Message-ID: <9513677.1075856649506.JavaMail.evans@thyme>
Date: Wed, 6 Dec 2000 02:59:00 -0800 (PST)
From: tanya.tamarchenko@enron.com
To: rudi.zipter@enron.com, stephen.stock@enron.com
Subject: Re: UK VaR calibration
Cc: david.port@enron.com, wenyao.jia@enron.com, vince.kaminski@enron.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Bcc: david.port@enron.com, wenyao.jia@enron.com, vince.kaminski@enron.com
X-From: Tanya Tamarchenko
X-To: Rudi Zipter, Stephen Stock
X-cc: David Port, Wenyao Jia, Vince J Kaminski
X-bcc: 
X-Folder: \Vincent_Kaminski_Jun2001_6\Notes Folders\All documents
X-Origin: Kaminski-V
X-FileName: vkamins.nsf

Rudi and Steve,
factor loading analysis was performed at the end of September for 17 
following UK curves:

PPPWD1-PPPWD6, PPPWE1-PPPWE6,  PE, PEO, NBP, SYSTEMDAY and SYSTEMNIGHT.

This analysis was based on historical forward prices information from 
RisktRac.
I sent the results to Vlady and Naveen for review on 10/4/00.

The results look decent for 3 curves only: NBP, SYSTEMDAY and SYSTEMNIGHT.
"Decent" looks as follows:


"Decent" means: the factors are rather smooth, the first one makes the 
largest contribution, the second one changes sign once.

The factors for the rest curves do not look good, for example for PE:



This messy picture means that the prices for different maturity contracts for 
PE are not highly correlated and
factor loadings analysis is meaningless in this case.

My conclusions: 
1.We should use NBP, SYSTEMDAY, SYSTEMNIGHT factors and map all other UK 
curves into one of those.
2. Winston's code is ready.
3. The mappings have to be identified by Risk Control and communicated to 
Winston.

Thank you,

Tanya..
