MarketThoughts.com Home Page
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups  StatisticsStatistics   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

the portfolio optimisation process

 
Post new topic   Reply to topic    MarketThoughts.com Forum Index -> Research Papers, Articles, and Speeches
View previous topic :: View next topic  
Author the portfolio optimisation process
Real
Newbie
Newbie


Joined: 24 Nov 2009
Posts: 1

PostPosted: Tue Nov 24, 2009 3:57 pm    Post subject: the portfolio optimisation process Reply with quote

Hi guys, what is the frame (3,4,5 etc years) of historical date should be used to estimate the inputs to the portfolio optimisation process and why?
Thanks
Back to top
View user's profile Send private message
Post new topic   Reply to topic    MarketThoughts.com Forum Index -> Research Papers, Articles, and Speeches
Author the portfolio optimisation process Replies
nodoodahs
Moderator
Moderator


Joined: 06 May 2005
Posts: 2408

PostPosted: Tue Nov 24, 2009 9:13 pm    Post subject: Reply with quote

So far as what is "optimal," I have no idea.

You probably need to provide more detail to get the answer you're after, because I don't know what YOU are trying to do, in terms of estimating inputs for the portfolio optimization process.

* do you have strategy backtest data that you're trying to incorporate?
* or are the components all individually-traded securities?
* are you making any drastic simplifying assumptions, such as normality or independence?
* etc.

The more detail (without getting anything you don't want disclosed, disclosed), the better to answer your question.

I am GUESSING here at what you're after ... the type of portfolio optimization used by firms like BARRA in their multi-factor models, for risk analysis, portfolio optimization, and historical performance attribution.

Assuming that's what you're after ... as of 1998, when Barra calculated their covariance matrix, they used an exponential weighting based on a "half-life," wherein a period one "half-life" ago gets 1/2 the weight of the current period, and they used a half-life of 48 months, based on their internal testing.

I've found from looking at S&P 500 returns that the descriptive statistics (average, standard deviation, skewness, kurtosis) vary from year to year and decade to decade. I've also found that correlations change, often abruptly, over time. Therefore I'm skeptical about the illusion of exactness given by the process. That said, I think the simplifying assumptions sometimes used (normal and independent returns) are just as dangerous as a false illusion of exactness - well, maybe LESS dangerous, if the person using them is aware of the flaws (and the person using the complicated method isn't).
_________________
I haven’t seen a beatin’ like that since somebody stuck a banana in my pants and turned a monkey loose.
Back to top
View user's profile Send private message
nodoodahs
Moderator
Moderator


Joined: 06 May 2005
Posts: 2408

PostPosted: Tue Nov 24, 2009 8:30 pm    Post subject: Reply with quote

Hi Real, welcome to the board!

Note that we hold first posts in moderation to reduce spam; I'll delete the duplicate posts. I thought this subforum was the proper one for the question.
_________________
I haven’t seen a beatin’ like that since somebody stuck a banana in my pants and turned a monkey loose.
Back to top
View user's profile Send private message

Please log in to view without the ad banners
Display posts from previous:   
Post new topic   Reply to topic    MarketThoughts.com Forum Index -> Research Papers, Articles, and Speeches All times are GMT - 6 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB