Vince:

Have you or a member of your group had a chance to look over the forwarded e-mail yet?  We have a meeting on Thursday with another potential client who plans to use his own capital.  His capital budgeting issues are similar to those of the client who prompted my original e-mail.

The more DSM projects that we can help the client get approved, the more DSM savings we can share with the client.  These DSM projects can turn a retail commodity contract with low profitability into a substantially more profitable contract for us.

Would one of your GARCH analyses provide us with a quantitative solution?  One of your experts in differential equations might have devised something.  We'd really like to have something quantitative if you think it's at all possible.

Thanks,

Jeff

---------------------- Forwarded by Jeff Gray/HOU/EES on 04/17/2001 12:16 PM ---------------------------


Jeff Gray
04/09/2001 04:41 PM
To:	Vince J Kaminski/HOU/ECT@ECT
cc:	Michael Moore/HOU/EES@EES, Jay Sparling/HOU/EES@EES 
Subject:	DSM projects as real options   

Vince:

Could you spare a few moments to look over the attached three-page Power Point presentation?  At a minimum, we're trying to demonstrate the following:  In high-volatility electricity environments, for those EES customers who employ their own capital, their usual capital-project IRR hurdles may be lowered slightly in the special case of DSM projects, to allow for the extra value derived from reduced risk exposure stemming from reduced power consumption/dependence.

I'm pushing the boundaries of real-option theory (and good sense) and viewing a DSM project as a series of options, the values of which are additive.  First is the option to engage in the DSM project itself; which, if we exercise the option immediately, can be valued intrinsically through a simple NPV analysis.  Second is a strip of put options to generate "negawatts," extending over the useful life of a DSM installation, the cumulative value of which is an extrinsic value associated with the DSM project.  These options are exercised sequentially--over the life of the DSM installation--whenever the customer experiences a dramatic, stochastic spike in electricity price, associated with a lognormal price distribution.  I'm calling the value of this strip a "residual time value" associated with the original DSM option.  If we can add this residual time value to the NPV and derive a total value that is quantitatively higher than a simple NPV alone,  we may be able to help the customer get more projects approved, even at the original high IRR hurdle.

Alternatively, and more feasibly, we'd like to give these same customers a "qualitative" tool with which they can convince their finance gatekeepers to lower--for DSM projects--their standard capital-project IRR hurdle .  The attached presentation is a first attempt at this qualitative argument.

We intend to use this only as a sales and marketing tool, to allow at least one of our customers who has an onerously high hurdle rate to manage around his company's internal capital-budgeting requirements.  He has set aside a large sum of money for DSM projects, but will only be able to spend a small portion of it under his company's current capital budgeting methodology, which does not take into account forward commodity price volatility.

In your opinion, is there any hope in devising a quantitative justification, or will we have to stick with the qualitative argument as described in the attached presentation?  In other words, is it possible to quantify "residual time value" as I've defined it above? Or, even better, are you aware of a more practical way of conceptualizing this problem?

Thanks,

Jeff Gray

 

 



Vince J Kaminski@ECT
01/05/2001 03:26 PM
To:	Jeff M Gray/NA/Enron@ENRON
cc:	Vince J Kaminski/HOU/ECT@ECT, Stinson Gibner/HOU/ECT@ECT, Alex Huang/Corp/Enron@ENRON, Gary Hickerson/HOU/ECT@ECT, Michelle D Cisneros/HOU/ECT@ECT 
Subject:	Power Plant Model

Jeff,

A few comments on the model:

1. We have a few reservations about some features of the model but would like to
discuss it internally and make the improvements without giving the benefit of our insights to the consultant.
In general, the model is not unreasonable but the devil is always in the details and in the inputs and
calibration. The same model may produce drastically different results depending
on the quality of inputs. 

2. We don't have a separate pool of programmers in the Research  Group. We were told that you
would provide an IT resource. Alex would supervise this person.


Vince






<Embedded StdOleLink>