Message-ID: <2833966.1075842460177.JavaMail.evans@thyme>
Date: Sun, 13 Aug 2000 05:14:00 -0700 (PDT)
From: drew.fossum@enron.com
To: steven.harris@enron.com, mary.miller@enron.com, maria.pavlou@enron.com, 
	jeffery.fawcett@enron.com, lindy.donoho@enron.com, 
	lorraine.lindberg@enron.com
Subject: tw indeces
Cc: donald.vignaroli@enron.com, richard.hanagriff@enron.com, 
	susan.scott@enron.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Bcc: donald.vignaroli@enron.com, richard.hanagriff@enron.com, 
	susan.scott@enron.com
X-From: Drew Fossum
X-To: Steven Harris, Mary Kay Miller, Maria Pavlou, Jeffery Fawcett, Lindy Donoho, Lorraine Lindberg
X-cc: Donald Vignaroli, Richard Hanagriff, Susan Scott
X-bcc: 
X-Folder: \Drew_Fossum_Dec2000_June2001_1\Notes Folders\Discussion threads
X-Origin: FOSSUM-D
X-FileName: dfossum.nsf

Please take a look at Vig's effort to detect price arbitrage behavior on TW 
using the same method that Steve Kirk used on Northern.  You'll recall that 
Steve's chart on Northern showed an very close correlation between the 
calculated arb opportunity and the direction and size of the aggregate 
customer imbalances, by month.  Proving once again that TW is just a little 
different, that method shows no correlation at all on TW.  I agree with Vig's 
conclusion that we should not use this data in the 637 filing.  

I tried one alternative reading of the data to see if I could find a 
correlation.  Steve's method evenly weights the value of arb opportunities 
that occur early and late in a month.  Please tell me if you agree, marketing 
folks, but I think a more logical approach would risk-weight the apparent arb 
opportunites.  In other words, an apparent opportunity to short TW and sell 
the gas for a $.25/mmbtu profit on the 5th of a month is unlikely to generate 
arb behavior because a trader has no idea on the 5th what the cash out index 
will be at the end of the month.  In contrast, an opportunity to make $.25 on 
the 25th of a month, when a trader can calculate pretty closely what the 
monthly index will be, is basically free money.  Using that logic, I ignored 
any apparent arb opportunites in the first half of the month.   I eyeballed 
the timing and size of the arb opportunites (i.e., there's a $.70-.90 arb 
opportunity to short TW from the 23d to the end of May, 2000) rather than 
using Steve's mathematical approach.   I still didn't see the kind of 
correlation I'd expect.  The correlation is directionally positive in only 4 
out of 12 months, and the size of the imbalances doesn't even make sense in 
the months that correlate on direction.  I give up, and am removing the arb 
argument from the daily imbalance section of the draft.  I'll route that 
separately.  

Thanks to Vig, Susan, Lorraine and Richard for hustling on this analysis.  
"Its better to try [to find a correlation] and fail than never to try at all. 
. .."          
---------------------- Forwarded by Drew Fossum/ET&S/Enron on 08/13/2000 
11:55 AM ---------------------------
Donald Vignaroli   08/11/2000 04:23 PM

To: Susan Scott/ET&S/Enron@ENRON, Drew Fossum/ET&S/Enron@ENRON
cc:  

Subject: tw indeces


