Message-ID: <633400.1075857580625.JavaMail.evans@thyme> Date: Wed, 7 Jun 2000 01:43:00 -0700 (PDT) From: john.arnold@enron.com To: jeffrey.shankman@enron.com Subject: Re: using new FF vols Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-From: John Arnold X-To: Jeffrey A Shankman X-cc: X-bcc: X-Folder: \John_Arnold_Dec2000\Notes Folders\Discussion threads X-Origin: Arnold-J X-FileName: Jarnold.nsf ---------------------- Forwarded by John Arnold/HOU/ECT on 06/07/2000 08:43 AM --------------------------- Tanya Tamarchenko 06/07/2000 08:33 AM To: John Arnold/HOU/ECT@ECT cc: Grant Masson/HOU/ECT@ECT Subject: Re: using new FF vols Hi, John, following up the discussion with you on Friday we talked with Risk Control people who are not excited to use that FF vol curve you sent to me. Also in order to use your curves we would have to have them for all the locations. The suggested alternative solution was to calculate the Forward Forward vol curves from historical data. I implemented this solution based on 18 last business days forward price curves for NG and all basis locations. I used exponential weights with 0.97 decay factor. I enclose these curves in the spreadsheet below. And here are the VAR numbers based on these curves: 5/30/00 5/31/00 AGG-STORAGE (production) 3,027,000 4,516,000 AGG-STORAGE (test, 0.97) 2,858,543 3,011,761 AGG-GAS (production) 36,627,200 40,725,685 AGG-GAS (test, 0.97) 29,439,969 31,207,225 You see that the numbers are stable, lower than the official numbers. I suggest that we use 0.94 decay factor as recommended by Risk Metrics which would give more weight to recent data. We need to test this approach for a period of time and also to collect backtesting data for an educated choice of decay factor. Tanya. John Arnold 06/07/2000 07:40 AM To: Tanya Tamarchenko/HOU/ECT@ECT cc: Subject: Tanya: On Friday I emailed a new vol curve to use for VAR testing. I was under the impression that you could apply this vol curve to the price book and storage book and have a new experimental VAR number by Monday. I have not received any response. Please reply with status of this project. John