EXHIBIT 99.384


Introduction:

This document expands upon the PX's initial response to the ISO's Congestion
Management Reform Recommendation. Whereas previous comments were limited to
concerns with the ISO's proposal, this is intended to provide solutions to those
concerns.

The document describes revisions to specific portions of the ISO's
recommendations. It is not intended to replace the ISO's entire proposal.
Instead, it describes how certain modifications to the ISO's recommendations
need to be implemented.

The complications involved in introducing changes to a complex, interdependent
model are understood. However, the reforms detailed below are specifically
designed to interact seamlessly with existing aspects of the ISO's proposal.

Much of the Comprehensive Market Reform design can and should be retained. The
decisions to build upon the existing California market structure and to uphold
the original fundamental design principles have significant value, as does the
proposed use of a consistent network model throughout the Real-Time and forward
markets. The introduction of a single commercial network model should eliminate
most of the inconsistencies that currently exist between the forward and spot
markets. The introduction of additional LPAs along with the enforcement of
nomograms in the forward markets should facilitate the reduction of intra-zonal
congestion in the forward markets thereby reducing Real-Time intra-zonal
congestion. However, the ISO should release the details of this model as soon as
possible and give stakeholders the opportunity to provide feedback.

Design consistency should extend beyond network model. When finalizing the
details of their recommendation, the ISO should make certain that policies are
uniform across markets and do not discriminate against any single class of
stakeholder.

Forward congestion will be better managed with the introduction of nomograms in
CONG (the ISO's congestion management software). The creation of new LPAs and
enforcing nomograms will allow the market to eliminate much of the current
intra-zonal congestion in the forward markets. The use of consistent nomograms
in the forward and Real-Time markets should further contribute to a reduction in
Real-Time congestion.

Utilizing recallable transmission will potentially further enhance congestion
management efficiency. We support the concept of a product that provides the
market access to non-firm transmission capacity that is reserved for Existing
Transmission Contract (ETC) holders. However, more details of the proposed
market are required to fully judge its impact.

Details of the proposed concepts are essential if the market is to accurately
evaluate the ISO's proposal. While an idea may seem desirable in concept, if it
is not properly implemented much of the potential value may be lost. When
defining market mechanics it is important that the ISO pursue the simplest
workable solution. Unnecessary levels of


                                                                     Page 1 of 1

complexity are significant barriers to market participation and thus, market
liquidity. The ISO should also seek to avoid market proliferation. Increasing
the number of markets into which the same capacity can be sold fragments the
markets and may adversely affect liquidity. Those aspects of the ISO's proposal
that have been identified as overly complex or fragmented are simplified in the
enclosed recommendations.

Comments on the Long Forward Market:

The Long Forward market design doesn't consistently conform to the ISO's
self-stated goals of enhancing market efficiency through decentralized decision
making. Instead, the market structure includes processes that abandon market
solutions - even in situations when the marketplace is best suited to
efficiently resolve reliability requirements.

One such process is Local Reliability Service (LRS) procurement. The ISO's
proposal relies on a two-day-ahead "auction" to procure generation within each
Local Reliability Area (LRA). Unlike other markets within California, loads
within LRAs do not submit bids to purchase this generation. Instead, the ISO
procures capacity on the behalf of loads based in its forecasts and requires
that portions of the acquired generating capacity be scheduled in Day-Ahead
based on nomograms. Capacity purchased by the ISO in the 2-Day-Ahead market, but
not identified as "Minimum Reliability Energy" (and thus required to be
scheduled in Day-Ahead) may be dispatched by the ISO as needed through
Real-Time.

Two distinct and separate products are being acquired by the ISO in this
process. Capacity required to be scheduled as Minimum Reliability Energy (MRE)
in the forward market is not capacity. It is instead, a firm energy purchase and
should be treated as such. Committed capacity that is reserved by the ISO and is
able to be dispatched through Real-Time would be better obtained through
existing processes such as the Day-Ahead and Hour-Ahead Ancillary Services
markets

Because the decision to schedule MRE is a forward energy allocation decision,
the ISO should let SCs complete this task if it is to remain consistent with the
goals of maintaining the decentralized decision making process. Thus, SCs should
be responsible for scheduling load and generation within LRAs according to
current Day-Ahead and Day-Of timelines, removing the need for a 2-Day-Ahead
auction.

All nomogram constraints will be placed in CONG, which will validate whether
enough generation within each LRA has been scheduled to meet the daily load
requirements. If insufficient generation has been scheduled CONG will either
increment the necessary generators based on their adjustment bids, or it will
decrement those loads that exceed the nomogram limitations.

CONG will ensure that scheduled generation and load satisfy each nomogram on a
market-wide basis rather than by SC. CONG would determine "usage charges" for
flows across nomogram constraints. SCs would pay usage charges for bringing
power across


                                                                     Page 2 of 2

nomograms. Conversely, SCs whose schedules allow other SCs to import more across
nomograms would be paid the usage charges for their counter-flow. This is
necessary to conform to nomogram requirements efficiently.

Qualified market operators could facilitate the forward procurement of MRE and
CC by running markets to help participants develop schedules that meet nomogram
requirements and avoid the associated usage charges. Buyers and sellers would be
able to trade MRE and CC using many different contract types and would
self-provide the resulting positions in the Day-Ahead market.

Market efficiency would be further enhanced if nomogram rights were traded like
transmission rights. If this were the case, loads could purchase the rights to
hedge against the Locational Price Areas (LPAs) and avoid the cost of expensive
in-area generation. Nomogram rights, like FTRs would be both physical and
financial.

If this structure were introduced, loads within LRAs would have sufficient
incentives to acquire nomogram rights and to schedule in the forward markets in
order to hedge against ISO charges. The ISO would also remain consistent with
its stated goals.

If after running Day-Ahead and / or Day-Of congestion management, the ISO felt
that an insufficient amount of in-area generation had been scheduled (or if the
ISO expected additional load within a LRA to appear in Real-Time), it would
acquire what it refers to as committed capacity. This process would be
facilitated by the introduction of committed capacity (CC) as an additional
Ancillary Service.

The Committed Capacity ancillary service will be treated in the same fashion as
existing AS types with the exception that CC will be purchased on a zonal / LPA
basis rather than a market-wide basis. SCs will be allowed to self-provide the
service according to the same rules that govern self-provision today.
Additionally, SCs who have available capacity will be able to submit bids to
supply CC to the ISO, who would procure it along with its other Ancillary
Services. As with minimum reliability energy, SCs will have the incentive to
self-provide this service in order to hedge against ISO charges.

Because some level of market power will exist within LRAs until new generation
or transmission is built, market power issues must be addressed. Market power
issues are the same regardless if the ISO adopts the PX's recommendations or
not: a price must be determined that protects loads from undue market power,
provides sufficient revenues to generators and provides locational price signals
in accordance with FERC requirements. The PX feels that the ISO and its
stakeholders are capable of developing a workable solution to this issue and
thus does offer a specific price cap proposal at this time.

While this approach differs significantly from the one proposed by the ISO, it
contains enough benefits to warrant adoption. This model's primary advantage is
its greater consistency with the original California design principles. In it,
the ISO's primary roles are information supplier (to facilitate the efficient
scheduling of load and generation) and supplier of last resort (in situations
where insufficient generation was scheduled or


                                                                     Page 3 of 3

additional load was expected to appear in Real-Time). Physical scheduling is
accomplished through decentralized decision making.

In the ISO's approach, however, the ISO is responsible for both forecasting and
obtaining MRE and CC. Although the right for SCs to self-provide these services
has been added to their proposal, the ISO's timelines make it more likely that
the majority of scheduling decisions will be made by the ISO. This is because
their proposal requires SCs to finalize LRA schedules one day in advance of the
Day-Ahead market. In addition, it is not clear that the ISO's proposed cost
allocation mechanism would provide sufficient incentives for self-provision.

The PX's approach relies on market forces to efficiently allocate LRA resources
within existing timelines. This ensures that procurement of MRE and CC does not
interfere with or distort other Day-Ahead scheduling decisions. When the amount
of energy and capacity that is likely to be allocated through this process is
considered, the importance of adopting an approach that compliments, rather than
supplants existing market structures is made clear. ISO statements that the
procurement of MRE and CC could even extend to competitive LPAs in certain
situations only underscores the importance of adopting the PX's approach.

Comments on FTRs:

The ISO is to be commended for responding to stakeholder input and limiting the
term of FTRs to one year. This will provide required price certainty as the
market becomes familiar with the implications of new zonal boundaries.
Additionally, annual FTRs are more simple to link to zone creation /
redefinition policies. As boundaries solidify, the sale of longer-term rights
should again be examined, however.

Some modifications to the ISO's proposed transmission rights policies are
required though. The ISO should release all ATC as FTRs in their annual primary
auction. The specific percentage should depend on the amount of rights that can
be released with a predetermined annual level of certainty. If this is done,
secondary markets will develop to reallocate transmission rights. The ISO can
compliment this process by releasing shorter-term seasonal FTRs, as they become
available.

Shorter-term FTRs may help the market adjust for seasonal variances in
transmission capacity. However, questions must be answered before the nature and
amount of shorter-term FTRs to be released is determined. How significant a
barrier are shorter-term FTRs to the development of liquid secondary markets?
How often is transmission capacity significantly affected by weather or other
considerations (are monthly auctions required or are longer terms such as
quarterly better)? How much of a change in transmission capacity is needed and
how long must the new levels be sustained before the ISO determines that the
capacity in question can be auctioned as an FTR? These questions will be able to
be answered only if the ISO releases its criteria for determining Available
Transmission Capacity.


                                                                     Page 4 of 4

Of additional concern is the ISO's proposed method of allocating costs to FTR
holders. The ISO has stated that in the event of topology or operating transfer
limit changes they intend to "keep market participant's financially whole by
creating a balancing account of net gains and losses." Balancing accounts are to
be cleared by each applicable PTO on a monthly basis. The details of this
proposed arrangement need to be provided, but an initial examination of the
information reveals deep concerns.

If the ISO publishes a "library" of shift factors as is stated in their
proposal, and the factors are published in advance of the primary auction, SCs
should be capable of planning for and reacting to system changes. For instance,
SCs could purchase enough FTRs in the primary auctions to cover the range of
potential shift factor changes. Those who were unwilling or unable to purchase
rights in the primary auction could bid to purchase required FTRs in the
secondary markets that will be provided by such qualified exchanges as the PX.

If SCs are given information about shift factor changes in time to adjust their
schedules and transmission rights, they are capable of making educated business
decisions and dealing with the resulting financial implications. Having a
central balancing account where profits are used to offset others losses is not
only unnecessary, it is contrary to the most fundamental free-market concepts.

Instead of pursing this course, the ISO should treat financial rights in the
same manner as they propose to resolve scheduling priorities: SCs should be
responsible for procuring additional FTRs when the ISO implements new sets of
shift factors. Those who choose not to hedge by purchasing additional FTRs on
the secondary market would expose portions of their schedules to both financial
and physical delivery risks.

Comments on the Forward Market:

We support the efforts of the ISO to preserve the original design concepts of
the deregulation process, which we feel are well represented in the proposed
enhancements to the forward market. However, certain changes are required to
further improve the recommended modifications.

A change that is supported is the introduction of Inter-SC adjustment bids.
While the implementation of this functionality is outside the scope of the
congestion reform project, its effects are not. The ISO's market separation
study clearly shows that the existing market structure is an efficient one.
However, Inter-SC trade adjustment bids can only enhance efficiency for those
SCs who choose to use them. When the efficiency of the existing market structure
is combined with the fact that decentralized decision making remains a
cornerstone of California's proposed deregulation model, arguments against
additional relaxation of market separation become overwhelming.



                                                                     Page 5 of 5

There is a proposed enhancement that would threaten market efficiency, however.
Implementation of the ISO's proposed congestion activity rule would create as
many, if not more, issues than it would resolve. While it is understood that
this rule was designed to provide the ISO with the opportunity to always select
the lowest cost scheduling iteration, it is not the appropriate way to insure
against potential market behavior. Rather than reducing the potential for
damaging market behavior, the congestion activity rule would simply increase the
likelihood that both iterations would become distorted. The level of uncertainty
that the implementation of the activity would introduce increases this tendency.
Stakeholders will be less willing to enter into new, binding transactions if
there is the danger that the ISO will force them to be cancelled with any degree
of frequency. If the ISO is concerned about limiting the potential gaming
affects of having two congestion management iterations, it should either adopt
different activity rules or consider eliminating the second iteration entirely.

Comments on Real-Time:

Based on our current understanding, the PX supports the ISO's proposed Real-Time
changes. However, more information is required to fully evaluate them. For
instance, the anticipated dispatch model hasn't been adequately defined and
should be released to the marketplace for examination. Additionally, since the
ISO is proposing to move to a transmission constrained economic dispatch
optimization model to dispatch Real-Time energy, references to Real-Time bid
stacks should be removed from the draft proposal to eliminate confusion. The PX
support of a Real-Time optimization model is predicated on the understanding
that only Real-Time bids will be optimized.

If the optimization model is used to re-dispatch previously scheduled energy
that doesn't have explicit bids (other than in system or local emergencies) many
of the fundamental principles of California's deregulation process will be
undermined.

Conclusion:

The PX commends the ISO for their efforts to develop a comprehensive congestion
management solution. However, we urge the ISO to incorporate these modifications
into their proposal. The integration of these concepts with the ISO's existing
proposal will ensure that the maximum benefits of any reform are enjoyed by all
Californians. Failure to do so, however, will significantly blunt the benefits
of any changes.



                                                                     Page 6 of 6