Message-ID: <20881330.1075856239095.JavaMail.evans@thyme>
Date: Wed, 3 Jan 2001 01:46:00 -0800 (PST)
From: vince.kaminski@enron.com
To: amitava.dhar@enron.com
Subject: Re: CreditDotCom Projects
Cc: vasant.shanbhogue@enron.com, vince.kaminski@enron.com, ben.parsons@enron.com, 
	tanya.tamarchenko@enron.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Bcc: vasant.shanbhogue@enron.com, vince.kaminski@enron.com, ben.parsons@enron.com, 
	tanya.tamarchenko@enron.com
X-From: Vince J Kaminski
X-To: Amitava Dhar
X-cc: Vasant Shanbhogue, Vince J Kaminski, Ben Parsons, Tanya Tamarchenko
X-bcc: 
X-Folder: \Vincent_Kaminski_Jun2001_1\Notes Folders\All documents
X-Origin: Kaminski-V
X-FileName: vkamins.nsf

Amitava,

Can you schedule a brainstorming session with Vasant, Tanya, Rakesh, myself
and Ben (if possible)? 

What about the trip on Monday. Has it been scheduled?


Vince





Amitava Dhar@ENRON
01/03/2001 08:30 AM
To: Vince J Kaminski/HOU/ECT@ECT, Ben Parsons/LON/ECT@ECT
cc: Vasant Shanbhogue/HOU/ECT@ECT 
Subject: CreditDotCom Projects

The following is a good summary from Vasant about the projects and what is 
ahead of us.
Ben, please let me know when things are ready; I can plan my trip accordingly.
Thanks, 
Amitava



EnronCredit.com is working on 3 main models --- the FMC model, the Placement 
model, and the Movement model.
FMC Model : provides credit default swap price curves for 16 ratings with 33 
industry offsets.  Each day, this model takes the previous day's trader 
curves and adjusts them for the movement in Bloomberg yields and observed 
swap prices.
Placement Model : uses linear regression to credit default swap prices using 
dummified ranges for Market Capitalization, Liquidity, KMV Score, etc.  One 
main issue here is that the dataset is biased towards investment grade 
names.  Alternate approaches looked at developing separate models for 
separate sunsets of names.
Movement Model : provides an alert system for news.  This is under 
development.

Current involvement of Houston Research Group members is in the 
development/modifications of the Placement Model, especially in thinking 
through the various single vs multi model approaches.  We can help in the 
brainstorming for better models for subsets.  One main project right now is 
to get a separate model for subsidiaries, and to exclude subsidiary names 
from the main regression.

Value-at-risk is based on historical variance-covariance approach.

The portfolio model is under development.

Agenda for Houston Research Group members : 

Once sufficient data is collected, start reviewing for potential analytic 
relationships.
Get weekly updates from London, and provide feedback.
Plan on a trip if the work warrants it, once the agenda is fairly well drawn 
out, and data is available for analysis.


