REGULATION

Advanced Search

ABOUT EEI

INDUSTRY ISSUES

PRODUCTS AND SERVICES

MEETINGS

MAGAZINE

NEWSROOM

home > magazine > editorial content

•	calendar
•	editorial index
•	guidelines

REGULATION

Sally Hunt is a special consultant and Hamish Fraser is a senior consultant with National Economic Research Associates, Inc. in New York. This is adapted from the forthcoming book, Making Competition Work in Electricity (New York: John Wiley & Sons. 2002).

Trading arrangement issues are in the Federal Energy Regulatory Commission's jurisdiction, and FERC has addressed them in Order 888 in 1996 and Order 2000 in 1999 and started a standard market design (SMD) process in 2001. [See the sidebar, "Standard Market Design Basics."] Until SMD, FERC's rules have been based on the wheeling model—a model that simply does not work for full generator competition and does not work particularly well anywhere. FERC must now provide the leadership in the design of trading arrangements hitherto lacking.

New trading arrangements have been established in the new competitive areas—namely, Pennsylvania-Jersey-Maryland (PJM), New York, New England, California, and Texas. The rules in each are different. The three in the Northeast developed from the old tight pools and have a family resemblance. They use the integrated trading model, wherein traders tell a system operator their physical, financial, and contractual parameters, and the operator does the trading automatically—dispatching generators in ascending order according to price and using this merit order to calculate spot and imbalance prices.

The tight pool model worked efficiently in the precompetition world, and the adaptations of it are working well in the new competitive world. In fact, the tight pools were the original blueprint for the development of competitive trading arrangements around the world. It is no accident that competition found its easiest home in the tight pool states. The access and pricing rules were changed to accommodate competition, but all this was fairly straightforward to arrange given the history of pooling.

California, on the other hand, had to invent a new trading system. The California trading arrangements were truly badly designed—deficient in virtually every respect—as were many other aspects of the market itself. Texas is the only other example in the United States; it only started up recently, although its independent system operator had been in place for several years (a feature, by our lights, worth emulating). The rest of the United States still uses the wheeling model.

Where new trading arrangements were developed, they were developed by the utilities that use them, in conjunction with the states that they serve. They are not the wheeling-based trading arrangements that FERC has been ordering in 888 and 2000. They certainly comply with FERC's requirements but have been made to fit—something akin to shaving the corners of a square peg so it will fit in a round hole. And although FERC said in 1997, "we have seen the future, and the future is PJM," the commission has not yet actually ordered the rest of the country to adopt the PJM model.

Wrong Question
The wheeling model asks the wrong question and therefore comes to the wrong answer: "What is the right charge for a delivery of a scheduled flow of power over a utility's transmission lines?" It starts from a delivery schedule (which, given the nature of electricity, is a fiction anyway) and piles on a dozen more unanswerable questions, creating serious inefficiencies and major operational headaches:

First it makes a preliminary cut—the right charge is a proportion of the investment being used.
Then it has to define a contract path to decide whose investment is being used. The contract path is a fiction, but once we have decided to work with it, it has to be made operational, and traders have to request a reservation of a path. Since the utility gets first priority for its own native load, the utility has to decide which paths are actually available for wheeling.
Then there have to be rules for priority among those who want to use the available paths and physical rights for the winners to use the system. And, since the contract paths bear little relation to reality, someone has to decide what the utility is permitted to do in real time when the real system does not correspond to the imaginary system of contract paths. Can the utility just curtail transactions; or, if it has to reorganize (redispatch) its own plants, what can it charge for that?
Then, almost as an afterthought, it has to worry about appropriate charges for losses, imbalances, and ancillary services, at regulated prices.

Each of these steps is a problem. Electricity is not like a carpet—you don't order one and complain if they deliver you a different one. Electricity is exactly the same everywhere. It is generated and flows over the transmission network where it will. Customers do not get the same electricity their supplier sends, and no one believes they do. They get whatever electricity happens to be flowing their way at the time. We just talk as if there is a delivery schedule that gets one producer's electricity to his own customers. We imagine a single transaction as a delivery with an origin and destination. But basing trading arrangements on this fiction is what leads to errors. Metaphors can be useful, but when taken literally (as the metaphorical "delivery schedule" was in California) whole edifices can be constructed on an erroneous concept.

Once the delivery schedule is the centerpiece of the design, it leads to paths and reservations and physical rights. These cause inefficiencies and operational problems.

The utility will obviously have incentives to over-reserve transmission capacity for itself rather than be caught short (and have to redispatch).
The priorities for transmission usage established on a first-come, first-served basis will not ensure the highest value and best use of the transmission.
Redispatch requires a utility system operator to have its own plants to adjust in the event of congestion (yet competitive markets should separate the role of system operator from generator, to assure a lack of discrimination and that generators face market forces).
Imbalances, losses, and ancillary services also have to be continually provided or absorbed by the system operator at regulated prices (and soon the traders want to provide their own losses, imbalances, and ancillary services—this quickly becomes unmanageable.)

The United States has another problem with the wheeling model because of the large number of utility control areas. Transactions between control areas are indeed "scheduled" and thus suffer from the contract path fiction. Congestion management, imbalances, ancillary services, scheduling, and dispatch are all made more difficult and the contract path fiction more troublesome when system operator control areas are small.

The right approach is to ask, "When competition has done the job of making the best use of the existing generating plants and the best use of the transmission, what will the system look like—who will be generating and how loaded will the transmission lines be?"

Actually, we know the answer to this: It is just what the utilities have been doing for years by command and control—making the best use at lowest cost of the existing system. Now that we have in mind what the end point will be, the question is: "How do we get there from here? Can we set up incentive-compatible rules to ensure that it gets there by the traders' own actions, without command and control?" The answer to the last question is yes. This is exactly what the integrated trading model does.

Instead of thinking of a path or a bridge or a pipeline, we have to think of a network, because that is what we have. Hundreds of generators and millions of customers are scattered around the network. In real time we only have to worry in aggregate about the generators and having sufficient power produced to serve the entire load. We do not have to worry about who is individually supposed to be serving whom—that is a financial question, not a physical one, and is taken care of in the settlement process.

Acronym Soup
When FERC wrote Order 888, the many special terms created to tie the old way of doing things to the new immediately created problems.

For example, shortly after the adoption of Order 888, the North American Electric Reliability Council (NERC) recognized that contract-path scheduling created incentives to overload the transmission network. NERC adopted transmission loading relief (TLR) protocols to undo the damage whenever the system became constrained. In essence, NERC created an administrative unscheduling system to counteract the effects of FERC's scheduling system. The NERC system did not work well. However, something was necessary to keep the lights on.

How do the FERC and NERC methods work? First, Order 888 automatically granted (grandfathered) rights, called native load priority (NLP), to the local utility's transmission system for all native load customers.

STANDARD MARKET DESIGN BASICS

The Federal Energy Regulatory Commission (FERC) has unveiled the essential elements of a standardized transmission service and wholesale electric market design—the basics of a standard market design (SMD) notice of proposed rulemaking (NOPR) are due this summer.

According to the "Working Paper on Standardized Transmission Services and Wholesale Electric Market Design," developed by FERC staff, the commission would create a new network access tariff combining the flexibility and universal access of network service and the reassignment rights of point-to-point service. It would create tradable transmission property rights to designated sources (the location where a transaction originates, that is, the generator) and sinks (the location where a transaction terminates, that is, either the load or a trading hub). The access charge would recover embedded system costs.

Under the proposed SMD, transmission providers must run both day-ahead and hourly energy markets using locational (or nodal) marginal pricing. LMP is a means of pricing that adjusts auction prices to reflect congestion on the grid, creating price signals that reflect the time and locational value of electricity. These markets would be voluntary, bid-based (meaning that participants would provide prices over the range of quantities they offer or seek to buy), and security-constrained (with the market administrator, through the energy auction process, accounting for all system constraints required for reliability).

Congestion would be managed using LMP. Market participants could schedule bilateral transactions (short- and long-term), self-supply, or bid into the day-ahead market. The demand side could also participate in the market. Scheduling options would accommodate energy-limited resources (e.g., hydro and environmentally constrained thermal power) and be fuel-and technology-neutral.

Also, FERC would combine two control area functions in NERC's new functional model: the balancing authority and the transmission service provider. Because under LMP the imbalance and transmission markets must operate together, the FERC staff argues that it is more efficient to have one independent entity perform both functions. The transmission provider would handle requests for transmission services, administer the open access same time information system, schedule transactions, and administer the imbalance markets. Regional transmission organizations and independent system operators qualify, but vertically integrated public utilities that are not part of an RTO or ISO would have to contract with an independent entity.

The paper did not resolve whether an independent transmission company qualifies as a transmission provider. Left for future discussion were issues of how rates would be designed; who would pay embedded transmission costs; how transmission rights would be allocated; how existing transmission rights and contracts would be transferred to the new SMD; and how the demand response process would work.

It also provided utilities with a transmission reserve margin, called a capacity benefit margin (CBM), to ensure that sufficient capacity on interconnections between utility systems was available to serve the native load in case an emergency necessitated importing power from adjacent areas.

After allocating transmission capacity sufficient to cover NLP and CBM to utilities, the remaining available transmission capacity (ATC, obtained by subtracting NLP and CBM from total transmission capacity) could be allocated on a first-come, first-served basis to other wholesale transmission users.

ATC must be posted on the open access same time information system (OASIS) web page, and the utility's own traders must get their information from OASIS, not by internal communications.

Under this scheme, transmission customers can purchase "network" service, "point-to-point firm," or "point-to-point nonfirm" transmission service.

The fundamental problem with allocating ATC among the three types of service rights is its use of embedded cost pricing that bears little relationship to the economic value customers place on those rights. Therefore, transmission capacity is not allocated (except coincidentally) to those who value it most highly.

Equal access under the Energy Policy Act of 1992 meant equal treatment of competitors for wholesale sales only, with priority for native load. Native load is allocated first priority, which seems right, until you realize that all load is served anyway. What is really allocated first priority is the generating plant that is used to serve native load. This can have two inefficient effects: it can give priority to a generator who is not the least cost; and it can reserve too much of the transmission, leaving capacity unused.

Another problem exists when the ATC calculated and allocated in advance doesn't turn out to match the real-time available transmission capacity. When this happens—and it does because of the contract path fiction—some transactions must be curtailed, or the system must be redispatched, to relieve the constrained element. Since transmission capacity rights are not rationed economically under Order 888, the curtailment rules must follow the "natural" order of priority established by service type and level of firmness: Network service has highest priority, followed by firm service, and finally nonfirm service. Aside from the efficiency implications, the problem is that property rights are not well defined, since even an energy transaction that is covered by a long-term firm transmission right can be (partially) curtailed if the congestion problem is severe enough.

TLR procedures enable system operators to address transmission congestion in terms of the physical reality of power flows on the network, curtailing transactions in proportion to their contribution to the congestion problem on a particular transmission system element, which also meant they addressed congestion caused by loop flows. Under the open access regime, the substantial increase in wholesale power transactions between and across regions and the increased use of the transmission grid in general have made the problem a more frequent occurrence. Using the TLR procedures, a transmission provider whose system becomes congested as a result of loop flows can require the curtailment of the transactions creating the loop flows.

FERC Understands the Problems
Even so, the problems of the wheeling model were not lost on FERC when it wrote Order 888. A close reading finds an extensive discussion of the obstacles to electricity markets created by the need for instantaneous balancing and managing the complexities of transmission usage. In particular, FERC recognized that the traditional wheeling model was built on the contract path fiction. But FERC embraced the wheeling model for the expedient reason that at the time the commission did not have clear legal authority or support to go further. The result is that FERC has approved new rules, agreements, and protocols in several systems, but the rules are still in the context of the old wheeling methodology.

In Order 2000, FERC made great advances in responding to the need for market design that is suitable for full competition and recognizes the complexity of electricity markets. It defined the RTO concept and told utilities to form them. It required an RTO to have operational authority for all transmission facilities under its control and to be the security coordinator for its region. FERC's objective was for all transmission-owning entities in the United States to place their transmission facilities under the control of appropriate RTOs in a timely manner.

But although Order 2000 allowed all the things that would make for full competition, it did not require them. It seemed that it was written by two people—one who defined the necessary criteria for full competition, and the other who tied the requirements back to accommodate the constraints of the wheeling model, OASIS, and all the paraphernalia of defining paths through someone else's system. FERC did require some of the things that are undeniably prerequisites for full competition, but it did not eliminate the vestiges of the wheeling model, and indeed it continued to require some of them.

The need for standard cohesive trading arrangements conceived "in the public interest" (i.e., with efficiency, liquidity, and transparency as a major objective) is at the heart of the current institutional problems. It has held center stage for a decade and has not been resolved, deflecting corporate and regulatory energy from arguably more fundamental issues. The SMD initiative is an encouraging sign that FERC is ready to make major changes to solve these problems—as long, of course, as they are prepared to make a true shift from wheeling to the integrated trading model.

| | | | |