Interface Pricing Approach Whitepaper May 2015 This white paper represents MISO's current progress and assessment on the Interface Pricing issue. Additional analysis is underway and will be shared with stakeholders upon completion.
Executive Summary MISO and PJM, being neighboring RTOs, facilitate the interchange of energy across their borders using transactions. Interchange transactions, which have a source in one RTO and a sink in another RTO, are integrated into the Security-Constrained Unit Commitment models for each associated RTO. As such, interchange transactions affect each RTO s generator dispatch and pricing outcomes. In its 2012 State-Ofthe-Market report, Potomac Economics (MISO s Independent Market Monitor) identified a flaw with the current Market-to-Market (M2M) process for MISO and PJM interchange transactions. The identified flaw is the result of MISO and PJM modelling interchange transactions with two different source-to-sink paths. This can result in overlap of M2M constraint flow impacts and associated congestion pricing. This overlap flaw can lead to: Redundant M2M congestion pricing Congestion revenue inadequacy Inefficient price incentives Reduced system reliability Example of a M2M transaction from PJM to MISO as observed by Potomac 1 : PJM s Path: Source from PJM s Load Center Sink at PJM/MISO Seam MISO s Path: Source from PJM s Generator Center Sink at MISO s Load Center Overlap occurs between PJM s Load Center and the PJM/MISO Seam Potomac and PJM have proposed different solutions to the overlap flaw 2. The proposed solutions can be generalized as follows: Potomac s Solution: Model one full transaction path from PJM to MISO PJM s Solution: Model two transaction paths separated at the PJM/MISO Seam 1 Please note that the images in this executive summary are simplified representations of images found later in the whitepaper. 2 Each solution requires that MISO and PJM work collaboratively in order to fully address the flaw. PJM independently implemented their portion of their solution in June 2014.
Potomac s Solution PJM s Solution Implementing either solution would eliminate the overlap flaw. MISO conducted an analysis, as detailed in this Interface Pricing Approach Whitepaper, to determine which of the two solutions would be ideal to implement. MISO s analysis revealed that Potomac s solution has the advantages of: (1) Being more consistent with the current M2M process with respect to treatment of interchange transaction flow impacts on M2M constraints (2) Having less adverse impacts on MISO s modeling of non-m2m constraints for associated power flow representation and price calculations (3) Avoiding inefficient price volatility which could cause uneconomic transaction scheduling and subsequent reliability concerns These are the key reasons supporting MISO s current assessment that Potomac Economics solution is superior.
Table of Contents 1 Introduction... 5 1.1 Transaction Impact Decomposition Model... 5 1.2 Market-to-Market Process Relation to Interface Pricing... 8 2 Current Market-to-Market Implementation... 8 3 Functional Evaluation of Alternative #1: MISO IMM Approach... 10 3.1 Elimination of Transaction Incentive Over/Under Counting and Revenue Inadequacy... 11 3.2 Consistent with the current M2M Process... 12 3.3 Small impact to MISO modeling of non-m2m constraints... 13 4 Functional Evaluation of Alternative #2: Common Interface Approach... 15 4.1 Elimination of Transaction Price Incentive Over/Under Counting... 16 4.2 Commercial Flow M2M Process Needed to Support Revenue Inadequacy... 17 4.3 Unintended Impact on Non-M2M Constraints... 18 4.4 Inefficient Price Incentive Volatility... 19 5 Summary of Assessment... 21
1 Introduction In the 2012 and 2013 State of the Market Reports 3, the MISO Independent Market Monitor (IMM) described a flaw in the combined treatment of interchange transactions whereby MISO and PJM congestion charges (or credits) and their associated constraint flow impact for Marketto-Market (M2M) managed constraints overlap. This overlap of constraint impact pricing can lead to redundant M2M congestion pricing, create congestion revenue inadequacy, provide inefficient price incentives, and ultimately reduce system reliability. The impacts of this flaw in pricing, which overstates or understates the congestion prices for interchange transactions, include (1) incorrect price signals affecting constraint management and system reliability, and (2) creation of an opportunity for revenue inadequacy for the Non- Monitoring RTO (NMRTO) due to lack of a mechanism to recover congestion payments to the transactions. Additionally, the current mismatch in interface price definitions creates challenges for participants to arbitrage price differences between the two markets and, in the future, the Coordinated Transaction Scheduling (CTS) process 4. Inefficient interchange scheduling hinders the ability to lower overall operating cost for the joint region. Both the MISO IMM and PJM have proposed approaches to address the deficiencies of the current approach with the objectives of better capturing the economic benefits of constraint coordination, providing accurate representation of constraint impacts for reliability, and improving market price incentives and settlement amounts for interchange transactions. This paper provides analysis of the existing approach and the two proposed alternatives using the following criteria: Elimination of over or under statement of constraint impact on transaction pricing Elimination of systematic RTO revenue inadequacy resulting from interface modeling Accuracy and consistency of transaction price incentives Accuracy of non-m2m constraint models The paper concludes with an evaluation summary for both MISO and PJM s proposals and MISO s current preference for moving forward to modify the current approach. 1.1 and 1.2 of this section describe the modeling approach for representing transaction impacts on constraint flow and the Market-to-Market (M2M) constraint coordination process. 1.1 Transaction Impact Decomposition Model This section describes a model that is used in this paper to decompose the constraint impact of a transaction into components. These components are used to facilitate discussion of the current approach and the proposed alternatives. 3 2012 State of the Market Report for the MISO Electricity Markets. https://www.misoenergy.org/library/repository/meeting%20material/stakeholder/bod/markets%20committee/20 13/20130724/20130724%20Markets%20Committee%20of%20the%20BOD%20Item%2005%202012%20SOM%20 Report.pdf 4 https://www.misoenergy.org/library/repository/meeting%20material/stakeholder/msc/2015/20150331/2015033 1%20MSC%20Item%2007a%20CTS%20Proposal.pdf 5 P a g e
To provide background on the MISO-PJM Overlap Issue, consider a jointly optimized region encompassing MISO, PJM, and potentially additional electrically connected areas. With a jointly dispatched region, there would be no interchange transactions and the combined fleet of generators would be dispatched optimally to meet the system obligations considering transmission constraints. The resulting pricing would be based on the marginal resources determined by the joint dispatch. A jointly dispatched MISO and PJM region is not possible at this time. Interchange transactions are supported in both markets to provide the opportunity to capture the efficiencies associated with the transfer of energy between the RTOs. Each RTO settles the interchange transactions at its applicable interface price. For example, an interchange transaction from PJM to MISO would settle as an export transaction at PJM s interface price for MISO and as an import transaction at MISO s interface price for PJM. In a simple operating model, as seen in Figure 1, the RTO owning a constraint provides all of the constraint management redispatch and incurs the associated costs. Such a simple approach is not practical in an interconnected grid like the Eastern Interconnect. To recognize other parties use of the transmission facility and the equitable use of the transmission system, the Transmission Loading Relief (TLR) process is employed in the constraint management process. Under TLR, each market entity independently provides its share of redispatch relief measured by its constraint market flow: the flow impact on a constraint of the market generation serving market load. Figure 1 TLR provides equity in use, but leaves room for efficiency improvement. The M2M process allows market operators to pool their TLR required constraint flow relief to enable to the more cost effective entity to provide the relief. The M2M process includes the exchange of constraint shadow price information to help the RTOs determine which entity can provide the most economic relief. The M2M process shifts the redispatch effort for active constraints between the RTOs and continuously moves toward convergence of the RTO shadow prices and improved efficiency as system conditions and the associated prices change over time. Although the RTOs are each performing their own dispatch and pricing, the communicated constraint shadow price provides coupling to facilitate relief by the lower cost RTO. The RTOs participating in M2M coordination settle with one another to pay equitably for the use of the transmission facility. The M2M settlement is based on the NMRTO use: if the NMRTO uses more than its entitlement, it will pay the Monitoring RTO (MRTO) for its use above its entitlement and if the 6 P a g e
NMRTO uses less than its entitlement, the MRTO pays the NMRTO for the portion of its entitlement which was not used. From a congestion revenue standpoint, these conditions bring the NMRTO to a congestion funding level based on a constraint market flow corresponding to its entitled flow on the constraint. To facilitate discussion of the existing and proposed methods, their impact on constraint modeling and congestion revenue adequacy, the impact for a transaction from the NMRTO to the NMRTO can be divided into 3 components. These components are selected to describe the full impact from changes in the source and sink area generation and to coincide with a natural alignment with the market processes: Component1 (Figure 2): NMRTO generators to NMRTO load Component 2 (Figure 3): NMRTO load to MRTO load division of the constraint impact in this component is the key difference in the alternatives evaluated in this paper Component 3 (Figure 4): MRTO load to MRTO generators Figure 2 For the first component of the decomposition (Figure 2), using PJM as the NMRTO and MISO as the MRTO, the PJM generators to PJM load has the flow impact represented in the PJM market flow. The current market flow used in M2M coordination and regional coordination with the NERC IDC (Interchange Distribution Calculator) is a measure of internal generation serving internal load with exports (imports) by reducing generation (load). The pricing impacts of the constraint on PJM generation and load are incorporated in the PJM LMPs. Figure 3 For the second component of the decomposition (Figure 3), the interchange transaction needs to cover cost impacts of energy transfer from PJM load center to MISO load center. Changes in 7 P a g e
load in component 2 (Figure 3) offset the load changes in component 1 (Figure 2) and component 3 (Figure 4) providing zero net change at the loads. Figure 4 The third component of the decomposition (Figure 4) is similar to component 1. The MISO generators to MISO load flow impact is captured in the MISO market flow. The current market flow used in M2M coordination and regional coordination with the IDC is a measure of internal generation serving internal load with exports (imports) by reducing generation (load). The pricing impacts of the constraint on MISO generation and load are incorporated in the MISO LMPs. 1.2 Market-to-Market Process Relation to Interface Pricing The current M2M process incorporates the use of net market flow. Market flow is also used by the NERC IDC application to enable equitable use of the transmission system by regional markets and non-market transmission rights. An RTO s market flow represents the constraint impact attributed to its internal generation serving its internal load. To remove the impact of transactions, exports are removed from generation MWs and imports are removed from load MWs. Thus, the NMRTO market flow represents the market generation serving the market load as shown in Figure 5. B MRTO C D NMRTO E A MISO Constraint F Figure 5 2 Current Market-to-Market Implementation Currently, MISO and PJM employ different approaches to modeling and pricing interchange transactions. These differences cause inefficiencies in management of the joint constraint coordination process which adversely impact reliability and reduce operational efficiency. At the core of the differences are the RTOs interface definition and pricing differences. 8 P a g e
In concept, the constraint flow impact on an M2M coordinated constraint due to an incremental interchange transaction and the associated pricing is consistently represented when it captures the full impact caused by the net increased marginal generation in the source area and net decreased marginal generation in the sink area. This impact can be divided amongst the M2M constraint management partners, but a necessary condition for efficiency is that the division should cover the entire path from source to sink and cover every segment only once. This is not the case with the current implementations. For a transaction from PJM to MISO relative to a M2M coordinated constraint in MISO, MISO s current transaction pricing model, including the MISO interface definition for PJM and the price calculation, divides the components defined in Chapter 1 as follows: Component 1: PJM (not implemented). PJM marginal generators to PJM load-centered reference is included in the PJM market clearing and LMPs. Component 2(a): unaccounted. PJM load-centered reference to MISO s PJM generatorbased interface definition is not captured by MISO or PJM. Analysis presented in Chapter 3 shows this component is small. Component 2(b) and Component 3: MISO. MISO s PJM generator-based interface definition to MISO marginal generators is covered by MISO. Component 2 of the decomposition model is divided into two elements, the small 2(a) and 2(b). The MISO model assumption about PJM participation in congestion pricing does not match the PJM implementation. PJM s current implementation model divides the components as follows: Component 1 and Component 2(i): PJM. PJM marginal generators to selected generators near the electrical seam between the RTOs (PJM s interface for MISO) is covered by PJM. Component 2(ii) and Component 3: MISO. Selected generators near the electrical seam between the RTOs to MISO marginal generators is covered by MISO. In the current PJM implementation, component 2 is divided differently into elements 2(i) and 2(ii). The current PJM implementation assumption about MISO participation in congestion pricing does not match the current MISO implementation. The fact that neither RTO participates as the other s model assumes significantly contribute to the undesirable distortion of interchange transaction pricing and settlement incentives that characterize the MISO-PJM Congestion Overlap Issue. The overlapping pricing and incentives seen with each RTO applying its own model are evident when the current implementation is drawn pictorially. Figure 6 depicts the current MISO and PJM implementations for a transaction from PJM to MISO. The overlapping paths identify excess or deficient incentives where transactions transmission congestion incentives are double counted. 9 P a g e
Figure 6 The diagram (Figure 6) for a transaction from PJM to MISO shows the overlap in the transaction pricing impact between the MISO generator-based PJM interface and the seam: this is the source of many of the issues with the current implementation including distortion of the combined transaction incentives and RTO revenue inadequacy. The proposed approaches address the undesirable issues found in the current approach. At a fundamental level, the proposed alternatives define non-overlapping segments of the source marginal generation to sink marginal generation path. With this basic condition met, the proposals are evaluated on other criteria related to their implementation. 3 Functional Evaluation of Alternative #1: MISO IMM Approach The alternative proposed by the MISO IMM focuses the RTO coordination for M2M constraints at the NMRTO load-weighted reference used for the NMRTO s sensitivity calculations. Under the MISO IMM alternative, the division of the constraint coordination and modeling between the MRTO and NMRTO is as follows: The NMRTO manages its generation to load impact including constraints flow, pricing, and associated congestion revenue relative to its market flow and the M2M process. The MRTO (1) manages the total (physical) flow of the constraint including the impact of all transactions and (2) collects congestion revenues associated with its internal dispatch and with transactions from interfaces representing neighboring areas load-based reference. Through the existing M2M process, the MRTO provides the NMRTO with M2M relief requests and other data as input the NMRTO constraint management. This approach meets the criteria of accurate and consistent price incentives for transactions. For example, a transaction from the NMRTO to the MRTO under the MISO IMM alternative models the congestion impact and associated transaction price incentive of the transaction path once and only once. Using the decomposition defined earlier in this paper, the MISO IMM alternative can be described as the following: Component 1: NMRTO. NMRTO provides congestion price incentives from the NMRTO generation to NMRTO load through the congestion component of its LMP. 10 P a g e
Component 2: MRTO. MRTO pricing, using the MRTO constraint shadow prices, is applied for the segment from the NMRTO load-weighted reference to MRTO loadweighted reference. The NMRTO provides no interface pricing signal for the constraint. Component 3: MRTO. MRTO provides congestion price incentives from the MRTO generation to MRTO load through the congestion component of its LMP. The modeling for a non-m2m constraint would be similar except the first component (Component 1) would not be modeled by the NMRTO since non-m2m constraints are not coordinated. By nature of not being a M2M constraint, the impact of the NMRTO generation to load impact on the non-m2m constraint will normally be relatively small. The modeling for the MISO IMM alternative can be depicted by Figure 7 for a transaction from PJM to MISO where MISO is represented as blue and PJM as green. Figure 7 MISO s analysis of the MISO IMM alternative finds that this approach addresses the most important criteria for implementation and that it is compatible with the existing MISO processes and operating practices, reducing the effort and risk associated with implementation. Specifically, these criteria are the following: Eliminates congestion incentive overlap in transaction price incentives and a systematic cause of RTO revenue inadequacy Consistent with current M2M coordination implementation Small impact on current MISO modeling of non-m2m constraints The following sections describe the analysis of these key criteria for the MISO IMM alternative. 3.1 Elimination of Transaction Incentive Over/Under Counting and Revenue Inadequacy Interchange transactions are settled by RTOs based on the published interface prices. When interchange occurs between two neighboring RTOs, such as MISO and PJM, the transaction settles with both RTOs at the interface price representing the other RTO as a withdrawal from the source RTO and an injection in the sink RTO. The net financial congestion incentives provided by settlement with both RTOs should be commiserate with the net incremental constraint impact of the transaction. 11 P a g e
Providing this collective incentive should not systematically create a congestion revenue shortfall for either RTO. The MRTO manages the overall flow on the M2M constraint including use by the NMRTO and other parties. The MRTO adjusts its market limits as needed in realtime to maintain secure operations. With consistent limits in the Day-Ahead Market, the MRTO can manage its revenue adequacy. The NMRTO manages only its portion of the transmission constraint impact, i.e., its market flow. The real-time M2M settlement process brings the NMRTO congestion revenue on a M2M constraint to a level corresponding to flow equal to its effective real-time entitlement. By managing its Day-Ahead Market impact, the NMRTO can manage the MW flow to an estimate of its firm flow entitlement to manage the flow impacting its real-time net congestion revenue on the constraint. Using a 10-unit simulation model initiated by PJM, MISO performed analysis for several potential combinations of design alternatives. As depicted in the table below, this analysis showed the following when both RTOs use the MISO IMM alternative: The net interchange settlement incentives for a transaction between MISO and PJM is consistent with the value of the impact represented in the market clearing models. The Transactions Settlement Amount is equal to the Target Transaction Settlement Amount representing the incremental financial benefit of the transaction on transmission congestion. With the Congestion Overlap Issue, the current implementation does not provide a transaction settlement amount equal to the target represented by the current model. The target transaction settlement amount includes components for the energy price difference between the RTOs and for the constraint congestion component. The energy component of the transaction settlement is the same as in the current implementation. The table also shows that under the MISO IMM alternative, the congestion component associated with the transmission congestion is collected only by the MRTO (MISO in this example). Table 1 Scenario Current implementation Both MISO and PJM adopt MISO IMM alternative Transaction Settlement Amount Target Transaction Settlement Amount MISO Market Congestion Revenue PJM Market Congestion Revenue $2,511.11 $2,041.27 -$871.43 $469.84 $2,041.27 $2,041.27 -$871.43 $0 3.2 Consistent with the current M2M Process As described in the introduction, the current M2M process incorporates the use of net market flow. Market flow was designed for use in the NERC Interchange Distribution Calculator (IDC) 12 P a g e
application to enable equitable use of the transmission system by regional markets and nonmarket transmission rights. Its use was extended to the M2M process. An RTO s market flow represents the constraint impact attributed to its internal generation serving its internal load. To remove the impact of transactions, exports are removed from generation MWs and imports are removed from load MWs. This leaves the NMRTO market flow representing the market generation serving the market load as shown in Figure 5. The MISO IMM alternative calls for the NMRTO to manage the M2M transmission constraint pricing and revenues associated with its generation to load constraint impact through the LMP resulting from its market dispatch. The NMRTO generator to load constraint impact modeled through the NMRTO market clearing model and LMP are similar to the generator to load constraint impact defined by the current market flow. This alignment facilitates improved revenue adequacy for the NMRTO. The following diagrams, which focus on the NMRTO, show the conceptual similarities and compatibility of the two approaches. Figure 8 With the MISO IMM alternative, the NMRTO s interface price for the MRTO does not include any congestion component for the M2M constraint; the NMRTO is revenue neutral with respect to interchange transactions and congestion revenues in the real-time market. To maintain revenue adequacy between the real-time and day-ahead markets, the NMRTO should ensure that the day-ahead market flow is at or below its anticipated real-time entitlement since the realtime M2M settlement process will provide the NMRTO with congestion funding for a level of flow equal to its real-time entitlement. 3.3 Small impact to MISO modeling of non-m2m constraints M2M constraints are an important part of an RTO s constraint management process. Non-M2M constraints which are solely managed by one RTO are also an important aspect of interface pricing. An empirical study was performed by MISO based on historical data to evaluate differences in shift factors and LMPs with respect to MISO s non-m2m constraints. The constraints selected for this study had significant impacts from import/export/wheel-through 13 P a g e
transactions involving the PJM interface. The results are summarized in Table 2. It should be noted that these statistics are based on absolute values of the shift factors. MISO s Existing PJM Interface refers to the existing MISO definition for the PJM interface that comprises all PJM generators with the congestion modeled from the PJM generator-based interface to the MISO load-weighted location. MISO IMM s Proposal refers to an alternative PJM interface definition that uses all external PJM loads with the congestion is modeled from the PJM load-weighted location to the MISO load-weighted location. The PJM Gen vs. Load Difference is calculated as the difference between MISO s Existing PJM Interface and MISO IMM s Proposal with the congestion modeled from the PJM generator-based interface to the PJM load-weighted location. As shown by the Percentage Change, the absolute shift factor difference between the generator and load based definitions is within 3% on average, and with a small standard deviation. It would therefore be reasonable to infer that the modeling difference of transaction congestion impact on non-m2m constraints would be small by employing the MISO IMM approach. MISO s Existing PJM Interface PJM Gen MISO Load [A] Table 2 MISO IMM s Proposal PJM Load MISO Load [B] PJM Gen vs. Load Difference Average 0.0341 0.0333 0.0014 2.35% StdDev 0.0275 0.0266 0.0015 5.54% Percentage Change [A-B]/A The study also examined the shift factor changes on a single non-m2m constraint. For this analysis, two extreme cases with both the minimum PJM Gen vs. Load Difference and the maximum PJM Gen vs. Load Difference were studied and summarized in the following table. It should also be noted that these statistics are based on absolute values of the shift factors. MIN and MAX values of shift factors were found for two different non-m2m constraints. The maximum difference is about 3%. MISO s Existing PJM Interface PJM Gen MISO Load [A] Table 3 MISO IMM s Proposal PJM Load MISO Load [B] PJM Gen vs. Load Difference MIN 0.0226 0.0226 0.0001-0.22% MAX 0.0630 0.0670 0.0040-6.41% Percentage Change [A-B]/A Analysis of the LMP prices was also conducted to investigate the difference between a PJM interface defined at the PJM generation-based location or the PJM load-weighted location. The real-time data in 2014 were collected and analyzed statistically including all MISO generators LMP values and the LMP values of the load-weighted reference (which is equal to the system marginal energy cost). The study results are summarized in the following table. The analysis shows that the average price of the load-weighted location is close to the average price of 14 P a g e
generator-based location and that about 75 percent of the price difference data points are within the plus and minus five dollars. Table 4 Generator-Based LMP Load-Weighted LMP Difference Average $33.34 $37.10 -$3.76 StdDev $26.09 $29.74 $8.45 4 Functional Evaluation of Alternative #2: Common Interface Approach The PJM proposed alternative defines a common interface near the seam between the RTOs and uses this interface definition to facilitate RTO coordination for M2M constraints. For MISO- PJM coordination, this interface serves both as PJM s MISO interface and MISO s PJM interface for all constraint impacts, both M2M and non-m2m constraints. This leads to an interface pricing approach where both the MRTO s and NMRTO s interface congestion pricing represents the congestion between its own load center (i.e., its reference) and the common interface. Under this approach, the division of the constraint coordination and modeling between MRTO and NMRTO is as follows: NMRTO manages its generation to load impact plus the transaction impact modeled between its load reference and the common interface. MRTO manages the total (physical) flow of the constraint for reliability, but for congestion revenue and pricing, the MRTO collects congestion impacts associated with its internal dispatch plus the transaction impact between MRTO load reference and the common interface. This alternative has symmetry between the treatment by the MRTO and NMRTO, but it is not consistent with the current definition of market flow used in the M2M process. The PJM alternative meets the criteria of accurate and consistent price incentives for transactions. For example, a transaction from the NMRTO to the MRTO under this alternative models the congestion impact and associated transaction price incentive of the entire path once and only once. Using the decomposition defined earlier in this paper relative to an M2M constraint, the PJM alternative divides the congestion management as follows: Component 1: NMRTO. NMRTO provides congestion price incentives from the NMRTO generation to NMRTO load through the congestion component of its LMP. Component 2: Shared. The component between the RTO load centers is divided into two non-overlapping segments at the common interface with each RTO responsible for pricing and collecting revenues for impacts between its load-center and the common interface. o Component 2(i): NMRTO. NMRTO pricing, using NMRTO constraint shadow prices, is applied for the segment between the NMRTO load-weighted reference and the common interface. 15 P a g e
o Component 2(ii): MRTO. MRTO pricing, using MRTO constraint shadow prices, is applied for the segment between the MRTO load-weighted reference and the common interface. Component 3: MRTO. MRTO provides congestion price incentives from the MRTO generation to MRTO load through the congestion component of its LMP. The model for a non-m2m constraint would use the same interface definition. Under the PJM alternative for a non-m2m constraint, Component 1 and 2(i) are not included in the transactions non-m2m constraint price incentives even though it does contribute the constraint physical flow managed by the MRTO. The magnitude of the non-m2m impact of this model will be analyzed in this chapter. The modeling for the PJM alternative can be depicted by Figure 10 for a transaction from PJM to MISO where MISO is represented as blue and PJM as green. Figure 9 MISO s analysis of the PJM alternative finds that this approach addresses the basic congestion price overlap causing the MISO-PJM Congestion Overlap Issue, but has several issues that would require significant effort to resolve (if the issues can be addressed at all). Specifically, these issues include the following: Inconsistent with current M2M coordination with market flow requiring M2M coordination use a new flow measure called constraint commercial flow Unintended impact on non-m2m constraints Price incentive volatility leading to inefficient system operation The following sections describe the analysis of these key issues of the PJM alternative. 4.1 Elimination of Transaction Price Incentive Over/Under Counting Similar to the analysis performed for the MISO IMM proposed alternative, the PJM proposed alternative can also provide a transaction from the NMRTO to the MRTO with consistent price signals which avoid over or under counting a transaction s impact on a M2M managed constraint and without creating a systematic RTO revenue shortfall. Under the PJM alternative, the pricing of the component of the flow impact from the NMRTO load to the MRTO load is divided between the two RTOs and the M2M settlement process changes from being based on 16 P a g e
market flow to a basis of commercial flow where commercial flow includes the RTO generation to load flow impacts plus the impact of transactions at the common interface for both the MRTO and NMRTO. (Commercial flow is discussed further in the next section.) As with the MISO IMM proposed alternative, the RTOs can maintain congestion revenue adequacy if they maintain constraints in the Day-Ahead Market consistent with the modeling and limits applied in real-time. To do this, the NMRTO should observe commercial flow constraints in the Day-Ahead Market with limits consistent with the commercial flow firm flow entitlement which will be used in the real-time M2M process. Simulating the PJM alternative with the 10-unit simulation model initiated by PJM shows that when both RTOs use the PJM proposed alternative, net interchange settlement incentives for a transaction between MISO and PJM is consistent with the target value of the impact represented in the market clearing models. This simulation shows that when both MISO and PJM adopt the PJM alternative, the Transactions Settlement Amount experienced by the transaction is equal to the Target Transaction Settlement Amount representing the incremental financial benefit of the transaction on transmission congestion. Note that due to the modeling differences between the MISO IMM alternative (Table 1) and the PJM alternative (Table 5), the transaction settlement amounts resulting from the methods are not equal. The component of the Transaction Settlement Amount due to energy is the same in both alternatives. The difference is due to modeling and valuation assumption differences but both are internally consistent and provide acceptable net results. The simulation also shows that under the PJM alternative, the congestion component associated with the transmission congestion revenue is collected by both RTOs where it was collected only by the MRTO under the MISO IMM alternative in Table 1. The PJM alternative is compared with the current implementation in Table 5 where MISO is the MRTO. Table 5 Scenario Current implementation Both MISO and PJM adopt PJM alternative Transaction Settlement Amount Target Transaction Settlement Amount MISO Market Congestion Revenue PJM Market Congestion Revenue $2,511.11 $2,041.27 -$871.43 $469.84 $1,711.11 $1,711.11 -$1,671.43 $469.84 4.2 Commercial Flow M2M Process Needed to Support Revenue Inadequacy The current use of market flow as the basis for M2M settlement is not consistent with the PJM alternative. As mentioned previously, market flow is a measure of the constraint flow impact attributed to a RTO s generation serving its load. Under a common interface approach in market clearing with the continued use of the current market flow in M2M coordination, the NMRTO would need to balance congestion revenues (or payments) derived from commercial flow in the market settlement with generation to load flow modeled in the M2M process. These 17 P a g e
flows differ by the constraint impact of the transactions between the NMRTO load center and the common interface. This difference in congestion revenue could be positive or negative depending on the system conditions and the constraint. When the circumstances cause the revenue imbalance derived from the difference in these two flow models to be negative, the NMRTO does not have sufficient congestion revenues to make the necessary payments. This can cause a systematic shortfall in congestion revenue and a revenue inadequacy issue for the NMRTO. The PJM alternative recognizes the importance of revenue adequacy, and the alternative includes the conversion of the M2M process from a market flow based process to a commercial flow based process. With the flow definitions used in both the market clearing and M2M processes both representing commercial flow, the NMRTO can better manage its revenue adequacy. The Table 8shows the NMRTO s matching flow representations for commercial flow in market clearing and commercial flow in M2M coordination used with the PJM alternative. Figure 10 The approach of managing Day-Ahead Market constraint limits consistent with real-time entitlement is similar to the MISO IMM alternative. Under the PJM alternative, the target entitlement is relative to commercial flow rather than market flow. Under the PJM alternative, the M2M process would also change such that the M2M settlement is based on commercial flow. This enables the associated congestion revenue (or payments) resulting from the RTO s internal market settlement and the M2M settlement to offset one another. The use of commercial flow in M2M settlements will require the RTOs to develop auditable realtime commercial flow calculations. This change could also produce unintended cost shifts between the RTOs. Since there is no change to the IDC, the current market flow calculation will still be required. Commercial flow will be an additional new real-time calculation to be managed by each RTO. 4.3 Unintended Impact on Non-M2M Constraints A similar empirical study conducted on non-m2m constraints for the MISO IMM alternative was performed by MISO based on historical data to evaluate difference in shift factor and LMP with 18 P a g e
respect to the non-m2m constraints under the PJM alternative. The results are summarized in the following table. It should be noted that these statistics are based on absolute values of the shift factors. Between Weighted Loads (MISO IMM Proposal) refers to the model under the MISO IMM alternative that uses the congestion modeled from PJM load-weighted location to MISO load-weighted location. MISO s Common Interface refers to the common interface model that uses the same 10 generator units PJM uses for MISO interface with the congestion modeled from the common interface to MISO load-weighted location. Between Weighted Loads (MISO IMM Proposal) PJM Load MISO Load [A] Table 6 MISO s Common Interface PJM Seams MISO Load [B] Percentage Change [A-B]/A Average 0.0333 0.0633 90% StdDev 0.0266 0.0522 96.2% As shown by the Percentage Change, the shift factor of MISO s PJM interface pricing using PJM approach is significantly different (average value changed by 90%, standard deviation changed by 96.2%) from the expected value of the impact calculated from the network model and used in MISO s current model. This distortion of the shift factor representing the constraint impact will distort the congestion impact on non-m2m constraints compared to the current implementation. This study also examined the shift factor changes on a single non-m2m constraint. For this analysis, two extreme cases with both the minimum weighted load difference and the maximum weighted load difference were studied and summarized in the following table. These statistics are based on absolute values. MIN and MAX values of shift factors were found for two different non-m2m constraints. The maximum Percentage Change goes up to 150% indicating a significant shift factor distortion on the non-m2m constraints by employing the common interface approach. Between Weighted Loads (MISO IMM Proposal) PJM Load MISO Load [A] Table 7 MISO s Common Interface for PJM Seams MISO Load [B] Percentage Change [A-B]/A MIN 0.0002 0.0005 150% MAX 0.0670 0.1081 61% 4.4 Inefficient Price Incentive Volatility With the PJM alternative, the aggregate price incentive for a M2M constraint seen by a transaction between the RTOs is calculated from both RTO s constraint shadow prices and common interface shift factors. This is different than the MISO IMM alternative where the transaction price incentive is solely a function of the MRTO shadow price and its load-to-load shift factor. 19 P a g e
When the RTO s real-time shadow prices are the same for the M2M constraint, the PJM alternative price incentive would be the same as the MISO IMM alternative, all other factors being equivalent. Although the M2M process is designed to bring the RTO shadow prices closer together in a static system, the RTO shadow prices are often different as the RTOs adapt to their own dynamic system changes. When the shadow prices for the M2M constraint are not equal, the differences in shadow prices will cause a different transaction price incentive under the PJM alternative. Analysis has shown that the price incentive can be much different between the two alternatives. In fact, the resulting financial incentive can have a sign opposite from the direction expected based on the net shift factor impact of the transaction. The reversal of the incentives results in prices which oppose reliability. The transaction price incentive differences between the approaches can be demonstrated using the COOK_PALISADES345_TWINBRANCH_ARGENTA345 constraint between the dates of January 31 February 25, 2014. During this time period, 226 binding hours were tabulated by the MISO IMM. The average injection-based shift factors in the following table are used in this analysis. Common Interface to MISO Weighted Load [A] Table 8 Common Interface to PJM Weighted Load [B] PJM Weighted Load to MISO Weighted Load [C] = [A] [B] 0.042 0.051-0.009 The net constraint flow impact of a transaction from the PJM weighted load to the MISO weighted load (-0.009) relieves flow on the constraint. A positive congestion price incentive for this constraint when it is binding is consistent with the reliability need to reduce flow on the constraint. Combining the shift factors in the table above with the MISO and PJM hourly shadow prices for the sampled hours, the following table summarizes the transaction price incentives for the MISO IMM and PJM alternatives. Table 9 MISO IMM Alternative Price Incentive PJM Alternative Price Incentive Average ($/MWh) 1.49-0.41 Standard Deviation 1.88 6.78 Negative Incentives None 50% Consistent with reliability needs, the MISO IMM alternative results in a non-negative transaction price incentive in each hour. Application of the PJM alternative to calculate the combined transaction price results in half of the samples providing transaction incentives opposite to reliability needs due to the differing MISO and PJM shadow prices. The average incentive under the PJM alternative is also 20 P a g e
opposite to reliability needs. The volatility (standard deviation) of the price incentive for the PJM alternative is also significantly larger than the MISO IMM alternative. Figure 12 shows the Day-Ahead Market price incentives for a PJM to MISO transaction from the COOK_PALISADES345_TWINBRANCH_ARGENTA345 constraint. An IMM study shows that the resulting transaction incentives are within 100% of efficient pricing target (great than zero and less than twice the efficient target) in only 3% of the hours. Figure 11 5 The impact of the more volatile prices anticipated with the PJM alternative could reduce the efficiency of the market pricing signals for transactions. The risks of price volatility may discourage economic transactions and the potential reversal of price incentives may encourage uneconomic transactions. Both impacts can lead to reliability concerns and can result in increased operating costs and less efficient joint RTO operations. 5 Summary of Assessment The results of the analysis indicate that both of the proposed alternatives could eliminate the over or under counting of congestion seen with the current approach. However, based on its analysis of the alternatives, MISO believes that the PJM alternative has more potential for issues in its fundamental design, such as inaccurate (even opposing) and more volatile price incentives and distortion of the pricing and flow calculations for transmission constraints not coordinated through the Market-to-Market process. These design issues would be expected to be exhibited in the operation of the PJM alternative and would continue to persist as risks and/or operational inefficiencies in the market solution. The MISO IMM alternative appears to be a superior approach. 5 Graph source: Potomac Economics presentation to MISO-PJM Joint and Common Market (JCM) Meeting, 2/19/2015 21 P a g e