2001 Annual Report on the New York Electricity MarketsPresented to:Federal Energy Regulatory CommissionDavid B. Patton, Ph.D.Independent Market AdvisorJune 26, 2002Executive Summary of 2001 Annual Report-3-Executive SummarySummary of Trends and Findings from 2001•Lower fuel prices and reduced generation outages in Eastern New York led to lower energy prices statewide and substantially less congestion.•Analysis of the market conduct of both suppliers and load-serving entities indicates that the markets have been workably competitive.•Price convergence between the day-ahead and real-time has improved, due in part to the implementation of virtual trading in the day-ahead market.•Virtual trading activity grew initially before leveling-off in the Spring 2002. The vast majority of virtual offers and bids have been price-sensitive with no evidence of attempts to strategically manipulate day-ahead prices. •The frequency of price corrections and reservations has decreased considerably in 2001, improving the price certainty provided by the NYISO markets.-4-Day-Ahead Energy Price and Fuel Price TrendsJanuary 2000 to December 2001$0.00$15.00$30.00$45.00$60.00$75.00Jan &FebMar &AprMay& JunJuly &AugSept& OctNov &DecJan &FebMar &AprMay& JunJuly &AugSept& OctNov &DecEnergy Price ($/MWh)050100150200250Natural Gas Price Index (Jan 2000 = 100) West PricesEast PricesNatural Gas Prices2000 2001 -5-0.00%2.00%4.00%6.00%8.00%10.00%12.00%14.00%Pricing Intervals Corrected (%)JanFebMarAprMayJunJulAugSepOctNovDecPercentage of Real-Time Prices CorrectedJanuary 2000 to May 2002200020012002Annual Average of Intervals Corrected: 2000 = 3.09 % 2001 = 0.37 % 2002 = 0.19 % (Year-to-Date)-6-* Includes hours beginning at 1pm through 5pm, Monday through Friday.Relationship of Deratings to Actual LoadDay-Ahead Market --East New York Summer 2001 --Peak Hours*2,0003,0004,0005,0006,00012,00014,00016,00018,00020,00022,000Actual Load (MW)Derated Capacity (MW)-7-Executive SummaryOngoing Improvements to Address Market Issues •A number of modeling changes have been made to address poor convergence of prices produced by the BME (hour-ahead scheduling model) and the SCD (real-time dispatch models) under peak conditions.This pricing inconsistency can cause inefficient scheduling of external transactions and off-dispatch generation. The effects of this issue on the markets in 2001 included inefficiently low prices in peak hours and inflated uplift under peak conditions. The longer-term solution is an improved real-time scheduling and dispatch system to be implemented in 2003. •Modeling changes were made in June 2002 to manage constraints within NYC, which has improved the locational prices signals and decreased uplift. Out-of-merit dispatch of generation to manage congestion within NYC resulted in depressed prices in peak hours and inflated uplift in NYC in 2001.The net effect on customers in NYC would include the net of the changes in locational prices, uplift amounts, TCC revenues, and NYC capacity prices.Out-of-merit dispatch of generation has been reduced by 90 percent.These changes were made in conjunction with mitigation measures to address potential locational market power. -8-* Includes hours beginning at 1pm through 5pm, Monday through Friday.Relationship of Price Differences to Actual LoadHour Ahead Prices Minus Real Time PricesNew York City --2001, Peak Hours*-$500$0$500$1,0004000500060007000800090001000011000Actual Load (MW)HAM-RT Price Difference ($/MWh)-9-Executive SummaryOngoing Improvements to Address Market Issues (cont.) •Changes have also been made to facilitate trading in the Northeast, including:Implementing reserve sharing with ISO-NE;Development of the Collaborative Scheduling System to improve the scheduling process and communications between PJM and New York;Allowing multi-hour block transactions that allow transactions to be submitted in New York with a minimum run-time to reduce scheduling risks;Establishing a pre-scheduling process in New York to reduce scheduling risks associated with long-term transactions;Implementing an inter-ISO congestion mgt. pilot program with PJM to allow redispatch of generation in one ISO to resolve congestion in the adjacent system; ISO-NE modifying its rules for scheduling short-notice exports to New York; and Revising the Hour-Ahead model to schedule external transactions more efficiently.•Given the timing of the changes, many of these improvements would not be reflected in the 2001 results.•While progress has been made in eliminating barriers, impediments to trading in the Northeast remain, and implementing proposed rules changes and systems to facilitate efficient trading should remain a high priority.-10-Executive SummaryRemaining Issues and Recommendations •Relatively low participation in the ancillary services markets remains an issue that can create significant inefficiencies under peak conditions –pricing reforms are recommended to improve incentives, including:Setting prices at the system marginal cost of procuring the reserves –this would cover the opportunity costs borne by suppliers held out of the energy market.Establishing a demand curve for reserves that would prevent the ISO from taking excessively costly actions to maintain low value reserves and, more importantly, set reserves andenergy prices at efficient levels during capacity shortages.Implementing a multi-settlement system (day-ahead and real-time) for reserves.Conditionally lifting the offer cap currently in place for 10-minute non-synchronous reserves based on our competitive analysis of this market.•The ICAP results in New York City were not consistent with competitive expectations –prices did not reflect the substantial capacity surpluses that emerged in the winter 2001-2002.Consideration of modifications to the current capacity market is recommended in the context of the ongoing work to standardize capacity markets in the Northeast. -11-*10 minute reserves includes only capability in Eastern New York due to locational reserve requirements.Ancillary Services Capability and Offers05001000150020002500Excl.PURPAAllUnitsExcl.PURPAAllUnitsExcl.PURPAAllUnitsExcl.PURPAAllUnits10 Min Spin*10 Min Nspin*Regulation30 Min ReservesRegulation and 10 Minute Reserves (MW)0300060009000120001500030 Minute Reserves (MW)AverageCapabilityApproximateDemandAverageOffer2001 Annual Report on the New York Electricity Markets-13-Introduction to the Annual Report•This presentation provides highlights from the Annual Report on the New York electricity markets for 2001.•The market assessment addresses the following areas: Energy market prices and outcomesMarket participant bidding patternsInstalled capacity marketExternal transactions Ancillary servicesMarket Prices and Outcomes-15-Energy Prices in the Day-Ahead Market•The following chart shows average prices during all hours in 2000 and 2001.•The trend in electricity prices in 2000 and 2001 were driven by the trends in fuel prices.Electricity prices in New York decreased 52% from January to December 2001.This is primarily due to substantial decreases in the prices of input fuels over the same period, 40% for fuel oil and 70% for natural gas.•Load was higher in 2001, particularly during the summer. However, both summers exhibited capacity shortages that resulted in price spikes.These shortage prices play an important role in sending efficient signals to the market –prices during 3 hours on August 9, 2001 raised the average price for the month by 20 percent.•The average price in the East remained 30% higher than in the West due to continued congestion on the Central-East Interface and into New York City.-16-Day-Ahead Energy Price and Fuel Price Trends January 2000 to December 2001 $0.00 $15.00 $30.00 $45.00 $60.00 $75.00 Jan & Feb Mar & Apr May & Jun July & Aug Sept & Oct Nov & Dec Jan & Feb Mar & Apr May & Jun July & Aug Sept & Oct Nov & Dec Energy Price ($/MWh) 050 100 150 200 250 Natural Gas Price Index (Jan 2000 = 100) West Prices East Prices Natural Gas Prices 2000 2001 -17-Congestion Revenue •The following chart compares the monthly congestion revenue generated in 2000 with 2001.•Congestion revenue is net revenue collected from loads and bilaterals (net of congestion payments to generators).•The reduction in congestion is due to:The return of Indian Point 2 (1000 MW) in Eastern New York;Increased imports from New England;Lower oil and gas prices in 2001; andReduced limit on imports over the HQ proxy bus.-18-$0$20$40$60$80$100$120Monthly Cost (Million $)JanFebMarAprMayJunJulAugSepOctNovDecTotal Congestion Costs: 2000 vs. 2001200020012000 Total = $517 Million2001 Total = $310 Million-19-Energy Price Corrections•All real-time energy markets are subject to some level of price corrections to account for metering errors and other input data problems –these problems should not be frequent.•Corrections can also result from software flaws that cause pricing errors under certain conditions –it is important to resolve these flaws as quickly as possible to maximize price certainty.•The following figure summarizes the frequency of price corrections in the real-time energy market from January 2000 to May 2002. •The figure shows that after the initial transition to the NYISO energy markets, the frequency of the price corrections decreased markedly.-20-0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% Pricing Intervals Corrected (%) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Percentage of Real-Time Prices Corrected January 2000 to May 2002 2000 2001 2002 Annual Average of Intervals Corrected: 2000 = 3.09 % 2001 = 0.37 % 2002 = 0.19 % (Year-to-Date)-21-Summary of Mitigation in 2001•Day Ahead MarketMitigation by the AMP occurred on 4 days: July 23rdand 24th, August 2ndand August 10th. Mitigation occurred only on Long Island on two of the days and in a broader area on the other two days.Mitigation under the ConEd Mitigation for New York City occurred in all but 17 days for energy bids and all but 4 days for reserves bids.•Real-Time MarketMitigation in the real-time energy market not related to Thunderstorm Alerts occurred on 6 days.Mitigation in the real-time related to Thunderstorm Alerts occurred on 7 days. All thirteen cases occurred during the Summer 2001.-22-Price Impacts of Out-of-Merit Dispatch in NYC•One of the key changes to the market for 2002 involves making changes to SCUC, BME, and SCD to recognize the transmission constraints within NYC.•The addition of new constraints will raise the locational prices in the load pockets and reduce the out-of-merit dispatch previously used to manage the constraints.•The following chart shows my estimate of the price impact of these modeling changes in the real time market. The chart includes:An adjusted price for NYC, computed by estimating a locational price in each load pocket equal to the highest priced (lower of ref. price or bid) generator dispatched out-of-merit in the pocket.The local reliability uplift for NYC on a MWh basis, computed by dividing the real time uplift by the total MWhs consumed.•These changes will provide more accurate price signals to generators within the NYC load pockets and a greater ability for participants to hedge these costs.•The net effect on customers in NYC would include the net of the changes in locational prices, uplift amounts, TCC revenues, and NYC capacity prices.-23-Estimated Price Increase in NYC Due to Load Pocket ModelingJune to December 2001$0$10$20$30$40$50$60$70$80$90 June July Aug Sept Oct Nov DecDollars per MWhEstimated price increase due to modeling changes.Local Reliability Uplift for NYC per MWhActual NYC Real-Time PriceAdjusted Real-Time Price-24-Summer 2002 Price Forecasts•The following figure provides a forecast for prices this summer, given the changes in supply and demand that have occurred over the past year.•This analysis is based on the NYISO’s load forecasts for the summer –the extreme weather case increases the peak demand by 900 MW over the normal forecast. •Fuel prices are assumed in each of the forecasts to be unchanged from last summer since the current fuel prices are close to last summer’s average levels.•A load pocket effect on NYC of 12.5 percent is included based on the prior analysis (the anticipated reduction in uplift costs is not included). •While changes in overall loads levels are also important, the figure shows that frequency of price spikes is a key determinant of the price levels.-25-$0$10$20$30$40$50$60$70$80Price ($ per MWh)2001Actual2002NormalWeather2002ExtremeWeather2001Actual2002NormalWeather2002ExtremeWeatherSummer 2002 Energy Price Forecast June to August --All HoursLoad Pocket EffectBase PriceNew York StateEast New York3% Increase30% Increase28% Increase2% Increase Price Spike Hours*2001 Actual = 6 hours2002 Normal = 2 hours2002 Extreme = 16 hours* Price spike hours are defined as hours with projected prices greater than $500 per MWh. Hours shown are for East New York.Sources: NYISO actual day-ahead price data and load forecasts; Potomac Economics analysis. All Prices shown are load-weighted.Market Performance-27-Supply Conditions and Prices in New York•The typical ―L‖ shape of the supply curve should cause prices in a well-functioning market to rise sharply under high load conditions when excess capacity is close to zero (i.e., shortage conditions).•I define excess capacity as the derated capability minus scheduled energy, ancillary services, and economically unavailable resources.This formula incorporates the effects of scheduled exports and imports. Economically unavailable resources are those whose offer prices were substantially above accepted offer prices during workably competitive periods. •Therefore, all substantial increases in prices should occur when the excess capacity quantities are very low, which has been the case during 2001.-28-Relationship of Day-Ahead Prices to Excess CapacityEast New York --Peak Hours*January 1 to December 31, 2001$0$200$400$600$800$1,00002,0004,0006,0008,00010,00012,000Excess Capacity (MW)Day-Ahead Price ($/MWh)* Includes hours beginning from 1pm to 5pm, Monday through Friday.-29-Day-Ahead and Real-Time Energy Prices•The following three charts show monthly average day-ahead and real-time energy prices in West NY, East NY outside NYC, and NYC.•The results show that a slight premium in the day-ahead market remains in in all three areas.•Price convergence improved in 2001 compared to 2000, particularly in the fall of 2001. Contributing factors likely included:Lower congestion in the State and overall price volatility.The introduction of virtual bidding in November and increased activity in price capped load bidding.Lower fuel prices in the fall of 2001.-30-Monthly Average Day Ahead and Real Time PricesWest New York --2001$0$10$20$30$40$50$60$70$80 Jan Feb March April May June July Aug Sept Oct Nov DecPrice/MWhDay AheadReal Time-31-Monthly Average Day Ahead and Real Time PricesEast New York Above NYC --2001$0$10$20$30$40$50$60$70$80 Jan Feb March April May June July Aug Sept Oct Nov DecPrice/MWhDay AheadReal Time-32-Monthly Average Day Ahead and Real Time PricesNew York City --2001$0$10$20$30$40$50$60$70$80 Jan Feb March April May June July Aug Sept Oct Nov DecPrice/MWhDay AheadReal Time-33-Energy Price Statistics•The following table shows annual price statistics for prices in the West, Capital and NYC zones for 2001.•The results show a 10 percent premium in day-ahead prices versus real time prices in the West, while premiums exist in the other two zones of less than 1 percent.These premiums likely reflect the higher risk to loads of purchasing in the more volatile real-time market and lack of TCC’s to hedge congestion in real-time, as well as the outage risk of generators associated with day-ahead schedules. •Convergence between day-ahead and real-time prices improved between 3 and 6 percent in the three zones (e.g., the price difference in the Capital zone fell from a 7% premium in 2000 to less than 1% in 2001)•The table also shows the standard deviations, which are the average of the monthly standard deviations for each hour of the dayThe standard deviations have fallen from 2000 levels.The volatility in the real-time prices remains roughly twice as high as day-ahead.-34-* Average of standard deviations calculated by month and hour of day.Table 1Day-Ahead and Real-Time Pricing Statistics for Selected ZonesJanuary to December 2001New York CityCapital ZoneWest ZoneDay-AheadReal-TimeDay-AheadReal-TimeDay-AheadReal-TimeMean$44.67$44.49$39.90$39.69$33.87$30.76Compared with 2000-$4.16-$5.85-$4.92-$2.36-$0.59$0.88Avg. Std. Deviation*$12.25$30.68$10.68$24.64$9.15$15.82Compared with 2000-$4.58-$15.38-$6.20-$2.60-$0.12-$3.15Minimum$0.11-$169.37$0.10-$167.80$0.10-$152.35Maximum$1,024.91$1,034.01$976.15$1,078.35$912.28$949.50-35-Hour-Ahead Prices and Uplift•These charts show the difference between hour-ahead and real-time prices at actual load levels during 2001.•The difference in Hour-Ahead and Real-Time prices derive from the differences in the models:SCD does not model all of the constraints in NYC that are modeled in the SCUC and BME models.30 minute reserves are treated differently in the BME and SCD models.External transactions and other resources scheduled hourly can set prices in the BME, but not in the SCD.•The Eastern New York chart show that the largest differences occur under the highest load conditions.•However, the NYC chart shows that substantial differences occur at mid-load levels. This likely reflected the out-of-merit dispatch that was used in SCD to manage constraints that are modeled in the BME (138 kv load pocket).-36-* Includes hours beginning at 1pm through 5pm, Monday through Friday.Relationship of Price Differences to Actual LoadHour Ahead Prices Minus Real Time PricesEast New York Above NYC --2001, Peak Hours*-$500$0$500$1,000250030003500400045005000550060006500Actual Load (MW)HAM-RT Price Difference ($/MWh)-37-* Includes hours beginning at 1pm through 5pm, Monday through Friday. Relationship of Price Differences to Actual Load Hour Ahead Prices Minus Real Time Prices New York City --2001, Peak Hours* -$500 $0 $500 $1,0004000 5000 6000 7000 8000 9000 10000 11000 Actual Load (MW) HAM-RT Price Difference ($/MWh)-38-Price Divergence and Uplift•Solutions to eliminate these differences are a high priority because they contribute to:•Higher uplift costs; and•Inefficiently low real-time energy prices in peak hours;•The subsequent chart shows how the sizable uplift payments—especially for externals—coincide with periods when prices diverge significantly.•The uplift costs related to externals occur when BME accepts expensive imports that are uneconomic relative to the real-time price.•Similarly, uplift costs related to internals occur when BME commits high cost generation that is uneconomic relative to the real-time price.•The modeling changes and reserve market changes to address this divergence are planned prior to Summer 2002.-39-0.0%0.5%1.0%1.5%2.0%2.5%3.0%3.5%4.0%4.5%5.0%Percent of Total Market ExpensesJanFebMarAprMayJunJulAugSepOctNovDecReal-Time Non-Reliability Uplift Expenses in 2001Uplift to External SuppliersUplift to Internal SuppliersUplift Generated on Peak Days*Uplift Generated on Peak Days** Days where the HAM price exceeded $1000/MWh for more than one hour in a zone other than Long Island.Analysis of Bidding Patterns-41-Analysis of Offer Patterns•Due to the nature of the supply in wholesale power markets, the incentive to withhold is generally highest under peak demand conditions when prices are most responsive to shifts in supply.Suppliers in a competitive market should increase offer quantities during higher load periods to sell more power at the higher peak prices;Suppliers with market power will have an incentive to offer less at peak load levels when the market impact is the largest.•Therefore, the following charts show the correlation of potential withholding to actual load levels. Potential withholding is defined as offers exceeding the mitigation thresholds for economic withholding and deratings for physical withholding.•The trends shown in these charts are consistent with the hypothesis that the New York markets have been workably competitive during 2001. -42-Relationship of Actual Load to Offers Exceeding Economic Withholding ThresholdsDay-Ahead Market --East New York January 1 to December 31, 2001 --3pm Hour05001,0001,5002,0002,5008,00010,00012,00014,00016,00018,00020,00022,000Actual Load (MW)Offers Above Threshold (MW)-43-* Includes hours beginning at 1pm through 5pm, Monday through Friday. Relationship of Deratings to Actual Load Day-Ahead Market --East New York Summer 2001 --Peak Hours* 2,000 3,000 4,000 5,000 6,00012,000 14,000 16,000 18,000 20,000 22,000 Actual Load (MW) Derated Capacity (MW)-44-Analysis of Load Bidding Patterns•The NYISO also monitors the bidding patterns of load-serving entities as specified in the mitigation plan.•The following charts show the load bidding patterns during 2001 in the entire state and in New York City•These charts show the following:Price-capped load bidding was much more active than in 2000, particularly in November and December, coinciding with the implementation of virtual bidding.The percent of the actual load supplied by physical bilaterals has been relatively constant.Virtual loads scheduled in November and December substantially closed the gap between day-ahead load scheduling and actual load that had been present in prior months.•A similar chart for Eastern New York, and a chart comparing the 2001 bidding patterns to 2000 are contained in the Appendix.-45-Composition of Day Ahead Load Bids as a Proportion of Actual LoadNew York State --20010%20%40%60%80%100%120%140%JanFebMarchAprilMayJuneJulyAugSeptOctNovDecPercent of Actual LoadPrice-Capped Bid Load -UnscheduledNet Virtual PurchasePrice-Capped Bid Load -ScheduledDay-Ahead Bid LoadPhysical Bilaterals-46-Composition of Day Ahead Load Bids as a Proportion of Actual LoadNew York City and Long Island --20010%20%40%60%80%100%120%140%JanFebMarchAprilMayJuneJulyAugSeptOctNovDecPercent of Actual LoadPrice-Capped Bid Load -UnscheduledNet Virtual PurchasePrice-Capped Bid Load -ScheduledDay-Ahead Fixed LoadPhysical Bilaterals-47-Analysis of Bid Patterns•Some had raised concerns that load-serving entities may intentionally under-bid their loads to cause the day ahead market to clear at depressed prices.•The following scatter diagram shows the load bidding patterns during 2001 in three areas.•This chart shows the following:Under-bidding by load is least pronounced as one moves from west to east with loads in NYC purchasing in excess of their actual load on average.Under-bidding did not increase under peak load conditions.-48-Percentage of Load Scheduled Day-Ahead versus Real-Time LoadEast New York --2001, Peak Hours70%80%90%100%110%120%130%70001000013000160001900022000Actual LoadPercentage Scheduled Day-AheadMean = 99.4%Two Standard Deviations-49-Percentage of Load Scheduled Day-Ahead versus Real-Time LoadNew York City and Long Island --2001, Peak Hours70%80%90%100%110%120%130%40006000800010000120001400016000Actual LoadPercentage Scheduled Day-AheadMean = 101.8%Two Standard Deviations-50-Percentage of Load Scheduled Day-Ahead versus Real-Time LoadWest New York --2001, Peak Hours70%80%90%100%110%120%130%40005000600070008000900010000Actual LoadPercentage Scheduled Day-AheadMean = 91.1%Two Standard Deviations-51-Virtual Bidding Patterns•Virtual bidding was introduced in November to allow participation in the day-ahead market by entities other than LSE’s and generators.•The following figures show the quantities of virtual load and supply quantities that have been offered and scheduled on a monthly basis in the State and in NYC.•This chart shows the following:Virtual load bids rose initially in both areas and have leveled off in March.Virtual suppliers have become much more active in the spring of 2002.Virtual loads have been larger than supply, raising the total day-ahead schedules as shown in the prior figures and, in part, displacing some of the price-capped load bids by the LSEs.-52-Hourly Virtual Bidding of Load and Supply, Scheduled and UnscheduledNew York State --November 2001 to March 200202004006008001,0001,2001,400LoadSupplyLoadSupplyLoadSupplyLoadSupplyLoadSupply Nov 2001 Dec 2001 Jan 2002 Feb 2002 Mar 2002MegawattsNet VirtualPurchasesScheduled QuantitiesUnscheduled Quantities-53-Hourly Virtual Bidding of Load and Supply, Scheduled and UnscheduledNew York City --November 2001 to March 20020100200300400500600VirtualLoadVirtualSuppVirtualLoadVirtualSuppVirtualLoadVirtualSuppVirtualLoadVirtualSuppVirtualLoadVirtulSupp Nov 2001 Dec 2001 Jan 2002 Feb 2002 Mar 2002MegawattsNet VirtualPurchasesScheduled QuantitiesUnscheduled Quantities-54-Virtual Bidding Patterns•Virtual bidding allows participants to arbitrage differences between the day-ahead and real-time energy prices, or hedge day-ahead risks.•Virtual bids submitted for this purpose will be price-sensitive and should improve the convergence of the day-ahead and real-time prices.•Price-insensitive virtual bids could be used to attempt to manipulate day-ahead prices.•The following chart uses fixed thresholds of $100 for virtual load bids and $5 for virtual supply offers to determine the shares of virtual bids that have been price insensitive on a monthly basis.•This chart shows that the vast majority of the bids have been price sensitive in the early months of virtual trading.•Virtual trading will continue to be monitored closely –initial conclusions regarding its operation should include an evaluation of the Summer 2002 to include peak conditions when the market is tight. -55-Price Sensitivity of Scheduled Virtual Load Bids and Supply OffersNew York State --November 2001 to March 20020%10%20%30%40%50%60%70%80%90%100%LoadSupplyLoadSupplyLoadSupplyLoadSupplyLoadSupply Nov 2001 Dec 2001 Jan 2002 Feb 2002 Mar 2002Percent of Scheduled Bids and OffersNot Price SensitivePrice SensitiveImports and Exports-57-Assessment of Imports and Exports•This section provides an update of the analysis for 2001 of external transactions provided in the Annual Assessment for 2000.•The analysis in this section has two focuses:First, it seeks to assess the extent to which the interfaces with neighboring markets in the Northeast are rationally utilized; andSecond, it analyzes the results of the NYISO’s import and export scheduling process to determine whether the NYISO market design has been an impediment to trading.-58-Utilization of the Interfaces•The following three charts plot the hourly difference in prices between New York and neighboring markets against the available import capability during hours when transmission constraints are not binding.•The price differences plotted against the left axis are always computed by subtracting the external price from the New York price (i.e., positive price differences mean prices are higher inside New York).•The available import capability is computed in the following manner: Total Transfer Capability -Net Scheduled ImportTherefore, when the NYISO is exporting (net scheduled import is negative), the available import capability will exceed the total transfer capability;The vertical dashed line is shown at the approximate TTC level for each interface --so higher points (to the right) generally represent exports while lower points (to the left) generally represent imports.•The counter-intuitive net schedules are a) net exports when NYISO prices exceed the adjacent market or b) net imports when NYISO prices are lower than adjacent prices.-59-* Price at PJM Western HubDifference Between West Zone and PJM Price* During Unconstrained HoursHour-Ahead Market --January to December 2001-$100-$50$0$50$1000500100015002000250030003500400045005000Available Import Capability (MW)NY Price -PJM Price ($/MWh)Month Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Mean $4.16 $2.53 $5.68 $4.00 $6.84 $7.79$22.57$28.24$16.23 $5.07 $2.91 $3.10Monthly Price StatisticsStd Dev$22.49$18.34$42.35$18.98$56.86$12.63$112.78$130.14$60.49$6.92$8.21$7.59Net ExportsNet Imports-60-* Price at PJM Western HubDifference Between West Zone and PJM Price* During Unconstrained HoursDay-Ahead Market --January to December 2001-$100-$50$0$50$1000500100015002000250030003500400045005000Available Import Capability (MW)NY Price -PJM Price ($/MWh)Month Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Mean $6.16 $4.33 $0.25 $1.32 $5.39 $2.27 $4.13 $0.72 $5.62 $3.31 $3.46 $1.72Monthly Price StatisticsStd Dev$12.54$8.30$11.07$9.79$10.01$7.44$10.14$46.41$3.79$4.49$4.40$5.08Net ExportsNet Imports-61-Difference Between New York and ISO-NE Price in Unconstrained HoursReal-Time Prices vs. Hour-Ahead Schedules --January to December 2001-$100-$50$0$50$1000300600900120015001800210024002700Available Import Capability (MW)NY Price -NE Price ($/MWh)Month Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Mean-$12.27 $1.11 -$1.21 $0.92 $11.84 -$1.66-$12.37 $11.24 $1.32 $0.24 $0.57 -$0.25Monthly Price StatisticsStd Dev$29.61$35.01$26.77$17.98$79.34$27.82$109.53$83.01$22.97$10.60$12.03$13.41Net ExportsNet Imports-62-NYISO Scheduling of External Transactions•The following bar charts show a comparison between the hourly average import transactions from New England that were scheduled or unscheduled in each month during 2000 and 2001. •Those that are unscheduled are divided between those that are uneconomic versus those not scheduled for other reasons (e.g., failed checkout process).•The chart shows that imports have fallen from the prior year.•However, much larger quantities are being offered. This suggests an improvement in the attempt to arbitrage substantial price differences, which will benefit both New York and New England.•Because the Central-East interface has been less congested, the prices in Eastern New York have provided lower incentives for imports.-63-0100200300400500MegawattsJan to AprMay to AugSep to DecJan to AprMay to AugSep to Dec2000 2001Imports from New EnglandHour-Ahead Market, 2000 vs. 2001ScheduledImportsUnscheduled Offers -Not EconomicUnscheduled Offers -Other-64-Assessment of Imports and Exports•These results continue to suggest that impediments to efficient arbitrage remain with adjacent markets, particularly in real-time.•Substantial changes are being made to improve scheduling with adjacent markets:Changes in the short-notice transaction rules in New England;Changes in BME that will improve export scheduling in peak hours;Seams agreement with PJM;Implementation of pre-scheduling provisions and multi-hour block transactions;•The analysis of bidding and curtailments reveals the following:The vast majority of the unscheduled transactions are unscheduled because they are not economic. Virtually none of the unscheduled transactions in the DAM and a very small share in the HAM are not scheduled for reasons other than economics and are often the result of transactions being withdrawn in one of the areas.Installed Capacity Market-66-Installed Capacity Market•The capacity market is intended to provide an economic signal to provide an efficient incentive for new investment and for capacity that is seldom utilized to remain in operation (e.g., peaking capacity). •The following chart shows the capacity amounts designated in New York State (exclusive of capacity sold externally), as well as the ―Rest of State‖ capacity prices.•The amounts shown as ―unsold‖ were not sold within or outside of New York.•This figure shows that when a capacity surplus existed statewide, capacity prices fell sharply as expected.-67-3200034000360003800040000MegawattsMayJuneJulyAugSeptOctNovDecJanFebMarAprInstalled Capability Market -New York StateMay 2001 to April 2002Internal Capacity UnsoldExternal CapacitySCR and Load Mgt.Internal Capacity SoldSummer 2001 Capability PeriodWinter 01-02 Capability Period$0$2$4$6$8$10MayJuneJulyAugSeptOctNovDecJanFebMarAprICAP Price ($/MW-Mo.)Strip PriceMonthly Price-68-Installed Capacity Market•The following chart show the capacity amounts designated in New York City (not limited to amounts sold to meet the NYC requirement), as well as the NYC capacity prices.•The increase in sales shown during Summer 01 corresponds to the new GTs that were installed in NYC.•In contrast to the statewide results, this chart shows that when surpluses existed during the Winter 01-02 in NYC, capacity prices remained at levels substantially higher than the likely marginal cost of the capacity.Marginal cost of providing capacity should not to be confused with the marginal cost of producing energy –it should generally equal the expected costs of the ICAP obligations accepted by the generator.This result is not consistent with competitive expectations.•New proposals should be considered regarding the structure of this market. •For example, establishing a forward capacity requirement that can be met by existing capacity or new investment may improve the competitive performance of the market.-69-700075008000850090009500MegawattsInstalled Capability Market -New York CityMay 2001 to April 2002Internal Capacity UnsoldSCR and Load Mgt.Internal Capacity SoldSummer 2001 Capability PeriodWinter 01-02 Capability Period$0$3$6$9$12MayJuneJulyAugSeptOctNovDecJanFebMarAprICAP Price ($/MW-Mo.)Strip PriceMonthly PriceAncillary Services Markets-71-Ancillary Services•The following chart shows the share of the total market expenses that are accounted for by ancillary services.•These expenses are within expected levels based on experience in other markets.•The increase in regulation costs as a share of market expenses is due in part to the fact that market expenses in total decreased sharply in the fall of 2001. •The chart also shows the cost increases for ancillary services that occur associated with peak conditions as these markets must compete for resources with the energy market.-72-0.0%1.0%2.0%3.0%4.0%Percent of Total Market ExpensesJanFebMarAprMayJunJulAugSepOctNovDecReserves and Regulation Expenses in 2001ReservesRegulationCosts without the Peak Days**Two days in July and four days in August where the HAM price exceeded $1000/MWh for more than one hour outside of Long Island.-73-Ancillary Services•Ancillary services markets are generally not tight because offers to supply typically exceed approximate demand:For 30 minute reserves, offers typically exceed approximate demand by 380 percent (almost five times the demand);For 10 minute NSR, offers typically exceed approximate demand by 160 percent --although this market currently is subject to a requirement to sell and a bid cap;For regulation and 10 minute spinning reserves, offers typically exceed approximate demand by 75 percent –but ignores the fact that some 10 minute spinning reserves can be purchased in the West;•However, since these markets are jointly optimized and the same resources are offered in multiple markets, energy and other AS markets can bid resources away from a given service resulting in relatively tight conditions.-74-*10 minute reserves includes only capability in Eastern New York due to locational reserve requirements. Ancillary Services Capability and Offers 0 500 1000 1500 2000 2500 Excl. PURPA All Units Excl. PURPA All Units Excl. PURPA All Units Excl. PURPA All Units 10 Min Spin* 10 Min Nspin* Regulation 30 Min Reserves Regulation and 10 Minute Reserves (MW) 03000 6000 9000 12000 15000 30 Minute Reserves (MW) Average Capability Approximate Demand Average Offer-75-Ancillary Services Summary•A substantial amount of capability is routinely not offered in the reserve markets.•Conditions become tight in the ancillary services markets in peak hours as the energy market and reserve markets compete for the same resources. This can occur in off-peak hours when a large share of the capacity is offline.•When this occurs, the shortage of reserves offers raises energy prices by causing relatively economic energy supplies to be diverted into the ancillary services markets.•The sustained nature of the low offers into the reserve markets indicates that the incentives to offer in these markets are inadequate. -76-Ancillary Services RecommendationsBased on the analysis in this report, I am recommending that the NYISO:•Modify the pricing rules for ancillary services to set prices at the marginal system cost of procuring the ancillary services. The marginal cost of ancillary services or ―shadow price‖ is equal to highest ancillary service cost of a resource designated to provide the service (defined as the highest total availability bid plus opportunity cost of not providing energy).This form of pricing would cover all suppliers’ opportunity cost of being held out of the energy market and would send a more accurate price signals to the market.•Implement a multi-settlement system (day-ahead and real-time) for reserves that would compliment the current multi-settlement system for energy. •Establish a demand curve for reserves, which would:Prevent the ISO from taking excessively costly actions to maintain low value reserves; Set reserves andenergy prices at efficient levels during capacity shortages.•Conditionally lift the offer cap currently in place for 10-minute non-synchronous reserves based on a competitive analysis presented in full annual report.