Backtesting the PP in Other Countries (Economies)

General Discussion on the Permanent Portfolio Strategy

Moderator: Global Moderator

aj76er
Junior Member
Junior Member
Posts: 5
Joined: Sun Mar 06, 2016 1:28 pm

Backtesting the PP in Other Countries (Economies)

Post by aj76er » Sun Apr 23, 2017 1:10 pm

Hi All,

For more empirical evidence of how the PP would respond to various economic conditions is there any (vetted) data on how it would have worked in other countries? Some example scenarios I'd like to see:
1. Long, deflationary conditions (Japan)
2. Transition from supplier of world's reserve currency (Britain)
3. Hyper-inflation (Argentina)
4. Economic collapse (Iceland)

These are (or were) all developed countries and so I feel like similar situations could occur in the U.S.. Thus, I'd like to see how the PP or GB holds up against real-world data as a validation of the theory.

And for FWIW, I'm considering transitioning to a GB-inspired portfolio:

1. 40% Stocks (in 2:1 split between U.S. and ex-U.S.)
2. 20% Gold
3. 20% LTT
4. 20% Cash (including I/EE-bonds)

Thanks!
aj76er
Junior Member
Junior Member
Posts: 5
Joined: Sun Mar 06, 2016 1:28 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by aj76er » Sun Apr 23, 2017 1:11 pm

In searching the forums, I found a link to an old blog:

http://europeanpermanentportfolio.blogs ... eland.html

It appears to no longer be hosted, or has moved somewhere
User avatar
frugal
Executive Member
Executive Member
Posts: 925
Joined: Sat Nov 10, 2012 12:49 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by frugal » Sun Apr 23, 2017 1:36 pm

Hi

Interesting but i dont know the answer.

Our PP friends know and will answer.

I think PP worked well on those countries...

Rgds
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Sun Apr 23, 2017 5:01 pm

I've been doing some tinkering, and am starting to get the hang of modeling international portfolios. Here's a sneak preview of one country on your wish list. Long story short -- the PP is really resilient and is more than just a US portfolio.

Image
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
aj76er
Junior Member
Junior Member
Posts: 5
Joined: Sun Mar 06, 2016 1:28 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by aj76er » Sun Apr 23, 2017 6:12 pm

Wow, awesome work. I can't wait for the final product! Out of curiosity, how many country data sets are you working on? I promise I won't divulge outside this thread :)

I must admit that your site is what clued me in to the PP concept :)
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Sun Apr 23, 2017 6:59 pm

I'm still experimenting, and nothing is set in stone. But I've long been interested in international data as well, and I'll be sure to let everyone know if/when I finally get some robust tools working.
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
User avatar
Hal
Executive Member
Executive Member
Posts: 856
Joined: Tue May 03, 2011 1:50 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Hal » Sun Apr 23, 2017 7:28 pm

Well done !

I have long been interested in how (or if) gold hedges currency inflation of smaller nations.....

And soon I shall be "enlightened"
AnotherSwede
Senior Member
Senior Member
Posts: 117
Joined: Mon May 11, 2015 10:24 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by AnotherSwede » Mon Apr 24, 2017 9:08 am

I did a backtest on Japan. Thinking that (demographics) is what we are all gonna go through.

I used ten year bonds and it was a decent return above cash and low risk.
User avatar
frugal
Executive Member
Executive Member
Posts: 925
Joined: Sat Nov 10, 2012 12:49 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by frugal » Mon Apr 24, 2017 3:20 pm

Tyler wrote:I'm still experimenting, and nothing is set in stone. But I've long been interested in international data as well, and I'll be sure to let everyone know if/when I finally get some robust tools working.
TYLER

you are the best! ;)

Please don't forget to do it for EUROPE in EUROS.

Thank you my friend!
grapesofwrath
Associate Member
Associate Member
Posts: 46
Joined: Thu Jul 14, 2016 5:57 am

Re: Backtesting the PP in Other Countries (Economies)

Post by grapesofwrath » Mon Apr 24, 2017 4:09 pm

Tyler, I would like to say I love your website and tools. So simple, so clean, so insightful !

PS. I would love to see Japan since its past is such a different beast (but one that not too improbably Europe and even US could follow...)
tim47
Junior Member
Junior Member
Posts: 8
Joined: Fri Sep 16, 2011 3:07 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by tim47 » Mon Apr 24, 2017 6:37 pm

Tyler,
I seldom post, but I really want you to get what a breath of fresh air you are in a community really stuck on and attached to their personal view of asset allocation. I really honor your transparency and your obvious commitment that each of us use our own basic intelligence to dinvest in a way that works best for us. I sense no hidden agenda with you. You are a really amazing soul... bless you and thank you for all you have provided me....
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Mon Apr 24, 2017 8:43 pm

Thanks guys. It's my pleasure.

Maybe it's the Harry Browne libertarian influence, but I've learned that trying to convince people to do something against their nature is a losing battle. Arming people with good information so that they can make educated choices is not only more constructive but also a lot more sustainable.

BTW, the beauty of the Permanent Portfolio is that the data speaks for itself. It's a really nice portfolio, and I think the consistency surprises a lot of people.
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
Kike Moreno
Associate Member
Associate Member
Posts: 43
Joined: Thu Jul 17, 2014 3:36 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Kike Moreno » Tue Apr 25, 2017 6:55 am

Hi Tyler, I join the congratulations, your work is really really useful for us.

As I suggestion, if you could consider adding these assets/indexes I think they would be very useful to evaluate a Euro Permanent Portfolio and Euro Golden Butterfly:

- MSCI EMU index (EUR)
- 30 year German bond
- MSCI Small Cap World (EUR), for the Euro Golden Butterfly

For the MSCI indexes the source of data is https://www.msci.com/end-of-day-data-search

I could not find the MSCI Small Cap World values for 1999 and 2000 but, since MSCI World SC is very similar to MSCI ACWI SC, this index could be used for these 2 years.

Thanks again!
Last edited by Kike Moreno on Wed Apr 26, 2017 2:39 am, edited 1 time in total.
fi50@fi2023
Associate Member
Associate Member
Posts: 36
Joined: Sun Apr 05, 2015 1:56 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by fi50@fi2023 » Tue Apr 25, 2017 9:38 pm

I almost never post here, but I have to jump in here on the Tyler praise bandwagon. I have been a long-time fan and love the mathematical approach to investing that his site promotes. The data is very helpful. I have a blog -- www.bk2fi.com -- and constantly send my readers to Tyler. I follow what I call the Golden Merriman - with 60% of my portfolio in Paul Merriman's Ultimate Buy and Hold portfolio and the balance equally divided between LTT, Gold, and STT. I could not have arrived at this portfolio without Tyler's excellent analytical tools. Thanks Tyler!
User avatar
frugal
Executive Member
Executive Member
Posts: 925
Joined: Sat Nov 10, 2012 12:49 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by frugal » Thu Apr 27, 2017 1:22 pm

Kike Moreno wrote:Hi Tyler, I join the congratulations, your work is really really useful for us.

As I suggestion, if you could consider adding these assets/indexes I think they would be very useful to evaluate a Euro Permanent Portfolio and Euro Golden Butterfly:

- MSCI EMU index (EUR)
- 30 year German bond
- MSCI Small Cap World (EUR), for the Euro Golden Butterfly

For the MSCI indexes the source of data is https://www.msci.com/end-of-day-data-search

I could not find the MSCI Small Cap World values for 1999 and 2000 but, since MSCI World SC is very similar to MSCI ACWI SC, this index could be used for these 2 years.

Thanks again!
+1

^-^
Kike Moreno
Associate Member
Associate Member
Posts: 43
Joined: Thu Jul 17, 2014 3:36 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Kike Moreno » Fri Apr 28, 2017 1:40 pm

This is a possible source of data for the German bonds:

https://index.db.com/dbiqweb2/index/ibo ... ny_15_year

It is in fact a index, and not exactly the 30-year bund, but could be used a reasonable substitute.

Another option is to use the value that Brownehead has calculated in his amazing blog (the second column, "Bonos"):

http://www.carterapermanente.es/evoluci ... ermanente/

He has derived the values using the rates evolution plus the yearly coupon:

http://www.bundesbank.de/Navigation/EN/ ... skms_it03b

Source: http://www.carterapermanente.es/evoluci ... ment-42220
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Fri Apr 28, 2017 2:35 pm

Thanks for the info.

I have a pretty good handle on bonds from many countries (including Germany), as I've figured out how to replicate real-world bond index funds from yield curves (similar to what Brownehead is doing, but accounting for the maturity ranges as well).

Still working on it, but I'm getting close.
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
Kike Moreno
Associate Member
Associate Member
Posts: 43
Joined: Thu Jul 17, 2014 3:36 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Kike Moreno » Fri Apr 28, 2017 2:47 pm

That sounds great, thanks!

Will you add support for other currencies like the euro?

Just having the end of the year conversion from the USD to the EUR (or any other currency) should be enough to generate all the assets values in EUR and evaluate the portfolios using assets in EUR.
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Fri Apr 28, 2017 3:00 pm

That's the plan. :)
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Sat Apr 29, 2017 5:43 pm

Alright -- this should keep everyone busy for a while. O0

https://portfoliocharts.com/2017/04/29/ ... ng-global/

Let me know if you spot any issues or have any questions.
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
grapesofwrath
Associate Member
Associate Member
Posts: 46
Joined: Thu Jul 14, 2016 5:57 am

Re: Backtesting the PP in Other Countries (Economies)

Post by grapesofwrath » Sat Apr 29, 2017 8:13 pm

wow !
User avatar
frugal
Executive Member
Executive Member
Posts: 925
Joined: Sat Nov 10, 2012 12:49 pm

Re: Backtesting the PP in Other Countries (Economies)

Post by frugal » Sun Apr 30, 2017 12:54 am

Tyler wrote:Alright -- this should keep everyone busy for a while. O0

https://portfoliocharts.com/2017/04/29/ ... ng-global/

Let me know if you spot any issues or have any questions.

OK BOSS!

Thank you.

I will read it now!
Kike Moreno
Associate Member
Associate Member
Posts: 43
Joined: Thu Jul 17, 2014 3:36 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Kike Moreno » Tue May 02, 2017 4:12 am

Hi Tyler, thanks a lot for the multi-country support. This is really great for us living outside the US.

I have 2 suggestions for you:

1) Instead of splitting the analysis by "countries", I think it would be more useful if splitted by "currencies". I understand that the generalization of Harry Browne Permanent Portfolio beyond the US and the USD should be done at the currency level, not a the country level, in this way:
- Stocks: index of all the stocks trading under the currency area (example: MSCI EMU). Why limit stocks to our own country if we can have a broader index without the currency risk?
- Bonds: safest long term bond (example: German 30 year bund). Why use our own country bond if we can use the safest bond (which behaves better in deflations) without any currency risk?

Then the only thing that is more useful at the country level than the currency level is the inflation. So having 2 selectors, one for the currency which will select all the applicable assets, and another for the country which will select only the inflation to calculate real returns, would be for me the optimum solution. Another option would be to use currency zone inflation (for example the eurozone inflation).

2) Could you add "World Small Caps" to the assets list?

Thanks again!
User avatar
Tyler
Executive Member
Executive Member
Posts: 1970
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: Backtesting the PP in Other Countries (Economies)

Post by Tyler » Tue May 02, 2017 11:57 am

Kike Moreno wrote: I understand that the generalization of Harry Browne Permanent Portfolio beyond the US and the USD should be done at the currency level, not a the country level, in this way:
- Stocks: index of all the stocks trading under the currency area (example: MSCI EMU). Why limit stocks to our own country if we can have a broader index without the currency risk?
- Bonds: safest long term bond (example: German 30 year bund). Why use our own country bond if we can use the safest bond (which behaves better in deflations) without any currency risk?
Yeah, I understand how investing only in home country stocks doesn't always make sense for smaller economies. The calculators account for that by offering 6 diversified international options, and I think building a PP with a heavy dose of a more global stock fund is a fine idea.

I don't like focusing so much on the EMU as an independent entity because the data is irrelevant prior to 1999. Also, from what I've seen in the data just because countries share a common currency does not mean they share the same inflation. IMHO, having a separate inflation setting just gets... messy. But I see where you're coming from.

The reason I've kept bonds purely to domestic markets for now is that the majority of international bond index funds tend to be hedged. That may be trivial across the European Union today, but when you think about outside countries and older timeframes before the euro existed it gets really complicated. So it's something I'll have to tackle later.

World small caps are on my short list for things I'd like to add. The issue is that data is really hard to come by. As you already pointed out, even MSCI only has ACWI SC data back to 1994. I'm exploring different sources and models for backtesting options, but it will take time to get right.
Mechanical engineer, history buff, treasure manager... totally not Ben Gates
Kike Moreno
Associate Member
Associate Member
Posts: 43
Joined: Thu Jul 17, 2014 3:36 am

Re: Backtesting the PP in Other Countries (Economies)

Post by Kike Moreno » Tue May 02, 2017 12:13 pm

OK, I understand that having 2 separate settings, one for the currency and another for the inflation could be too messy. But maybe you could just add a new "country" called "eurozone" with this configuration:
- data only from 1999. I know the timeframe is short, but it's better than nothing :)
- stocks: MSCI EMU
- bonds: German 30-year bund (the same currently used for German configuration)
- cash: German short term (the same currently used for German configuration)
- inflation: eurozone inflation (it is published monthly)

From what I see in your engine it looks like all these are configurable parameters except for the start year. Currently all the countries work on datasets starting on 1970, so maybe it is somehow hardcoded in the tool...
Post Reply