Contour plots of Sharpe Ratio for various asset mixes

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
fundtalker123
Posts: 886
Joined: Tue Feb 27, 2007 4:18 am

Contour plots of Sharpe Ratio for various asset mixes

Post by fundtalker123 » Sun Jun 03, 2007 12:39 am

I calculated Sharpe Ratio for continuous mixes of stocks (US, International, REITS, Small Cap Value, etc) and bonds (5 year treasury bills) using the 1972-2006 data posted here: http://www.diehards.org/forum/viewtopic ... 9219#29219.

Sharpe ratio is used to estimate portfolio "efficiency" measured by reward to variability (risk) ratio: http://en.wikipedia.org/wiki/Sharpe_ratio
Higher is better (reddest area in each plot is "best")

US Stock vs. EAFE Stock vs. 5 Yr Treasury MIXES:

Image
US Stock - EAFE+EM Stock - 5 Yr Treasury MIXES:
Image
US Stock - US Small Value Stock - 5 Yr Treasury MIXES:
Image
US Stock - REIT - 5 Yr Treasury MIXES:
Image
US Stock - EM Stock - 5 Yr Treasury MIXES:
Image

Do these calculations look right (did I make a mistake)?
* 30-50% stocks (rest bonds) is "good"
* 30-70% EAFE stock tilt is "good"
* 70-100% Small Cap Value tilt is "good"
* 70-100% REIT tilt is "good"
* 60-100% EM tilt, plus ~70% bonds is "good"

Trev H
Posts: 1874
Joined: Fri Mar 02, 2007 10:47 pm

Morning...

Post by Trev H » Sun Jun 03, 2007 8:20 am

FundTalker...

I have played around with maxing out Sharpe considering using International Market (EAFE/EM) only on the International Side.

If you look at combinations of 500 Index and Intl Market the Sharpe maxes out at around...

60% 500 Index
40% Intl Market
515,464.10 = 10K Growth
11.92 = CAGR
17.10 = StDev
.4243 = Sharpe

If you mix (other asset classes) with Intl Market searching for the highest Sharpe Ratio... you have to go heavy on REIT, SV and a bit of LV.

05% LV
50% REIT
20% SV
25% Intl Market
1,017,482.94 = 10K Growth
14.12 = CAGR
15.34 = StDev
.5981 = Sharpe

10K Growth nearly doubled, with StDev near 3 points lower and Sharpe .17 higher.

I could not find a combination with > 25% Intl Market that pushed the Sharpe Higher considering LV REIT SV combinations on the US Equity Side.

I did not try mixing in InterTerm Treasury but if you do I doubt you will find a much more efficient mix than something similar to this from 1972-2006 using International Market for Intl Exposure.

Fun to look at...

Trev H

Trev H
Posts: 1874
Joined: Fri Mar 02, 2007 10:47 pm

With 5 Year Treasury Notes

Post by Trev H » Sun Jun 03, 2007 9:31 am

.
Keeping the ratios the same...

30% 500 Index
20% Intl Market
50% InterTerm Treasury
319,644.24
10.41 = CAGR
9.56 = StDev
.5005 = Sharpe

2.5% LV
25% REIT
10% SV
12.5% Intl Market
50% InterTerm Treasury
441,098.31
11.43 = CAGR
8.79 = StDev
.6539 = Sharpe


Trev H

User avatar
dave.d
Posts: 935
Joined: Mon Mar 19, 2007 10:30 pm
Location: Richmond, VA

Post by dave.d » Sun Jun 03, 2007 10:03 am

This is all great... so long as the next 35 years are a play-by-play replay of the last 35. 95% of equity in REIT, Intl, and SV? Only if the market is drastically inefficient. Color me dubious. This sort of game is a way to lose your shirt.

--Dave

Valuethinker
Posts: 37429
Joined: Fri May 11, 2007 11:07 am

Post by Valuethinker » Sun Jun 03, 2007 10:37 am

dave.d wrote:This is all great... so long as the next 35 years are a play-by-play replay of the last 35. 95% of equity in REIT, Intl, and SV? Only if the market is drastically inefficient. Color me dubious. This sort of game is a way to lose your shirt.

--Dave
It reminds me, perhaps incorrectly, of how if you plug historic data into a Markowitz optimizer you get an asset allocation which tilts heavily towards 'past favourites'.

Since asset classes have a tendency to mean revert in performance, that usually gives an *underperforming* portfolio in the future.

REITs in particular I would like to see how/where the outperformance has come. REITs were not a big asset class until the ?mid 90s? and they have done very well in the last 10 years.

I don't see where the scope for a further rerating by REITs against common stocks comes from.

User avatar
stratton
Posts: 11082
Joined: Sun Mar 04, 2007 5:05 pm
Location: Puget Sound

Post by stratton » Sun Jun 03, 2007 12:03 pm

I don't see where the scope for a further rerating by REITs against common stocks comes from.
David Swensen likes them for his suggested portfolio with 20% real estate in a 70/30 portfolio. However, I wouldn't put the entire REIT allocation in the VG REIT Index. I'd diversify into foreign real estate and something like the TIAA Real Esate Product if I had access to it. I'd seriously think about buying a couple of individual timber REITs too.

But, buying the typical REIT index for a huge ~30% of my asset allocation: no!

Paul

User avatar
Cb
Posts: 274
Joined: Wed Feb 21, 2007 12:39 am

Post by Cb » Sun Jun 03, 2007 12:04 pm

dave.d wrote:This is all great... so long as the next 35 years are a play-by-play replay of the last 35. 95% of equity in REIT, Intl, and SV? Only if the market is drastically inefficient. Color me dubious. This sort of game is a way to lose your shirt.

--Dave
Relax, Dave...I think these guys are just doing "what if's"...have any of them suggested that they'll revamp their portfolios tomorrow? It's a useful tool. It simply demonstrates the benefits of diversification. As Simba explains on the Excel document's README tab:

"Please use this spreadsheet only for learning purposes and as a need for diversification. Don't change your AA purely based on historical returns of a fund.
Past returns are not indicative of future returns."

User avatar
Topic Author
fundtalker123
Posts: 886
Joined: Tue Feb 27, 2007 4:18 am

Post by fundtalker123 » Sun Jun 03, 2007 12:05 pm

I calculated side by side plots of Sharpe and Return (CAGR), and indicated where I am with a black dot. I'm low on bonds (15%) because I'm greedy for high return, and low on international (10%) because I've never been comfortable with it. But the plot suggests lowering bonds and upping international would have lowered risk and increased or maintained high returns. BUT, in the past 35 years international did better than the US. If that reverts, maybe I want to stay where I am (maybe still increase bonds).

Image

Valuethinker
Posts: 37429
Joined: Fri May 11, 2007 11:07 am

Post by Valuethinker » Sun Jun 03, 2007 12:32 pm

stratton wrote:
I don't see where the scope for a further rerating by REITs against common stocks comes from.
David Swensen likes them for his suggested portfolio with 20% real estate in a 70/30 portfolio. However, I wouldn't put the entire REIT allocation in the VG REIT Index. I'd diversify into foreign real estate and something like the TIAA Real Esate Product if I had access to it. I'd seriously think about buying a couple of individual timber REITs too.

But, buying the typical REIT index for a huge ~30% of my asset allocation: no!

Paul
- international REITs are not, by and large, cheaper than American REITs.

There are exceptions (Japan perhaps). And the global upswing in real estate has been so universal, that I suspect historic correlation coefficients will be of less relevance

- timber

Too many people are talking timber, for my taste. And the 'you can't lose' argument that time means trees grow, and hence in value, is also pretty common.

I think the AIM market here is about to have a £250m timber vehicle listing.

Timber *is* interesting. The Chinese are driving demand through the roof. The pine bark beetle in BC has hopped the Rockies and is now killing trees in Alberta.

But travel with care.

User avatar
Topic Author
fundtalker123
Posts: 886
Joined: Tue Feb 27, 2007 4:18 am

Effect of Slice-Dice

Post by fundtalker123 » Sun Jun 03, 2007 3:23 pm

Here's the effect of using a Slice-Dice on the US Stock portion (20% each of LC, LCV, SC, SCV, REIT)

US Slice-Dice vs. International vs. Bonds (5 yr treasuries)

Image
Same, but with stronger tilt to emerging markets (50% EAFE, 50% EM)
Image

However, be sure to re-read Bogle's Little Book on Common Sense Investing . . .

WWV
Posts: 107
Joined: Fri Feb 23, 2007 8:58 am
Location: Northwest of Chicago

Just to say

Post by WWV » Mon Jun 04, 2007 8:03 am

Thanks for posting these plots. I agree with the cautions posted by others-but I do find these representations quite helpful in understanding the relationships.

Bob
"The average investor has only 11,000 more genes than a worm"-New York Times

jms969
Posts: 156
Joined: Tue Feb 20, 2007 5:56 pm
Location: South Central Wisconsin

Post by jms969 » Mon Jun 04, 2007 2:52 pm

This would be awsum added to Simba's spreadsheet!!!!!!!!!
I want to die peacefully in my sleep like my grandfather, not screaming like the passengers in his car.

User avatar
Random Musings
Posts: 5349
Joined: Thu Feb 22, 2007 4:24 pm
Location: Pennsylvania

Post by Random Musings » Mon Jun 04, 2007 3:06 pm

Be interesting to see the same contour plots EOY 1999 - to see what they looked like then.

RM

User avatar
stratton
Posts: 11082
Joined: Sun Mar 04, 2007 5:05 pm
Location: Puget Sound

Post by stratton » Mon Jun 04, 2007 3:18 pm

Valuethinker wrote: - international REITs are not, by and large, cheaper than American REITs.

There are exceptions (Japan perhaps). And the global upswing in real estate has been so universal, that I suspect historic correlation coefficients will be of less relevance
The one thing going here is most of the foreign real estate isn't in REIT forms. International RE as an investment has been switching to REITs over the last few years and hasn't actually become legally available in countries such as Germany yet. So there is a lot future activity that hasn't happened yet.
Valuethinker wrote:- timber

Too many people are talking timber, for my taste. And the 'you can't lose' argument that time means trees grow, and hence in value, is also pretty common.

I think the AIM market here is about to have a £250m timber vehicle listing.
This is a difficult area from what I can see. There is Plum Creek Timber and Rayonier then a couple of smaller timber REITs. Then there are a couple of companies like Weyerhauser that might turn into REIT. To get any real diversification you need private timber investments, Canroys, and some other foreign timber interests. Expensive and investment quality on some of the players isn't real great.

From what I can tell just buying a tiny bit of PLC and RYN will get most of the market cap in this area, most of the diversification, and the top investment quality. Timber is only around 2% of the REIT market caps so adding timber means over weighting like a SV tilt.

Paul

Post Reply