|
|
Enron Mail |
Kim,
I wanted to forward you this email that I sent to Vince. He had come back to Richard and I to resolve an issue with the dropping VaR associated with TW. Vince was asking if this was related to the processes that we have automated. In speaking with Vernon, the date of the ERMS problem with the San Juan basis discussed below happened to coincide with the date or near proximity of the unwinding the Wholesale positions that TW had held. Consequently, there were substantial changes with the VaR increasing during this period. When the rolling price data dropped off, it rolls for 30 days, the VaR decreased substantially. If you have any questions please give me a call. Sean 713-287-1014 -----Forwarded by Sean Bolks/HOU/Dynegy on 02/12/2002 05:01PM ----- To: vincent.strohmeyer@enron.com From: Sean Bolks/HOU/Dynegy Date: 02/12/2002 04:58PM cc: Richard Riehm/HOU/Dynegy Subject: Issue with San Juan Historic Volatility Vince, I left you a message concerning the issue with the VaR. The problem is associated with the ERMS Price that came on December 19. The basis was .23 instead of -.23. This is represented in SJ Price Quote File below. Consequently the inclusion of this date in the rolling 30 day bucket for the Historical Volatility calculation essentially changed the volatility for the entire Winter 2002-2003 strip. Any I-I deal involving the San Juan was thus flawed and shown in the VaR. As this did occur on December 19th, this should have been caught in the December time period for any VaR report done on a daily basis since that time. For example, the San Juan Historical Volatility was .395 on December 17, .394 on December 18, .521 on December 19 and .657 on December 20. The VaR went up in the same way it came down-- on Feb 4 it was .66, on Feb 5 it was .52, on Feb 6 it was .357 and on Feb 7 it was .343. This is shown in the attached file San Juan Hist Vol. As Richard and I do the base calculations, which are correct, we would never catch this mistake. We are somewhat surprised that when this volatility was included on December 19th the VaR report did not show a substantial increase from the previous day. In the same way the VaR dropped by half, we would have expected it to double over a period of a couple of days in that week of December. All in all, this issue took 10 hours for Richard and I to resolve. As it is essentially out of the perview of setting up the SAS automation process--the process did run correctly, it was a little frustrating for us to devote the time to identifying this particular issue. It would also appear to be a troubling issue from a financial perspective given the current Enron environment. I hope this resolves the problem. Sean
|