PP Periodic Table

[Clive]

?

[Jan Van]

Cool!
I'd also like to see that graph with the losing investments in red, and gaining in green. Just to see winners vs. losers...

[Pkg Man]

Very nice Clive.  Thanks.

[KevD]

Great chart, for sure. 

Would love to hear people's comments on any general patterns that come out of this.  Not that it matters from a PP standpoint, but maybe offers clues to overweights for the variable part.

[MediumTex]

Would love to hear people's comments on any general patterns that come out of this. 

I once had a quilt with a pattern that looked a bit like that.

[Pkg Man]

Great chart, for sure. 

Would love to hear people's comments on any general patterns that come out of this.  Not that it matters from a PP standpoint, but maybe offers clues to overweights for the variable part.

For me it shows that trying to divine what may be the hot segment next year is fruitless.  Looks rather random to my eyes.

[craigr]

I love it!

[moda0306]

Pkg Man,

If you use pure probability vs actual results, you'll see that there is a repetative trend... even though it doesn't appear so.

I did an analysis (I counted, so I may be off by 1 count +/-) of only the 3 volatile assets (gold/LT/Stocks) and here's what I found.

In 38 years, 17 times (45%) has the best asset also been the best the year before. (33% chance of that happening)
In 38 years, only 4 times (12.5%) has the worst asset done the best the next year. (33% chance)
In 38 years, only 9 (24%) times has the best asset done the worst the next year. (33% chance)
In 38 years, 16 times (42%) has the worst asset done the worst in the next year. (33% chance)
In 38 Years, the assets have performed in the same order as the prior year 11 times (26%). (11% chance)
In 38 Years, the assets have performed in the Opposite order of the prior year 3 times (8%). (11% chance)

Like I said, this is just based on the 3 volatile assets and my counting, but compared to if it was completely random, the PP does tend to favor previous years' best asset.  Not worth throwing all your cards at, but maybe a VP consideration.

[MediumTex]

It also demonstrates what a secular bull market in an asset looks like--muscular year over year gains for long periods.  Look at stocks in the 1990s.  Pretty impressive.

After each of those huge years I'm sure people were saying "that's got to be it--it can't go any higher."  But then it would go higher.

Right now we are in such a market for gold.  Perhaps people will look back on this period and say "I can't believe gold was available for under $1,400 and people weren't buying all they could."

But then again after a secular bull market ends some of those gains look sort of silly.  I'm sure that through the stock market bull in the 1980s and 1990s people were wondering how they could have had such strong gold fever in the 1970s.  It's just a shiny piece of metal after all.

It's funny how we create the narrative after the fact and imagine that people who lived through it were thinking in those terms the whole time.

[Pkg Man]

I see your point moda.  But as you said, I am not sure it shows anything that can be acted upon.  Also, if you play around with random numbers you will often see what appears to be a pattern, but later the pattern fails to hold up.

In other words, I am suggesting that the difference between 33% and 45%, as in the first case listed below, may not be statistically significant.  Actual outcomes generated by a random process won't always equal the expected value of the random process; i.e., if I flip a coin 38 times it wouldn't be a terrible surprise if I don't get 19 heads and 19 tails.  Of course this is particularly true over a small number of trials.

That said, I am not suggesting the returns on stocks, bonds, and gold is a purely random process.  

Nice work, and thanks for posting.

[moda0306]

I think if you flipped a 3 sided coin (you know what I mean) 38 times, you'd play hell getting that far off probability.  But it's easy to look at the past and see trends that are nothing but coincidences that won't relive themselves.  Like I said, it's not enough to bet the bank on, but it's enough to think that maybe a VP "chasing performance" candidate is worth trying.

[cowboyhat]

Very interesting.

I did a hand tally of rank order while drinking my coffee (might have made a mistake or two) and these were the results:

      G    TSM  LT    ST
1  12    20    5      3
2    4    5      18    11
3    7    6      7      18
4  15    8      8      6

The median ranking is TSM, LT, ST, G.

I think it is interesting that G is typically either doing best or worst while ST is almost exactly opposite (always second or third). Seems like this makes them a complementary long term pairing. TSM and LT are just winners in the series.

[moda0306]

Clive,

On top of the growth gain that you mentioned, you also get smoothing of returns, which looking back may not seem important, but in the midst of a 30% drawdown it feels huge and should.  That's what's so great about the PP.  The fact that the assets all return reasonably well (nominally), and compound on that lack of correlation when combined, make the PP return even better than stocks in some instances.  If you tweak the PP for more stocks, you do get bigger drawdowns, but you'd probably have a risk/return ratio still head and shoulders above the Bogleheads folks that don't like the PP.

[Roy]

A reasonable variation IMO of the FTM would be instead of perhaps something like 30% SCV and 70% Treasury's, to maybe allocate 10% to each of SCV, Emerging Markets and maybe the above timed gold or stock (TSM) method (i.e. hold gold if it was ranked one or two last year otherwise hold TSM).  This in effect still holds 'spicier' average gain potential, but in a more diversified manner and potentially with an even lower overall volatility.  Perhaps for the 70% 'cash' side, a 5 year treasury ladder might be held.

Larry often holds a small amount of commodities (CCFs, though), representing about 10% of his equities.  He does not style drift his equities or alter the overall percentage, but did move out of Large Caps in 1998 (to his "Fat Tails Minimized" SV-based, internationally diversified + some EM to about 25%-33% of foreign), and has stayed so ever since with this portfolio concept.  Of course, he uses ST nominal fixed income with a shifting maturity (like DFA) or TIPS as Real yields favor.

I was looking at some of the years going way back to the time one needed to hold individual stocks (no indexes).  Just comparing SV and TSM since so much Boglehead debate lies there:

• TSM and SV both had 24 years where the returns were Negative.
• 18 of those 24 years both TSM and SV were Negative (same time).
• There were 6 years where TSM was Negative and SV Positive.
• There were 6 years where SV was Negative and TSM Positive.
• There were NO extended periods where SV was Negative and TSM Positive.
• The years where SV was Negative and TSM positive = 1939,1948,1953,1960,1987,1998.
• The years where TSM was Negative and SV positive = 1970,1977,1981,2000,2001,2007.
• There were no 2 consecutive years where SV went down and TSM up.
• There was 1 string of 2 years where TSM went down and SV up (2000-2001).

Overall, his portfolio seems to accomplish much of what the PP does (Returns and Risk).  But that might be said of any low-Beta portfolios that also have lots of AAA Treasuries.

[Roy]

Two 'riskier' assets that individually have higher yearly average gains, but also above average volatility/risk (standard deviation) might individually produce a comparable annualised reward to that of a less risky alternative.  When however those assets are not perfectly correlated then both the annualised can rise and the standard deviation (risk) decline.

As you say Roy these are all just similar overall to any other 'lots of treasury' style (Nassim Taleb's etc.). I suspect the three speculative allocations I suggested might work as equally as well as any other choice... 

The PP's rebalance method is such that from a relative high (such as the 1979 peak after a series of sequential good years) can result in progressively adding more into gold during its decline in price down to a low-and-stay-low type state that then stays there for the longer term (20+ years). 

Hi, Clive,

You bring up a number of good points we've discussed. 

We can chat about correlations of riskier asset classes and how they can combine to form more efficient portfolios thanks to Markowitz, who gets the Nobel Prize for establishing this and other concepts.  Hard to imagine a time when this was not commonly known, yet many seem still not to know it.

I worry about the Gold too.  I held GLD (which is both a great convenience and a concern) and won't hold physical Gold (just too inconvenient);  there are lots of Gold players these days due to the ETFs and I wonder also about the potential effects of that.  But perhaps nothing remarkable will occur.

The rebalance-into-losers drag is something that will afflict every portfolio type (Domestic vice International equities;  Small and Large Caps, etc.), though I imagine given a 25% weighting to a volatile asset may be the crux of your concern?  But if Treasuries take a prolonged dive, that 70% assigned to a FTM portfolio will seem a draining drudgery too, especially if using IT vice ST.  I suppose, this "problem" derives from its strength—the way the PP works with its volatile components.  Another way is to just hold a portion always, maybe 5%-10% and perhaps using CCFs and Gold, but that is not the PP anymore.  So many love the PP concept, and then figure ways to change it!

There is going to be style drift by unconventional portfolios, perhaps severe drift.  This is also true of the FTM portfolio types, though to a lesser degree, perhaps.  Everyone loves it when the Left Tail is minimized but many will grumble on the Right—and that may last a few years.  The last 3 years have been as good as it gets for the PP—outstanding Bear and Bull performance (like '73-'76).  No other portfolio type has responded this way (with large moves up in both cycles), including FTM.  Maybe this is the PP new normal, but counting on that runs the risk of illumination at a later time.   No easy answer in this emotional realm.  One either accepts this risk going-in or gets bushwhacked by it at a bad time.  I believe, like Monty Python's "Spanish Inquisition" skit, that nobody expects this when it does happen, including those who claim otherwise.  It is the quintessence of "easy to say, hard to do."  Heck, look at the number Bogleheads who can't even stick with their IT/TBM fixed income at this time, and that's not even style drift, just normal market stuff with lots of rumors.  Wait til they see what inflation and actual rising rates do!  That aside, genuine style drift is a real risk and trying to time a way around it, perhaps riskier still.

There have been lots of ways to go with FTM, including your approaches.  At center there is low Beta and lots of Treasuries to all these approaches, with volatile hard assets at the fringes (or as a major player with the PP).  It all comes down to finding some approach to which commitment can avoid later capitulation.  And as we've seen from the famous "Plan B" threads, that isn't easy, even for the true faithful.

[cowboyhat]

Thoughtful comment, Roy.

Perhaps it is too much to ask for one system of investment to cover all circumstances. The PP seems to be a good way to invest in uncertain times. The reason I am attracted to it is that it seems to offer me a rational way to invest in gold.

The weakness in the PP is that when gold is in one of its long periods of being out of favor, generally prosperity reigns, and therefore stocks are producing great returns (the lost decades of Japan are an instructive counter example to this generality). As Roy points out, in boom times a PP investor faces the really serious problem of losing faith in the PP due to tracking error.

If you accept that you can not plan for all seasons with one investment system, maybe an approach is to use Harry Browne's VP idea to your advantage. If gold has had two consecutive years of declines then maybe that is reason to believe an era of prosperity is dawning and it is time to open a VP focused on stocks. If gold has had two consecutive years of out performance maybe that is a clue that there is rough weather ahead and it is a good time to pull your bets off the table and go back to a 100% PP.

I think what happened in Japan was that when their economy went into the tank the rest of the world was booming. Consequently gold did not perform and Japanese investors were left without good domestic options. To me the lesson of Japan is that if you have a PP in a country that does not dominate the world's economic system you should not count on gold to save your bacon. Might be a good idea to invest part of your cash and stocks internationally.

[cowboyhat]

Clive,

Do the numbers in your Periodic Table include re-balancing or are they year-over-year return numbers?

The reason I ask is that when I use your numbers as year-over-year return numbers and set up a PP in Excel I noted that all of the re-balancing events in my Excel sheet were triggered by either stocks or gold. Furthermore, all the <15% re-balancing events were triggered by poor gold performance.

Thinking about this I asked my self what a "Let Sleeping Dogs Lie PP" would return. In this PP you ignore all the <15% signals and only re-balance when something gets >35%. Return was about 8% better over the entire series in Let Sleep Dog's Lie due to skipping two re-balances during gold's fallow years. Let Sleeping Dogs Lie might be a way to mitigate some of the problems with periods of gold under-performance.

My curiosity was roused by this observation so I also looked at a "Contrary PP" where only the <15% signal was used to trigger a re-balance.  In other words buying the dogs. Performance over the series was significantly enhanced (32%) by this approach.

[Roy]

Perhaps it is too much to ask for one system of investment to cover all circumstances. The PP seems to be a good way to invest in uncertain times. The reason I am attracted to it is that it seems to offer me a rational way to invest in gold.

The weakness in the PP is that when gold is in one of its long periods of being out of favor, generally prosperity reigns, and therefore stocks are producing great returns (the lost decades of Japan are an instructive counter example to this generality). As Roy points out, in boom times a PP investor faces the really serious problem of losing faith in the PP due to tracking error.

If you accept that you can not plan for all seasons with one investment system, maybe an approach is to use Harry Browne's VP idea to your advantage

Hi, Cowboy,

Note that the tracking error I'm discussing will also afflict other non-conventional portfolios, to varying degrees (those that deviate strongly from Total Markets investing).  Small Value tilters did not enjoy cocktail parties with S&P investors in '98-'99, assuming anyone was investing that way then.

The PP has also done well in good times.  In most of the big S&P years, the PP had "good" returns.

I never understood the Variable PP, probably because I have no interest in getting a gambling fix.  And if extending it for reasons other than play, I have even bigger problems with it.  Thus, were I to use the PP, I would use it exactly as designed, always.

Back to Gold again, I wrote in the original thread:

Gold had 17 down years averaging -7.84%
In those years the HB PP averaged +6.75%

• The worst Gold loss was -32.8%. In that year the portfolio returned -3.9%
• The second worst Gold loss was -22.7%. In that year the portfolio returned +8.3%
• The third worst Gold loss was -21.5%. In that year the portfolio returned +7.5%

Thus, if you were not watching the daily news, and rebalanced yearly, (or as HB suggested), you could do OK emotions-wise.

[cowboyhat]

Roy,

I wasn't thinking about the PP's decent performance when gold was getting crushed, and that's an excellent point.

Really, probably, the greatest strength of the PP is that when followed it addresses the human investor's psychological vulnerabilities. Harry Browne's idea of a reasonably sized VP as an emotional outlet if your acquaintances are getting you down with their stock market talk is another good idea. Have your fling. Lose a little money. Sober up. Crawl home to lick your wounds. If you don't bleed once in awhile you aren't alive.

Not letting a <15% gold position trigger a re-balance also doesn't seem to meaningfully impact returns. So if buying into that ugly gold market when nothing else in the portfolio is working very well is too much to bear emotionally, there's probably no harm in skipping it.

Not re-balancing out of your winners does hurt returns though.

[julian]

Clive,

TLT lost 20%+ in 2009 not 12%?

[MediumTex]

Simba's spreadsheet uses VUSTX for LT treasurys.

That's probably where the 12% came from.

[julian]

But if we all suggest using TLT then alot of the info we share using VUSTX is incorrect. There was a huge disparity between the 2 in 08 & 09

[MediumTex]

That's the problem with Simba's spreadsheet when discussing PP-related matters.

[Roy]

One FTM investor might opt to hold all that cash in perhaps a 1 to 5 year treasury ladder.  Another might opt to use a ST and LT barbell.  Either or both might also opt to hold some foreign currencies as a domestic currency crisis hedge, or perhaps some gold instead.

Over the mid to longer term likely both might achieve somewhat similar average rewards, but over shorter periods equally likely is that there will be differences seen in the two sets of yearly gains.  If there was a clear and evident case of one being better than the other then those differences could be arbitraged profitably - which would have the effect of closing that gap down.

The particular funds used in various strategies (DFA, Vanguard, etc.) will also affect returns.  The nice thing about the PP is you don't have to worry about that (being a broad, index-like based strategy), unless one makes it a problem.  Same thing for employing shifting maturity strategies (like DFA uses) and such, vice using a set allocation.  PP is set as it gets;  it isn't "hampered by information," as one of my friends likes to say when complex improvements threaten otherwise good ideas.  It does great in Bear markets, and, contrary to some opinions, it usually does well in Bull markets, and when Gold lags;  this would have to be true to get that sort of long-term return.  But it isn't until you scrutinize what's really happening til you see it.  Twenty years from now I'm guessing they'll be saying much the same things about it.
---


The Crawlingroad PP Returns from 1972 (last I saw them) appears to use VUSTX also.  Has this Returns page been updated yet to 2010? And where is it?  What I have is this:

https://web.archive.org/web/20160324133 ... l-returns/

edit:  The 2008 LT Bonds return appears to be TLT, but the ones before match with VUSTX.


And also this old one which differs somewhat:

http://harrybrowne.org/PermanentPortfolioResults.htm

[julian]

Craig,

what r u using for lt bonds for the historical returns?