Forums  > Risk Management  > Systematic Sector Determination  
     
Page 1 of 1
Display using:  

HitmanH


Total Posts: 488
Joined: Apr 2005
 
Posted: 2017-06-26 21:51
Has anyone done any work on systematic / data driven methods to determine sector classifications, as opposed to relying on ICB/GICS/Bloomberg?

Sure i read a paper on this on SSRN, but having trouble finding it or anything like it

mtsm


Total Posts: 239
Joined: Dec 2010
 
Posted: 2017-06-26 22:29
Do you mean clustering?

HitmanH


Total Posts: 488
Joined: Apr 2005
 
Posted: 2017-06-27 00:05
I've seen the WQ/Kakushadze paper on it, using clustering, and that's a great start (and would love more / similar ideas) , but was also wondering about assessment at a systematic look at fundamentals too - if anyone has even done?

Tradenator


Total Posts: 1598
Joined: Sep 2006
 
Posted: 2017-06-27 10:01
I think the GICS methodology would be a good place to start. Pick earnings apart, and systematize it? From a fundamentals perspective, any company could have a curve-like GICS profile = {10: 50%, 15: 20%, 20: 1%, etc, etc, ..., 60: 0%} rather than one single number.

Hansi


Total Posts: 303
Joined: Mar 2010
 
Posted: 2017-06-27 10:51
We had a graduate student rotating with us doing some work on this using unsupervised learning to determine clustering using return data. Some interesting results but ultimately a bit flawed (may have been implementation rather than ideaology).

mtsm


Total Posts: 239
Joined: Dec 2010
 
Posted: 2017-06-27 16:44
I find it interesting that a lot of pretty prominent places seem to be deploying students and interns on pretty important topics and then scratch their head when nothing comes out of it.

In this case, this is a critical ingredient that goes into the holy mix. Putting an expert on the case can make all the difference in my experience.

Martinghoul


Total Posts: 866
Joined: Oct 2008
 
Posted: 2017-06-27 17:00
Experts? We live in an age when "expert" is a swear word!

Insofar as I may be heard by anything, which may or may not care what I say, I ask, if it matters, that you be forgiven for anything you may have done or failed to do which requires forgiveness...

Hansi


Total Posts: 303
Joined: Mar 2010
 
Posted: 2017-06-27 22:19
@mtsm: We are not in control of the grad program and they just rotate through if there is a good fit for their skill, they want to join the team and we accept mentoring them. The reason some of these kind of projects wind up with students or juniors is because the benefit is not clear up front so they get employed to do the initial exploratory checks for greenfield stuff with support from a senior researcher. The senior researcher will then generally have to do a lot of cleaning up and review to get concrete results and that's where the project is at currently but that same researcher also had multiple other projects with a more clear benefit that get prioritized over things like this.

Zoho


Total Posts: 28
Joined: Feb 2018
 
Posted: 2018-07-02 00:56
also attempted to do systematic/data-driven methods that help to determine sector classification... in my case good results occured when the clustering methods were constructed in accordance with the methods to discern "good/bad" companies within the sectors - something very similar to things mentioned in this topic
http://www.nuclearphynance.com/Show%20Post.aspx?PostIDKey=187134

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-08 13:53
deleted

"amicus Plato sed magis amica Veritas"

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-08 13:53
Deleted

"amicus Plato sed magis amica Veritas"

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-08 13:53
deleted

"amicus Plato sed magis amica Veritas"

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-08 13:53
Deleted (sorry I don't know what happened with these multiple copies of the same message)

"amicus Plato sed magis amica Veritas"

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-08 13:53
Pardon me if I don't know that much about it, but what is the reason to classify and cluster by sector if you already know what sectors each stock belongs to?

"amicus Plato sed magis amica Veritas"

nikol


Total Posts: 850
Joined: Jun 2005
 
Posted: 2019-05-08 14:25
@rickyvic

ENRON before collapse turned into pure M&A arb engine.
Few years back GE had financial profits larger than from sales of products (not sure about now).

etc etc.

Somewhere, in 2005/06, to validate margin model for my bank's Prime Brokerage I checked Bloomberg Classification of Hedge Funds "strategy" applying clustering metrics over monthly NAV returns. Of course, Bloomberg just mapped HF prospects.

I did it in two ways:
- constructing clusters of returns of HF's and then mapping it to BB-classes. (direct)
- taking BB-classes and then mapping to "most suitable" cluster. (reverse)

Basically, substitution of BB-classes with clusters did not change margins applied to specific FoHFs in our existing portfolio, so the conclusion about the use of BB-classes as information source was positive.

rickyvic


Total Posts: 188
Joined: Jul 2013
 
Posted: 2019-05-09 12:12
In fact that would be my guess, especially in stocks there might be some companies that have overlapping sectors but really how many.
The risk that instead you pick a statistical correlation that in fact is not meaningful is probably higher.
No study behind it just educated guess.

"amicus Plato sed magis amica Veritas"

nikol


Total Posts: 850
Joined: Jun 2005
 
Posted: 2019-05-09 12:38
> but really how many

Systematic trading embeds Systematic risk measurement and study.

To understand Risk one have to generate full (i.e. complete) list of risk factors and then one by one estimate if it can be neglected.

If you single out those "not many", then you have to figure out if it is an opportunity or it is an unnecessary risk.

Feel pedantic a bit, sorry :)

eeng


Total Posts: 27
Joined: Dec 2014
 
Posted: 2019-06-28 12:27
Read this some time ago:
https://www.winton.com/research/systematic-methods-for-classifying-equities

First they cluster the covariance matrix (nothing new) and then they mine the 10K Section 1 (business description) which sounded more interesting.

quantmatters.wordpress.com
Previous Thread :: Next Thread 
Page 1 of 1