April 10, 2019| Markets committee

1. April 10, 2019| Markets committee Andrew Gillespie 413.540.4088 | agillespie@iso-ne.com Discussion of a market-based solution to improve energy security in the region ENERGY SECURITY IMPROVEMENTS: MARKET-BASED APPROACHES

2. Energy Security Improvements WMPP ID: 125 Proposed Effective Date: 2024 In accordance with FERC’s July 2, 2018 orders in EL18-182-000, the ISO must develop improvements to its market design to better address regional fuel security and file by October 15, 2019 Key Projects – Energy-Security Improvements • Discussion paper 2019-04-09 and 2019-04-10 MC A00 ISO Discussion Paper on Energy Security Improvements – Version 1

3. Scope of Today’s Discussion • This presentation focuses on some of the material in the discussion paper, and is intended to highlight particularly important concepts and insights in the discussion paper, and facilitate discussion Please note: ‘Stakeholder questions’ embedded in this presentation and marked this way are a paraphrasing of questions previously asked by two or more stakeholders

4. Today’s Presentation Agenda First, we will briefly review important aspects and components that will be covered in subsequent revisions to the paper and in subsequent presentations Second, we will examine Problem 1 in more depth, including a simple example to demonstrate the problem Third, we will examine the ‘energy on call’ concept for the new ancillary services Finally, we will revisit the example to show the application of the ‘energy on call’ concept, and how it is expected to provide the incentives needed

5. Overview and context

6. Three Conceptual Components • Multi-day ahead market. Expand the current one-day-ahead market into a multi-day ahead market, optimizing energy (including stored fuel energy) over a multi-day timeframe and producing multi-day clearing prices for market participants’ energy obligations • New ancillary services in the day-ahead market. Create several new, voluntary ancillary services in the day-ahead market that provide, and compensate for, the flexibility of energy ‘on demand’ to manage uncertainties each operating day • Seasonal forward market. Conduct a voluntary, competitive forward auction that provides asset owners with both the incentive, and necessary compensation, to invest in supplemental supply arrangements for the coming winter

7. Why the Shift Away from the EIRC Approach? Below are some of the challenges we encountered with the Energy Inventory Reserve Constraint (EIRC) input approach: • How to choose and precisely define the input ‘product(s)’ needed to ensure electrical energy security Stakeholder questions to the ISO regarding EIRC approach: • What are the performance characteristics needed? Parameters such as lead-time, minimum run-time • Where is the cut-off for those who can and cannot ‘perform?’ Continued on next slide

8. Why the Shift Away from the EIRC Approach?– Continued • How to determine failure-to-deliver, and the consequences for failure Stakeholder questions to the ISO regarding EIRC approach: • What are the obligations to perform? • What precisely is the resource expected to do? • What is the quantity to be procured? Stakeholder questions to the ISO regarding EIRC approach: • How would two resources compare/compete to provide performance characteristics? For example, one with 1 MWh for seven days vs. one with 7MWh for one day? • What happens if there is significant under-clearing of demand DA? Continued on next slide

9. Why the Shift Away from the EIRC Approach?– Continued The EIRC design sought to compensate for resource inputs rather than for the actual services (resource outputs) the power system requires • The challenges above are daunting, not robust to technological change, and miss the primary focus: electrical energy security • Hence, the shift to the new ancillary services, which focus on outputs (i.e., electrical energy capabilities) to address electrical energy security

10. What About the Seasonal Forward Market? • This is a longer term forward concept, intended to facilitate investments in energy supply arrangements well in advance of the Multi-day Ahead Market (M-DAM) market horizon • A successful forward market requires a corresponding spot market for the same good or service • The new ancillary services procured in the M-DAM could be, in this context, the ‘spot’ market for this longer term forward market • This component may entail a substantial re-vamping of the existing Forward Reserve Market, making it the longer term forward market for the suite of new ancillary services • Stakeholder discussions on this topic will be held on a separate track

11. Where is the Gap in the Current Products? Stakeholder question: What is it that the market is not currently providing/procuring? Where is the gap in the current products? • Generally speaking, the market currently assumes there will be something else available (in real-time, something without a day-ahead position) to fill the energy ‘gap’ created if, for example, another unit cannot fulfill its day-ahead obligations • This may not be a safe assumption as the system becomes increasingly reliant on just-in-time energy supply resources Continued on next slide

12. Where is the Gap in the Current Products?– Continued Stakeholder question: Doesn’t the information about unit supply offers and characteristics (notification time, for example) provide enough information to meet these needs? • The current suite of energy market products does not provide adequate financial incentives for resource owners in some cases (see prior question above) to make energy supply arrangement investments in advance • Investments that would ensure the resource can provide energy to fill the energy gap when and if needed, and that would be cost-effective and benefit the power system • Unit characteristics (i.e., flexibility) are not fully valued in the current construct • Supply offer information alone does not provide any incentive to make advanced preparations, nor does it provide any assurance the resource can fill any energy gap

13. Three Interrelated Energy-Security Problems • Incentives and Compensation. Market participants whose resources face production uncertainty may have inefficiently low incentives to invest in additional energy supply arrangements, even though such arrangements would be cost-effective from society’s standpoint as a means to reduce reliability risks • Operational Uncertainty. There may be insufficient energy available to the power system to withstand an unexpected, extended (multi-hour to multi-day) large generation or supply loss, particularly during cold weather conditions • Inefficient Schedules. The power system may experience premature (inefficient) depletion of energy inventories for electric generation, absent a mechanism to coordinate and reward efficient preservation of limited-energy supplies over multiple days

14. Notes on the M-DAM Component • M-DAM addresses Problem 3 - Inefficient Schedules • This problem is, to some degree, a consequence of Problem 1 • Solving Problem 3 helps, but will not completely solve, Problem 1 • M-DAM should not be confused with the new ancillary services - generally speaking: • The M-DAM address how to schedule limited supply resources efficiently over a longer day-ahead market horizon (resources that are typically ‘in-rate’ day-ahead) • The new ancillary services address how to ensure there is sufficient energy to fill the energy gap (by resources that are not typically ‘in-rate’ day-ahead) Continued on next slide

15. Notes on the M-DAM Component - Continued • Problem 2 - Operational Uncertainties requires a ‘margin’ to deal with uncertainty, which is the role of (and a high-level way to interpret) the new ancillary services • The margin provided by the new ancillary services is on top of, and does not take the place of better multi-day resource scheduling • The new ancillary services do not address, by themselves, how to use an ‘in-rate’ limited supply resource most efficiently over a longer market horizon; an M-DAM does • The M-DAM, by itself, does not create sufficient incentives to make up-front supply arrangements for resources that are not typically ‘in-rate’ – the new ancillary services accomplish that

16. Problem 1. A More Detailed look

17. Problem 1. Incentives and CompensationA Closer Look • Resources that face production uncertainty may have inefficiently low incentives to invest in additional energy supply arrangements, even though such arrangements would be cost-effective from society’s standpoint, and reduce reliability risk • Resources that do not regularly clear in the day-ahead market, may be called upon by the ISO to operate in real-time, face production uncertainty (e.g., the higher heat rate units and units with low capacity factors) • These resources may not be able to operate unless they have made costly supplemental fuel/energy supply arrangements in advance

18. A Case of Misaligned Incentives • Making supplemental fuel/stored energy arrangements in advance requires and investment (a cost is incurred) • The value that society places on making the supplemental arrangements is based on the high price it avoids with the investment • The value the generator places on the same arrangement is based on the lower price it receives in the energy market (with the investment) • Misaligned incentives: This value difference results in a divergence between the social and private benefit of the investment in supplemental fuel/stored energy arrangements

19. Production Uncertainty: Catch-22 • The very act of investing in a costly supplemental fuel/stored energy arrangement that reduces the risk of shortages can simultaneously prevent high market prices that, at present, are the only means by which a competitive generation owner would receive a return on such an investment

20. Example This section will walk-through a simple numerical example showing the misaligned incentives problem (Problem 1) This example is simplified as much as possible to focus on the essentials

21. Example 1. Setting and Assumptions • This example is a case where the fixed cost of arranging fuel is lower than society’s expected cost savings, but the fixed cost exceeds the generator’s expected profit • As a result, the competitive generator’s rational decision is to not make advance fuel/supply arrangements • But the system would be better off (more efficient) if it did • This example considers a single generator with 1 MW of capacity, for a single hour, without a day-ahead market award • A note on units of measurement: • In this presentation we denote costs in dollars ($) and prices as rates ($/MWh) • Because the example is based on a 1 MW generator for one hour (i.e., 1 MWh) we can add/subtract these terms Continued on next slide

22. Example 1. Setting and Assumptions - Continued • The generator must decide in advance whether or not to incur the cost of supplemental fuel/supply arrangements, knowing that it is uncertain whether the generator will actually be called upon to operate (in real-time) • ‘In advance’ meaning however far in advance as is necessary (a day, a week, a month, or a season) • Note: The use of the term ‘fuel’ in the remainder of the presentation is used for brevity, and should be interpreted as the neutral ‘fuel/energy supply’ term Continued on next slide

23. Example 1. Setting and Assumptions - Continued • If the generator makes supplemental fuel arrangements in advance, an up-front fixed cost of $40 is incurred • This cost will be incurred regardless of whether or not the unit operates • This up-front fixed cost is in addition to the unit’s marginal cost of $70/MWh, which is only incurred if the generator operates

24. Example 1. Possible Market Outcomes • With advanced fuel arrangements the unit may operate • If demand is high (20% probability) the real-time LMP is $120/MWh, and the unit operates • If demand is low (80% probability) the real-time LMP is $50/MWh, and the unit does not operate • Without advanced fuel arrangements the unit will not (cannot) operate • if demand is high (same 20% probability) the real-time LMP is instead $400/MWh • If demand is low (same 80% probability) the real-time LMP is still $50/MWh

25. Cost and Price Assumptions(from ISO Discussion Paper on Energy Security Improvements – Version 1)

26. Example 1. The Generator’s Decision • With advanced fuel arrangements the unit may operate; the cost of advanced fuel arrangements is $40, whether or not the unit operates • If demand is high (20% probability) the real-time LMP is $120/MWh, and the unit operates • If demand is low (80% probability) the real-time LMP is $50/MWh, and the unit does not operate • The generator’s expected profit, if it arranges fuel in advance is a $30 loss • Without advanced supplemental fuel/energy supply arrangements the unit will not (cannot) operate • The generator’s expected profit, if it does not arrange for fuel is $0

27. Expected Net Revenue(from ISO Discussion Paper on Energy Security Improvements – Version 1)

28. Example 1. The Generator’s Decision Takeaways • Accounting for uncertainty, the generator’s expected net revenue is a loss for the up-front $40 investment in fuel arrangements • There is only a 20% chance of earning a gross margin to cover the up-front cost • Arranging fuel in advance is not financially prudent for the generator

29. Example 1. Society’s Preferred Outcome High Demand High demand is expected to occur 20% of the time • When it does, having fuel arrangements made in advance means incurring a marginal cost of $70/MWh instead of $400/MWh • In this example the expected cost savings to the system is $66/MWh 20% x ($400/MWh - $70/MWh) = $66/MWh • From society’s perspective, incurring the $40 cost of the supplemental arrangements is beneficial: $66/MWh - $40 = $26/MWh

30. Example 1. Modify the Assumption Reserve Deficiency Slightly modify the assumptions: • With advanced fuel arrangements the unit may operate • If demand is high (20% probability) the real-time LMP is $120/MWh, and the unit operates • If demand is low (80% probability) the real-time LMP is $50/MWh, and the unit does not operate • Without advanced fuel arrangements the unit will not (cannot) operate • if demand is high (same 20% probability) the real-time LMP is instead $1,400/MWh(high demand and reserve deficiency) • If demand is low (same 80% probability) the real-time LMP is still $50/MWh

31. Example 1. Society’s Preferred Outcome Reserve Deficiency Having fuel arrangements made in advance means incurring a marginal cost of $70/MWh instead of $1,400/MWh • In this variant the expected cost savings to the system is $266/MWh 20% x ($1,400/MWh - $70/MWh) = $266/MWh • From society’s perspective, incurring the $40 cost of the supplemental arrangements is even more worthwhile $266/MWh - $40 = $226/MWh However, things are the same from the generator’s standpoint! It is never paid the high LMP in the scenario where it doesn’t operate

32. Example 1. Society’s Preferred OutcomeTakeaways • The problem of misaligned incentives not only has adverse efficiency consequences, it can also have adverse reliability consequences • The divergence between society’s and the generator’s incentives gets worse (high demand → reserve deficiency) • In both cases society would be better off, and the system’s reliability risk lower, if the generator makes advanced fuel arrangements • However, in both cases arranging fuel in advance is not financially prudent for the generator

33. Example 1. Takeaways • The market price for energy is impacted by the supplier’s investment in advanced fuel arrangements • Investments in energy supply arrangements lowersthe system’s expected total cost to achieve equally (or more) reliable outcomes • However, under the current market design making such fuel supply arrangements is not in the generator’s commercial interest – misaligned incentives! • This misalignment problem will not solve itself – the current energy market cannot produce efficient market outcomes

34. What about Pay-for-Performance? Pay-for-Performance helps, but it does not fully solve Problem 1 • To illustrate why we extend the example, but with only a 1% chance of high demand (i.e., a 1% chance of a Capacity Scarcity Condition), absent the advance fuel arrangements Additional Pay-for-Performance assumptions: • The generator has a Capacity Supply Obligation of 1 MW • The balancing ratio (BR) is 0.8 • The performance payment rate (PPR) is $3,500/MWh

35. Example 1. Pay-for-PerformanceThe Generator’s Situation • With advanced fuel arrangements the unit may operate • If demand is high (1% probability) the real-time LMP is $120/MWh, and the unit operates • If demand is low (99% probability) the real-time LMP is $50/MWh, and the unit does not operate • Without advanced fuel arrangements the unit will not (cannot) operate • if demand is high (1% probability) the real-time LMP is $1,400/MWh, and the generator will incur a non-performance charge • If demand is low (99% probability) the real-time LMP is $50/MWh

36. Example 1. Pay-for-PerformanceThe Generator’s Decision • The generator’s most prudent financial decision is the loss-minimizing one • Arranging fuel involves an expected net loss of $39.50 • Not arranging fuel involves an expected net loss of $28.00 • The prudent course of action is to notincur the up-front cost of fuel arrangements

37. Expected Net Revenue with Pay-for-Performance(from ISO Discussion Paper on Energy Security Improvements – Version 1)

38. Example 1. Pay-for-PerformanceTakeaways • Pay-for-Performance does not fully solve Problem 1 • Society faces a lower reliability risk (as before) if the generator arranges fuel in advance of the operating day, but those arrangements are not consistent with the generator’s commercial interest • As this example illustrates, when the risks of a reserve shortage are low, the performance incentive is significantly reduced, perhaps to the point that the generator is not induced to incur the up-front investment in fuel arrangements • Even though the investment would be efficient (cost effective for the system overall)

39. Root causes

40. Problem 1. Root Causes Example 1 should be viewed as a broader problem with the current energy market construct, with three root causes: • Uncertainty over whether the generating unit will be in demand • Fixed costs of making arrangements for fuel, which must be incurred in advance of learning whether or not the generator will be in demand (asked to operate) • Energy supply arrangements that matter, in the sense that if the generator does not make arrangements for fuel in advance, then with some probability the real-time price for energy will be higher, or reliability will be worse, than if it did Continued on next slide

41. Problem 1 - Root Causes- Continued • In the past, if a unit wasn’t able to operate (for fuel or any other reason) there was (assumed to be) some other unit to dispatch up in its place • This assumption may not be valid in an increasingly energy-limited system, especially during cold winter conditions • The decision to make advanced fuel arrangements could dramatically impact the energy price, especially if demand turns out to be high (or cause a deficiency)

42. Problem 1 - Takeaways

43. Problem 1. Incentives and Compensation Takeaways • Resources that face production uncertainty may have inefficiently low incentives to invest in additional energy supply arrangements, even though such arrangements would be cost-effective from society’s standpoint and would tend to reduce reliability risk • Misaligned incentives:The value difference between what society avoids and what generators are paid (with the investment in advanced fuel arrangements) results in a divergence between the social and private benefit of the investment • Energy supply arrangements matter: If the generator does not make arrangements for fuel in advance, then with some probability the real-time price for energy will be higher, or reliability will be worse, than if the generator did make advance fuel/supply arrangements

44. Key Concepts and Properties of New Ancillary Services This section provides an overview of the key concepts and properties that are similar between the various new day-ahead ancillary services This overview is intended to facilitate stakeholder understanding of these common fundamentals prior to discussing the different detailed design elements of each new ancillary service (in a subsequent presentation)

45. Categories of New Ancillary Services A multiple-day ahead market for energy, co-optimized with three new ancillary service products: • Replacement Energy Reserves (RER) – New product in the day-ahead markets, not in real-time • A ‘call on energy’ product from on-line or off-line gen • 1+ hour ramp-up/start-up capability • Generation Contingency Reserves (GCR) – Forward form of existing TMSR, TMNSR, and TMOR products • Energy Imbalance Reserves (EIR) – New product in day-ahead markets • A ‘call on energy’ product to cover the load forecast in excess of day-ahead cleared load (net virtual bids/offers) Combined, these provide the ‘margin for uncertainty’ in an increasingly energy-limited system

46. Design Objectives for a Market-Based Solution • Risk Reduction. Minimize the heightened risk of unserved electricity demand during New England’s cold winter conditions by solving Problems 1, 2, and 3 • Cost Effectiveness. Efficiently use the region’s existing assets and infrastructure to achieve this risk reduction in the most cost-effective way possible • Innovation. Provide clear incentives for all capable resources, including new resources and technologies that can reduce this risk effectively over the long term

47. Design Principles for a Market-Based Solution • Product definitions should be specific, simple, and uniform. The same well-defined product or service should be rewarded, regardless of the technology used to deliver it • Transparently price the desired service. A resource providing an essential reliability service (for instance, a call on its energy on short notice) should be compensated at a transparent price for that service • Reward outputs, not inputs. Paying for obligations to deliver the output that a reliable system requires creates a level playing field for competitors that deliver energy reliably through cold-weather conditions • Sound forward markets require sound spot markets. Forward-market procurements work well when they settle against a transparent spot price for delivering the same underlying service • Compensate all resources that provide the desired service similarly.

48. Common Parameters of New Ancillary Services (Energy Options)

49. Goals Two broad goals in creating a market product to solve the misalignment problem (and Problem 2) • Compensate the supplier such that it will be willing to incur the fixed costs of arranging energy supplies in advance when that would be cost-effective from the system’s standpoint • This first sub-section covers the construct generally (product definition concept, quantities, etc.) • Tie compensation to consequences, so that the suppler will be induced to incur the fixed costs of arranging energy supplies whenever efficient (for the system) • The second sub-section covers the settlement rules, which show the consequences of not covering the call option

50. New Ancillary ServicesProduct Definitions: Concept • At a high level, a day-ahead seller of these ancillary services is providing the ISO with a call option on its resource’s energy during the operating day • Different time-related parameters (e.g., generator ramp or startup times) will be relevant to the awards of different ancillary service products • The resource capabilities necessary to provide (i.e., be awarded) each day-ahead ancillary service will differ by product (GCR, RER, and EIR) • The clearing process would be designed to select the most valuable assignment of offers to awards • For a given MWh the obligation may be for energy, or GCT, etc.