
Date: Thu, 24 May 2001 06:54:00 -0700 (PDT)
From: paul.dawson@enron.com
To: richard.shapiro@enron.com
Subject: FW: Frontier Economics' Whitepaper
Cc: 
Bcc: 

Rick,

Some comments from Nick are attached below.

Nick's first comment reflects a general concern I had about the paper which
was it's treatment of imbalance settlement.  I would see this as one of the
essential elements of market design.  I've always felt that the
physical/financial distinction was slightly artificial because even with
"physical" contracts, power systems must have a means for cash settlement of
imbalances (in the absence of universal controls which limit actual takes to
contract volumes).  However, the paper does not really pick this up and in
section 3.1 implies that cash settlement might only be "feasible" in some
market structures.  I would have added imbalance settlement to section 2.1
which describes the fundamentals required for any market.  It would also help
to draw a brighter distinction between spot markets and imbalance settlement.

At the same time I would have dropped the mention of "binding market rules"
from 2.1.  I'd see this as a subset of the need for an ISO to take over the
operation of the system in real time rather than as a separate fundamental
principle.  It's also not made clear in what sense these commitments are
binding.  Are they binding in a financial sense (in which case they are
covered by imbalance settlement) or somehow physically (either this is
unfeasible since units trip etc or it requires some form of balancing penalty
which is ruled out later in the paper).

I felt that the paper also compounded issues such as LMP's/transmission
rights with other non-locational elements of market design, this makes it
difficult to follow in places.  It would be clearer to have the locational
elements addressed separately as an adjunct to the basic principles.

Centralised unit commitment is put up as an element of the model proposed.
As Nick hints at below, we wouldn't see this as an essential element of
market design.  Under NETA, individual participants commit their own units
with the SO only taking over full control at 3.5 hours out.  Even with
centralised commitment, I'm not convinced that simple bidding would be
superior to multi-part bids.  Much here depends on timescales, eg, a
day-ahead market with simple bids is more exposed to allocative inefficienc
than a continuous hourly market with simple bids.  Simple bidding was often
proposed for the Pool, but I think a single daily big would have increased
prices if two-shifting generators had not been able to reflect their start-up
and no-load costs separately.  (These elements were also gamed, although this
game eroded as competition increased.)

Specific UK sensitivities which you may want to finesse in the paper:

? we've consistently opposed locational pricing because of our Teesside
positions.  Publicly we've argued that the proposals do not provide good
economic signals over sufficiently long time-periods to enhance allocative
efficiency and that the transactions costs outweigh any possible efficiency
gains (constraints only cost about o20m pa here).  Despite arguing our
commercial position, I'm quite happy with this approach and am genuinely
sceptical about the benefits of locational pricing in the UK.  This position
would of course change in much larger, sparse and less-interconnected
networks such as exist in parts of the US.
? Our fall back position is that any locational pricing scheme in the UK must
preserve pre-existing implied rights for generators, ie, grandfathering.  I
recognise that this can be a barrier to competition in some networks with
imcumbent market power, but I don't think that this is as big an issue in the
UK (where competition already exists) as it is on the Contintent and in the
US.

Hope these help, let me know if you want to discuss further.

Paul

