definitely not all-inclusive of everything I think might be useful.  For example, I think TIPS are nice products and would love to include them on the site if I can find an accurate method to simulate their performance prior to their introduction in 1997. It's a difficult problem to solve, and I'm still researching models behind the scenes. So I encourage people to use the data to expand their thinking rather than artificially limit it.

How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio?

Since there's no single portfolio suitable for all people, to a large extent I think the answer is personal.  If you really hate a specific asset like gold, then choose a portfolio that doesn't have it!  There are lots of good options.  My one overarching piece of advice, however, is to try to stop tasting each portfolio ingredient in isolation and instead think about its important contribution to the overall recipe.  Nobody thinks baking soda tastes good, but there's a reason it's in the cookies.

These are really good points. I also think @ChpBstrd has pointed out a really interesting point on asset allocation.  A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 18, 2019, 12:52:10 PM
A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.

Say what you want about gold, but it's been a valuable means of exchange for thousands of years longer than any other asset on the list.  So I seriously doubt it's going to suddenly vanish in our lifetimes.  ;) 

I hear ya, though.  I appreciate your perspective about the uncertainty of the best performers in the future, and I agree that nobody has a crystal ball.  Personally, that's one reason I value diversification!  Eggs, baskets, and all that.

For the purposes of this discussion, just keep in mind my previous point (https://forum.mrmoneymustache.com/investor-alley/portfolio-design-idiots-v-gurus/msg2479041/#msg2479041) about how Radagast's SWR analysis accounts for consistency over all timeframes rather than simply high returns over one timeframe (also read his original "no bullshitting" (https://forum.mrmoneymustache.com/investor-alley/portfolio-design-idiots-v-gurus/msg2471613/#msg2471613) explanation in Note 1). It's an important distinction that implies the order of the list is not as random as you may believe if you're only accustomed to thinking about maximizing returns. 

Title: Re: Portfolio Design: Idiots v. Gurus
Post by: BicycleB on October 18, 2019, 05:05:48 PM
Re-reading the OP, I finally notice something that's probably important:

"Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant."

Oh. Ah. Ding. Light bulb. (Man, how did I miss this?) So...all the naive portfolios are about 70-75% stock. Where they differ from a lot of expert portfolios is having more "real assets" (I guess real estate, gold and commodities) than some of the expert portfolios, and possibly having more diversification with the broad categories listed in the quote. They're kind of a consistent allocation of their own, seemingly.

An allocation that worked well, evidently. But... so does that lead us back to questioning whether a good past allocation is a good future one?

I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.


Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Tyler on October 18, 2019, 09:53:47 PM
I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.

Now you're talking!  That basic idea is one goal of the Portfolio Finder.  Try this:

https://portfoliocharts.com/portfolio/portfolio-finder/

It will help you model hundreds of naive portfolios simultaneously, and you can explore the effects of diversification breadth vs. depth by controlling the assets under consideration.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: chevy1956 on October 19, 2019, 03:33:48 AM
A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.

Say what you want about gold, but it's been a valuable means of exchange for thousands of years longer than any other asset on the list.  So I seriously doubt it's going to suddenly vanish in our lifetimes.  ;) 

I hear ya, though.  I appreciate your perspective about the uncertainty of the best performers in the future, and I agree that nobody has a crystal ball.  Personally, that's one reason I value diversification!  Eggs, baskets, and all that.

For the purposes of this discussion, just keep in mind my previous point (https://forum.mrmoneymustache.com/investor-alley/portfolio-design-idiots-v-gurus/msg2479041/#msg2479041) about how Radagast's SWR analysis accounts for consistency over all timeframes rather than simply high returns over one timeframe (also read his original "no bullshitting" (https://forum.mrmoneymustache.com/investor-alley/portfolio-design-idiots-v-gurus/msg2471613/#msg2471613) explanation in Note 1). It's an important distinction that implies the order of the list is not as random as you may believe if you're only accustomed to thinking about maximizing returns.

I understand where you are coming from in relation to portfolio design. I like diversification as well but maybe the message should be buy the most diversified low cost index funds across different sectors. So real estate, bonds, stocks and commodities. Then possibly pick your asset allocation among those four sectors. We can all come to different conclusions based on the data. I think recognizing the principles of portfolio design and WR's is really important to understand but there isn't one solution to fit all.

Your data and portfolio analysis is really good. Your site is unreal. Thank you for that.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 19, 2019, 11:30:59 PM
How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio? Crypto seems like a bad "investment" to me, but... well, where is the line?  "The answer is stocks" seems too glib. Even "the ones Tyler wisely chose because of a bunch of data" just seems...well, probably good, but I feel like we're missing something. Even if the categories on portfoliocharts.com are the "right" ones historically, it seems likely to me that they probably include something (stocks?) that are ready for a permanent fall, and exclude something with a shorter history or other exclusion reason that will perform much better. Is diversifying amoung Things That Did Well Up To Now really our best option?

Sorry if I'm rambling. This part seems really confusing to me.
I was recently reading Taleb, who suggests that the time an idea or concept can be expected to endure in the future is proportional to the time it has already existed. Apparently it originated with Broadway shows, a show that played for a week would be expected (on average) to end in a week, one that lasted a year would continue another year. He extended that to ideas: Socrates has been around 2500 years and will probably be around another 2500, Machiavelli has been around six hundred and will probably be around another 600, Arendt has been around 70 and will still be read in another 7 years. Also possibly for species. The fact that something abstract has been around a long time indicates it has enduring value that will be relevant to the future.

Extrapolate those to investments. "Real" assets like land, gold, wheat, pork bellies would be expected to continue to be traded for thousands of years in the future because they have already been traded for that long. Petroleum "rock oil" has been big for maybe 150 years, and will likely mirror that and disappear in another 150 years. Bitcoin is the oldest and safest of cryptos and might be expected to last another decade, while a crypto introduced a year ago will probably fade by next year.

Of course that is a median expectation, not a perfect explanation. Obviously new things appear, while Cats left Broadway and lead plumbing is a terrible idea even though it was used for millennia. But it seems like a good starting point.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 19, 2019, 11:48:46 PM
Re-reading the OP, I finally notice something that's probably important:

"Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant."

Oh. Ah. Ding. Light bulb. (Man, how did I miss this?) So...all the naive portfolios are about 70-75% stock. Where they differ from a lot of expert portfolios is having more "real assets" (I guess real estate, gold and commodities) than some of the expert portfolios, and possibly having more diversification with the broad categories listed in the quote. They're kind of a consistent allocation of their own, seemingly.

An allocation that worked well, evidently. But... so does that lead us back to questioning whether a good past allocation is a good future one?

I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.
Yeah I probably understated that. "Is very significant" might be more appropriate. Although one of the "Naive" portfolios is 90% stock 10% cash and I don't recall it standing out. I do think that the ratio in Portfolio Charts is a pretty good guess about the future though.

Another point is that I first thought of this idea a few years ago and even did a few tests using Portfolio Visualizer, which uses a totally different set of assets over a different time. It also has very different backtesting tools. PV has a lot more and different types of bonds and would have given more bonds and lower stocks and fewer "real" assets. So this was more of a test which supported the hypothesis.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: kenmoremmm on October 20, 2019, 01:30:33 AM
How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio? Crypto seems like a bad "investment" to me, but... well, where is the line?  "The answer is stocks" seems too glib. Even "the ones Tyler wisely chose because of a bunch of data" just seems...well, probably good, but I feel like we're missing something. Even if the categories on portfoliocharts.com are the "right" ones historically, it seems likely to me that they probably include something (stocks?) that are ready for a permanent fall, and exclude something with a shorter history or other exclusion reason that will perform much better. Is diversifying amoung Things That Did Well Up To Now really our best option?

Sorry if I'm rambling. This part seems really confusing to me.
I was recently reading Taleb, who suggests that the time an idea or concept can be expected to endure in the future is proportional to the time it has already existed. Apparently it originated with Broadway shows, a show that played for a week would be expected (on average) to end in a week, one that lasted a year would continue another year. He extended that to ideas: Socrates has been around 2500 years and will probably be around another 2500, Machiavelli has been around six hundred and will probably be around another 600, Arendt has been around 70 and will still be read in another 7 years. Also possibly for species. The fact that something abstract has been around a long time indicates it has enduring value that will be relevant to the future.

Extrapolate those to investments. "Real" assets like land, gold, wheat, pork bellies would be expected to continue to be traded for thousands of years in the future because they have already been traded for that long. Petroleum "rock oil" has been big for maybe 150 years, and will likely mirror that and disappear in another 150 years. Bitcoin is the oldest and safest of cryptos and might be expected to last another decade, while a crypto introduced a year ago will probably fade by next year.

Of course that is a median expectation, not a perfect explanation. Obviously new things appear, while Cats left Broadway and lead plumbing is a terrible idea even though it was used for millennia. But it seems like a good starting point.

hmm. maybe it's late and i'm not comprehending, but this taleb concept makes no sense to me.

why is the expected duration based on the year in which i look at that commodity to estimate how much time it has left? if bitcoin has been around 10 years, and is expected to go kaput in 10 more, how do i logic this out when bitcoin has been around for 15 years? do i say it only has 5 more years to go, or another 15?

it seems like this is an infinity loop.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 20, 2019, 11:33:28 AM
When it's been around 15 years its self-life is expected to be another 15.

However long it has lasted is a rough rule of thumb for how much longer it will last.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: kenmoremmm on October 20, 2019, 11:55:01 AM
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: dragoncar on October 20, 2019, 12:09:11 PM
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yeah I’m struggling to figure out how this rule minimizes prediction error.  Let’s say a new play opens in year 1 and ends in year 10.  At the end of year 1, the rule predicts a two year run, error 8.  In year two, prediction is 4, error 6.  So each (year; prediction; error):

1; 2; 8
2; 4; 6
3; 6; 4
4; 8; 2
5; 10; 0
6; 12; 2
7; 14; 4
8; 16; 6
9; 18; 8
10; 20; 10

Average  error 5, so presumably average error is half of the life of the play. 

Even a naive “the play will last one more year” rule would have an average error of 3.7.  Im sure there are even better rules
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: arebelspy on October 20, 2019, 06:44:01 PM
I think it's more for older things, saying "Socrates has been around a long time, it's a good bet he'll still be around in 500+ years (2k+ according to his rule of thumb" and "this thing is very new, I shouldn't predict it should last forever."

I think it's mostly to remind you that the old stuff isn't going away, and the new stuff that looks great may not be around for as long as you think. Sort of a mental model to help ward against recency bias.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 20, 2019, 10:48:42 PM
Think of it as mean time to failure, not a definite time it will end. If it is relevant after ten years you estimate there is a 50/50 chance it will still be relevant relevant in another 10. Not it will end then, but that on average you would expect something that has endured that long to double it’s longevity. Half would fail sooner and half later.

Or you could invert it and estimate the odds it will become worthless next year. If something started ten years ago you estimate there is a 1/10 chance it will be gone next year. If it has been around 5,000 years and is still widely used, you estimate there is just a 1/5000 chance it will suddenly become irrelevant next year.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: mrmoonymartian on October 21, 2019, 02:09:38 AM
Yeah I’m struggling to figure out how this rule minimizes prediction error.
Probably talking about different errors. I see that if you don't take the absolute value of the error and keep the vector, then the errors cancel out. Meaning for a particular event predicted regularly, the mean of all of predictions was in fact accurate. They just weren't particularly precise, as you demonstrated.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: talltexan on October 21, 2019, 09:54:37 AM
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yesterday, I thought it was going to be dead at the end of today with 50% probability. Since it's clear that hasn't happened, I need to update my beliefs with the fact that it didn't die and recalculate.
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: ChpBstrd on October 21, 2019, 12:00:56 PM
The durability of cars and Thanksgiving turkeys and virtually anything physical doesn’t work this way. IDK about ideas/memes though, such as the idea that gold/cryptocurrency has value or that one is supposed to eat turkey on Thanksgiving. Technology, culture, and industries are changing at a faster pace now than at any point in history. I would be wary of claims that people in 20-30 years will behave the same way people do now, or even want the same things (e.g. car ownership is starting to decline, people now spend more money and time on cell phones than TVs, veganism is taking off, stock trade commissions are free, and all this was not even sci-fi a few years ago.).
Title: Re: Portfolio Design: Idiots v. Gurus
Post by: Radagast on October 21, 2019, 09:38:51 PM
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yesterday, I thought it was going to be dead at the end of today with 50% probability. Since it's clear that hasn't happened, I need to update my beliefs with the fact that it didn't die and recalculate.
Right, you continually update your expectations of the unknown. In 2010 I expect bitcoin to be forgotten in a year, if anybody even knew of it. In 2011 I guess there is a 50/50 chance bitcoin will disappear by 2013. In 2013 I think there is a 50/50 chance it will be relevant in 2017. In 2017 it is still around so I estimate there is a 50% chance it will exist in 2025. Right now in 2019 we guess it has 50/50 of making it to 2029.

The durability of cars and Thanksgiving turkeys and virtually anything physical doesn’t work this way. IDK about ideas/memes though, such as the idea that gold/cryptocurrency has value or that one is supposed to eat turkey on Thanksgiving. Technology, culture, and industries are changing at a faster pace now than at any point in history. I would be wary of claims that people in 20-30 years will behave the same way people do now, or even want the same things (e.g. car ownership is starting to decline, people now spend more money and time on cell phones than TVs, veganism is taking off, stock trade commissions are free, and all this was not even sci-fi a few years ago.).
Easy come, easy go.

It is supposed to apply to things without a defined expiration date. A mosquito, rat, or human would individually die on a fairly predictable schedule. Mosquitoes, rats, and humans can be expected to last another 200 million, 50 million, and 100,000 years.

Anyhow getting off topic.