Just a heads-up - a long-winded email, but please see highlighted paragraphs. 
I find it hard to believe the scenario portrayed here, but stranger things 
have happened - will keep you posted.
---------------------- Forwarded by Fernley Dyson/LON/ECT on 29/11/2000 08:44 
---------------------------


James New
28/11/2000 13:42
To: Ted Murphy/HOU/ECT@ECT, David Port/Market Risk/Corp/Enron@ENRON
cc: Mike Jordan/LON/ECT@ECT, Coralie Evans/LON/ECT@ECT, Fernley 
Dyson/LON/ECT@ECT 

Subject: Status update on the issues surrounding Continental Power 
volatilities and correlations

As you know for some time we have been trying to get the Continental Power 
traders to increase the volatilities they use for valuing their option 
portfolio and for use in calculating their VAR as it has been obvious that 
the market  has become more volatile but this was not being reflected in 
their mark.

In looking at the impacts of this, various other issues came to light. 
Firstly it raised issues over the correlation matrix used in the VAR model as 
this was way out of date. We have worked with Jitendra and have now put in a 
matrix that both the traders, risk and RAC are happy with. Secondly, as you 
may be aware, in Portcalc swaptions currently use the volatility input on the 
deal ticket and not off any volatility curve. 

The current portfolio is 90% swaptions and so this represents a serious issue 
for us in that it is not practical to manually change all the volatilities at 
the deal ticket level every time the implied voaltilities change. We cannot 
use the option functionality as this only values hourly or daily exercise 
options and not fixed expiry swaptions or options. We also have not found any 
way in Portcalc of valuing Asian options so we are having to 'force' in the 
few we have. We have had the Houston based Portcalc IT team work on a 'new' 
piece of swaption valuation code but currently it contains a number of 
worrying bugs. These are primarily that the proposed code does not take the 
correct underlying forward price and uses the volatility of (approximating) 
the last day of the underlying rather than the implied volatility of the  
underlying. 

They are currently being hampered in that they do not have any IT personnel 
who have sufficient option knowledge (I understand that the two most 
experienced coders resigned to go to another Houston based energy company). 
This is extremely worrying and I find it hard to believe that the whole of 
Enron's power business is having to use the methodology of inputting the 
volatility at the deal ticket level. This will almost certainly mean that 
different volatilities are being used in the VAR as are being used to value 
the swaption portfolio globally. As well as these issues we have concerns 
over the use and accuracy of the smile adjustment in Portcalc (worryingly it 
does not seem that you can 'switch' this off). Briefly it seems that Portcalc 
calculates the delta used to adjust the volatility by comparing the strike 
price to the forward price of the last day of the underlying rather than the 
forward price of the underlying itself. 

We have engaged London based  IT and are looking at the code we have used in 
Power 99 to se if there is a way we can get something done quickly to start 
to value all option products off a curve and to be able to book Asian options.

Once we get the above resolved we still have no way or extracting the vega or 
gamma risk from Portcalc in it's current state as the information is just no 
there (so we are told by IT). The system 'seems' (we cannot be 100% sure 
given all the other errors) to produce a vega and gamma P&L number but what 
the risk is remains a mystery. Again this is very worrying in that we have 
Power option traders all over the world who can't get their underlying vega 
or gamma risk position from the global valuation system. We are trying to 
work on something manual to draw a line in the sand but I would really like 
to know how the American based traders cope as they have this problem 
(knowingly or not)  for years. What sign offs were gained for the 
implementation of Enpower because it seems to us that using the current 
system it is not possible to comply with the Global Risk policy. How do we 
validate the inputs to the valuation if we do not know the risk ?

Despite the above we have now persuaded the traders to save out newer higher 
volatilities a week ago and you should have seen a large move up in their 
VAR. However, as it stands at the moment they are notable to save out a daily 
remark as Jitendra is saying that he is not happy with their 'model' / vol 
curve generator (I have today asked for a full explanation, details of any 
testing etc etc and time lines so we can move to getting over this hurdle). 
The swaptions have also not yet had the ticket volatility changed but the 
impacts are expected to be positive as we are long volatility.

I would appreciate your comments on the above as it seems that we are pushing 
the boundaries here on areas which we really expect to have already been 
covered in Houston some time ago and could do with some help. We are having 
to use a system which seems to us to have numerous bugs and short comings and 
are having to spend an enormous amount of IT time overcoming the inadequacies 
of Enpower wich I do not think most people are aware of. Consequently we have 
not been able to deliver all the improvements we were required to do in 
gaining the increased limits so 3 months ago but I would hope you agree that 
it is not for a lack of trying.

Sorry it's so long but if there are any questions or any perceived 
inaccuracies in any of the above then please let me know.

James