Why power plants should pay for grid upgrades.

Do we make all generators equal-using affirmative action to give rights to merchants that are "comparable" to utility-owned plants?

Or, do we let the locational price signals shine through-trusting all plant developers, whether regulated or not, to act in self-interest?

These questions capture the choice now pending before the Federal Energy Regulatory Commission (FERC) in its rulemaking case on generation interconnection,¹ as FERC ponders who should pay for expanding the transmission grid to accommodate new power plant projects.

The Problem

Let's look at a simple example that compares variable operating costs between two power plants located in different states.

One developer builds a gas-fired, combined-cycle power plant in Southern California. A second developer builds a plant in Arizona. In operational terms, the two plants are similar. Both projects have similar characteristics: minimum up/down time, forced outage rates, maintenance rate, capacity, and heat rate. They both have identical fuel prices, as they are both tied to the same gas transmission pipeline.

Nevertheless, the two plants differ significantly in actual performance, if we compare the two on the basis of unit dispatch and variable costs (ignore capital cost recovery and emission constraints).²

Consider capacity factor-the degree to which each plant is used. In our example, for 2003, the Southern California project will show a capacity factor of 80 percent, while the Arizona project comes in lower at 42 percent.³ For 2004, the two projects post capacity factors of 78.4 percent and 40.6 percent, respectively. By contrast, a re-power project located in the Los Angeles basin, which pays an additional intrastate gas transportation fee of about four mils per kWh ($4 per MWh), but has lower capital costs, will post a capacity factor of 63.3 percent. Why the difference in utilization?

The reasons for the difference in capacity factors between the Arizona and the California projects stem from transmission constraints into Southern California and the additional transmission losses (three to six percent, on average) associated with wheeling energy from Arizona to California. Bottom line: The Arizona unit is not making its expected rate of return.

Viewed another way, this example shows that congestion costs and losses provide short-run locational signals that new generation should locate in California. Yet that is not the case.

Despite this obvious price signal, more new generators are expected to continue to come online in the Arizona region, increasing congestion. At the same time, the gap in capacity factors between these locations is expected to widen.

Why would generators continue to build in a constrained area that is relatively far from the load, and when the plant loses money?

In our case, the answer must lie not with variable costs, but with fixed costs. From the perspective of the power plant developers, the fixed costs (land, environmental mitigation, permitting) must be lower in Arizona. But this focus is too narrow. It ignores another important locational factor-namely, the long-run marginal costs. By siting their plants in Arizona, far from load, the developers ignore the long-run costs of transmission network upgrades and expansions.

Historically, these costs have been socialized (rolled into the embedded cost of transmission) and collected from ratepayers. If this policy continues, the RTO (regional transmission organization) will require payment for network upgrades and bill the cost through transmission rates. In essence, generators will be able to make consumers pay for costs that flow from siting choices. Over time, as congestion increases, generators expect that the RTO or the existing transmission owners (TOs) will expand the transmission system and eliminate the congestion-again, at ratepayers' expense.

Is this the result that FERC wants?

Certainly, the time has come for a thoughtful, coordinated, efficient, and fully participatory siting process. We need to know who pays for the transmission grid beyond the point of interconnection:

• Who pays to maintain system reliability?
• Who pays to add capacity to the grid to accommodate the full output of the new plant?

FERC, in creating RTOs, must recognize that generation siting is an integral element to ensuring efficient electric transmission planning and investment. At a minimum, FERC interconnection policy should mesh well with its vision for a standard market design (SMD) for wholesale power transactions.⁴

SMD supports transmission congestion management based on locational marginal pricing (LMP) and each node is associated with full marginal transmission losses. These elements of SMD provide market participants with the short-run marginal cost price signals of using the transmission system. They provide market participants with the operational costs associated with different locations on the grid.

The Current Policy

Much of the FERC Notice of Proposed Rulemaking on Interconnection Policy (NOPR) is geared to ensure comparability between merchant generation and generation owned by transmission owners. This feature reflects FERC's desire to prevent alleged interconnections abuse, where vertically integrated utilities charge higher interconnection costs for merchant generation than they charge themselves.

If one were to accept FERC's premise-that comparability is the paramount concern-then, yes, one might indeed favor FERC's current interconnection policy. But other factors also warrant consideration.

Under its current policy, FERC requires the generator to pay all interconnection costs up front, but then it is given a credit on its Open Access Transmission Tariff (OATT). In essence, the generator is fully reimbursed for all upgrades except for the generation tie, i.e., the facilities required to connect the generator to the grid. Comparability, when TOs also own generation and are not in a RTO, is ensured by simply removing reliability and deliverability network facilities' costs from the interconnection mix by having the TO and ratepayers ultimately pay for all network upgrades.

However, as discussed above, this solution is an inefficient policy in the new paradigm of RTOs and LMP. If interconnection costs associated with network facilities are socialized (rolled into the existing rate base), the locational signal is limited to only the costs of facilities used to tie into the network. This policy will lead to inefficient siting decisions.

Instead, by directly assigning network upgrade costs to the generator, the RTO can send the appropriate long-run marginal cost locational signal to the generator.

To see what that would mean, recall our example where the combined-cycle plants in California and Arizona had identical operational characteristics. If FERC instead would assign network upgrades to generators, it would become more efficient to locate the plants in California. In that case, the additional transmission expansion is no longer needed to alleviate congestion. Ratepayers would face lower costs in the long run.

Moreover, with an RTO, the issue of comparable interconnection charges between merchant generation and generation owned by transmission owners is no longer an issue. The RTO, as the independent system operator with responsibility for grid planning, has no incentive to discriminate between generators based on ownership. Comparable price signals are the main issue in the RTO/LMP market design. Interconnection costs provide market participants with the long-run marginal costs associated with different locations.

A Partial Fix

One solution would allow an RTO to bill generators for the cost of upgrades to network transmission facilities that the RTO would not require in its transmission expansion plan "but for" the generator's new project. That would eliminate the problem of socialized costs for at least one category of grid upgrades. And, to its credit, FERC has recognized in a recent case that for an RTO with LMP it may be appropriate to assign costs directly to generators for such "but for" network facilities.

In June, pending the outcome of the NOPR on generation interconnection, FERC gave a temporary OK to a "but for" rule proposed by the California Independent System Operator (CAISO), as part of its interconnection policy embodied in Tariff Amendment 39.⁵ The new rule would cover the allocation of two types of costs.

First, the rule would cover reliability upgrades beyond the first point of interconnection that were not already in the CAISO transmission expansion plan. Second, the rule would cover so-called "deliverability" upgrades that relieve congestion, including economic transmission projects and network facilities costs.

In each case, it appears that the tariff would allocate at least some costs to generators, as project beneficiaries. That's because CAISO tariff sec. 3.2.72 states that, for transmission additions or upgrades, "costs shall be borne by the beneficiaries, in approximate relative proportions by which they benefit."⁶ Certainly, it is inconceivable that the new generator would not be considered as one of the beneficiaries of the expansion.⁷

For any additional capacity created by the upgrade, the market participants in the project would receive financial transmission rights (FTRs). Some industry experts see FTRs as a logical, alternative way to finance grid expansion, because they reward grid investors for increasing the transfer capability of the network.⁸

And in the CAISO example, awarding FTRs to plant developers would work better than requiring generators to pay up front for all network upgrades and then reimbursing them through rate credits against future payments for transmission service transmission. That's because generators do not pay directly for transmission services under the CAISO model.

Also, rate credits were designed for the pre-RTO paradigm. Rate credits reflect the old assumption that generators interconnect with individual TOs under a company-specific OATT approved under Order 888. That policy was designed to get around the rule against "and" pricing, which protects transmission customers from having to pay simultaneously for both the network upgrade and use of the existing transmission network.

It should be clear that the new CAISO policy, now officially approved on an interim basis by FERC, is a consistent policy with respect to both short-run and long-run price signals. It aligns costs and benefits, and only as a last resort (backstop) is the ratepayer asked to foot the bill.

Yet even with CAISO's "but for" solution, which assigns some upgrade costs directly to generators, several problems will remain:

• Financial Capability. Private funding may be prohibitive for an individual generator, yet the upgrade is beneficial to ratepayers.
• Queue Anomalies. Lower-queued generators may get benefits without paying (the "free rider" problem) since the first generator to interconnect at a specific location (the developer with the highest queue position) is usually the one that contributes all the upgrade costs.
• Incumbent Bias. New power plants will face costs that existing generators never had to pay, raising a barrier to entry and creating a new version of the comparability problem.

The Likely Result

What happens to energy prices if FERC changes policy and the RTO assigns grid upgrade costs directly to new power plant developers?

Let's reconsider our original example with the two plant sites in California and Arizona. If a new FERC policy prompts the generator to to change plans and instead build the plant in California, where fixed costs are higher, could the generator pass those costs on to the ratepayer by charging a higher-market-clearing price (MCP) for energy?

To answer this question, consider how energy prices are set in a competitive market, and whether generators can influence that price.

In a competitive market, the marginal generator sets the MCP for all generators. A new generator that is more efficient will find that price exceeds their operating costs. Thus, the new generator collects revenues above its operating costs. These revenues help pay the new generator's fixed costs and expected rate of return. But MCP does not change to reflect a change in these fixed costs.

Bid caps present a special case.

Ordinarily, our new generator has no incentive to increase its bid price. However, in a period of constrained resources or excess demand, the generator can charge up to its bid cap (or a "damage control" cap, if one exists). If bid caps do not exist, our new generator can submit bids that reflect the scarcity value ("rent") of the energy. In that case, the price of energy is no longer related to the cost of the product.

Now let's return to our example.

To the extent that our new generator entails higher sunk costs in siting his plant in California, instead of Arizona, those costs will not affect dispatch decisions or the MCP. But the change in siting might still make a difference. For example, if the siting change produces higher operating costs, like paying intrastate gas transmission rates, these costs will affect the bid price, dispatch decisions, and the MCP.

Also, a direct assignment of upgrade costs to generators could prompt some developers to delay projects until the MCP rises to offset the upgrade costs and ensure profitability. In order for the new entrant to meet its hurdle rate, expected net revenues must increase. As the supply and demand balance moves up the existing supply curve, the MCP rises. The longer entry is delayed, the higher the MCP.

In summary, if FERC should agree to modify its interconnection policy to assign upgrade costs directly to generators, then we might see higher energy prices. These potentially higher market-clearing prices must be balanced, however, against the higher access charges for transmission service that ratepayers otherwise will assuredly pay if the cost of additional grid facilities is assigned to load.

The content of this article reflects solely the views of the authors, and are not necessarily those of Southern California Gas Company, which has no responsibility for its content.

1. Interconnection NOPR, FERC Docket No. RM02-1-000.
2. The comparison is conducted using Henwood's Prosym. This analysis represents a detailed production cost simulation model with multi-area modeling capability. The power system modeled represented the Western Electricity Coordinating Council (WECC), formerly known as Western System Coordinating Council (WSCC). Approximately 39,000 MW of new thermal generating units were assumed to be added to the existing system between 2001 through 2004. In Prosym, power plants are grouped into their geographical regions with each region having its unique, electrical load, and individual power plant characteristics (fuel, generating capacity, heat rate, outages, and operation limits). The regions are interconnected with transmission lines, and transmission line constraints are enforced. Transmission line limits include line capacity ratings, average transmission losses, and wheeling costs. Buying and selling among regions is permissible either through contracts or economic transactions. The model dispatches power plants economically to meet load demands in all regions while meeting all operation limits.
3. As a point of reference, a unit with these physical characteristics, a 20-year payback, and a 15% rate of return must operate at about a 60% capacity factor.
4. Electric Market Design and Structures, FERC Docket No. RM01-12-000.
5. FERC Interconnection NOPR, Docket No. RM02-1-000, at 29-30. For the CAISO, generators are not charged an OATT. Instead, the load pays the transmission access charge.
6. CAISO Tariff § 3.2.7.2. If a specific beneficiary cannot be identified, then the upgrade cost is reflected in the access charge. See Tariff § 3.2.7.3.
7. In a fully regulated, cost-of-service world, one could argue that the grid was built for the load. However, in an electric industry with a competitive wholesale generation market, this argument no longer holds. Generation is built at a specific location because of expected profits predicated on access to the market. To the extent an upgrade increases the generator's access to the market, the generator should be viewed as a beneficiary of the upgrade.
8. Michael D. Cadwalader, Scott M. Harvey, and William W. Hogan, Review of the California ISO's MD02 Proposal, June 4, 2002 at p. 25.

Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.