European PP, update 6 months after starting

General Discussion on the Permanent Portfolio Strategy

Moderator: Global Moderator

johntaylor
Junior Member
Junior Member
Posts: 10
Joined: Fri Jun 01, 2018 8:33 pm

Re: European PP, update 6 months after starting

Post by johntaylor » Sun Jun 03, 2018 12:17 pm

- Stocks: why you use this unknown ETF ?
I presume you mean DE0001134922. I don't actually use these german bonds, because my brokers don't offer them. But I used them for my calculations because it is the only bonds option I know which exists and offers data since 2000. If anyone knows an european bond fund with longterm data I really like to hear.

I am actually using EXX6 and E20Y (50/50) for the bonds portion. EXX6 is <20y restrained to Germany, E20Y is 20+y and covers a greater part of Europe. I prefer long bonds from Germany, because the Netherlands and Germany are economically close related and are expected to move on together if the eurozone would collapse.
- CASH - can you get 1% in a Deposit? Seems too much...
well it took a while to find it ;) but there is a catch, because you get a penalty if you want to end the deposit before the end date (5yr).
- How much you pay for gold storage per year?
0.12%/yr
I believe your numbers are correct. Everyone says USPP is better than EUPP.
I know. But I am especially interested if I am right with my calculations, because the cagrR I calculated will determin the monthly income I will offer myself with the PP. It also determins if and when it can/will be time to give my emplorer the middle finger.

Have anyone ideas why my caluation of 2002-2017 EU PP cagrR is 3.5% while Tyler calculates 5.5% for the german PP in 2002-2017 ?
http://www.carterapermanente.es/cartera ... -con-etfs/

<quote>Our problem is exchange rate... :'( :'(</quote>
Don't understand what you mean exactly. The exchange rates euro->dollar has been:
2001-12-28: E1=$0.8851
2017-12-30: E1=$1.2005
that is a growth of 135.6%. So an european cagrR of 3.5% in euro's in this period should be compared to a cagrR of 3.5%*135.6% => 4.8% ? Is this correct ?

I don't see a big difference between USPP and GerPP on portfoliocharts.com, resp 5% and 5.5%. Does this mean that both are demoninated in dollars ? This should indeed explain the bigger part of the differences between my calculation (3.5%) and Tyler's (5.5%). Please feedback if I undertand things right ?

Please comment.
User avatar
Tyler
Executive Member
Executive Member
Posts: 1638
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: European PP, update 6 months after starting

Post by Tyler » Sun Jun 03, 2018 12:47 pm

johntaylor wrote:
Sun Jun 03, 2018 12:17 pm
Have anyone ideas why my caluation of 2002-2017 EU PP cagrR is 3.5% while Tyler calculates 5.5% for the german PP in 2002-2017 ?
http://www.carterapermanente.es/cartera ... -con-etfs/
Here's my asset assumptions that I would use in the PC calculators from reading your previous posts. From punching in the numbers, I'm pretty confident it's what you're also using.

Stocks: Europe
LTT: Long term German bunds
Cash: German Tbills
Gold: Gold

I can see a few things here.

First, the real CAGR for this portfolio from Jan 2002 - Dec 2017 is 4.98%. Your 5.5% number is from 2003-2017. So make sure you're comparing apples-to-apples on the start date.

Second, I see that you're using a fixed rate of return for cash. Tbills don't work like that. For example, 3-month German bills had an annualized return of 4.6% in 2008. For more accurate results, find a fund that tracks short term bunds or simply use the 3-month German bill rate from OECD.

And third, make sure that you're converting the numbers correctly to German inflation and currency. I use known German inflation and exchange rates every year to represent the real purchasing power in the home country.

Finally, it's always possible that different data sources have different returns. 2% a year is a big difference, though, so I suspect something else is going on from a methodology perspective.
PortfolioCharts.com : a picture is worth a thousand calculations
johntaylor
Junior Member
Junior Member
Posts: 10
Joined: Fri Jun 01, 2018 8:33 pm

Re: European PP, update 6 months after starting

Post by johntaylor » Sun Jun 03, 2018 10:01 pm

Thanks for replying @Tyler, I am sure I can learn things from you ;)
First, the real CAGR for this portfolio from Jan 2002 - Dec 2017 is 4.98%. Your 5.5% number is from 2003-2017. So make sure you're comparing apples-to-apples on the start date.
I took the 5.5% from your '15-yr Rolling returns' graph. As far as I read the graph well the bar with start year 2002 has time frame 2002-2017. But please correct me if I misread the graph.
Image
Second, I see that you're using a fixed rate of return for cash. Tbills don't work like that. For example, 3-month German bills had an annualized return of 4.6% in 2008. For more accurate results, find a fund that tracks short term bunds or simply use the 3-month German bill rate from OECD.
Thanks for pointing me to the OECD site. I found a nice datafeed there:
https://data.oecd.org/interest/short-te ... -rates.htm
I did a quick and dirty check with these data, but that does not seem to change my total real cagr very much. I will update my program and take care to put some proper data for the cash portion. As soon as I have an update I will post a new table with the results here.
And third, make sure that you're converting the numbers correctly to German inflation and currency. I use known German inflation and exchange rates every year to represent the real purchasing power in the home country.
I will update my program with german cpi data also. I found a source here:
https://www.quandl.com/data/RATEINF/CPI ... ex-Germany
But I can't imagine they will be very different from the dutch cpi data I used so far, but we will see.
Finally, it's always possible that different data sources have different returns. 2% a year is a big difference, though, so I suspect something else is going on from a methodology perspective.
Same feeling here, the difference is to big and I like to find out where I am missing essential things. Can it not be caused that we both calculate in different currencies (euro's on my side) ?
User avatar
Tyler
Executive Member
Executive Member
Posts: 1638
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: European PP, update 6 months after starting

Post by Tyler » Mon Jun 04, 2018 12:07 pm

johntaylor wrote:
Sun Jun 03, 2018 10:01 pm
I took the 5.5% from your '15-yr Rolling returns' graph. As far as I read the graph well the bar with start year 2002 has time frame 2002-2017. But please correct me if I misread the graph.
I think it’s just a matter of reference dates. The column you’re pointing to shows the 15-year annualized return starting at the beginning of 2002 through the end of 2016. 2002-2017 is actually 16 years. So just make sure you’re comparing apples to apples.

BTW, I also get my inflation data from OECD if you want to take that variable out of the equation.

Also, one other thing you might want to keep an eye on is the annual rebalancing. All of my numbers assume that the portfolio is rebalanced every year. If you just let the assets ride or use a different rebalancing scheme, the numbers could vary quite a bit.
PortfolioCharts.com : a picture is worth a thousand calculations
Pet Hog
Senior Member
Senior Member
Posts: 128
Joined: Tue May 28, 2013 4:08 pm

Re: European PP, update 6 months after starting

Post by Pet Hog » Mon Jun 04, 2018 2:38 pm

Tyler wrote:
Mon Jun 04, 2018 12:07 pm
The column you’re pointing to shows the 15-year annualized return starting at the beginning of 2002 through the end of 2016.
Tyler, if you're using monthly data, which I believe you are, how is this range not 14 years and 11 months? Perhaps one extra month won't affect the CAGR much, but it's not the same as 15 years.
User avatar
Tyler
Executive Member
Executive Member
Posts: 1638
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: European PP, update 6 months after starting

Post by Tyler » Mon Jun 04, 2018 3:04 pm

I’m in the process on getting monthly data and incorporating it into the calculators, but that will take time. Right now all of the returns are annual numbers as of December 31st.

So it’s definitely a full 15 years. The way the rolling returns chart is formatted, the labeled year is associated with the first full year of returns.
PortfolioCharts.com : a picture is worth a thousand calculations
johntaylor
Junior Member
Junior Member
Posts: 10
Joined: Fri Jun 01, 2018 8:33 pm

Re: European PP, update 6 months after starting

Post by johntaylor » Mon Jun 04, 2018 4:49 pm

Tyler wrote:
Mon Jun 04, 2018 12:07 pm
I think it’s just a matter of reference dates. The column you’re pointing to shows the 15-year annualized return starting at the beginning of 2002 through the end of 2016. 2002-2017 is actually 16 years. So just make sure you’re comparing apples to apples.
oops, thanks for noticing that. It's quite a challenge to compare apples to apples, but I hope to make some progress ;)
Image
The final real cagr is now 4.4% I took care 1. inflation is now german 2. cash is now with short term interest rate of OECD 3. length is 15yr (2002-2016) 4. the portfolio is rebalanced every end of the year to 25/25/25/25. The rebalancing is done by 1. calculate the total return (both nominal and real) = average of the returns of the 4 portions. 2. calculate total growth (both nominal and real) from total return 3. calculate total cagr (both nominal and real) from total growth.
Also, one other thing you might want to keep an eye on is the annual rebalancing. All of my numbers assume that the portfolio is rebalanced every year. If you just let the assets ride or use a different rebalancing scheme, the numbers could vary quite a bit.
Thanks again for pointing me to that. I changed the scheme for that again.

Thanks for helping me on.

[Update] I suddenly realise that the way I work the portfolio is nèver rebalanced and gets more unbalanced every year. You can see it by comparing the values for growth for each portion with each other. For example in 2013 the (un) balance is stocks:13%, bonds:26%, gold:16% cash:44%. So there is still more work to do :)
[Update2] I removed the graph. I will replace it in a few hours with a new graph in which I took care of a yearly rebalancing of the portfolio. As far as I now know it will rise the final cagrR from 4.2% to 4.4%. But I want to do some more testing before placing the graph.
[Update3] I replaced the graph. The new graph is taking care of rebalancing each end of the year. I replaced my reaction on Tyler's quote.
Last edited by johntaylor on Wed Jun 06, 2018 10:40 am, edited 16 times in total.
User avatar
Tyler
Executive Member
Executive Member
Posts: 1638
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: European PP, update 6 months after starting

Post by Tyler » Mon Jun 04, 2018 4:59 pm

johntaylor wrote:
Mon Jun 04, 2018 4:49 pm
I don't see why I should need to do something extra to make the scheme balanced. Every year end I just calculate the average of the individual growth of each of the 4 portfolio parts. With the resulting average growth I calculate the cagr for the total. Is this not the right method ?
That's pretty much how I do it, too. Sounds reasonable.
PortfolioCharts.com : a picture is worth a thousand calculations
User avatar
Tyler
Executive Member
Executive Member
Posts: 1638
Joined: Sat Nov 12, 2011 3:23 pm
Contact:

Re: European PP, update 6 months after starting

Post by Tyler » Wed Jun 06, 2018 9:26 am

By the way, if you want an example of a spreadsheet that calculates the real return from a collection of different assets (in addition to a bunch of different metrics) you might try downloading the Simba Spreadsheet from the Bogleheads forum. https://www.bogleheads.org/wiki/Simba%2 ... preadsheet
PortfolioCharts.com : a picture is worth a thousand calculations
johntaylor
Junior Member
Junior Member
Posts: 10
Joined: Fri Jun 01, 2018 8:33 pm

Re: European PP, update 6 months after starting

Post by johntaylor » Wed Jun 06, 2018 10:15 am

I don't see why I should need to do something extra to make the scheme balanced. Every year end I just calculate the average of the individual growth of each of the 4 portfolio parts. With the resulting average growth I calculate the cagr for the total. Is this not the right method ?
I removed this reaction, because I realised I indeed forgot to take care of rebalancing. I just replaced my scheme in my former post with a new one. The error I made in the first graph was that I calculated total growth by averaging the growths of the individual portions. In the new graph I calculate total return by averaging the returns of the individual portions and calculate total growth from total return.
Tyler wrote:
Wed Jun 06, 2018 9:26 am
By the way, if you want an example of a spreadsheet that calculates the real return from a collection of different assets (in addition to a bunch of different metrics) you might try downloading the Simba Spreadsheet from the Bogleheads forum. https://www.bogleheads.org/wiki/Simba%2 ... preadsheet
Thanks for pointing me to that. I already extensively studied the Simba spreadsheet and learned a lot from it and continue to do so.
johntaylor
Junior Member
Junior Member
Posts: 10
Joined: Fri Jun 01, 2018 8:33 pm

Re: European PP, update 6 months after starting

Post by johntaylor » Mon Jun 11, 2018 5:56 pm

I found a new subject that is causing the differences between the results of my own calculations and those of portfoliocharts and the simba spreadsheet. I am using (free) data of the morningstar data stream:
http://globalquote.morningstar.com/glob ... ode=DIRECT

From earlier experiments I already learned that the data stream from finance.yahoo.com is not always reliable, especially in the way it calculates adj_close values. The free data stream from morningstar does not offer adj_close values, but it separately offers the bare dividend values itself. So with this data it is quite easy to calculate the adj_close values. I used the method described on this website:
https://joshschertz.com/2016/08/27/Vect ... th-Python/
_
Image

If I calculate the adj_close for VTSMX I get results that are pretty close to the data that morningstar presents on it's own graph:
For the period 1992-12-31 / 2017-12-29 the return values of VTSMX are:
my calculation on morningstar data: cagrN: 9.4% cagrR:7.0% growth: 944%
simba spreadsh: cagrN: 9.6% cagrR:7.2% (total growth: 994%, calculated from Data_TR_USD)
growth on morningstar graph: 981%
_
Image
Image
_
But if I calculate the adj_close for VFITX I get very different results then the results morningstar presents on its own graph:
my calculation on morningstar data: cagrN: 4.6% cagrR: 2.3% total growth: 293%
simba spreadsh: cagrN: 5.5% cagrR: 3.2% (total growth: 372% calculated from Data_TR_USD)
growth on morningstar graph: 373%
_
Image
Image
_
For VFITX I compared in more detail my own calculations with the results on the morningstar chart. One of the periods in which the results very quickly drift away from each other is 1992-12-29 to 1993-01-22.
_
Image
_
Underneath is a print screen of the calculations I made. In the graph next to it you see how the results from my calculations drift away from the results which I copied from the morningstar chart itself.
_
Image
_
I have no clue why I get such different results. Am I making a calculation error ? Or can't I trust the morningstar data ?

If anyone have any hints or clues I would appreciate very much !
Last edited by johntaylor on Tue Jun 12, 2018 8:35 am, edited 2 times in total.
User avatar
frugal
Executive Member
Executive Member
Posts: 887
Joined: Sat Nov 10, 2012 12:49 pm

Re: European PP, update 6 months after starting

Post by frugal » Tue Jun 12, 2018 12:10 am

Hi,

Do you admit to change from EUPP to USPP in function of these true results?

Why is it so important to you to know this real cagr?

My best regards.
Post Reply