STATE OF THE MARKET annual state of the... · 2019-05-15 · The following members of the . Market Monitoring Unit . contributed to this report: Keith Collins . Jodi Woods . ... 2.5

STATE OF THE MARKET 2018

published

May 15, 2019

Southwest Power Pool, Inc. Market Monitoring Unit

State of the Market 2018

Acknowledgement

The following members of the Market Monitoring Unit

contributed to this report:

Keith Collins Jodi Woods John Luallen

Greg Sorenson Kevin Bates

Jason Bulloch Jared Greenwalt

Esat Serhat Guney David Hurtado James Lemley

Mark Rouse Cristie Turner

Will Vestal Cuiping Xu

Report lead: Greg Warren


State of the Market 2018

Southwest Power Pool, Inc. Table of contents Market Monitoring Unit

State of the Market 2018 i

TABLE OF CONTENTS

Table of contents .................................................................................................................................. i

List of figures ........................................................................................................................................ v

1 Executive summary .................................................................................................................1 1.1 Market highlights .................................................................................................................. 1 1.2 Overview ................................................................................................................................ 3 1.3 Day-ahead and real-time market performance ................................................................. 4 1.4 Transmission congestion and hedging .............................................................................. 6 1.5 Uplift costs ............................................................................................................................. 6 1.6 Competitiveness assessment .............................................................................................. 7 1.7 Structural issues .................................................................................................................... 9 1.8 Recommendations ............................................................................................................ 10

1.8.1 New recommendations for 2018 ............................................................................. 10 1.8.2 Clarification of previous recommendations ............................................................ 13 1.8.3 Outstanding recommendations ............................................................................... 14

2 Load and resources .............................................................................................................. 17 2.1 The Integrated Marketplace ............................................................................................. 18

2.1.1 SPP market footprint .................................................................................................. 18 2.1.2 SPP market participants............................................................................................. 19

2.2 Electricity demand ............................................................................................................. 21 2.2.1 System peak demand ................................................................................................ 21 2.2.2 Market participant load ............................................................................................. 22 2.2.3 SPP system energy consumption ............................................................................. 24 2.2.4 Heating and cooling degree days ........................................................................... 25

2.3 Installed generation capacity ........................................................................................... 28 2.3.1 Capacity additions and retirements ......................................................................... 31 2.3.2 Generation interconnection...................................................................................... 32

2.4 Generation.......................................................................................................................... 34 2.4.1 Generation by technology ........................................................................................ 34 2.4.2 Generation on the margin ........................................................................................ 38

2.5 Growing impact of wind generation capacity ................................................................ 40 2.5.1 Wind capacity and generation ................................................................................. 40 2.5.2 Wind impact on the system ...................................................................................... 44 2.5.3 Wind integration ........................................................................................................ 47 2.5.4 Challenges with non-dispatchable variable energy resources ............................ 50

2.6 External transactions ......................................................................................................... 51 2.6.1 Exports and imports .................................................................................................. 51 2.6.2 Market-to-market coordination ................................................................................ 56


State of the Market 2018 ii

3 Unit commitment and dispatch processes ....................................................................... 61 3.1 Commitment process ........................................................................................................ 62

3.1.1 Resource starts ........................................................................................................... 63 3.1.2 Demand for reliability ................................................................................................ 66 3.1.3 Quick-start resources commitment .......................................................................... 69 3.1.4 Generation scheduling .............................................................................................. 74

3.2 Dispatch .............................................................................................................................. 77 3.2.1 Scarcity pricing ........................................................................................................... 77 3.2.2 Ramping ...................................................................................................................... 83 3.2.3 Ramp capability product ........................................................................................... 88

3.3 Must-offer provision .......................................................................................................... 94 3.4 Virtual trading .................................................................................................................... 95

4 Prices.................................................................................................................................... 105 4.1 Market prices and costs .................................................................................................. 106

4.1.1 Energy market prices and fuel prices .................................................................... 106 4.1.2 Hub prices ................................................................................................................. 111 4.1.3 Energy price volatility .............................................................................................. 113 4.1.4 Day-ahead and real-time price convergence ....................................................... 115 4.1.5 Negative prices ........................................................................................................ 122 4.1.6 Operating reserve market prices ........................................................................... 126 4.1.7 Market settlement results ........................................................................................ 132

4.2 Make-whole payments .................................................................................................... 136 4.2.1 Make-whole payment allocation ............................................................................ 141 4.2.2 Regulation mileage make-whole payments ......................................................... 142 4.2.3 Potential for manipulation of make-whole payments .......................................... 144 4.2.4 Jointly-owned unit make-whole payments ........................................................... 145

4.3 Total wholesale market costs and production costs ................................................... 147 4.4 Long-run price signals for investment ........................................................................... 149

5 Congestion and transmission congestion rights market .............................................. 153 5.1 Transmission congestion ................................................................................................ 154

5.1.1 Pricing patterns and congestion ............................................................................ 155 5.1.2 Congestion by geographic location ...................................................................... 156 5.1.3 Transmission expansion .......................................................................................... 158 5.1.4 Transmission constraints ......................................................................................... 160 5.1.5 Planned transmission projects ................................................................................ 164 5.1.6 Geography and marginal losses ............................................................................ 164 5.1.7 Frequently constrained areas and local market power ....................................... 166 5.1.8 Market congestion management ........................................................................... 167 5.1.9 Congestion payments and uplifts .......................................................................... 168 5.1.10 Distribution of marginal loss revenues (over-collected losses) .......................... 171


State of the Market 2018 iii

5.2 Congestion hedging market .......................................................................................... 173 5.2.1 Market design ........................................................................................................... 173 5.2.2 Market transparency ................................................................................................ 175 5.2.3 Funding ..................................................................................................................... 181 5.2.4 Intra-auction sales and bulletin board transactions ............................................. 186 5.2.5 Issues, progress, and next steps ............................................................................ 188

6 Planning process ................................................................................................................ 191 6.1 Resource adequacy ......................................................................................................... 193

6.1.1 Capacity additions and retirements ....................................................................... 193 6.1.2 Generation capacity compared to peak load ....................................................... 194

6.2 Integrated transmission plan .......................................................................................... 196 6.2.1 2020 ITP generator retirement assumptions ........................................................ 197 6.2.2 2020 ITP renewable capacity accreditation assumptions ................................... 199 6.2.3 Other topics .............................................................................................................. 202

6.3 Assessment ....................................................................................................................... 203 6.4 Recommendations going forward ................................................................................. 204

7 Competitive assessment ................................................................................................... 207 7.1 Structural aspects of the market .................................................................................... 209 7.2 Behavioral aspects of the market ................................................................................... 216

7.2.1 Offer price markup .................................................................................................. 216 7.2.2 Mitigation performance and frequency ................................................................ 219 7.2.3 Output gap (measure for potential economic withholding) ............................... 223 7.2.4 Unoffered generation capacity (measure for potential physical withholding) . 227

7.3 Offer behavior due to mitigation threshold ................................................................. 231 7.4 Start-up and no-load behavior ....................................................................................... 231 7.5 Generator retirements .................................................................................................... 232 7.6 Competitive assessment summary ................................................................................ 234

8 Recommendations ............................................................................................................. 237 8.1 New recommendations .................................................................................................. 237

8.1.1 Limit the exercise of market power by creating a backstop for parameter changes .................................................................................................................................... 237 8.1.2 Enhance credit rules to account for known information in assessments ........... 238 8.1.3 Develop compensation mechanism to pay for capacity to cover uncertainties ...... .................................................................................................................................... 239 8.1.4 Enhance ability to assess a range of potential outcomes in transmission planning .................................................................................................................................... 240 8.1.5 Improve regulation mileage price formation ....................................................... 241

8.2 Clarification of previous recommendations ................................................................. 243 8.2.1 Address inefficiency caused by self-committed resources ................................. 243 8.2.2 Enhance commitment of resources to increase ramping flexibility ................... 245 8.2.3 Further enhance alignment of planning processes with operational conditions .... .................................................................................................................................... 246


State of the Market 2018 iv

8.3 Address previous recommendations ............................................................................ 248 8.3.1 Develop a ramping product ................................................................................... 248 8.3.2 Enhance market rules for energy storage resources ........................................... 249 8.3.3 Address inefficiency when forecasted resources under-schedule day-ahead . 250 8.3.4 Convert non-dispatchable variable energy resources to dispatchable ............ 251 8.3.5 Address gaming opportunity for multi-day minimum run time resources ....... 252 8.3.6 Replace day-ahead must offer, add physical withholding provision ................. 253

8.4 Recommendations update ............................................................................................. 253

Common acronyms ........................................................................................................................ 255

The data and analysis provided in this report are for informational purposes only and shall not be considered or relied upon as market advice or market settlement data. All analysis and opinions contained in this report are solely those of the SPP Market Monitoring Unit (MMU), the independent market monitor for Southwest Power Pool, Inc. (SPP). The MMU and SPP make no representations or warranties of any kind, express or implied, with respect to the accuracy or adequacy of the information contained herein. The MMU and SPP shall have no liability to recipients of this information or third parties for the consequences that may arise from errors or discrepancies in this information, for recipients’ or third parties’ reliance upon such information, or for any claim, loss, or damage of any kind or nature whatsoever arising out of or in connection with:

i. the deficiency or inadequacy of this information for any purpose, whether or not known or disclosed to the authors;

ii. any error or discrepancy in this information;

iii. the use of this information, and;

iv. any loss of business or other consequential loss or damage whether or not resulting from any of the foregoing.


State of the Market 2018 v

LIST OF FIGURES

Figure 2—1 SPP market footprint and RTO/ISO operating regions ....................................... 19 Figure 2—2 Market participants by type .................................................................................... 20 Figure 2—3 Capacity by market participant type ...................................................................... 20 Figure 2—4 Monthly peak system demand ............................................................................... 21 Figure 2—5 Cleared demand bids in day-ahead market ......................................................... 22 Figure 2—6 System energy usage .............................................................................................. 23 Figure 2—7 System energy consumption, monthly .................................................................. 24 Figure 2—8 Load duration curve................................................................................................. 25 Figure 2—9 Heating and cooling degree days ......................................................................... 26 Figure 2—10 Degree days and loads compared with a normal year ....................................... 27 Figure 2—11 Generation nameplate capacity by technology type .......................................... 29 Figure 2—12 Aggregate supply curve, typical summer day ...................................................... 30 Figure 2—13 Capacity additions ................................................................................................... 31 Figure 2—14 Capacity retirements ............................................................................................... 32 Figure 2—15 Active generation interconnection requests, megawatts ................................... 33 Figure 2—16 Executed generation interconnection requests, on-schedule ........................... 34 Figure 2—17 Generation by technology type, real-time, annual .............................................. 35 Figure 2—18 Generation by technology type, real-time, monthly ............................................ 36 Figure 2—19 Implied heat rate ...................................................................................................... 37 Figure 2—20 Generation on the margin, real-time, annual ....................................................... 38 Figure 2—21 Generation on the margin, real-time ..................................................................... 39 Figure 2—22 Generation on the margin, day-ahead .................................................................. 40 Figure 2—23 Wind speed map...................................................................................................... 41 Figure 2—24 Wind capacity and generation, annual ................................................................. 42 Figure 2—25 Wind capacity factor ................................................................................................ 43 Figure 2—26 Wind capacity factor by month, real-time, three year ......................................... 44 Figure 2—27 Wind generation as a percent of load ................................................................... 45 Figure 2—28 Wind production curve ........................................................................................... 46 Figure 2—29 Net demand by hour of day.................................................................................... 47 Figure 2—30 Dispatchable and non-dispatchable wind generation ........................................ 48 Figure 2—31 Manual dispatches for wind resources .................................................................. 49 Figure 2—32 Exports and imports, SPP system ........................................................................... 52 Figure 2—33 Exports and imports, ERCOT interface .................................................................. 53 Figure 2—34 Exports and imports, Southwestern Power Administration interface ................ 53 Figure 2—35 Exports and imports, MISO interface..................................................................... 54 Figure 2—36 Exports and imports, Associated Electric Cooperative interface ....................... 54 Figure 2—37 Imports and export transactions by type, day-ahead .......................................... 56 Figure 2—38 Market-to-market settlements ................................................................................ 57 Figure 2—39 Market-to-market settlements by flowgate ........................................................... 58 Figure 3—1 Commitment process timeline ............................................................................... 62 Figure 3—2 Start-up instructions by resource count ................................................................ 63 Figure 3—3 Start-up instructions by resource capacity ............................................................ 64

Southwest Power Pool, Inc. List of figures Market Monitoring Unit

State of the Market 2018 vi

Figure 3—4 Origin of start-up instructions for gas resources .................................................. 65 Figure 3—5 Average hourly capacity increase from day-ahead to real-time ........................ 67 Figure 3—6 Instantaneous load capacity required ................................................................... 68 Figure 3—7 Instantaneous load capacity upper bound default levels ................................... 69 Figure 3—8 Deployment of quick-start resources .................................................................... 70 Figure 3—9 Operation of quick-start resources ........................................................................ 72 Figure 3—10 Day-ahead market offers by commitment status and participant type ............. 74 Figure 3—11 Day-ahead market offers by commitment status ................................................. 74 Figure 3—12 On-line capacity as a percent of load .................................................................... 75 Figure 3—13 Generation outages ................................................................................................. 76 Figure 3—14 Short-term generation outages, fall months ......................................................... 77 Figure 3—15 Scarcity intervals and marginal energy cost, after scarcity implementation ..... 80 Figure 3—16 Scarcity events by interval of the hour ................................................................... 81 Figure 3—17 Spinning reserve shortages .................................................................................... 83 Figure 3—18 Frequency of net load change, real-time .............................................................. 84 Figure 3—19 Volatility of interval-to-interval net load change .................................................. 85 Figure 3—20 Average up-rampable capacity by month ............................................................ 86 Figure 3—21 Average down-rampable capacity by month ....................................................... 87 Figure 3—22 Average up-rampable capacity by hour ............................................................... 87 Figure 3—23 Average down-rampable capacity by hour .......................................................... 88 Figure 3—24 Interval length of short-term price spikes ............................................................. 89 Figure 3—25 Interval length of short-term price spikes, percentage ....................................... 90 Figure 3—26 Cleared virtual transactions as percent of real-time load ................................... 96 Figure 3—27 Cleared virtual offers by settlement location type ............................................... 97 Figure 3—28 Cleared virtual bids by settlement location type ................................................. 98 Figure 3—29 Virtual offers and bids, day-ahead market ............................................................ 99 Figure 3—30 Virtual profits with distribution charges .............................................................. 100 Figure 3—31 Profit and loss per cleared virtual, average ........................................................ 101 Figure 3—32 Virtual portfolios by market participant ............................................................... 102 Figure 4—1 Energy price versus natural gas price, annual .................................................... 106 Figure 4—2 Day-ahead energy price versus natural gas cost ............................................... 107 Figure 4—3 Fuel price indices and wholesale power prices ................................................. 108 Figure 4—4 Fuel-adjusted marginal energy cost .................................................................... 110 Figure 4—5 Hub prices, annual average .................................................................................. 111 Figure 4—6 Hub prices, day-ahead, monthly .......................................................................... 112 Figure 4—7 Hub prices, real-time, monthly ............................................................................. 112 Figure 4—8 Comparison of RTO/ISO average on-peak, day-ahead prices ......................... 113 Figure 4—9 System price volatility ............................................................................................ 114 Figure 4—10 Price volatility by market participant.................................................................... 115 Figure 4—11 Day-ahead and real-time prices ........................................................................... 116 Figure 4—12 Difference between day-ahead and real-time prices ........................................ 116 Figure 4—13 Price divergence .................................................................................................... 118 Figure 4—14 On-peak marginal energy prices, excluding scarcity hours ............................. 119 Figure 4—15 Off-peak marginal energy prices, excluding scarcity hours ............................. 120


State of the Market 2018 vii

Figure 4—16 Average hourly real-time generation incremental to day-ahead market ........ 121 Figure 4—17 Negative price intervals, day-ahead, monthly .................................................... 122 Figure 4—18 Negative price intervals, real-time, monthly ....................................................... 123 Figure 4—19 Negative price intervals, day-ahead, by hour ..................................................... 124 Figure 4—20 Negative price intervals, real time, by hour ........................................................ 124 Figure 4—21 Negative price intervals, real-time, by asset owner ........................................... 125 Figure 4—22 Operating reserve product prices, real-time ...................................................... 126 Figure 4—23 Regulation-up service prices ................................................................................ 127 Figure 4—24 Regulation-down service prices ........................................................................... 128 Figure 4—25 Spinning reserve prices ......................................................................................... 128 Figure 4—26 Supplemental reserve prices ................................................................................ 129 Figure 4—27 Regulation mileage prices, real-time ................................................................... 130 Figure 4—28 Regulation mileage payments and charges ....................................................... 132 Figure 4—29 Energy settlements, load ...................................................................................... 133 Figure 4—30 Energy settlements, generation ........................................................................... 134 Figure 4—31 Operating reserve product settlements .............................................................. 135 Figure 4—32 Make-whole payments by fuel type, day-ahead ................................................ 137 Figure 4—33 Make-whole payments by fuel type, reliability unit commitment .................... 137 Figure 4—34 Make-whole payments, commitment reasons .................................................... 138 Figure 4—35 Make-whole payments for voltage support ........................................................ 139 Figure 4—36 Concentration of make-whole payments by resource ...................................... 140 Figure 4—37 Market participants receiving make-whole payments ....................................... 141 Figure 4—38 Make-whole payments by market uplift allocation, real-time ........................... 142 Figure 4—39 Regulation mileage make-whole payments ....................................................... 143 Figure 4—40 All-in price of electricity and natural gas cost ..................................................... 147 Figure 4—41 Production cost, daily average, day-ahead......................................................... 148 Figure 4—42 Production cost, daily average, real-time ........................................................... 149 Figure 4—43 Net revenue analysis assumptions ....................................................................... 150 Figure 4—44 Net revenue analysis results ................................................................................. 151 Figure 4—45 Net revenue analysis by zone ............................................................................... 151 Figure 5—1 Price map, day-ahead and real-time market ...................................................... 155 Figure 5—2 Marginal congestion cost map, day-ahead market ........................................... 157 Figure 5—3 SPP transmission expansion in service, 2018 ..................................................... 159 Figure 5—4 SPP transmission expansion plan ......................................................................... 160 Figure 5—5 Congestion by shadow price, top ten flowgates ............................................... 161 Figure 5—6 Central Kansas congestion ................................................................................... 162 Figure 5—7 Southwest Missouri congestion ........................................................................... 163 Figure 5—8 Top ten congested flowgates with projects ....................................................... 164 Figure 5—9 Marginal loss component map, day-ahead ........................................................ 165 Figure 5—10 Breached and binding intervals, day-ahead market .......................................... 167 Figure 5—11 Breached and binding intervals, real-time .......................................................... 168 Figure 5—12 Annual congestion payment by load-serving entity .......................................... 169 Figure 5—13 Total congestion payments .................................................................................. 170 Figure 5—14 Over-collected losses, real-time ........................................................................... 172


State of the Market 2018 viii

Figure 5—15 Total congestion and hedges ............................................................................... 176 Figure 5—16 Net congestion revenue by market participant .................................................. 177 Figure 5—17 Transmission outages by reporting lead time .................................................... 179 Figure 5—18 Transmission outages by duration ....................................................................... 180 Figure 5—19 Transmission outages by market ......................................................................... 180 Figure 5—20 Transmission congestion right funding levels, annual ...................................... 182 Figure 5—21 Transmission congestion right funding levels, monthly .................................... 183 Figure 5—22 Transmission congestion right funding, daily .................................................... 184 Figure 5—23 Auction revenue right funding levels, annual ..................................................... 185 Figure 5—24 Auction revenue right funding levels, monthly .................................................. 186 Figure 5—25 Intra-auction sales and bulletin board transactions ........................................... 187 Figure 6—1 Capacity by age of resource ................................................................................. 193 Figure 6—2 Capacity additions and retirements by year ....................................................... 194 Figure 6—3 Peak available capacity percent ........................................................................... 195 Figure 6—4 Average age of retirements for coal and gas resources ................................... 198 Figure 6—5 Average age of generators in the SPP market ................................................... 199 Figure 6—6 MMU calculation for wind and solar capacity factors ........................................ 201 Figure 7—1 Market share of largest supplier .......................................................................... 210 Figure 7—2 Market concentration level, real-time .................................................................. 211 Figure 7—3 Herfindahl-Hirschman Index ................................................................................. 212 Figure 7—4 Herfindahl-Hirschman Index, annual ................................................................... 213 Figure 7—5 Market participants with a marginal resource, real-time ................................... 214 Figure 7—6 Hours with at least one pivotal supplier .............................................................. 216 Figure 7—7 Average offer price markup of marginal resource, monthly ............................ 217 Figure 7—8 Average on-peak offer price markup by fuel type of resource, monthly ........ 218 Figure 7—9 Average offer price markup, annual .................................................................... 219 Figure 7—10 Mitigation frequency, day-ahead market ............................................................ 221 Figure 7—11 Mitigation frequency, real-time market ............................................................... 222 Figure 7—12 Mitigation frequency, start-up offers ................................................................... 222 Figure 7—13 Output gap, monthly ............................................................................................. 225 Figure 7—14 Output gap, SPP footprint .................................................................................... 226 Figure 7—15 Output gap, Lubbock frequently constrained area ........................................... 226 Figure 7—16 Unoffered economic capacity .............................................................................. 228 Figure 7—17 Unoffered economic capacity at various load levels, SPP footprint ................ 229 Figure 7—18 Unoffered economic capacity at various load levels, Lubbock area ............... 230 Figure 7—19 Mitigation start-up offer mark-up by fuel category ............................................ 232 Figure 8—1 Annual State of the Market recommendations update ..................................... 253

Southwest Power Pool, Inc. Executive summary Market Monitoring Unit

State of the Market 2018 1

1 EXECUTIVE SUMMARY

The Southwest Power Pool (SPP) Market Monitoring Unit’s (MMU) Annual State of the Market

report for 2018 presents an overview of market design and market outcomes, assesses

market performance, and provides recommendations for improvement. The purpose of this

report is to provide SPP market stakeholders with reliable and useful analysis and information

to use in making market-related decisions. The MMU emphasizes that economics and

reliability are inseparable and that an efficient wholesale electricity market provides the

greatest benefit to the end user both presently and in the years to come.

1.1 MARKET HIGHLIGHTS

The following list identifies key observations in the SPP marketplace over the past year.

• SPP market results were workably competitive, with infrequent mitigation of offers and

high resource participation levels.

• Total wholesale market costs—including energy, operating reserve, and uplift

payments—averaged around $28/MWh in 2018, which was about 15 percent higher

than in 2017.

• Day-ahead and real-time prices both averaged around $25/MWh for the year, up from

$23/MWh in 2017.

• While the annual peak load of 49,926 MW was two percent lower this year compared

to last year, total electricity consumption was up about six percent.

• The incidence of negative prices in 2018, 3.5 percent of intervals, was about half of

the 2017 level of seven percent.

• While natural gas-fired resources frequently set prices in the SPP market, other factors

such as load increases, had a large impact on market prices. Natural gas prices

decreased about two percent in 2018 compared to 2017.

• When system prices are controlled for changes in fuel prices, they averaged about 13

percent higher in 2018 compared to 2017.



• In 2018, there were 167 intervals with operating-reserve scarcity. This was roughly

five times as many intervals of operating reserve scarcity as was seen in 2017.

• Day-ahead and real-time congestion costs totaled over $450 million in 2018, a four

percent decrease from 2017.

• Wind generation peaked at 16.3 GW and peak wind penetration was almost 64

percent of load in December. Wind capacity increased to almost 20.6 GW in 2017, up

about 17 percent from 2017.

• Wind generation totaled 24 percent of all generation in 2018, up slightly from 23

percent in 2017. Coal generation fell from 46 percent in 2017 to 42 percent in 2018.

• New capacity additions were just over 2,300 MW at nameplate capacity, with wind

representing 97 percent of the new capacity. Capacity retirements increased to just

under 2,000 MW in 2018, split about evenly between coal and gas resources.

• The generator interconnection process includes just over 84 GW of additional

resources, of which 99 percent are renewable or storage.

• SPP continues to have significant excess capacity at peak loads. The MMU estimates

that capacity at peak is 35 percent higher than the peak demand level in 2018.

• Market prices themselves do not signal new investment in generation. Furthermore,

MMU analysis shows that market revenues do not support going forward costs for coal

resources.

• Market uplifts remained low at about $73 million in 2018, which was up slightly from

2017 levels.

• Combined ancillary reserve costs totaled $76 million last year, a decrease of three

percent from 2017.

• Auction revenue rights were funded at 145 percent in 2018, down from just over 160

percent in 2017.

• Transmission congestion rights funding was virtually unchanged from 2017 to 2018, at

94 percent.



1.2 OVERVIEW

Overall, the SPP market produced highly competitive market results with total market costs

around $28/MWh. As with previous years, the largest component of total wholesale costs

remains energy costs, which represented almost 98 percent of total costs in 2018. While total

costs increased by 15 percent in 2018 compared to 2017, a main driver for the increase in

energy costs was an increase in load, as well as several other factors. When adjusted for fuel

prices, average SPP marginal energy prices increased by 13 percent.

While the annual peak load of 49,926 MW was two percent lower this year compared to last

year, total electricity consumption was up about six percent. Over 97 percent of the 2,300

MW increase in nameplate generation capacity last year was from wind resources. This

continues a pattern that has occurred over the past several years. Wind generation as a

percent of total generation continued to increase as it represented 24 percent of system

generation, up from 23 percent in 2017 and 18 percent in 2016. Conversely, coal generation

continued to decline, representing around 42 percent of total generation last year, down

from 46 percent in 2017 and 49 percent in 2016. Prior to 2016, coal represented 55 percent

or more of generation in SPP.

Significant market events in 2018 included:

• A second circuit from Woodward to Mathewson was added.1 Our analysis shows that

this transmission addition helped shift power further east in the SPP footprint and

reduce the frequency of negative prices.

• Market participants continued to express concerns with the effectiveness of the

auction revenue rights process to allow participants to receive sufficient hedges for

congestion.2 We observe that while many participants were able to manage

congestion, a handful of participants did not have sufficient hedges. There were

multiple reasons for this, including the nature of congestion patterns, outages, and

market participant strategies.

1 See Section 5.1.3 for further discussion. 2 Auction revenue rights and transmission congestion rights are covered in Section 5.2.



• SPP launched the Holistic Integrated Tariff Team effort. This group comprised

members, Board of Directors, Regional State Committee members, and supported by

SPP staff, focused on developing problem statements, reviewing analysis, and

proposing solutions to related areas of concern including transmission cost allocation,

auction revenue right allocation, market enhancements, and load integration. The

findings of the group are anticipated to be presented to the April 2019 Board

meeting.

1.3 DAY-AHEAD AND REAL-TIME MARKET PERFORMANCE

Overall, energy prices were about $2/MWh higher in 2018 compared to 2017. This can

primarily be attributed to several factors, including increased loads, increased exports,

transmission expansion, lower wind capacity factors, and increased generator outages.

These offset other factors that would typically lower prices, including lower gas prices and a

large and increasing reserve margin. Historically, day-ahead prices have been higher than

real-time prices. However, in 2017 real-time prices were higher than day-ahead prices in nine

months primarily because of higher real-time price volatility. This trend reversed in 2018 as

eight out of twelve months had higher day-ahead average prices.

While load participation in the day-ahead market continued to be strong in 2018, generation

participation, particularly from wind resources was substantially less in the day-market market

compared to the real-time market. For instance, the average level of participation for the

load assets was between 99 percent and 101 percent of the actual real-time load. However,

we found that on average for the year, wind generation was over 1,000 MW higher in the

real-time market compared to the amount scheduled in the day-ahead market on an hourly

basis. This represents an increasing challenge to the market as wind generation has

continued to increase substantially over the past few years.

Virtual bids and offers may theoretically offset the under-scheduling of renewable supply in

the day-ahead market, however, in net they did not as they averaged around 500 MW of net

virtual supply. Furthermore, it is important to recognize that even if virtual transactions were

to match the quantity of under-scheduled renewables, the prices associated with the virtual

offers are not likely to fully represent the offer prices of the renewable resources in order to

preserve a profit margin.



In general, virtual transactions were profitable in the SPP market. However, total profits

decreased in 2018 to about $44 million, down from $54 million in 2017. When transaction

fees are included, net profit for virtual transactions was $18 million in 2018. Net virtual profits

were highest in January ($4.9 million) and October ($4.0 million) when wind generation is

typically high and loads are low.

Self-commitment of generation continues to be a concern because it does not allow the

market software to determine the most economic market solution. Furthermore, it can

contribute to market uplifts and low prices. Some of the reasons for self-committing may

include contract terms for coal plants, low gas prices that reduce the opportunity for coal

units to be economically cleared in the day-ahead market, long startup times, and a risk-

averse business practice approach. Generation offers in the day-ahead market averaged

almost 53 percent as “market” commitment status followed by “self-commit” status at 30

percent of the total capacity commitments for 2018.3 These levels almost exactly match

those in 2017, however the overall trend is still downward, as 2016 had 48 percent as

“market” commitment status, and 35 percent as “self-commit” status. While the overall

increase in market commitments and decrease in self-commitments highlights an

improvement, self-commitments still represent over 30 percent of generation, a trend that

has existed since the Integrated Marketplace began in 2014. In order to improve market

commitment in the SPP market, we recommend that SPP and stakeholders look to find ways

to address this issue.

Scarcity events in 2018 were generally in line with 2017 levels, with the exception of October.

October 2018 experienced significant wind volatility, which among other factors, led to an

uptick in operating reserve scarcity events. Additionally, looking at the intervals where

scarcity occurs each hour shows that 30 percent of all scarcity events, and two-thirds of

regulation-down scarcity events, in 2018 occurred in the first interval of the hour. One

potential reason for this trend is that SPP does not preposition regulating resources to be at

their regulating effective maximum and minimum limits prior to the period that the resource

is cleared for regulation. Unlike some other RTO/ISO markets, the current SPP model does

not account for forecasted ramping needs. Our analysis shows that accounting for ramping

needs would greatly assist in preparing and compensating generation for both anticipated

3 Other resource commitment statuses are “reliability”, “not participating”, and “outage” at two percent, three percent, and 11 percent, respectively.



and unanticipated ramping needs. As such, we continue to recommend, as discussed below,

that SPP and stakeholders complete design for a ramping product in 2019.

1.4 TRANSMISSION CONGESTION AND HEDGING

Locational marginal prices reflect the sum of the marginal cost of energy, the marginal cost of

congestion, and the marginal cost of losses for each pricing interval at any given pricing

location in the market. Although the SPP market currently maintains a high reserve margin,

certain locations of the footprint experience significant price movements resulting from

congestion caused by high wind generation and transmission limitations.

Areas that experienced the highest congestion costs in 2018 were in central Kansas,

southeastern Oklahoma, and southwest Missouri. Western Nebraska experienced the lowest

congestion costs for the year. The frequently constrained area study for 2018 saw the

addition of the central Kansas and southwest Missouri frequently constrained areas, while the

Texas panhandle frequently constrained area was removed.

In total, congestion costs were just over $450 million in 2018. This was down slightly from

$465 million in 2017. While most load-serving entities were able to successfully hedge their

congestion exposure with auction revenue rights and transmission congestion rights, a

handful of participants were under-hedged. In 2018, the total of all transmission congestion

right and auction revenue right net payments to load-serving entities of $458 million was

more than the total day-ahead and real-time markets congestion costs of $381 million.

However, on an individual basis, some participants were over-hedged, whereas others were

under-hedged. The largest amount over-hedged was by just over $16 million, while the

largest amount under-hedged was just under $10 million.

1.5 UPLIFT COSTS

Generators receive make-whole payments to ensure that they receive sufficient revenue to

cover energy, start-up, no-load, and operating reserve costs for both market and local

reliability commitments. Make-whole payments are additional market payments in cases

where prices result in revenue that is below a resource’s cleared offers. These payments are

intended to make resources whole to energy, commitment, and operating reserve costs.

In 2018, total make-whole payments were approximately $73 million, up slightly from $68

million in 2017. Make-whole payments averaged about $0.26/MWh in 2018, which was



nearly the same as in 2017. In comparison to other RTO/ISO markets, SPP’s make-whole

payments were in the range of uplift costs, which varied from $0.08/MWh to $0.29/MWh in

2017 and 2018.4

Day-ahead make-whole payments constituted about 38 percent of the total make-whole

payments in 2018. SPP pays about 80 percent of all make-whole payments to gas-fired

resources.

In December 2017, FERC found SPP’s quick-start pricing practice may be unjust and

unreasonable, and argued that pricing enhancements would reduce uplift payments.5 Our

analysis shows that uplift payments remain fairly low. Furthermore, our assessment of FERC’s

proposal suggests that, if anything, FERC’s proposal would transfer, if not increase, uplifts.6

FERC has yet to rule on SPP’s quick-start pricing practices.

1.6 COMPETITIVENESS ASSESSMENT

The SPP market provides effective incentives and mitigation measures to produce

competitive market outcomes even during periods when the potential for the exercise of

local market power could be a concern. The MMU’s competitive assessment using structural

and behavioral metrics indicate that market results in 2018 were workably competitive and

that the market required mitigation of local market power infrequently to achieve competitive

outcomes.

Structural competitiveness metrics—which review the structural potential for the exercise of

market power—indicate minimal to moderate potential structural market power in SPP

markets outside of areas that are frequently congested. With the merger of Great Plains

Energy and Westar, creating Evergy, Inc., the percent market share of the largest supplier

saw an increase, with 35 percent of hours for the year exceeding the 20 percent threshold

4 2017 ISO NE State of Market Report https://www.potomaceconomics.com/wp-content/uploads/2018/06/ISO-NE-2017-EMM-SOM-Report_6-17-2018_Final.pdf, page 39; 2018 State of the Market Report for PJM http://www.monitoringanalytics.com/reports/Presentations/2019/IMM_MC_SOM_20190321.pdf, page 9. 5 Order Instituting Section 206 Proceeding and Commencing Hearing Procedures and Establishing Refund Effective Date, 161 FERC ¶ 61,296 (2017). 6 Southwest Power Pool Market Monitoring Unit Reply Brief, Docket No. EL18-35-000, March 13, 2018.

https://www.potomaceconomics.com/wp-content/uploads/2018/06/ISO-NE-2017-EMM-SOM-Report_6-17-2018_Final.pdf


http://www.monitoringanalytics.com/reports/Presentations/2019/IMM_MC_SOM_20190321.pdf



that indicates potential market power, with a high value of 29 percent. This is up from 2017,

when no hours were above the 20 percent threshold, and the highest value was 17 percent.

An additional measure of structural market power is the Herfindahl-Hirschman Index (HHI).

This analysis, based on actual generation, indicates that 12 percent of hours in 2018 had

values between 1,000 and 1,800, which indicates a moderate level of concentration. The

market had been considered unconcentrated since the addition of the Integrated System in

October 15, up until the creation of Evergy in June 2018. Prior to the addition of the

Integrated System, approximately 45 percent of all hours were considered moderately

concentrated.

While moderately concentrated hours increased following the creation of Evergy, an increase

in market share and HHI in themselves does not pose a threat to the structural

competitiveness of the SPP market. Other relevant market data including pivotal supplier

hours and local market power mitigation must also be evaluated for competitive assessment.

For the two frequently constrained areas, where potential for concerns of local market power

is the highest, existing mitigation measures serve well to prevent pivotal suppliers from

unilaterally raising prices.

Behavioral indicators—which assess the actual exercise of market power—show low levels of

mitigation frequency. Mitigation of day-ahead energy, operating reserve, and no-load offers

each occurred less than 0.2 percent of the time and real-time mitigation occurred about

0.015 percent of the time. The overall mitigation frequency of start-up offers was the lowest

since the market began in 2014, as it decreased in 2018 relative to 2016 levels to just under

three percent.

The decline in mitigation may be the result of declining offer price mark-ups. Both off-peak

and on-peak average offer markups were at the lowest levels since implementation of the

Integrated Marketplace at around −$2.05/MWh and −$2.41/MWh, respectively. Although a

lower offer price markup level in itself would indicate a competitive pressure on suppliers in

the SPP market, the observed continuous downward trend may raise questions about the

commercial viability of generating units and the possibility of generation retirements.

The monthly average output gap—which measures economic withholding—shows very low

levels of economic withholding in all months in 2018. Specifically, there was miniscule



amounts of measurable output withheld in the frequently constrained areas. These low levels

of economic output withholding reflect highly competitive participation in the market.

Another method of competitive assessment is unoffered generation capacity for potential

physical withholding. Specifically, any economic generation capacity that is not made

available to the market through derates, outages, or otherwise not offered to the market is

considered for this analysis.

Annually for the SPP footprint, the total unoffered capacity (as a percent of total resource

reference levels) equaled 2.0 percent in 2016, 1.9 percent in 2017, and 3.1 percent in 2018.

When short and long-term outages are removed, the remaining unoffered capacity was 0.22

percent, 0.23 percent, and 0.39 percent, respectively. The majority of the outages were long-

term outages due to maintenance during the shoulder fall and spring months. From a

competitive market perspective, the results indicate reasonable levels of total unoffered

economic capacity and are consistent with the results in other RTO/ISO markets.

1.7 STRUCTURAL ISSUES

Installed generation capacity in the SPP market has grown rapidly over the past several years.

This has contributed to high levels of capacity at peak loads. Specifically, the MMU estimates

that capacity was 35 percent higher than the peak load in 2018. SPP’s current annual

planning capacity requirement is 12 percent.

Wind capacity has more than doubled from 8.6 GW in 2014 to 20.6 GW in 2018. At the same

time, wind generation has constituted a growing and significant part of the total annual

generation, from around 12 percent in 2014 to 24 percent in 2018. Furthermore, the

interconnection process includes just over 84 GW of additional resources, of which 99

percent are renewable resources.

The shift in generation mix towards renewable resources is a significant development and

carries both market and operational challenges. Furthermore, these challenges are further

exacerbated by the fact that currently 29 percent of the total wind capacity is non-

dispatchable.

A recent wave of generator retirements, particularly of coal-fired generation, has been widely

observed throughout the country. The SPP market should be expected to follow this trend



because of excess capacity, aging fleet, and cost disadvantages of certain types of

generation technologies vis-à-vis the prevailing market prices.

The MMU believes that SPP stakeholders should prepare for the challenges these changes,

and potential changes, to the market present. However, additional changes from planning to

operations needs to be developed to improve market outcomes. As such, we make several

recommendations to address these growing market concerns.

1.8 RECOMMENDATIONS

One of the primary responsibilities of a market monitoring unit is to evaluate market rules and

market design features for market efficiency and effectiveness. When we identify issues with

the market, one of the ways to correct them is to make recommendations on market

enhancements. These recommendations are highlighted in detail in Chapter 8. Below is a

summary of our 2018 recommendations.

1.8.1 NEW RECOMMENDATIONS FOR 2018

• Limit the exercise of market power by creating a backstop for parameter changes –

The market is currently vulnerable to the exercise of market power by the

manipulation of resources’ non-dollar based parameters. Non-dollar-based

parameters should not be manipulated to affect market clearing. Although SPP’s tariff

and protocols have well-defined expectations and precise limitations for the basis of

dollar-based offer components in the presence of local market power, the

expectations for the basis of non-dollar-based offer components are much less clearly

defined and the limitations are much less precise.

The MMU recommends that SPP strengthen the language regarding non-dollar-based

parameters so that the expectation for the basis of these values is clear and the

exercise of market power is limited.

• Enhance credit rules to account for known information in assessments – In 2018,

GreenHat Energy, a financial-only market participant in PJM, defaulted on its portfolio

of congestion hedging products in the PJM markets. Current estimates of the default



exposure exceed $400 million.7 While the SPP market is different from the PJM

market, SPP’s credit policy is similar to PJM’s in some respects. For instance, SPP’s

financial security requirements for transmission congestion rights are based only on

historic congestion patterns even when significant transmission upgrades are planned

or have occurred.

The MMU recommends that changes to the SPP credit policy be developed to protect

stakeholders from risks of default, particularly with respect to information that is

known such as transmission expansion.

• Develop compensation mechanism or product to pay for capacity to cover

uncertainties – Because of unexpected variations in wind output, SPP operations often

needs to manually commit resources (often in excess of 50 units) in order to meet

instantaneous load capacity requirements. Often, however, these resources receive

real-time wake whole payments and are not compensated specifically for the need.

While the MMU recognizes that SPP operators may need to commit units to account

for unforeseen circumstances, manual capacity commitments occur often enough that

systematic solutions should be developed. The large number of manual capacity

commitments may indicate that there is a need for a new uncertainty product that

might better address problems in the 30-minute to 3-hour time frame.8

The MMU recommends and supports developing products to reduce the need for the

large number of manual capacity commitments to ensure appropriate compensation

is being provided for the reliability services provided.

• Enhance ability to assess a range of potential outcomes in transmission planning –

SPP’s transmission planning process develops an annual look-ahead plan. This plan

evaluates transmission needs over a 5- and 10-year time horizon. The plan typically

7 https://pjm.com/-/media/committees-groups/committees/mic/20190206/20190206-item-01a-informational-update-ferc-order-denying-waiver-request.ashx. 8 The exact time periods of potential uncertainty products should be determined through the stakeholder process.

https://pjm.com/-/media/committees-groups/committees/mic/20190206/20190206-item-01a-informational-update-ferc-order-denying-waiver-request.ashx




evaluates two scenarios.9 The first case is a base case, and the second case is an

emerging technology case. The process could consider a third scenario. In 2018,

stakeholders requested that SPP staff consider an additional case. The third case

would have considered a shift in environmental regulations—such as a carbon tax or

adder—and a change in technologies/market trends including accelerated

deployment of storage devices or electric vehicles, higher levels of generation

retirement, and higher penetration of renewables.10

We recommend that SPP enhance their study process to allow the ability to study a

range of potential outcomes. If such range of potential outcomes are not captured in

a third case study, the MMU recommends that they be factored into either an existing

case study, as part of the 20-year assessment, or other assessment.

• Improve regulation mileage price formation – In addition to regulation capacity

payments, resources that are deployed for regulation also receive payments for costs

incurred when moving from one set point instruction to another. These payments,

known as mileage, are paid directly through regulation-up and regulation down

payments in the day-ahead market. The MMU is concerned that the mileage clearing

price does not correctly reflect price formation of mileage given that this price does

not represent the expected or realized mileage deployment. Furthermore, the MMU

is concerned that participants with resources frequently deployed for regulation will

have an incentive to inflate the mileage prices by offering in $0 regulation offers and

high mileage offers. The MMU also identified systematic overpayment of regulation

mileage in the day-ahead market, which appears to be the result of the mileage factor

being set consistently too high relative to actual mileage deployed.

We recommend that SPP staff review the performance of regulation mileage, and

develop potential approaches to improve regulation mileage price formation.

9 These scenarios are also known as futures cases. 10 This scenario was similar to a scenario used in the MISO transmission plan, see https://cdn.misoenergy.org/MTEP19%20Futures%20Summary291183.pdf.

https://cdn.misoenergy.org/MTEP19%20Futures%20Summary291183.pdf



1.8.2 CLARIFICATION OF PREVIOUS RECOMMENDATIONS

The MMU has provided recommendations to improve market design in our previous annual

reports. The following are clarifications made to those recommendations from the prior

report.

• Address inefficiency caused by self-committed resources – The MMU recommends

that SPP and its stakeholders continue to explore and develop ways to reduce the

incidence of self-commitment of resources outside of the market solution. With

regards to the development of multi-day forecasting of prices or schedules or a multi-

day unit commitment process, the MMU supports these as attempts to reduce self-

commitment of generation, but we will carefully weigh the benefits and costs before

we support any specific proposal especially in light of the accuracy of forecasted

information and the potential changing generation mix. We continue, however, to

view reducing self-commitment of generation as a high priority for SPP and its

stakeholders as this will enhance market efficiency and improve price signals.

• Enhance commitment of resources to increase ramping flexibility – As noted in last

year’s annual report, ramping flexibility is becoming increasingly important to

integrate higher levels of renewable generation.11 Over-commitment of resources in

real-time reduces flexibility, suppresses prices, and leads to increased make-whole

payments. This can be caused by changing conditions between the time a resource is

locked into a commitment by the market software and the time the resource actually

comes on-line. The MMU recommends that SPP and its stakeholders address this

issue by enhancing its markets rules to enhance the commitment of resources to

increase resource flexibility. Last year, the MMU described this as enhancing

decommitment of resources. However, having explored this issue further, the issue is

not just about decommitment of resources, but is also about improving how and when

resources are committed.

• Further enhance alignment of planning processes with operation conditions –

Enhancing the accuracy of planning processes with operational realities enables SPP

and its members to more effectively plan for future system needs and

11 2017 Market Monitoring Unit Annual State of the Market Report, https://www.spp.org/documents/57928/spp_mmu_asom_2017.pdf , page 194.

https://www.spp.org/documents/57928/spp_mmu_asom_2017.pdf



conditions. Many of the challenges outlined in this report—including congestion,

negative prices, and low generator net revenues—as well as improvements—such as

transmission additions—are, in part, a reflection of planning decisions. The more the

planning process can learn from and incorporate operational information, the more

planning can identify and address concerns in advance of market operations. SPP and

stakeholders over the past few years have worked to improve and align the planning

and operational processes.

Based on our experience in this process over the last year, we recommend that SPP

staff continue to collaborate with the MMU to build upon the work in 2018, by

reviewing data and best practices in other markets, factoring in market conditions and

trends in decision-making process, and performing analysis to support and inform

assumptions. Doing so will help to improve assumptions and ultimately study results.

1.8.3 OUTSTANDING RECOMMENDATIONS

The MMU has provided recommendations to improve market design in our previous annual

reports. Overall, SPP and its stakeholders have found ways to effectively address many of our

concerns. However, there are a number of recommendations that remain outstanding. A

description of each of these outstanding recommendations are outlined below.

• Develop a ramping product – A ramping product that incents actual, deliverable

flexibility can send appropriate price signals to value resource flexibility. This

resource flexibility can help prepare the system for fluctuations in both demand and

supply that result in transient short-term positive and negative price spikes. This is a

high priority initiative in the SPP stakeholder process. We concur and continue to

recommend that a ramping product design be completed in 2019.

• Enhance market rules for energy storage resources – With the increase in wind

penetration in the SPP market, there is not only a need for resource flexibility, but also

for storage. Stored energy resources have the potential to address both the need for

flexibility and reduce the incidence of negative prices. The MMU views integration of

storage resources in the SPP markets as an ongoing high priority as several

outstanding items beyond compliance with FERC Order No. 841 need to be

addressed in order to fully integrate electric storage resources in the SPP markets.



• Address inefficiency when forecasted resources under-schedule day-ahead –

Systematic under-scheduling of wind resources in the day-ahead market can

contribute to distorting market price signals, suppressing real-time prices, and

affecting revenue adequacy for all resources. Therefore, the MMU continues to

recommend that SPP and its stakeholders address this issue through market rules

changes that address the consequences of under-scheduling of forecasted supply of

resources in the day-ahead market. We consider this a high priority as it helps to

enhance market efficiency and improve price signals.

• Convert non-dispatchable variable energy resources to dispatchable – Non-

dispatchable variable energy resources exacerbate congestion, reduce prices for

other resources, increase the magnitude of negative prices, cause the need for

market-to-market payments, and force manual commitments of resources that can

increase uplifts. Going forward, resource flexibility is essential to integrate an

increasing volume of wind generation in the SPP market. While the MMU’s preferred

solution did not pass the SPP stakeholder process, an acceptable amended solution

passed the board and was submitted to FERC in late 2018. FERC approved the filing

in April 2019.

• Address gaming opportunity for multi-day minimum run time resources – Resources

with minimum run times greater than two days have the opportunity to game the

market. The current implementation of the market rules limit make-whole payments

to the as-committed market offers for the first two days of a resource’s minimum run

time. However, after the second day, no rule exists to limit make-whole payments for

a resource that increases its offers from the third day onward until the resource’s

minimum run time is satisfied. For resources with minimum run times greater than two

days, the market participant knows that the resource is required to run and can

increase their market offers after the second day to increase make-whole payments.

The SPP board passed a proposal12 at the July 2018 meeting. However, subsequent

to board approval of the proposal, SPP legal staff identified internally inconsistent

tariff language that was revealed by but not addressed in the revisions. The MMU

strongly recommends that SPP and its stakeholders clean up the tariff language

12 Revision request 306 (2014 ASOM MWP MMU Recommendation [3-Day Minimum Run Time])



necessary to implement the changes to address the gaming issue and provide an

anticipated implementation date for the system changes. Addressing this matter in a

timely manner is a high priority for the MMU.

• Replace day-ahead must-offer, add physical withholding provision – FERC rejected

SPP’s proposal to remove the day-ahead must offer requirement and indicated that it

would consider removal of the requirement if it were paired with additional physical

withholding provisions.

While the MMU remains concerned with the current day-ahead must offer

requirement, we recommend that further consideration of this issue be a low priority.

The MMU will continue to track market performance concerns related to this provision

and will consider raising the priority on this matter if further issues are identified or

current issues are exacerbated. Otherwise, we regard other matters as having higher

impacts to the market and priority for development at this time.

Southwest Power Pool, Inc. Load and resources Market Monitoring Unit


2 LOAD AND RESOURCES

This chapter reviews load and resources in the SPP market for 2018. Key points from this

chapter include:

• Total system energy consumption was up six percent from 2017 to 2018.

• Variations in demand continue to trend with seasonal temperature changes and

departures from normal temperatures

• Nearly 2,300 MW of wind generation was added to the market in 2018. Wind

generation now accounts for 23 percent of installed nameplate capacity in the SPP

market.

• The generation interconnection queue continues to grow, with over 83,000 MW of

projects in the queue. Only 312 MW is from fossil fuel generation, with the remainder

renewable or storage resources.

• Of total energy produced in 2018, 42 percent was from coal resources, while wind

resources produced nearly 24 percent of energy produced. For comparison, in 2014

coal generation represented 60 percent of the total, and wind generation accounted

for 12 percent of the total.

• SPP remained a net exporter for 2018 increasing from an hourly average of just over

350 MW in 2017 to just over 500 MW in 2018.

• Exports to MISO and AECI decreased from 2017, but tight supply in ERCOT over the

summer drove exports higher than previous years.

• September experienced the highest amount of exports when SPP’s real-time price of

$21/MWh was lower than both ERCOT and MISO prices.

• Dispatchable transactions in the day-ahead increased, but still only contributed to 10

percent of total volume for the year.

• Market-to-market payments from MISO to SPP totaled almost $25 million for 2018

with almost half occurring in the first two months of the year.



2.1 THE INTEGRATED MARKETPLACE

SPP is a Regional Transmission Organization (RTO) authorized by the Federal Energy

Regulatory Commission (FERC) to ensure reliable power supplies, adequate transmission

infrastructure, and competitive wholesale electricity prices. FERC granted RTO status to SPP

in 2004. SPP provides many services to its members, including reliability coordination, tariff

administration, regional scheduling, reserve sharing, transmission expansion planning,

wholesale electricity market operations, and training. This report focuses on the 2018

calendar year of the SPP wholesale electricity market referred to as the Integrated

Marketplace, which started on March 1, 2014.

The Integrated Marketplace is a full day-ahead market with transmission congestion rights,

virtual trading, a reliability unit commitment process, a real-time balancing market, and a

price-based operating reserves market. SPP simultaneously put into operation a single

balancing authority as part of the implementation of the Integrated Marketplace. The

primary benefit of the introduction of a day-ahead market was to improve the efficiency of

daily resource commitments. Another benefit of the new market includes the joint

optimization of the available capacity for energy and operating reserves.

2.1.1 SPP MARKET FOOTPRINT

The SPP market footprint is located in the westernmost portion of the Eastern

Interconnection, with Midcontinent ISO (MISO) to the east, Electric Reliability Council of Texas

(ERCOT) to the south, and Western Electricity Coordinating Council (WECC) to the west.

Figure 2—1 shows the current operating regions of the nine RTO/ISO markets in the United

States and Canada, as well as a more detailed view of the SPP footprint. The SPP market also

has connections with other non-RTO/ISO areas such as Saskatchewan Power Corporation,

Associated Electric Cooperative, and Southwestern Power Administration.13

13 Southwestern Power Administration belongs to the SPP RTO, Reliability Coordinator (RC), and Reserve Sharing Group (RSG) footprints. Associated Electric Cooperative belongs to the SPP RSG.



Figure 2—1 SPP market footprint and RTO/ISO operating regions

2.1.2 SPP MARKET PARTICIPANTS

At the end of 2018, 226 entities were participating in the SPP Integrated Marketplace. SPP

market participants can be divided into several categories: regulated investor-owned

utilities, electric cooperatives, municipal utilities, federal and state agencies, independent

power producers, and financial only market participants that do not own physical assets.

Figure 2—2 shows the distribution of the number of the 77 resource owners registered to

participate in the Integrated Marketplace.



Figure 2—2 Market participants by type

The number of independent power producers is high because most wind producers are

included in this category. Market participants referred to as an “agent” represent several

individual resource owners that would individually be classified as different types, such as

municipal utilities, electric cooperatives, and state agencies.

Figure 2—3 shows generation nameplate capacity owned by the type of market participant.

Investor-owned utilities and cooperatives own two-thirds of the nameplate generation

capacity in the SPP market.

Figure 2—3 Capacity by market participant type

Cooperatives7

Investor owned utilities7

Independent power producers

45

Marketer3

Municipal9

Industrial1

Agent1

State agency3

Federal agency1

77 resource owners totalas of December 31, 2018

Cooperatives11,503

Investor owned utilities47,630

Independent power producers

10,477 Marketer2,197

Municipal3,690

Industrial20

Agent4,957

State agency6,094

Federal agency2,598

89,166 total nameplate capacity (MW) as of December 31, 2018



Although investor-owned utilities represent only a small percent of the total market

participants at nine percent, they own the majority of the SPP generation capacity at 53

percent. This is in contrast to the “independent power producer” category, which has a large

number of participants (58 percent) representing only a small portion (12 percent) of total

nameplate capacity.

2.2 ELECTRICITY DEMAND

2.2.1 SYSTEM PEAK DEMAND

One way to evaluate load is to review peak system demand statistics over an extended

period of time. The market footprint has changed over time as participants have been added

to or withdrawn from the market. The peak demand values reviewed in this section are

coincident peaks, calculated out of total generation dispatch across the entire market

footprint that occurred during a specific market interval. The peak experienced during a

particular year or season is affected by events such as unusually hot or cold weather, daily

and seasonal load patterns, and economic growth and change.

Figure 2—4 shows a month-by-month comparison of peak-day demand for the last three

years. The monthly peak demand in eight months in 2018 was higher than the monthly peak

demand in both 2016 and 2017. Heating and cooling demand increases contributed to this

change (see Section 2.2.4).

Figure 2—4 Monthly peak system demand

20,000

30,000

40,000

50,000

60,000

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Peak

load

(MW

)

2016 2017 2018



The SPP system coincident instantaneous peak demand in 2018 was 49,926 MW, which

occurred on July 12 at 3:35 PM. This is 2.5 percent lower than the 2017 system peak of

51,181 MW.

2.2.2 MARKET PARTICIPANT LOAD

Load continued to participate in the day-ahead market at high levels in 2018 as shown in

Figure 2—5.

Figure 2—5 Cleared demand bids in day-ahead market

The average monthly participation rates in the day-ahead market for load assets on an

aggregate level were between 99 and 101 percent of the actual real-time load. Accurate

reflection of demand in the day-ahead market economically incents generation to participate

in the day-ahead market. Additionally, accurate reflection of the load helps to converge

prices. Load participation in the day-ahead market increased by about one percent in 2018

from 2017.

Figure 2—6 depicts 2018 total energy consumption and the percentage of energy

consumption attributable to each entity in the market.

94%

96%

98%

100%

102%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2016 '17 '18

Cle

ared

dem

and

as

a p

erce

nt o

f lo

ad

Off-peak On-peak



Figure 2—6 System energy usage 2016 2017 2018

Energy

consumed (GWh)

Percent of system

Energy consumed

(GWh)

Percent of system

Energy consumed

(GWh)

Percent of system

American Electric Power 42,746 17.2% 41,887 17.0% 43,109 16.6%

Oklahoma Gas and Electric 28,078 11.3% 27,747 11.3% 29,411 11.3%

Southwestern Public Service Company 25,658 10.3% 25,826 10.5% 27,359 10.5%

Westar Energy * 23,885 9.6% 23,845 9.7% 25,101 9.7%

Basin Electric Power Cooperative 17,859 7.2% 18,665 7.6% 20,165 7.8%

Kansas City Power and Light, Co * 15,528 6.3% 15,194 6.2% 16,049 6.2%

The Energy Authority, NPPD 13,248 5.3% 13,206 5.4% 12,932 5.0%

Omaha Public Power District 11,168 4.5% 11,066 4.5% 11,431 4.4%

Kansas City Power and Light, GMOC * 8,423 3.4% 8,275 3.4% 8,815 3.4%

Western Farmers Electric Cooperative 8,448 3.4% 8,046 3.3% 8,312 3.2%

Grand River Dam Authority 5,957 2.4% 5,581 2.3% 5,805 2.2%

Golden Spread Electric Cooperative Inc. 5,132 2.1% 4,817 2.0% 5,684 2.2%

Empire District Electric Co. 5,144 2.1% 4,984 2.0% 5,413 2.1%

Sunflower Electric Power Corporation 4,732 1.9% 4,693 1.9% 4,906 1.9%

Western Area Power Administration, Upper Great Plains 4,477 1.8% 4,534 1.8% 4,471 1.7%

Arkansas Electric Cooperative Corporation 3,708 1.5% 3,675 1.5% 4,258 1.6%

Lincoln Electric System Marketing 3,515 1.4% 3,441 1.4% 3,570 1.4%

The Energy Authority, CU 3,332 1.3% 3,227 1.3% 3,424 1.3%

Oklahoma Municipal Power Authority 2,857 1.1% 2,766 1.1% 2,846 1.1%

Kansas City Board of Public Utilities 2,427 1.0% 2,347 1.0% 2,530 1.0%

Midwest Energy Inc. 1,710 0.7% 1,715 0.7% 1,762 0.7%

Northwestern Energy 1,651 0.7% 1,632 0.7% 1,748 0.7%

Kansas Municipal Energy Agency 1,480 0.6% 1,473 0.6% 1,557 0.6%

Tenaska Power Service Company 1,363 0.5% 1,365 0.6% 1,418 0.5%

Missouri River Energy Services 1,260 0.5% 1,226 0.5% 1,309 0.5%

City of Independence 1,065 0.4% 1,030 0.4% 1,094 0.4%

Municipal Energy Agency of Nebraska 1,015 0.4% 1,022 0.4% 1,057 0.4%

East Texas Electric Cooperative 983 0.4%

Kansas Power Pool 860 0.3% 843 0.3% 866 0.3%

City of Chanute 482 0.2% 487 0.2% 497 0.2%

City of Fremont 441 0.2% 433 0.2% 450 0.2%

Missouri Joint Municipal Electrical Utility Commission 450 0.2% 430 0.2% 445 0.2%

Big Rivers Electric Corporation 310 0.1%

MidAmerican Energy Company 284 0.1% 280 0.1% 285 0.1%

South Sioux City, Nebraska 225 0.1% 261 0.1%

Harlan Municipal Utilities 19 0.0% 17 0.0% 18 0.0%

NSP Energy 4 0.0% 5 0.0% 5 0.0%

Otter Tail Power Company 41 0.0% 3 0.0% 1 0.0%

System Total 248,446 246,009 259,653

* Evergy, Inc. 49,965 19.2%

*Evergy was formed in June 2018 and is the corporate parent of Westar Energy, Kansas City Power and Light, and Kansas City Power and Light GMOC.



The largest five entities account for 56 percent of the total system energy usage. This

concentration is understandable as SPP’s market is primarily composed of vertically

integrated investor-owned utilities, which tend to be large. When Evergy is considered one

entity in the market, it represents the largest user of energy in the SPP market footprint at 19

percent of the total, and the four largest entities would then comprise 58 percent of energy

consumed in the market. Overall, the total system energy usage in 2018 was almost six

percent above the 2017 level. This was driven by increased loads in 2018 related to

increased heating and cooling demand.

2.2.3 SPP SYSTEM ENERGY CONSUMPTION

Figure 2—7 shows the monthly system energy consumption in thousands of gigawatt-hours.

Figure 2—7 System energy consumption, monthly

Energy consumed was higher in 2018 during the first six months of the year, and again in

November. August 2017 saw lower than normal temperatures across the footprint, as

evidenced by the lower level of energy consumption in that period. For the year, total system

energy consumption was up by over 6 percent in 2018 compared to 2017.

Figure 2—8 depicts load duration curves from 2016 to 2018. These load duration curves

display hourly loads from the highest to the lowest for each year.

10

14

18

22

26

30


GW

h (t

hous

and

s)

2016 '17 '18



Figure 2—8 Load duration curve

In 2018, the maximum hourly average load was 47,522 MW, down from both prior years. The

minimum hourly load for 2018 was 20,203 MW, which was slightly above both previous years.

Comparing annual load duration curves shows differentiation between cases of extreme

loading events and more general increases in system demand. If only the extremes are

higher or lower than the previous year, then short-term loading events are likely the reason.

However, if the load curve is higher than the previous year, as it was in 2018, it indicates that

total system demand has increased.

Reference percentage lines indicate higher load levels in 2018 at all but the highest

percentages, above five percent. As mentioned above, the 2018 load duration curve reflects

an increase in overall load compared to the previous two years.

2.2.4 HEATING AND COOLING DEGREE DAYS

Changes in weather patterns from year-to-year have a significant impact on electricity

demand. One way to evaluate this impact is to calculate heating degree days (HDD) and

cooling degree days (CDD). These values can then be used to estimate the impact of actual

weather conditions on energy consumption, compared to normal weather patterns.

15000

25000

35000

45000

55000

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000

Load

(MW

)

Hour2016 2017 2018

25%50%

75%

5%



To determine heating degree days and cooling degree days for the SPP footprint, several

representative locations14 were used to calculate system daily average temperatures.15 In this

report, the base temperature separating heating and cooling periods is 65 degrees

Fahrenheit. If the average temperature of a day at a location is 75 degrees Fahrenheit, there

would be 10 (=75–65) cooling degree days at that location. If a day’s average temperature is

50 degrees Fahrenheit, there would be 15 (=65–50) heating degree days at that location.

Using statistical tools, the daily estimated load impact of a single cooling degree day is much

higher than the impact of a single heating degree day. This is in part because of more

electric cooling than electric heating.

Figure 2—9 shows monthly heating and cooling degrees over the last three years compared

to the total monthly load.

Figure 2—9 Heating and cooling degree days

14 Amarillo TX, Topeka KS, Oklahoma City OK, Tulsa OK, and Lincoln NE. After October 1, 2015, Bismarck ND was added to represent SPP’s expanded market footprint. 15 Daily average temperature is calculated as the average of the daily lowest and highest temperatures. The source of temperature data is the National Oceanic and Atmospheric Administration (NOAA).

0

10,000

20,000

30,000

0

1,000

2,000

3,000

Jan

Feb

Mar

Ap

rM

ayJu

nJu

lA

ug Sep

Oct

No

vD

ec Jan

Feb

Mar

Ap

rM

ayJu

nJu

lA

ug Sep

Oct

No

vD

ec Jan

Feb

Mar

Ap

rM

ayJu

nJu

lA

ug Sep

Oct

No

vD

ec

2016 2017 2018

Load

(GW

h)

Deg

ree

day

s

Heating degree days Cooling degree days Monthly load



As shown in the chart, cooling degree days are more prevalent in the higher load months of

May through September, whereas heating degree days are more prevalent in the other

months. The SPP market footprint experienced much warmer temperatures in May, June,

and August 2018 compared to the previous year, which resulted in higher demand during

those months. July temperatures were milder than last year, explaining the slight drop in

market demand in July.

Figure 2—10 shows heating degree days, cooling degree days, and load levels from 2016

through 2018 compared to a normal year.16 Normal 2018 load was derived from a

regression analysis of actual footprint heating degree days, cooling degree days, weekends,

and holidays, substituting footprint normal temperatures.

The figures indicate loads are influenced by cooling demand in the late spring and summer

months, whereas late fall and winter loads are, to a lesser degree, influenced by heating

demand. Moreover, the figures show that cooling degrees in 2018 were higher than the

average but similar to 2016 in most months, whereas heating degree days were more similar

to average but higher than both 2016 and 2017 in most months.

Figure 2—10 Degree days and loads compared with a normal year

16 30 year normal temperatures are from the 1981-2010 U.S. Climate normals product from NOAA.

14,000

18,000

22,000

26,000

30,000


GW

h

SPP system load

2016 2017 2018 normal



2.3 INSTALLED GENERATION CAPACITY

Figure 2—11 depicts the Integrated Marketplace installed generation for the SPP market

footprint at the end of the year. Total installed nameplate generation in the SPP Integrated

Marketplace was 89,167 MW by the end of 2018, representing an increase of over two

0

1,000

2,000

3,000

4,000


deg

ree

day

scooling degree days

2016 2017 2018 normal

0

1,000

2,000

3,000

4,000


deg

ree

day

s

heating degree days

2016 2017 2018 normal



percent from 2017.17 This increase was driven by a 17 percent increase in nameplate wind

capacity in 2018.

Figure 2—11 Generation nameplate capacity by technology type

Fuel type 2016 2017 2018 Percent as of year-end 2018

Coal 26,939 25,717 25,064 28%

Gas, simple-cycle 24,024 23,737 22,846 26%

Wind 16,114 17,596 20,589 23%

Gas, combined-cycle 12,870 12,618 13,248 15%

Hydro 3,428 3,422 3,431 4%

Nuclear 2,107 2,061 2,061 2%

Oil 1,684 1,639 1,639 2%

Solar 215 215 215 0%

Other 74 74 74 0%

Total 87,453 87,079 89,167 Note: Capacity is nameplate rating at year-end.

Natural gas-fired installed generation capacity still represents the largest share of generation

capacity in the SPP market at 42 percent (gas simple-cycle 26 percent, gas combined-cycle

15 percent), with coal being the second largest type at 28 percent. Wind continues to

increase due largely to new additions, with a 2018 market share of 23 percent of total

capacity in the SPP market. In terms of nameplate capacity, coal resources decreased from

17 The change in total generation capacity from year to year includes additions, retirements, and nameplate rating changes that occur during the year.

Coal28%

Gas, simple cycle26%

Wind23%

Gas, combined cycle15%

Hydro4%

Nuclear2%

Oil2%

Solar0% Other

0%

Year end 2018



30 percent of total nameplate capacity in 2017 to 28 percent in 2018. Wind nameplate

capacity more than offset the decline in nameplate coal capacity, increasing from 20 percent

of nameplate capacity in 2017 to 23 percent in 2018.

Figure 2—12 shows the total SPP aggregate real-time generation supply curves by offer price,

peak demand, and average demand for the summers of 2016 to 2018. Resources in

“outage” status were excluded from the supply curve. To calculate the summer supply

curves, a typical summer day was used for each analysis year. We calculated the aggregate

generation supply curves using the real-time offers of non-wind resources and wind forecast

data for wind resources.

Figure 2—12 Aggregate supply curve, typical summer day

Total aggregate real-time generation supply for summer 2018 was 65,295 MW, compared to

62,252 MW for summer 2017, an increase of nearly five percent. Just over 2,300 MW of

capacity was added to the SPP market in 2018, and most of the 1,900 MW of retirements

occurred after the summer season. The system peak-demand of 2018 was the lowest in the

last three years at about 2.5 percent lower than 2017. On the other hand, there was a six

percent increase of average demand of 2018 compared to 2017. Based on the heating and

cooling degree days analysis in Section 2.2.4, the SPP market footprint experienced much

warmer temperatures in May, June, and August 2018, which resulted in higher demand

during those months as compared to the previous year. July temperatures were milder than

last year, explaining the slight drop in market peak demand in 2018.

-$100

$0

$100

$200

$300

0 10 20 30 40 50 60 70

$/M

Wh

GW

2016 supply 2016 avg demand 2016 peak demand2017 supply 2017 avg demand 2017 peak demand2018 supply 2018 avg demand 2018 peak demand



Also evident is the approximately 24 GW gap between this maximum supply and the total

installed nameplate generation capacity on a typical summer day. This is primarily a result of

excluding the resources in “outage” status from the available supply (approximately 4 GW),

reduced summer capacity due to high ambient temperatures (approximately 6 GW), as well

as the difference between the wind forecast and installed capacity of wind resources

(approximately 13 GW).

The section of the offer curve below $0/MWh is mostly due to wind and solar energy and can

vary between 1,000 and 16,000 megawatts, based on wind and solar availability. Negative

offers typically reflect opportunity costs associated with state and federal tax incentives. The

sharp uptick in price at the top of the supply curves represents the transition from natural gas

units to oil units.

2.3.1 CAPACITY ADDITIONS AND RETIREMENTS

Figure 2—13 shows the capacity by the technology and the number of resources added in

2018.

Figure 2—13 Capacity additions

Just over 2,300 MW of generation capacity was added to the SPP market during 2018. Of the

new capacity added in 2018, 15 were wind resources accounting for 2,266 MW of capacity,

three were gas resources representing 66 MW of capacity, and one was a hydro resource

representing nine megawatts of capacity. All of the wind and gas capacity was new

construction.

Gas

Wind

Hydro

0 400 800 1,200 1,600 2,000 2,4000

4

8

12

16

20

Capacity (MW)

Res

our

ces



In 2018, the SPP market had generation retirements amounting to just over 1,900 MW of

installed capacity, shown in Figure 2—14.

Figure 2—14 Capacity retirements

Six medium-sized coal resources representing 954 MW of capacity, five gas resources

representing 924 MW of capacity, and one oil resource representing 54 MW of capacity were

retired in 2018.

A look at annual trends in additions and retirements can be found in Section 6.1.1.

2.3.2 GENERATION INTERCONNECTION

SPP is responsible for performing engineering studies to determine if the interconnection of

new generation within the SPP footprint is feasible, and to identify any transmission

development that would be necessary to facilitate the proposed generation. The generation

interconnection process involves a cluster study methodology allowing participants several

windows to submit requests for evaluation.18

Figure 2—15 shows the megawatts of capacity by generation technology type in all stages of

development. Included in this figure are interconnection agreements in the process of being

18 See Guidelines for Generator Interconnection Requests to SPP’s Transmission System http://sppoasis.spp.org/documents/swpp/transmission/studies/GuidelinesAndBusinessPracticesForGIP.pdf.

Coal

Gas

Oil

0 200 400 600 800 1000 12000

1

2

3

4

5

6

7

8

Capacity (MW)

Res

our

ces

http://sppoasis.spp.org/documents/swpp/transmission/studies/GuidelinesAndBusinessPracticesForGIP.pdf

http://sppoasis.spp.org/documents/swpp/transmission/studies/GuidelinesAndBusinessPracticesForGIP.pdf



created; those under construction; those already completed, but not yet in commercial

operation; and those in which work has been suspended as of year-end 2018.

Figure 2—15 Active generation interconnection requests, megawatts Prime mover 2016 2017 2018

Wind 32,690 28,147 55,118 Gas 1,080 3,122 312

Solar 3,770 15,306 24,583 Battery 40 1,135 4,305 Total 37,580 47,710 84,318

As can be seen in the table, generation capacity from renewable resources accounts for the

vast majority of proposed generation interconnection, with only 312 MW of non-renewable

generation interconnection requests in 2018. After a slight decrease of wind generation in

the queue from 2016 to 2017, the amount of wind generation of the queue nearly doubled

from 2017 to 2018. Interconnection requests for solar generation increased from 2017 to

2018, but not nearly as much as from 2016 to 2017. Battery interconnection requests also

increased from 2017 to 2018 with just over 4 GW in the queue at the end of 2018.

Development of renewable generation in the SPP region is expected to continue and the

proper integration of wind and solar generation is fundamental to maintaining the reliability

of the SPP system.

0

20

40

60

80

100

2016 2017 2018

Nam

epla

te c

apac

ity (G

W)

Wind Gas Solar Battery



Figure 2—16 Executed generation interconnection requests, on-schedule

As the chart above shows, at the end of 2018, just under 3 GW of generation is currently

slated to be added to the market in 2019. This is consistent with the levels of generation

additions in SPP in recent years. In the next three years, just over 100 MW of solar generation

is currently on-schedule to be developed. This would be the first solar generation added to

the market since 2016. It is important to note that generation can still be added or removed

from the list, even in the current year. However, there is more surety to the levels of

generation to go into production closer to the current year, and additions and deletions to

on-schedule projects are more typical in future years.

Development of renewable generation in the SPP region is expected to continue and the

proper integration of wind and solar generation is fundamental to maintaining the reliability

of the SPP system. Additional wind impact analysis follows in the Section 2.5.

2.4 GENERATION

2.4.1 GENERATION BY TECHNOLOGY

An analysis of generation by technology type used in the SPP Integrated Marketplace is

useful in understanding pricing, as well as the potential impact of environmental and

additional regulatory requirements on resources in the SPP system. Information on fuel types

and fleet characteristics is also useful in understanding market dynamics regarding

congestion management, price volatility, and overall market efficiency.

2019

2020

2021

Wind

Solar

0

2

4

6

Nam

epla

te c

apac

ity (G

W)

by year by fuel type



Figure 2—17 depicts annual generation percentages in the SPP real-time market by

technology type for the years 2007 through 2018.

Figure 2—17 Generation by technology type, real-time, annual

The long-term trend for coal-fired generation had been relatively flat through 2014 making

up around 60 to 65 percent of total generation, but has declined to under 50 percent

beginning in 2016. In 2018, coal generation accounted for 42 percent of total generation,

down from 46 percent in 2017. This decrease can primarily be attributed to increasing wind

generation and low natural gas prices.

The wind generation share continues to increase from three percent in 2007 to nearly 24

percent in 2018. Generation from simple-cycle gas units such as gas turbines and gas steam

turbines has seen a decline over the past few years, decreasing share from 13 percent in

2007 to just under five percent in 2017. However, with low gas prices during much of 2018,

the share increased to nearly eight percent. Gas combined-cycle generation has remained

relatively stable at about sixteen percent for the past three years, which can mostly be

attributed to low gas prices.

Some of the annual fluctuations in generation by technology type shares are driven by the

relative difference in primary fuel prices, namely natural gas versus coal. Gas prices in 2012,

and 2015 to 2018 were extremely low, resulting in some displacement of coal by efficient gas

generation, as can be seen in the higher generation from combined-cycle gas plants.

Another trend appears to be the increase in wind generation pushing simple-cycle gas

generation up the supply curve, making it less competitive.

0%

20%

40%

60%

80%

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Perc

ent

of t

ota

l gen

erat

ion

Coal Gas, simple cycle Gas, combined cycle Wind Hydro Nuclear



Retirement of older coal generation, environmental limits, along with competition from wind

and natural gas technologies are some of the factors that will continue to put pressure on

coal generation levels. Wind generation is expected to continue to increase in the years

ahead.

Figure 2—18 depicts the 2018 monthly fluctuation in generation by technology type.

Figure 2—18 Generation by technology type, real-time, monthly

Wind generation as a percentage of total generation is generally lowest in the summer

months, with peaks increasing above 30 percent in the highest wind generation months of

March and April. October is typically a high wind month, but wind generation was down in

October 2018, and was higher in November and December. The increase in wind

generation accompanied with low natural gas prices resulted in coal-fired generation market

share falling to below 40 percent from March through June 2018.

The monthly generation percentages show a general inverse relationship between monthly

levels of wind and coal generation in 2018. As wind generation increases, coal generation

decreases and vice versa. However, in November and December both coal and wind as a

percentage of total generation increased. This was primarily attributed to an increase in

natural gas prices and a subsequent decrease in natural gas-fired generation during the last

two months of 2018.

One method commonly used to assess price trends and relative efficiency in electricity

markets originating from non-fuel costs is the implied heat rate. The implied heat rate is

calculated by dividing the electricity price, net of a representative value for variable

0%

20%

40%

60%


Perc

ent

of t

ota

l gen

erat

ion

Coal Gas, simple cycle Gas, combined cycle Wind Hydro Nuclear



operations and maintenance (VOM) costs, by the fuel (gas) price.19 For a gas generator, the

implied heat rate serves as a “break-even” point for profitability such that a unit producing

output with an operating (actual) heat rate below the implied heat rate would be earning

profits, given market prices for electricity and gas. If the price of natural gas was $3/MMBtu,

and the electricity price was $24/MWh, the implied heat rate would be (24/3) = 8

MMBtu/MWh (8,000 Btu/kWh). This implied heat rate shows the relative efficiency required

of a generator to convert gas to electricity and cover the variable costs of production, given

market prices.

Figure 2—19 shows the monthly implied heat rate using real-time electricity prices for 2016 to

2018, along with an annual average for those years.

Figure 2—19 Implied heat rate

The chart shows a decline in implied heat rates in 2017 compared to 2016 and 2018. Of

note, wind generation and load affected electricity prices in 2017. For instance, high levels of

wind generation affected electricity prices in May and October, lowering average heat rates,

while high temperatures in July increased load and electricity prices. Implied heat rates in

December 2018 were low as a result of coal resources setting electricity prices more

19 For the implied heat rate calculation, natural gas units are assumed to be on the margin and accordingly, gas prices are taken as the relevant fuel cost. Emission costs are ignored in fuel cost as they rarely apply in the SPP market.

6,000

8,000

10,000

12,000


Imp

lied

hea

t rat

e (B

tu/K

Wh)

2016 2017 2018



frequently (see Section 2.4.2) as higher natural gas prices caused more gas resources to be

uneconomic.

2.4.2 GENERATION ON THE MARGIN

The system marginal price represents the price of the next increment of generation available

to meet the next increment of total system demand. The locational marginal price at a

particular pricing node is the system marginal energy price plus any marginal congestion

charges and marginal loss charges associated with that pricing node.

Figure 2—20 illustrates the frequency with which different technology types were marginal

and price setting. For a generator to set the marginal price, the resource must be: (a)

dispatchable by the market; (b) not at the resource economic minimum or maximum; and (c)

not ramp limited. In other words, it must be able to move to provide the next increment of

generation.

Figure 2—20 Generation on the margin, real-time, annual

This chart illustrates the dramatic shift in technology on the margin with natural gas

representing about 75 percent in the first year of an SPP market in 2007 to only about 52

percent in 2018. There is a corresponding shift in coal generation on the margin from about

25 percent in 2007, peaking at around 50 percent in 2015, then declining to 34 percent by

2018. With low natural gas prices, coal-fired plant owners are experiencing more daily

swings in dispatch level, which are reflected in this generation on the margin analysis.

0%

20%

40%

60%

80%

100%

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

% in

terv

als

on

the

mar

gin

Wind Coal Gas Gas, combined cycle Gas, simple cycle Other



It is worth noting the increase in wind generation being on the margin—from five percent in

2014 and 2015 to over 12 percent in 2018. With the growing amount of dispatchable wind

generation and an overall quantity of 23 percent of total nameplate capacity, wind

generation is increasingly becoming the marginal technology a higher percentage of the

time. At the end of 2017, 64 percent of wind capacity was dispatchable, compared to 60

percent at the end of 2016, 46 percent at the end of 2015, and 27 percent at the beginning

of the Integrated Marketplace in March 2014. The recommendation to convert non-

dispatchable variable energy resources to dispatchable variable energy resources is

discussed in Section 8.3.4.

Figure 2—21 shows monthly values for real-time generation on the margin for 2018.

Figure 2—21 Generation on the margin, real-time

Intervals with coal generation on the margin are typically lower in the spring and fall months,

resulting in more coal-fired units running as base load units with less cycling. The increased

wind generation is also affecting prices to some extent in every month of the year. The

higher wind generation on the margin values in the spring and fall are as expected given that

these periods are the windiest time of the year, as well as the lowest demand periods in the

SPP footprint.

Day-ahead generation on the margin, shown in Figure 2—22, is different from real-time in that

the day-ahead market includes virtual transactions. The real-time market does not include

0%

20%

40%

60%

80%

100%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec '16 '17 '18

% in

terv

als

on

the

mar

gin

Wind Coal Gas, combined cycle Gas, simple cycle Other



virtual transactions and is required to adjust to unforeseeable market conditions such as

unexpected plant and transmission outages.

Figure 2—22 Generation on the margin, day-ahead

Wind generation on the margin is comparable in the day-ahead and real-time markets with a

similar annual cyclical pattern. Both coal and gas generation on the margin in the day-ahead

market is noticeably lower than in the real-time market. The most significant difference is the

displacement of natural gas-fired generation by virtual offers in the day-ahead market. Virtual

energy offers on the margin have been declining over the past three years. However, virtual

energy offers represent 28 percent of the marginal offers in the day-ahead market in 2018,

compared to 25 percent in 2017 and 27 percent in 2016. While marginal virtual offers occur

at all types of settlement locations, 67 percent of marginal virtual offers are at resource

settlement locations, with a significant amount of activity at non-dispatchable wind

generation resource locations.

2.5 GROWING IMPACT OF WIND GENERATION CAPACITY

2.5.1 WIND CAPACITY AND GENERATION

The SPP region has a high potential for wind generation given wind patterns in many areas of

the footprint. Federal incentives and state renewable portfolio standards are additional

factors that have resulted in significant wind investment in the SPP footprint during the last

several years.

0%

20%

40%

60%

80%

100%


% in

terv

als

on

the

mar

gin

Wind Coal Gas, combined cycle Gas, simple cycle Other Virtual



Figure 2—23 is a wind speed map of the United States of America with the SPP footprint

outlined in black.

Figure 2—23 Wind speed map

Outside of coastal areas, much of the SPP footprint highlighted on the map is covered with

some of the highest wind speeds in the country. As has been discussed, there continues to

be a high potential for additional wind resource development in the SPP footprint going

forward.

Figure 2—24 depicts annual nameplate capacity and total generation of SPP wind facilities

since 2007.



Figure 2—24 Wind capacity and generation, annual

Total registered wind nameplate capacity at the end of 2018 was 20,589 MW, an increase of

17 percent from 2017. Wind generation output increased 10 percent in 2018 to nearly

65,000 GWh produced.

Wind resources comprise about 23 percent of the installed capacity in the SPP market,

behind only natural gas with 42 percent and coal with 28 percent. Consistent with previous

years, wind generation fluctuated seasonally as summer as usual was the low wind season,

while spring and fall were the high wind seasons. Also typical of wind patterns is lower

production during on-peak hours than off-peak. Furthermore, higher levels of wind

generation tend to coincide with the morning ramp periods.

Figure 2—25 shows the wind capacity factor. Note that the wind capacity factor is reported for

the entire month.20

20 Wind resources may be considered in-service, but not yet in commercial operation. In this situation, the capacity will be counted but the resource may not be providing any generation to the market.

0

20,000

40,000

60,000

80,000

0

6,000

12,000

18,000

24,000

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Gen

erat

ion

(GW

h)

Cap

acity

(MW

)

Capacity Generation



Figure 2—25 Wind capacity factor

The wind capacity factor in the real-time market dropped from 40.3 percent in 2017 to 39.5

percent in 2018, while the day-ahead wind capacity factor rose from 31.5 percent to 32.4

percent during the same period. Lower wind generation totals, coupled with increased wind

capacity, drove this drop in real-time capacity factor. The spread between the real-time and

the day-ahead wind capacity indicates a disconnect in the amount of wind in the real-time

market, compared to the forecasted wind in the day-ahead market.

Figure 2—26 shows the monthly real-time wind capacity factor for the past three years.

0%

10%

20%

30%

40%

50%

60%

Jan18

Feb18

Mar18

Apr18

May18

Jun18

Jul18

Aug18

Sep18

Oct18

Nov18

Dec18

2016 '17 '18

Win

d c

apac

ity fa

cto

r

Day-ahead capacity factor Real-time capacity factor



Figure 2—26 Wind capacity factor by month, real-time, three year

As shown above, the wind capacity factor for October and November 2018 was well below

the levels of the prior two years. These months, along with the spring, typically have the

highest wind generation of the year.

2.5.2 WIND IMPACT ON THE SYSTEM

Average annual wind generation as a percent of load continues to increase as shown in

Figure 2—27. The chart shows the trend for average and maximum wind generation as a

percent of load since 2007, illustrating the dramatic increase since the start of the SPP

markets.

0%

10%

20%

30%

40%

50%

60%


Win

d c

apac

ity fa

cto

r

2016 2017 2018



Figure 2—27 Wind generation as a percent of load

Average wind generation as a percent of load in the real-time market increased one percent

to nearly 25 percent in 2018. The growth of average wind generation as a percent of load

has leveled off after climbing much more sharply from 2015 to 2017. Wind generation

peaked at 16,329 MW in 2018 on a five-minute interval basis, an increase of over four percent

from 15,675 MW in 2017. Wind generation as a percent of load for any five-minute interval

reached a maximum value of 64 percent, which was up from 50 percent in 2016 and 57

percent in 2017.

Figure 2—28 shows wind production duration curves that represent wind generation as a

percent of load by real-time (five-minute) interval for 2016 through 2018.

0%

20%

40%

60%

80%

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Win

d g

ener

atio

n as

a p

erce

nt o

f lo

ad

average maximum



Figure 2—28 Wind production curve

The shift upward for the curve from year-to-year reflects an increase in total wind generation

on an annual basis. The wind production curves for 2017 and 2018 are nearly identical for

the lower two-thirds of the curve, while in the upper third of the curve, 2018 values climb

above 2017. Wind generation in both 2017 and 2018 served at least 23 percent of the total

load during half of the year, compared to 17 percent in 2016. It is also important to note that

the low end curve continues to gradually approach zero while the high end of the curve

increases very steeply.

Figure 2—29 below shows average hourly demand, along with wind generation, and net

demand (demand minus wind generation) for 2018.

0%

10%

20%

30%

40%

50%

60%

70%

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000

Win

d g

ener

atio

n as

per

cent

of l

oad

Wind production intervals

2016 2017 2018

10% 50%



Figure 2—29 Net demand by hour of day

With wind generation at the highest levels in the overnight hours, net demand climbs more

steeply than total demand, as wind generation begins to taper off when approaching the

peak hours of the day. This can have an impact on the market as generation needed from

traditional resources climbs more quickly than demand. When this occurs, ramp scarcity is

more likely. This is discussed in Section 3.2.1

2.5.3 WIND INTEGRATION

Wind integration brings low cost generation to the SPP region but does not count for much

accredited capacity. There are a number of operational challenges in dealing with

substantial wind capacity. For instance, wind energy output varies by season and time of day.

This variability is estimated to be about three times more than load when measured on an

hour-to-hour basis. Moreover, wind is counter-cyclical to load. As load increases (both

seasonally and daily), wind production typically declines. The increasing magnitude of wind

capacity additions since 2007, along with the concentration, volatility, and timeliness of wind,

can create challenges for grid operators with regard to managing transmission congestion

and resolution of ramping constraints (which began being reflected in scarcity pricing in May

2017) as well as challenges for short- and long-run reliability. Several price spikes occurred

because of wind forecast errors. Wind forecast errors are also the leading cause of day-

ahead and real-time price divergence.

In the SPP market, wind and other qualifying resources were allowed to register as non-

dispatchable variable energy resources, provided the resource had an interconnection

0

10

20

30

40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

GW

h

Hour ending

Demand Net demand Wind generation



agreement executed by May 21, 2011 and was commercially operated prior to October 15,

2012. Because 29 percent (5,933 MW) of the existing installed wind capacity is composed of

non-dispatchable variable energy resources, and these generally produce without regard to

price, SPP operators must still issue manual dispatch instructions to reduce or limit their

output at certain times.

Figure 2—30 illustrates dispatchable variable energy resources (DVERs) and non-dispatchable

variable energy resources (NDVERs) wind output since 2016, with dispatchable variable

energy resource output mirroring the increasing percentage of installed wind capacity.

March 2018 saw over 6,300 GWh of monthly wind production, which was the highest since

the start of the Integrated Marketplace and 30 percent of this output originated from non-

dispatchable variable energy resource capacity.

Figure 2—30 Dispatchable and non-dispatchable wind generation

Figure 2—29 also shows the amount of reduced output of dispatchable variable energy

resources below their forecast (grey line). This depicts the increase in reductions of

dispatchable variable energy resource dispatch output from 2016 to 2017 but fell back to

2016 levels in 2018. This change was likely the result of increased transmission that

alleviated congestion bottlenecks and higher loads. Reductions in dispatchable resource

generation also follow the seasonal pattern of lower wind output during the summer months,

resulting in the decrease in need to reduce dispatchable variable energy resource output

during these times. This increase in dispatchable wind capacity has helped in the

management of congestion caused by high levels of wind generation in some of the western

parts of the SPP footprint.

0

100

200

300

400

0

2,000

4,000

6,000

8,000


Red

uced

dis

pat

ch (G

Wh)

Gen

erat

ion

(GW

h)

Non-dispatchable Dispatchable Dispatchable resource reduced dispatch



Substantial transmission upgrades in the SPP footprint over the past few years have provided

an increase in transmission capability for wind-producing regions, helping to address

concerns related to high wind production, and resulting congestion. The increased

transmission capability directly reduces localized congestion, creating a more integrated

system with higher diversity and greater flexibility in managing high levels of wind

production. However, given the historical growth of wind capacity and indicators of future

additions in the generation interconnection queue, additional transmission upgrades may

entice further development of wind capacity.

Figure 2—31 shows the number of out-of-merit energy directives (manual dispatches) initiated

for dispatchable and non-dispatchable variable energy wind resources for the past three

years.

Figure 2—31 Manual dispatches for wind resources

As expected, manual dispatches are fewer during the lower wind output and higher demand

months of summer. In 2018, 20 percent of the 444 manual dispatches were for dispatchable

variable energy wind resources, whereas almost 49 percent were for non-dispatchable

variable energy wind resources. Line loading in excess of 104 percent, operating guides,

0

20

40

60

80


Mau

nal d

isp

atch

es

2016 2017 2018



and outages caused 75 percent of dispatchable manual dispatches. These same factors plus

transmission switching21 caused 80 percent of non-dispatchable manual dispatches.

SPP is at the forefront among RTOs in managing wind energy integration. The Integrated

Marketplace has reliably managed wind generation when it represented more than 60

percent of load. Even though the use of manual dispatch is limited and SPP continues to see

an expanding dispatchable wind generation fleet, ramping capability is needed because of

the variability of wind. Since May 2017, ramp shortages are reflected in prices. Section 3.2.1

discusses the pricing of ramp shortages.

2.5.4 CHALLENGES WITH NON-DISPATCHABLE VARIABLE ENERGY RESOURCES

A non-dispatchable variable energy resource is defined as “a variable energy resource not

capable of being incrementally dispatched down by the transmission provider.”22 This

definition does not delve into the requirements of a non-dispatchable variable energy

resource. However, the market design requires that these resources, barring absence of fuel

or mechanical limitations, follow close to their current output or forecast. This concept also

applies to dispatchable variable energy resources not receiving a signal to follow dispatch

and all resources in manual control status that are not in start-up or shutdown. Significant

deviation from the most recent actual or forecasted output causes market inefficiencies that

will be evaluated by the MMU.

Large swings in generation from non-dispatchable variable energy resources responding to

the ex-ante real-time price is known as “price chasing”. This behavior introduces oscillations

on constraints, adversely impacting prices and dispatch instructions for other resources as

well as an impact on regulation products. Price chasing occurs when non-dispatchable

variable energy resources or resources on manual control respond to prices by curtailing

output in response to lower prices and increasing production when prices rise. Such

behavior can cause operational and market issues. For instance, it can create breaches on

flowgates when these resources raise output in response to a price increase. This in turn

21 Transmission switching out-of-merit instructions are issued to accommodate switching of 345kV transmission lines, because of stability concerns during the switching process. Typically, these instructions last from two hours prior to switching to two hours after switching is completed, whereas the 345kV line may be out of service for a longer time frame. 22 SPP Tariff, Attachment AE, Section 1.1.



causes more relief than necessary and security constrained economic dispatch effectiveness

declines. Other impacts include additional volatility in the real-time market, more regulation

needs, and more output loss due to increased regulation. Operators have at times resorted

to reducing line ratings to ensure system reliability. As a result, out-of-merit energy directives

are issued to other resources, which means extra cost (uplift) to the system, which translates

into reduced market efficiency.

In addition to the inefficiencies introduced to the market because of price chasing behavior,

there are also inefficiencies introduced by non-dispatchable variable energy resources when

they are physically incapable of responding to dispatch signals or when they are acting as

“price takers”. These inefficiencies exist at times when a non-dispatchable variable energy

resource is operating uneconomically, even when considering state and/or federal subsidies.

This results in transmission congestion at times not relieved by the most price effective unit

possible, creating greater price differences and volatility.

The Market Working Group passed a proposal by SPP at the February 2018 meeting that

required full conversion of non-dispatchable variable energy resources. The proposal failed

to pass the Market Operations and Policy Committee meeting in April and was appealed to

the SPP board. The SPP board tabled discussion in April. In July, the Market Worked Group

passed a modified version of the proposal that allowed exemptions for resources exercising

rights under PURPA23 and for run-of-river hydro that could not be dispatchable. This

subsequently passed the July Market Operations and Policy Committee meeting and SPP

board.

SPP filed with FERC to require conversion of non-dispatchable resources to dispatchable

resources in late 2018 and received approval from FERC in April 2019.

2.6 EXTERNAL TRANSACTIONS

2.6.1 EXPORTS AND IMPORTS

The SPP Integrated Marketplace has greater than 6,000 megawatts of AC interties with MISO

to the east, 810 megawatts of DC ties to ERCOT to the south, and over 1,000 megawatts of

DC ties to WECC to the west. Additionally, SPP has over 1,500 megawatts of interties with

23 Public Utilities Regulatory Policies Act of 1978



the Southwestern Power Administration (SPA) in Arkansas, Missouri, and Oklahoma, and over

5,000 megawatts of AC interties the Associated Electric Cooperative (AECI) in Oklahoma and

Missouri.

As shown in Figure 2—32, SPP has been a net exporter in real-time since 2016. It is important

to note that because of the addition of the Integrated System to SPP in October 2015,

transactions that were external imports from Western Area Power Administration (WAPA)

became internal transactions within the SPP footprint.

Figure 2—32 Exports and imports, SPP system

Generally, SPP exports follow the wind production curve for the day. As wind generation

increases, exports typically increase. However, in 2018, exports were highest in September

and August, reaching almost 2,000 MW and 1,700 MW of gross exports, respectively. This

was driven by lower real-time prices in SPP during these summer months compared to higher

prices in ERCOT and MISO.

Exports to ERCOT were driven by tight supply conditions and high prices during the summer

months. Southwestern Power Administration hydro power is imported to serve municipals

tied to SPP transmission and is highest during on-peak hours, but is scheduled day-ahead.

MISO interchange generally follows wind production, while AECI interchange is coordinated

on an ad hoc basis. DC tie imports and exports are scheduled hourly, and the DC ties are not

responsive to real-time prices. Nonetheless, many exports and imports with ERCOT and

MISO are adjusted based on day-ahead price differences in the organized markets and

-1,500

-1,000

-500

0

500

1,000

1,500

2,000


Ave

rag

e M

W

Exports Imports Net



expectations of renewable generation. Interchange with SPA, AECI, and Western

Interconnection parties is less responsive to prices. Figure 2—33 through Figure 2—36 show

the data for the four most heavily used interfaces in real-time, namely ERCOT, SPA, MISO,

and AECI.

Figure 2—33 Exports and imports, ERCOT interface

Figure 2—34 Exports and imports, Southwestern Power Administration interface

-200

0

200

400

600


Ave

rag

e M

W

Exports Imports Net

-800

-600

-400

-200

0

200


Ave

rag

e M

W

Exports Imports Net



Figure 2—35 Exports and imports, MISO interface

Figure 2—36 Exports and imports, Associated Electric Cooperative interface

Interchange transactions in the SPP market can be scheduled in the real-time market, as well

as in the day-ahead market. The day-ahead market has three types of interchange

transactions:

• Fixed interchange transactions are physical transactions that bring energy into or out

of the SPP balancing authority. Energy prices are settled at the price at the applicable

external interface settlement location. Submitters of this type of transaction in the

Integrated Marketplace are price takers for that energy.

-500

-250

0

250

500

750

1000


Ave

rag

e M

W

Exports Imports Net

-200

-100

0

100

200

300

400


Ave

rag

e M

W

Exports Imports Net



• Dispatchable interchange schedules are physical transactions that bring energy into

or out of the SPP balancing authority and specify a bid or offer for an amount of

megawatts. These schedules are supported in the day-ahead market only and also

must meet all market requirements. Prices are determined in the day-ahead market at

the appropriate external interface settlement location representing the interface

between the SPP balancing authority and the applicable external balancing authority.

• An up-to-transmission usage charge (or up-to-TUC) offer on an interchange

transaction specifies both a megawatt amount and the maximum amount of

congestion cost and marginal loss cost the customer is willing to pay if the transaction

is cleared in the day-ahead market.

All interchange transactions cleared in the day-ahead market, regardless of type, become

fixed interchange transactions in the reliability unit commitment and real-time market.24

As shown in Figure 2—37, of the 1,657 MW of day-ahead import and export transactions in

2018, 90 percent were fixed in the day-ahead market, ten percent were dispatchable, and

none were up-to-TUC. Dispatchable transactions increased threefold from an average of 56

MW in 2017 to an average of 173 MW in 2018. Dispatchable transactions were highest in the

summer months peaking at almost 400 MW in August.

24 Per Market Protocols section 4.2.2.7 Import Interchange Transaction Offers.



Figure 2—37 Imports and export transactions by type, day-ahead

Some reasons for the fixed transactions that make up the vast majority of interchange

transactions include bilateral contracts with external entities, Southwestern Power

Administration hydro contracts, and generally lower prices of the SPP market compared to

other RTOs. To enhance market efficiency, market participants should consider further use of

the dispatchable and up-to-TUC imports and exports, which allow for a specific strike price to

be set, allowing for more economic imports and exports.

2.6.2 MARKET-TO-MARKET COORDINATION

SPP began the market-to-market (M2M) process with MISO in March 2015 as part of a FERC

requirement that also included regulation compensation and long-term congestion rights.

These were required to be implemented one year after go-live of the SPP Integrated

Marketplace. The market-to-market process under the joint operating agreement allows the

monitoring RTO and non-monitoring RTO to efficiently manage market-to-market constraints

by exchanging information (shadow prices, relief request, control indicators, etc.) and using

the RTO with the more economic redispatch to relieve congestion.25

25 Essentially, the RTO which manages the limiting element of the constraint is the monitoring RTO. In most cases, the monitoring RTO has most of the impact and resources that provide the most effective relief of a congested constraint.

0

500

1,000

1,500

2,000


MW

h

Fixed Dispatchable Up to TUC



Each RTO is allocated property rights on market-to-market constraints. These are known as

firm flow entitlements (FFE), and each RTO calculates its real-time usage, known as market

flow. RTOs exchange money (market-to-market settlements) for redispatch based on the

non-monitoring RTO’s market flow in relation to its firm flow entitlement. The non-monitoring

RTO receives money from the monitoring RTO if its market flow is below its firm flow

entitlement. It pays if above its firm flow entitlement. Figure 2—38 shows payments by month

between SPP and MISO (positive is payment from MISO to SPP and negative is payment from

SPP to MISO.)

Figure 2—38 Market-to-market settlements

For 2018, total market-to-market payments from MISO to SPP totaled almost $25 million,

while market-to-market payments from SPP to MISO totaled nearly $9 million, resulting in a

net payment of approximately $16 million from MISO to SPP for the year. Almost half of the

payments from MISO to SPP occurred in the first two months of 2018. Of the $12 million of

payments in January and February, $9 million is attributed to the Neosho – Riverton 161kV

and Nashua transformer constraints. These constraints are discussed more below.

Figure 2—39 shows market-to-market payments (over $100,000 either from SPP to MISO, or

MISO to SPP) by flowgate for 2018.

-$8

$0

$8

$16

$24

$32

-$2

$0

$2

$4

$6

$8


Mill

ions

Mill

ions

Payments from SPP to MISO Paymnents from MISO to SPP Net payments



Figure 2—39 Market-to-market settlements by flowgate

Seven flowgates had payments from MISO to SPP over $500,000. Three flowgates had

payments from SPP to MISO over $500,000 while the highest was just over $1 million. The

Neosho-Riverton 161kV flowgate was highly congested in the latter half of 2017 and first few

months of 2018. The line was uprated later in 2018, which alleviated some of the congestion

on this constraint. This constraint is impacted by wind and external flows and is discussed in

more detail in Section 5.1.4.2. Market-to-market payments from MISO to SPP in 2018 on this

flowgate totaled over $9 million which was half of the $18 million in payments in 2017. The

Nashua transformer market-to-market payments totaled over $4 million compared to over

$900,000 in 2017.

Market-to-market allows for a coordinated approach between markets to provide a more

economical dispatch of generation to solve congestion. In most cases, MISO is paying SPP to

help resolve congestion at a lower cost than what was available to MISO and in a few cases,

SPP pays MISO to help resolve congestion. The following are points of interest on the SPP

and MISO seam. The MMU feels continued improvement in these areas would lead to

improved benefits for both markets.

2.6.2.1 Transmission planning

SPP and MISO continue to discuss options to overcome hurdles in coordinated projects

along the seam. Differing cost allocation methodologies, assumptions for separate

transmission planning studies, and separate approval processes can lead to varying results

-$2

$0

$2

$4

$6

Mill

ions

Payments from MISO to SPP Payments from SPP to MISO

$9



within each RTO/ISO. Benefits are also viewed differently within each region due to the cost

allocation methodology differences.

2.6.2.2 Price/power swings

SPP and MISO implemented the ability to transfer monitoring and non-monitoring RTO roles

in December 2017. MISO and PJM have been utilizing this function to address constraint

volatility or power swings when the non-monitoring RTO may have more “effective control”

on certain constraints. This added feature helps alleviate power swings in cases where an

RTO has more effective control than the other. However, instances where both RTOs have

almost equal amount of flow on a constraint still results in oscillations. SPP and MISO also

implemented the ability to utilize shadow price overrides which will limit the non-monitoring

RTO’s upward swing when limited capability exists on a constraint.

2.6.2.3 Third party impacts

The transmission loading relief (TLR) process is used when tagged impacts or other external

market/non-market flows (other than MISO) impacts are present on an SPP constraint. The

market monitor believes that the transmission loading relief process is not needed when the

SPP and MISO markets have the majority of impacts, but is still needed when external impacts

from non-market (third party) entities are significant. Assuming interface price definitions

correctly reflect congestion, tagged transactions should respond to the market conditions

and either withdraw or delay submitting tags during congestion. Thus, this should alleviate

the need for transmission loading relief when impacts on the constraint are mostly between

SPP and MISO.

When third-party impacts exist, the MMU believes transmission loading relief is warranted to

subject the third party to redispatch.26 A scenario observed between SPP and MISO entailed

third-party firm network and native load impacts that are not subjected to redispatch by

either market. The third party does not have a market signal in the form of a price and by the

absence of a transmission loading relief will not have an obligation to provide relief on the

constraint. Transmission loading relief is not as efficient as a market using price and dispatch

to manage congestion on a constraint. Market-to-market is the preferred method in

addressing congestion along the seams, but until further development is made in areas

26 Third parties impacting the SPP-MISO seam typically include Tennessee Valley Authority, Associated Electric Cooperative Incorporated, and Southwestern Power Administration.



outside RTO markets, transmission loading relief is the current mechanism to manage

impacts between markets and non-markets.

2.6.2.4 Market flow methodology

SPP, MISO, and PJM calculate market flows differently. MISO and PJM use a marginal zone

methodology (although the margins are derived in different manners), and SPP uses a

tagging impact approach. This topic is not discussed by SPP and MISO, but because market

flow is a component used in market-to-market settlements, the MMU suggests this topic

should be revisited to ensure consistency and equitable measurements across RTOs.

Southwest Power Pool, Inc. Unit commitment and dispatch processes Market Monitoring Unit


3 UNIT COMMITMENT AND DISPATCH PROCESSES

This chapter covers unit commitment and dispatch, scarcity pricing, ramp, and virtual

transactions. Key points from this chapter include:

• In 2018, day-ahead commitments increased by six percent and self-commitments

decreased by four percent. This is a positive trend.

• During the fall months, short-term outages increased by 13 percent from 2017 to

2018.

• In 2018, there were 167 intervals with operating-reserve scarcity. The average scarcity

price for these events was $362/MW. This was roughly five times as many intervals of

scarcity as was seen in 2017. Sixty-four of the 167 intervals of operating reserve

scarcity happened in October.

• Over one-third of the regulation-down scarcity events happened in the first interval of

the hour. This trend has held since the inception of the SPP marketplace.

• SPP is in the process of designing a ramp product which should help to reduce the

frequency of scarcity events.

• Cleared virtual energy offers as a percentage of load were flat year-over-year;

however, cleared virtual energy bids as a percentage of load increased 12 percent

year-over-year.

• Virtual energy offers continue to transact most frequently at wind resources, whereas

virtual energy bids continue to occur most frequently at non-variable energy

resources.

• Virtual profits before fees decreased 19 percent year-over-year and virtual profits after

fees decreased 42 percent year-over-year.

• Average profit per cleared virtual megawatt decreased almost 50 percent in 2018.

• Virtual profits before fees exceeded total virtual losses by more than forty fold,

whereas virtual profits after fees exceeded total virtual losses by almost seven fold.



3.1 COMMITMENT PROCESS

The Integrated Marketplace uses centralized unit commitment to determine an efficient

scheduling and dispatch of generation resources to meet energy demand and operating

reserve requirements. The principal component of the commitment process is the day-ahead

market, which determines a least cost commitment schedule that meets day-ahead energy

demand and operating reserve requirements simultaneously. Most of the time it becomes

necessary to commit additional capacity outside the day-ahead market to ensure all reliability

needs are addressed and to adjust the day-ahead commitment for real-time conditions. This

is done through the reliability unit commitment (RUC) processes and manual commitments.

SPP employs five reliability commitment processes:

• multi-day reliability assessment;

• day-ahead reliability unit commitment (DA RUC) process;

• intra-day reliability unit commitment (ID RUC) process;

• short-term intra-day reliability unit commitment (ST RUC) process; and

• manual commitment instructions issued by the RTO.

Figure 3—1 shows a timeline describing when the various commitment processes are

executed.

Figure 3—1 Commitment process timeline

Multi-day reliability assessments are made for at least three days prior to an operating day.

This assessment determines if any long lead-time generators are needed for capacity for the

operating day. Any generator committed from this process is treated as a “must commit” in

the day-ahead market. The day-ahead closes at 0930 Central time and is executed on the

day before the operating day, with the results posted at 1400. The day-ahead reliability unit

commitment process is executed approximately 45 minutes after the posting of the day-

ahead market results. This allows market participants time to re-offer their uncommitted

resources.



The intra-day reliability unit commitment process is run throughout the operating day, with at

least one execution occurring every four hours. The short-term intra-day reliability unit

commitment may be executed as needed to assess resource adequacy over the next two

hour period as part of the intra-day process. SPP operators may also issue manual

commitment instructions for capacity, transmission, or local reliability during the operating

day to address reliability needs not fully reflected in the security constrained unit

commitment algorithm used in the day-ahead and reliability unit commitment processes.

Transmission operators occasionally also issue local reliability commitments.

3.1.1 RESOURCE STARTS

The SPP resource fleet, excluding variable energy resources, had 35,568 starts during 2018.

This is up 26 percent from 28,166 starts last year. This increase can primarily be attributed to

an increase in starts by gas combustion turbine resources. The following two tables and

graphs provide a breakdown of the origins of the commitment decisions for resources. For

all generation participation offers in the day-ahead market by commitment status see Figure

3—10.

Figure 3—2 Start-up instructions by resource count 2016 2017 2018 Day-ahead market 59% 68% 74% Self-commitment 19% 15% 11% Intra-day RUC 10% 8% 7% Manual, regional reliability 5% 4% 5% Short-term RUC 3% 2% 3% Day-ahead RUC 2% <1% <1% Manual, local reliability 2% 1% 1%

Day-ahead market74%

Self-commitment11%

Intra-day RUC7%

Manual, regional reliability

5%Short-term RUC

3%Day-ahead RUC

0%Manual, local reliability

1%2018 total starts: 35,568



Seventy-four percent of start-up instructions in 2018 were a result of the day-ahead market.

As shown in Figure 3—2, each year there has been an increase in day-ahead market starts and

a decrease in self-commitment starts. This is an encouraging trend as it leads to greater

market efficiency and suggests that resource owners may be gaining confidence in allowing

the market to start-up resources. However, a limiting factor on the number of day-ahead

commitments is that the optimization algorithm is restricted to a 48-hour window; hence,

large base-load resources with long lead-times and long run times may not appear economic

to the day-ahead market commitment algorithm. Some market participants choose to self-

commit these resources, which contributes to many self-commitments. Nonetheless, many

market participants have improved their operating practices to decrease the start-up times

on the units to take advantage of the market commitment process. The day-ahead, intra-day,

short-term, and manual reliability unit commitments represent 13 percent of the resource

start-ups.

Figure 3—3 is based on capacity committed and provides a slightly different look at the data

with the percentages based on capacity committed to start-up.

Figure 3—3 Start-up instructions by resource capacity 2016 2017 2018 Day-ahead market 59% 72% 75% Self-commitment 21% 15% 13% Intra-day RUC 9% 7% 5% Manual, regional reliability 4% 3% 4% Short-term RUC 3% 2% 2% Day-ahead RUC 3% 1% <1% Manual, local reliability 2% 1% 1%

Day-ahead market75%

Self-commitment13%

Intra-day RUC5%

Manual, regional reliability

4%Short-term RUC

2%Day-ahead RUC

0%

Manual, local reliability1%2018 total starts: 35,568



One important observation is the percentage differences between Figure 3—2 and Figure 3—

3. This is the result of larger resources either self-committed or committed by the day-ahead

market, and smaller resources with shorter lead times committed in the day-ahead reliability

unit commitment, intra-day reliability unit commitment, and manual commitment processes.

Within the operating day, commitment flexibility is limited by resource start-up times. As the

operating hour approaches, fewer resources are eligible to be started. Quick-start units can

be directly dispatched by the real-time balancing market and the reliability unit commitment

processes is not designed to start other units if the quick-start units are able to resolve the

problem. SPP issued 75 percent of all start-up instructions to gas-fired generators in 2018,

up from 72 percent of all start-ups in 2017, and 59 percent in 2016.

Figure 3—4 shows that almost all start-up instructions issued to combined-cycle generators

are the result of the day-ahead market. This result is expected given the lower variable costs

and different operating parameters for these resources relative to other gas units.

Figure 3—4 Origin of start-up instructions for gas resources

Commitment Process

Combined-cycle Simple-cycle,

combustion turbine Simple-cycle, steam turbine

2016 2017 2018 2016 2017 2018 2016 2017 2018 Day-ahead market

92% 94% 93% 59% 69% 76% 46% 64% 73%

Day-ahead RUC

1% 0% 0% 2% 0% 0% 16% 3% 2%

Intra-day RUC

1% 1% 2% 15% 12% 8% 18% 15% 14%

Short-term RUC

0% 0% 0% 6% 4% 4% 4% 4% 2%

Manual instruction

0% 0% 0% 11% 7% 7% 9% 2% 3%

Self 6% 5% 5% 7% 7% 5% 8% 11% 6%

SPP issued day-ahead starts for gas-fired generators with simple-cycle combustion turbine

technology accounted for 76 percent of their total starts. This is an increase from 59 percent

in 2016 and 69 percent in 2017. Steam turbine starts saw a marked increase in the day-

ahead market starts, with 73 percent in 2018, compared to 64 percent the year before. The

MMU is encouraged by the increase in units being started by the day-ahead market, and the

accompanying decrease in manual commitments, as well as self-commitments. Starts in the



day-ahead market are preferred since the market clearing engine will optimize starts and

dispatch to the most benefit to the market.

Some reliability unit commitments are made to meet instantaneous load capacity

requirements; however, this is not a product generators are directly compensated for by the

market. Therefore, reliability commitment processes, more often than the day-ahead market,

make commitments that may not be supported by real-time price levels. These situations

often lead to make-whole payments. The next section discusses the drivers behind reliability

commitments.

3.1.2 DEMAND FOR RELIABILITY

The previous section noted that eight percent of SPP start-up instructions by capacity

originated from SPP reliability unit commitment processes. To understand the need for the

reliability commitments it is useful to discuss the different assumptions, requirements, and

rules that are used in the reliability unit commitment processes versus the day-ahead market.

A fundamental difference is the definition of energy demand between the two studies. The

energy demand in the day-ahead market is determined by bids submitted by the market

participants, and averages between 99 to 101 percent of the real-time values, as shown in

Figure 2—5.

Another important difference between the two studies is virtual transactions. Market

participants submit virtual bids to buy and virtual offers to sell energy in the day-ahead

market. A virtual transaction is not tied to an obligation to generate or consume energy;

rather, it is a financial instrument that is cleared by taking the opposite position in the real-

time market. Because the reliability unit commitment processes must ensure sufficient

generation is on-line to meet energy demand, virtual transactions are not included in the day-

ahead, intra-day, or short-term reliability unit commitment algorithms.

The assumptions regarding wind generation differ as well. Eighty-five percent of the real-

time wind production cleared in the day-ahead market on an average hourly basis in 2018.

While the market participants determine the participation levels for their wind generators in

the day-ahead market through the use of supply offers, a wind forecast is used by the

reliability unit commitment processes. Import and export transaction data are updated to

include the latest information available for the reliability unit commitment processes.



These types of differences are referred to as resource gaps (i.e., a gap in meeting demand)

between the day-ahead and real-time markets. The resource gap is the difference between

the (1) excess supply between day-ahead and real-time markets, and the (2) excess demand

between the day-ahead and real-time markets.

The primary drivers for the negative resource gaps are differences in virtual supply net of

virtual demand, differences in real-time wind generation compared to wind cleared in the

day-ahead market, and real-time net exports exceeding day-ahead net exports. It is generally

true that real-time wind generation exceeds the clearing of wind in the day-ahead market.

The mismatch between real-time and day-ahead wind is partly because some market

participants with wind generation assets do not participate or offer the full amount of

forecasted capacity in the day-ahead market. Instead, they take real-time positions given the

uncertainty of the wind generation.

The resource gaps can help explain why additional commitments occur after the day-ahead

market has cleared. Figure 3—5 compares on-line capacity between the day-ahead and real-

time markets.

Figure 3—5 Average hourly capacity increase from day-ahead to real-time

The chart indicates that in 2018 there was, on average, around 1,800 MW of additional

capacity on-line during the real-time market relative to the capacity cleared in the day-ahead

market, a reduction of six percent compared to 2017. However, this is a 33 percent reduction

from 2016. The lower capacity increases in July through November can be primarily

$0

$10

$20

$30

$40

0

1,000

2,000

3,000

4,000

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2016 '17 18

Pric

e ($

/MW

h)

MW

Average capacity increase Real-time price Day-ahead price



attributed to resources producing closer to their day-ahead totals. The month of May also

had wind producing closer to the day-ahead schedule, but the load came in higher than what

was forecasted in the day-ahead market. This increased the total capacity in real-time.

One of the reasons for reliability commitments is the need for ramp capability. The

instantaneous load capacity constraint may commit additional resources to ensure there is

adequate ramping capacity to meet the instantaneous peak demand for any given hour. The

instantaneous load capacity constraint is defined as the greater of the forecasted

instantaneous peak load, or an SPP defined default value. A value is calculated for upper

bound (upward ramp) and a lower bound (downward ramp). Because the default value is

used the majority of the time, the MMU believes that instantaneous load capacity constraint

can contribute to reliability commitments in excess of the resource gaps. Figure 3—6 shows

the percentage of hours for which the default value is used for the upper bound and lower

bound of instantaneous load capacity.

Figure 3—6 Instantaneous load capacity required

The percent of hours at various upper bound default levels is shown in Figure 3—7. The most

frequent observations were from 500 MW to 599 MW at around 23 percent. Overall, the

2018 levels were less than the values observed after the change in August 2017. The lower

bound default values remained at 500 for all hours of the day.

0%

20%

40%

60%

80%

100%

Upper bound Lower bound

Perc

ent

of t

ota

l ho

urs

Minimum requirement Greater than minimum requirement



Figure 3—7 Instantaneous load capacity upper bound default levels

Resources committed to provide ramp capability, whether as a result of applying the

instantaneous load capacity constraint in a reliability commitment algorithm or a manual

process, can affect real-time prices. Without the appropriate scarcity pricing rules that reflect

the market value of capacity shortages due to ramp capability, the cost of bringing the

resource on-line may not be fully reflected in the real-time prices.

Reliability commitments, along with wind exceeding the day-ahead forecast, can dampen

real-time price signals, as is evidenced by 42 percent of make-whole payments made for

reliability unit commitments as shown in Figure 4-35.

3.1.3 QUICK-START RESOURCES COMMITMENT

A quick-start resource can be started, synchronized, and begin injecting energy within 10

minutes of SPP notification. This section will detail the ways quick-start resources are used,

which includes quick-start resources dispatched by the real-time market27. This section has

changed from prior years because of the inclusion of real-time market dispatches, which are

not considered commitments since they get dispatched by the real-time market rather than

committed. Quick-start resources do not show up in the current operating plan.

27 Add Protocol Section Reference for quick start logic

0%

5%

10%

15%

20%

25%

Perc

ent

of h

our

s

Head-room requirement (MW)



For 2018, the definition of a quick-start resource has been updated to include only resources

that were dispatched by the real-time market and were cleared by the market clearing

engine. This differs from the 2017 definition which was any resource that could be on-line in

less than 10 minutes. The new method results in 97 resources counting as quick-start

resources all of which have been offered and used by the real-time market at least once

during 2018 in this capacity.

The more encompassing term “deployment” is being used to account for both commitment

and dispatch of resources. Figure 3—8 summarizes the deployment method available for

quick-start resources along with the number of times each method was selected, lead time,

original commitment hours, and actual hours on-line.

Figure 3—8 Deployment of quick-start resources

Commitment process

Number of starts

Committed available

capacity (MW)

Lead time

(hours)

Hours in original

commitment

Actual hours

on-line

Day-ahead RUC 0 N/A N/A N/A N/A

Intra-day RUC 748 26,331 1.8 3.8 6.0

Short-term RUC 453 18,761 0.3 2.5 4.2

Manual 508 26,642 0.6 4.0 6.4

Day-ahead market 16,991 603,551 21.4 6.1 12.8

Real-time market 12,225 416,994

The real-time market is the only time when these resources are being deployed as quick-start

resources. There is some overlap in the numbers from the real-time market and day-ahead

market commitments because there are a number of the day-ahead market committed

resources having intervals where their commitment period is moved up or extended by the

real-time market.

The level of make-whole payments associated with the commitment of quick-start resources

in the reliability processes is noteworthy. In 2018, 76 percent of the reliability commitments

for quick-start units resulted in real-time make-whole payments, which is similar to 2017 and

2018 at 78 percent. In 2018, quick-start resources received $7 million in real-time make-

whole payments and $350,000 in day-ahead make-whole payments. This is a decrease from

their respective 2017 make-whole payments of $9.5 million and $485,000. The MMU

believes that the commitment costs of quick-start resources needs to be included in the



commitment decisions of these resources, which is consistent with changes developed

through the stakeholder process.28

With the inception of the short-term reliability unit commitment process in February 2016,

there has been a significant reduction in the number of day-ahead reliability unit

commitments for these units. The short-term reliability unit commitment can commit units in

as little as 15 minutes ahead, increasing certainty of the need for the unit. Before the short-

term reliability unit commitment process was implemented, units had often been committed

hours ahead of the actual start time—sometimes more than a day—ignoring the value of their

flexible capability. The 15 minute lead-time leaves time to commit these quick-start

resources when needed, but allows the commitment to be held off longer providing, more

certainty of the need of the resource. This also minimizes the time these units are at

minimum load levels with market prices below their marginal costs.

The Integrated Marketplace protocols29 describe the real-time market dispatch and

registration options of resources with quick-start capability. Resources use these quick-start

capabilities in their real-time offer, and registered as quick-start capable, can choose not to

receive start-up and/or shutdown orders out of the short-term, intra-day RUC.

Figure 3—9 shows the percent of time quick-start resources generated power and the

relationship of prices to their offer.

28 See MMU comments to FERC regarding pricing of quick-start resources, https://elibrary.ferc.gov/idmws/common/opennat.asp?fileID=14820032. 29 Integrated Marketplace protocols, Section 4.4.2.3.1 and Section 6.1.1.

https://elibrary.ferc.gov/idmws/common/opennat.asp?fileID=14820032



Figure 3—9 Operation of quick-start resources

Over the last three years, about 25 percent of the megawatt-hours produced by quick-start

resources30 had energy prices below real-time energy offers. In 2018, this is about the same

as the percentage of offers to energy price for other resources in the SPP footprint, which is

represented by the green line in Figure 3—9. Quick-start resources directly dispatched in

real-time using the quick-start logic are not eligible for a make-whole payment, nor is the

minimum run time respected.31 This is a concern that revision request 116 is designed to

address.

In order to deal with the issue of uneconomic production, SPP staff presented a new quick-

start design proposal that was well-received by stakeholders in May 2015. Subsequently, this

proposal was submitted to the Market Working Group (MWG) by Golden Spread Electric

Cooperative32 and was approved in September 2015. The MMU agrees that incorporating

30 Quick-start resources are defined as those resources with a 10 minute start-up time and a minimum run time of one hour or less. Solar and wind resources are not considered quick-start resources. 31 Protocols 4.4.2.3.1 states that only the offer curves are used to dispatch. 32 Revision Request 116 (Quick-Start Real-Time Commitment) by the SPP board, but was placed on hold in lieu of FERC’s Order issued December 21, 2017.

0%

20%

40%

60%

80%

100%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2016 '17 18Price above offer Price at or below offer Non-quick-start percent



commitment costs in the evaluation of the commitment will help reduce the incidence where

these resources are committed and prices are insufficient to cover costs.

While this change was pending FERC filing, FERC began a Section 206 proceeding33 in late-

December 2017 finding SPP’s current practices regarding pricing of quick-start resources

may be unjust and unreasonable. In response, the MMU filed an initial brief and reply

comments highlighting the following positions:

• Commitment costs should be evaluated as part of market optimization to ensure production costs are minimized.

• Quick-start resources should have a minimum run time of less than or equal to one hour.

• Economic minimum operating levels should not be reduced (relaxed) to zero because the approach would set prices based on resource flexibility and availability that is false (does not exist) and would likely require another optimization run that could put pressure on performance time and affect the real-time mitigation process.

• SPP’s screening run and subsequent pricing and scheduling runs are reflective of actual available resources and that the optimization logic uses appropriate operational and reliability constraints (i.e., relies on a resources’ physical offer parameters that determine how and when a resource should operate).

• The Commission’s proposal departs from the marginal pricing methodology upon which the SPP market is based and would likely result in average cost pricing.

• Uplift would likely be transferred under the Commission’s proposal and not necessarily reduced.

• Tariff language regarding the basis of parameters needs to be revised to ensure that information is accurate and based on physical and environmental factors.

FERC has yet to respond to comments and propose further direction on modeling of quick-

start resources. At this time, the MMU does not intend to take further action until FERC has

ruled on the Section 206 proceeding.

33 See FERC Docket No. EL18-35-000: https://www.ferc.gov/whats-new/comm-meet/2017/122117/E-5.pdf.

https://www.ferc.gov/whats-new/comm-meet/2017/122117/E-5.pdf




3.1.4 GENERATION SCHEDULING

The day-ahead market provides market participants with the ability to submit offers to sell

energy, regulation-up service, regulation-down service, spinning reserves, and supplemental

reserves, and/or to submit bids to purchase energy. The day-ahead market co-optimizes the

clearing of energy and operating reserve products out of the available capacity. All day-

ahead market products are traded and settled on an hourly basis.

In 2018, participation in the day-ahead market was robust for both generation and load.

Load-serving entities consistently offered generation into the day-ahead market at levels in

excess of the requirements of the limited day-ahead must-offer obligation. Participation by

merchant generation—for which no such obligation exists—was comparable to that of the

load-serving entities.

Figure 3—10 Day-ahead market offers by commitment status and participant type

Resource type Owner type Market Self Reliability Not participating Outage

Fossil fuel resources

Load-serving entity 51% 33% 2% 1% 13%

Merchant 74% 14% 0% 0% 12% Variable energy resources

Load-serving entity

53% 43% 0% 0% 4%

Merchant 48% 12% 0% 21% 19%

Figure 3—11 shows generation participation offers in the day-ahead market by commitment

status.

Figure 3—11 Day-ahead market offers by commitment status

0%

20%

40%

60%

80%

100%


Market Self Reliability Not participating Outage



The average percent of total offered capacity by commitment status remained consistent with

2017 levels with a slight decrease in the “self-commit” status and increase in the “outage”

status. The “market” commitment status averaged 53 percent while resources with

commitment statuses of “reliability” and “not participating” averaged one percent and four

percent, respectively. The “self-commit” status averaged 30 percent of total offered capacity

which was a slight decrease compared to 32 percent in 2017. 34 The “outage” commit status

averaged 12 percent up from 10 percent in 2017. While self-commits decreased from 2017,

they still constitute a large amount of the capacity offered into the market.

Compared with Figure 3—2 and Figure 3—3 in Section 3.1.1, which shows origins of only initial

starts, these values represent commitment status of all generation capacity offered including

those on-line. Initial starts show 11 percent being self-committed in 2018, whereas all

capacity offered averaged 30 percent. This can be attributed to desire to keep resources on-

line after its initial start even during low prices.

Figure 3—12 shows on-line capacity commitment as a percent of load.

Figure 3—12 On-line capacity as a percent of load

The capacity commitment as a percent of load has decreased significantly over the past few

years, declining from 122 percent in 2016 to just under 120 percent in 2018. Factors that

34 Of the self-committed resources, qualifying facilities (QF) account for three to four percent. Qualifying facilities often use self-commit status to exercise their rights under the Public Utilities Regulatory Policies Act of 1978 (PURPA).

110%

115%

120%

125%

130%


On-

line

cap

acity

as

per

cent

of d

eman

d

2016 2017 2018



contribute to lower levels of on-line capacity are fewer self-committed coal plants and the

continued growth of wind capacity and generation.

Generation outages for the past three years are shown in Figure 3—13. Long-term outages

consist of outages with a duration beyond seven days, while short-term outages have a

duration of seven days or less. Additionally, outages are categorized for the reason for the

outage, which includes either maintenance, derates, or forced outages.

Figure 3—13 Generation outages

Generation outages are shown broken down into shoulder35 months and peak months for

each year. As can be observed from the chart above, the vast majority of outages are taken

during the shoulder months, when loads are typically lower. From 2017 to 2018, outages in

the shoulder months increased by 27 percent, after declining by seven percent from 2016 to

2017.

Figure 3—14 shows short-term outages for just the fall period for the past three years.

35 Shoulder months are March, April, May, September, October, and November.

0

20

40

60

80

100

2016 2017 2018 2016 2017 2018

Peak months Shoulder months

MW

h (m

illio

ns)

Long-term derate Long-term forced Long-term maintenanceShort-term derate Short-term forced Short-term maintenance



Figure 3—14 Short-term generation outages, fall months

For the fall period, short-term outages increased by 13 percent from 2017 to 2018, after

increasing by 11 percent from 2016 to 2017. This change was driven by an increase in short-

term maintenance and short-term derates. Short-term forced outages were about the same

in 2018 as in 2017. The large increase in outages in the fall can have a wide-ranging impact

to the market, including having fewer units able to dispatch when additional generation is

required. The increased outages in the fall period were one of the factors leading to an

increase in scarcity events in fall 2018.

3.2 DISPATCH

The real-time market co-optimizes the clearing of energy and operating reserve products out

of the available offered capacity based on the offer price for each product while respecting

physical parameters. The real-time market clears every five minutes for all products. The

settlement of the real-time market also occurs at the five-minute level, and the settlement is

based on market participants’ deviations from their day-ahead positions.

3.2.1 SCARCITY PRICING

When an insufficient amount of regulation-up service, regulation-down service, or

contingency reserve is cleared, a scarcity price is set by a demand curve. This scarcity price

informs market participants that the product was short and incentivizes future provision of

that product. The scarcity of these products can be caused by a lack of capacity or a lack of

0

2

4

6

8

10

2016 2017 2018

MW

h (m

illio

ns)

Short-term derate Short-term forced Short-term maintenance



ramp. These scarcities are due to capacity when a resource cannot provide the product

because it is limited by its minimum or maximum operating limit. These scarcities are due to

ramp when a resource cannot provide the product because of its ramp rate. Though scarcity

prices reflect the value of the scarce product, they can also raise the price of other products

when multiple products compete for the same, limited capability of the resource.

Scarcity events in 2018 were in line with 2017, with the exception of October. October 2018

had significant wind volatility, which along with other factors led to an uptick in operating

reserve scarcity events. In fact, 64 of the 167 intervals of operating reserve scarcity that

happened in 2018 happened in October.

While operating scarcity events happened evenly across the hour, regulation scarcity events

do not. Almost two-thirds of the regulation down reserve scarcity intervals in 2018 occurred

in the first interval of the hour. The same trend occurred with regulation up reserve scarcity

events, but to a lesser degree. Almost 30 percent of the intervals of regulation up scarcity

intervals occurred in the first interval of the hour.

The Integrated Marketplace uses scarcity pricing demand curves that administratively set

prices during periods of shortage. In an efficient market, the price of a product is the

marginal cost to produce the next increment of that product. When the next increment

cannot be produced, then the value of that product must be determined by something else.

Demand curves are used to determine the value of a product in this situation. Prices on a

demand curve are set administratively to send economic signals to alert market participants

that future capacity may be needed in this area.

FERC Order No. 825 was released in June 2016. The order stated that: “[W]e require each

RTO/ISO to trigger shortage pricing for any interval in which a shortage of energy or

operating reserves is indicated during the pricing of resources for that interval.” 36 At the

time of the order, SPP did not price ramp related shortages because those events were

considered transient in nature.

Revision request 175 “Ramp Shortage Compliance” was developed to bring SPP in

compliance with FERC Order No. 825 and was implemented in May 2017. Revision request

36 FERC order number 825 Settlement Intervals and Shortage Pricing in Markets Operated by Regional Transmission Organizations and Independent System Operators ,https://www.ferc.gov/whats-new/comm-meet/2016/061616/E-2.pdf ,citation 162 (June 16, 2016) , page 90.





198 was introduced to minimize the impacts of these increased scarcity events and was

implemented in August 2017. The revision request put variable price demand curves in

place of the fixed price demand curves. These variable price demand curves buffer the

impacts of scarcity pricing.

Prior to revision request 198, each products’ scarcity demand curves used one scarcity price.

Now regulation-up, regulation-down, and contingency reserves have multiple price points

depending on how short the cleared operating reserve was of the requirement. For

example, regulation-up and regulation-down had one scarcity price of $600/MW prior to

August implementation. This meant that being short just one megawatt of regulation

capacity would cause a $600/MW price for that product.37 Now there are six graduating

price points on the regulation up and down reserves’ scarcity demand curves and three for

operating reserves.

As an outcome of SPP pricing the ramp shortages, the MMU has difficulty discerning capacity

shortages from ramp shortages. The clearing engine does not record the reason for the

scarcity. The MMU suggests that SPP capture the appropriate information so that the reason

for the scarcity will be transparent.

Figure 3—15 displays the number of scarcity intervals for 2018 and compares the annual

scarcity numbers with 2017. The 2017 average scarcity prices in the chart only reflect the

periods from May 11 to the end of the year in 2017. This is because SPP did not start pricing

ramp scarcity events prior to that time.

37 The 2017 Market Monitoring Unit Annual State of the Market Report provides more details around the development of the ancillary service requirements and the structuring of the variable offer curves. https://www.spp.org/documents/57928/spp_mmu_asom_2017.pdf , page 110.




Figure 3—15 Scarcity intervals and marginal energy cost, after scarcity implementation

In 2018, there were 425 intervals with regulation-up reserve scarcity, 117 intervals with

regulation-down scarcity, and 167 intervals with operating-reserve scarcity. The average

respective scarcity prices for the events were $197/MW, $135/MW, and $362/MW. Sixty-four

of the 167 intervals of operating reserve scarcity happened in October. This month had

volatile wind fluctuations, along with baseload resources on outage, as is typical during the

shoulder months. When baseload resources are on outage, fast moving resources often run

up towards their maximum limits to make up for the lost capacity. This means that when

contingencies occur, like unexpected wind drops, the fast moving resources may not be

available to respond.

The frequency of regulation-down scarcity events decreased substantially from a monthly

average of 24 intervals in 2017 to just under 10 intervals per month in 2018. Regulation-up

scarcity was similar to 2017, averaging 35 intervals a month compared to 38 intervals in 2017.

The biggest increase happened with operating-reserve scarcity with an average of 14

intervals per month compared to three intervals per month in 2017. However, as stated

above, the 64 intervals in October were one of the leading reasons for the average monthly

increase.

Regulation up and down price spikes happed more frequently at the beginning of each hour.

Figure 3—16 below illustrates a count of the 2018 scarcity events in the real-time market by

the 12 intervals of each hour.

$0

$300

$600

$900

$1,200

0

50

100

150

200

Mar

gin

al e

nerg

y co

st

($/M

Wh)

Scar

city

inte

rval

s

Regulation-up scarcity Regulation-down scarcity Operating reserve scarcityRegulation-up scarcity price Regulation-down scarcity price Operating reserve scarcity price



Figure 3—16 Scarcity events by interval of the hour

Almost two-thirds of the 117 regulation down reserve scarcity intervals in 2018 occurred in

the first interval of the hour. The same trend occurred with regulation up reserve scarcity

events, but to a lesser degree. Almost 30 percent of the 425 intervals of regulation up

scarcity intervals in 2018 occurred in the first interval of the hour. Operating reserves appear

to be equally dispersed across the hour. The trend for regulation scarcity events occurring at

the beginning of the hour is not unique to 2018. This trend has occurred since the inception

of the marketplace. However, the pricing impacts were less severe prior to May 2017, as

ramp scarcity was not priced.

One potential reason for the trend is because SPP does not preposition regulating resources

to be at their regulating effective maximum and minimum limits prior to the period that the

resource is cleared for regulation. For example, assume a resource is currently being

dispatched to its economic minimum of 100 megawatts and it cleared for 20 megawatts of

regulation down reserves in the next hour. In order for the unit to be able to provide the

regulation down reserves it will need to move up to 120 megawatts. However, if the resource

moves there prior to the hour it will deviate from its current dispatch instruction, which has

0

25

50

75

100

125

:00 :05 :10 :15 :20 :25 :30 :35 :40 :45 :50 :55

Inte

rval

s

Interval beginningRegulation up scarcity Regulation down scarcity Operating reserve scarcity



negative financial penalties.38 This is why resources often follow dispatch until the hour their

regulation commitment begins, contributing to shortages in the first interval of the regulating

commitment.

The reason that regulation down has such a higher percentage in the first interval than

regulation up is likely because there is more ramp down offered into the market than ramp

up, as wind resources can only be ramped down. In either case, there is an argument for SPP

to preposition resources to their regulating ranges prior to the regulation period. However,

opportunity costs could occur for these resources during the preposition period and

compensation may be needed.

Scarcity prices apply to regulation up, regulation down, and operating reserves. Spinning

reserve scarcity is currently not priced through scarcity demand curves. Instead, SPP uses a

violation relaxation limit (VRL) during periods of spin scarcity. The VRL for spinning reserve

scarcity is set at $200/MW. This means that SPP will keep dispatching resources to meet the

spinning reserve requirement up to the point when the redispatch cost reaches $200/MW.

At this point, the market will relax the spinning reserve requirement to the quantity of

spinning reserve megawatts that can be obtained at a redispatch cost under the $200/MW

shadow price. The spinning reserve price will be the price of the marginal resource cleared

at the relaxed limit for that product, plus any scarcity pricing from an operating reserve

scarcity event. Since spinning reserves are a higher priority product than supplemental

reserves the market will never set the spinning reserve price lower than the supplemental

reserve price.

In most instances when the VRL shadow price is reached the marginal clearing price is set

near $200/MW. However, in circumstances where the spinning reserve is scarce because of

competition with other products, the spinning reserve shadow prices can be set at $0/MW.

Even though the shadow prices are $0/MW, the actual spinning reserve clearing prices are

typically set by an operating reserve scarcity event price. Figure 3—17 shows a scatter plot of

2018 spinning reserve shortages.

38 Resources that deviate from their dispatch signals can receive Uninstructed Deviated Charges for the deviated megawatts. The median price per megawatt for deviation in 2018 was $0.54. However, the maximum price charged for deviation was $11.01. In addition to this charge, resources do not receive cost reimbursement for any deviated megawatts in the event energy prices are lower than energy cost.



Figure 3—17 Spinning reserve shortages

This figure plots spinning reserve shadow prices against the percentage of spinning reserves

short, during periods of shortage. As can be seen in the chart, prices most frequently get set

near the $200/MW shadow price limit. However, there are often instances where the shadow

prices can be set near or at $0/MW. In 2018, there were three intervals where shadow prices

were set at $0/MW. During these intervals the spinning reserve clearing prices were set at, or

just above, the supplemental reserve price. In two of those three intervals that supplemental

reserve price was set by the operating reserve scarcity demand curve.

Energy shortages are currently priced using the VRL process. There are three VRL constraints

associated with energy shortages. The current energy shortage VRL shadow prices are

$5,000/MW for ramp shortages, $50,000/MW for balancing resources dispatch to load

consumption, and $100,000/MW for resource capacity shortages.

3.2.2 RAMPING

The increase or decrease of the resource’s output to achieve the next dispatch instruction is

called “ramp”. The number of megawatts a resource can ramp in one minute is the

resource’s “ramp rate”.

In real-time, resources are increasing and decreasing output to meet changes in both load

and generation. These changes can be measured as changes in net load. Net load is load

net of both variable energy generation and the combination of imports and exports.

$0

$50

$100

$150

$200

$250

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Spin

sha

do

w p

rice

s ($

/MW

)

Spinning reserve short as a percent of requirement



Figure 3—18 shows the frequency of how often the net load changes from one real-time

interval to the next.

Figure 3—18 Frequency of net load change, real-time

About 15 percent of intervals have a net load change of over 200 megawatts when both

positive and negative changes are considered. Some of these net load changes peaked over

1,000 megawatts. While the percentage of total intervals is low, these greater deviations

cause greater price volatility. This difference must be provided by resources with a flexible

dispatch range, or ramp capability.

Figure 3—19 below shows that the volatility in net load change increased significantly each

year from 2015 through 2017, but decreased slightly in 2018.

0%

10%

20%

30%

Perc

ent

of i

nter

vals

Interval to interval net load change



Figure 3—19 Volatility of interval-to-interval net load change

Net load volatility increased about 10 percent from 2015 to 2016 and about 19 percent from

2016 to 2017. Then in 2018, the net load volatility dropped by about five percent. Even

though the volatility in 2018 is slightly lower than in 2017, volatility is trending upward over a

longer period.

The dispatch typically assumes that variable energy resources remain at the same output

level at which they were last observed. Practically, this will not be the case. This discrepancy

can require more ramp in the next interval. Nearly one-quarter of total generation was

produced by wind and solar resources in 2018. This was a significant source of net load

changes. With over 55 gigawatts of wind-powered generation and nearly 25 gigawatts of

solar generation with an active generation interconnection request, this problem will likely

increase in the future as volatility from variable energy resources increases.

In any interval, resources typically have a range either above or below their current operating

point that they can move to in the next interval. The rate at which they can move is their ramp

0

40

80

120

160

200

2015 2016 2017 2018

Net

load

cha

nge

(MW

) vo

latil

ity

(sta

ndar

d d

evia

tion)



rate. The total amount they can ramp is their rampable capacity. SPP operators count on this

rampable capacity to meet future energy needs and to protect against uncertainty.39

Rampable capacity in the up direction is lowest in the spring and fall seasons when averaged

by month. These seasons are known for high wind and outages. Figure 3—20 shows the

average up-rampable capacity by month after energy and operating reserve obligations are

accounted for.40

Figure 3—20 Average up-rampable capacity by month

The amount of rampable capacity in the down direction is much larger than in the up

direction when averaged by month, as shown in Figure 3—21. The monthly average does not

change very significantly throughout the year.

39 This ramp in not currently being systematically procured by the software or being compensated as stand-by capacity. See section 3.2.3 on ramping capability product for further detail on the MMU’s recommendation to address these concerns. 40 The figures showing average rampable capacity are approximations. The market clearing engine allows for ramp sharing and also allows for some products to go short so that higher priority products can clear. There can be different amounts of rampable capacity available depending on the product. These graphs average the amount of load increase or decrease that would cause a shortage on the product that is nearest to a shortage.

0

250

500

750

1,000




Figure 3—21 Average down-rampable capacity by month

Figure 3—22 shows the average rampable capacity in the up direction by hour of the day.

Rampable capacity in the up direction is lower following the morning ramp in hours

beginning 08, 09, and 10. From hour beginning 11 until hour beginning 21, the rampable

capacity increases slightly but remains lower. As resources move closer to their minimum

limits during the night, this rampable capacity increases.

Figure 3—22 Average up-rampable capacity by hour

-4,000

-3,000

-2,000

-1,000

0


0

250

500

750

1,000

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Ave

rag

e up

-ram

pab

le c

apac

ity

(MW

)

hour beginning



Figure 3—23 shows the average rampable capacity in the down direction by hour. As

previously mentioned, there is much more rampable capacity in the down direction than in

the up direction. Although this rampable capacity is less in the late evening hours, this

amount does not vary significantly throughout the day.

Figure 3—23 Average down-rampable capacity by hour

Ramp capability is needed to meet all of these changes in generation and load from interval

to interval. However, unlike capacity, ramp is not procured ahead of time. A resource’s ramp

rate is used to calculate its dispatch instruction. A resource will generally not be dispatched

beyond the capability of the resource’s ramp rate. However, the ramp rate is used only to

calculate the current dispatch instruction. Unlike capacity, the market clearing engine does

not specifically procure ramp ahead of time to meet these intra-hour net load changes,

although committed capacity usually comes with incidental ramp capability. Because the

committed capacity must have ramp capability, the market clearing engine should specifically

procure and incentivize rampable capacity.

3.2.3 RAMP CAPABILITY PRODUCT

In light of the scarcity events and the demonstrated need for ramp capability, a ramp

capability product would be beneficial to the market. A resource’s ability to ramp should be

planned for and should be valued by a price to the extent the ramp is beneficial to the

market. The market monitor prefers the use of a variable demand curve to accomplish this.

-4,000

-3,000

-2,000

-1,000

0

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Ave

rag

e d

ow

n-ra

mp

able

cap

acity

(M

W)

hour beginning



3.2.3.1 Ramping limitations affect market outcomes

The real-time dispatch does not consider future intervals. It simply calculates one value: a

dispatch instruction for the next interval. While the real-time balancing market considers a

resource’s ramp capability for the purpose of calculating the dispatch instruction for the next

interval, ramp is not considered for any interval after that. Ramp is not currently accounted

for in terms of the subsequent dispatch instructions even though ramp is the very capability

that allows a resource to get to future dispatch instructions.

When ramp capability is not considered for future intervals, then the market clearing engine

may not be able to procure enough energy to serve the load or provide sufficient operating

reserves in those future intervals. Even when enough capacity is available, a lack of ramp

renders that capacity unreachable. This often leads to short-term transitory price spikes,41 as

seen in Figure 3—24.

Figure 3—24 Interval length of short-term price spikes

This figure shows that a scarcity pricing event in real-time was most likely to occur for only

one five-minute interval. Very few scarcity pricing events last more than two intervals. This

pattern was consistent in both 2017 and 2018.

Figure 3—25 shows the interval length for the different types of scarcity events.

41 This is essentially temporal, or time-based, congestion.

0

50

100

150

200

250

1 interval 2 intervals 3 intervals 4 intervals 5 intervals 6 or moreintervals

Scar

city

eve

nts

Event length in intervals

Regulation up Regulation down Operating reserve



Figure 3—25 Interval length of short-term price spikes, percentage

Of all the regulation-up scarcity events, about 70 percent lasted for only one interval and

about 15 percent lasted for two intervals. For regulation down scarcity events, about 90

percent lasted for only one interval. Operating reserve scarcity event lengths were more

diverse with about 40 percent lasting for one interval, about 20 percent lasting two intervals,

and about 20 percent lasting three intervals. Operating reserve scarcity events lasted up to

eight intervals, though there were very few events that lasted longer than three intervals.42

Scarcity events in the SPP market have historically been short in duration, and 2018 was no

exception.

Where sufficient capacity cannot be dispatched, scarcity prices are invoked. Scarcity prices

are economic signals alerting market participants to the insufficient supply of a product.

Almost all of these intervals with scarcity pricing were due to a lack of cleared ramp and not a

lack of capacity. If sufficient ramp were reserved in advance for these scarcity intervals, then

these scarcities likely could have been avoided.

In addition, marginal energy prices can be elevated even when energy is not scarce. When

ramp in the up direction is short, energy will always be given the highest priority. If there is

not sufficient ramp to meet both energy and regulation-up, for instance, then the regulation-

42 In the 2017 Annual State of the Market report, energy scarcity was shown in a similar graph. In 2018, a violation relaxation limit was used to mitigate energy shortages instead of a demand curve.

0%

20%

40%

60%

80%

100%

Regulation up Regulation down Operating reserve

Scar

city

eve

nts

Event length in intervals

1 interval 2 intervals 3 intervals 4 intervals 5 intervals 6 or more intervals



up scarcity price will be reflected in the marginal energy price. This causes a high marginal

energy price even though there is no energy scarcity because the two products are

competing for ramp. This makes prices more volatile. The regulation-up scarcity price raised

the marginal energy price in this way for about 30 percent of all regulation-up scarcity

intervals in 2018. If sufficient ramp had been available, then regulation-up scarcity prices

would not have raised the marginal energy price. A ramp capability product can ensure that

more ramping is available to meet energy so that regulation-up scarcity prices can be

avoided. This helps to better reflect system conditions and reduces dispatch volatility.

When prices are high in real-time, analysis of settlement data reveals that resources pay for

most of this cost to the market.43

Loads are typically well-hedged in the day-ahead market, so when prices spike in real-time, a

load’s settlement is typically not affected much. This is because the real-time settlement is

based on the difference between the withdrawal in real-time and in the day-ahead market.

Because loads are well-hedged, the difference between the two markets is small. This small

difference causes the real-time price to have little overall effect on load settlements.

Imports and exports have a small effect on the market overall, so they do not pay much when

real-time prices are high.

Though virtual offers pay more for high real-time prices than the aforementioned categories,

they do not pay the most. When real-time prices are high, the amount virtual offers pay is

less than when prices are low. This indicates that, overall, virtual offers adjust their market

behavior so that they are affected less by the high real-time prices.

Resources that produce less in real-time than day-ahead are mostly paying for the resources

that produce more in day-ahead than real-time. Although load pays for energy in the day-

ahead market, resources that buyback in the real-time market typically pay for the real-time

energy price above what load paid in the day-ahead.

43 Ramping Product, February 19-20, 2019 MWG, (https://www.spp.org/Documents/59502/mwg%20agenda%20&%20background%20materials%2020190219%2020.zip).

https://www.spp.org/Documents/59502/mwg%20agenda%20&%20background%20materials%2020190219%2020.zip

https://www.spp.org/Documents/59502/mwg%20agenda%20&%20background%20materials%2020190219%2020.zip



3.2.3.2 Clarification of preferences on ramp product design

In the 2017 annual report, the market monitor recommended that SPP create a ramp

capability product. Previously, the market monitor recommended that a ramp capability

product include the following features:

• Two products: ramp capability up and ramp capability down;

• Co-optimization with energy and other products to ensure the most economical

solution;

• Opportunity cost basis for pricing;

• No limitations on resource type as long as the resource can reliably provide ramp in

the direction for which it is cleared; and

• Consideration of both expected and unexpected ramping needs.

SPP has started on this design and has discussed progress with SPP’s Market Working Group.

At the time this report was written, the Market Working group has discussed design choices

such as:

• A violation relaxation limit versus a demand curve;

• Interaction between the ramp product and instantaneous load capacity; and

• Transmission constraints on ramp.

Though the role of the market monitor is to advise and not drive policy choices, some design

elements are more preferable to the market monitor than others. In light of this, we share

some preferences below. These preferences do not indicate that there is one design that the

market monitor will find acceptable.

3.2.3.2.1 Demand curves reflect the value of ramp in the price

The market monitor prefers demand curves over violation relaxation limits because demand

curves allow the product’s value to be reflected in the price. Though there can be many right

answers to the value of an unavailable product, it is inappropriate to say that there is no

value. Relaxing the ramp constraint without pricing it informs the market that the product



was not valued and is not beneficial. A requirement that is unfulfilled without cost or

consequence is not a requirement.

Furthermore, the resources that supplied the ramp may be under-incentivized to supply it in

the future because no economic signal is sent to the market. When resources have no

incentive to supply ramp that is necessary to achieve dispatch, that necessary ramp must be

acquired suboptimally and without payment. For these reasons, the market monitor prefers a

demand curve to reflect the value of ramp.

Additionally, a variable demand curve is preferable. A single-value (flat) demand curve

reflects merely that a product is short, but a variable demand curve reflects how short the

product is. This can be thought of as ramp insurance. Based on historical observations of

ramp needs, the probability can be calculated that a certain amount of ramp will be needed

in the future. The cost to the market that is avoided by supplying the ramp can be used to

price the certainty that a ramp scarcity will be avoided. These probabilities and avoided costs

can be represented much like the demand bid curves that are used in the market today. In

this way, the price for being 10 megawatts short of the ramp requirement should be cheaper

than the price for being 100 megawatts short.

3.2.3.2.2 Intra-hour and inter-hour ramp products may be appropriate

The process known as instantaneous load capacity is ramp procurement without ramp

payment. The instantaneous load capacity currently ensures that sufficient rampable capacity

is committed to ramp from one average hourly load to the next. To the extent that a resource

is being dispatched uneconomically so that its ramp can be used in this way, the resource

should be paid for that value. If it is determined that more than one ramping timeframe is

needed for more than one ramping purpose,44 then the market monitor could support

multiple ramp products.

It is clear that for resources to be dispatched so that future ramp can be provided, sufficient

rampable capacity must be committed. A ramp requirement should be included in the

commitment process.

44 For instance, it may be appropriate to have a ramp product similar to instantaneous load capacity to address inter-hour ramping needs in addition to an intra-hour ramp product to address short-term load variability.



3.2.3.2.3 Considering transmission congestion when clearing ramp is not necessary at this

time

Though transmission constraints could keep cleared ramp from being useful, the market

monitor does not see this as an issue that needs to be resolved in the initial implementation

of the ramp product. It is possible that ramp could be cleared on a resource that would not

be able to deliver the ramp because it is behind a congested transmission element.

Currently, regulation clearing does not account for congestion. Even though the SPP market

has delineated reserve zones, there is currently no zonal requirement for regulation clearing.

Ramp should be a lower priority than regulation. While transmission constraints could cause

inefficient ramp clearing, it is not necessary to clear the lower-quality ramp more precisely

than the higher quality regulation.

3.3 MUST-OFFER PROVISION

The Integrated Marketplace has a limited day-ahead must-offer provision that was intended

to incentivize load-serving entities to participate in the day-ahead market. Market

participants that are non-compliant are assessed a penalty based on the amount of available

capacity available in the day-ahead market relative to the market participant’s real-time load.

The requirement is limited in the sense that only market participants with generation assets

that serve load are subject to the rules. Load-serving market participants that offer enough

generation, and/or provide scheduling information indicating a firm power purchase to cover

at least 90 percent of their real-time load are not subject to a penalty. An alternative way to

satisfy the provision is to offer all generation that is not on outage. In 2018, one day-ahead

must-offer penalty, totaling less than $500, was assessed. No penalties were assessed in

2017.

In 2014, the MMU recommended that SPP simultaneously eliminate the limited day-ahead

must-offer provision and revise the physical withholding rules to include a penalty for non-

compliance based on the premise that the recommended penalty provision would be

sufficient to ensure an efficient level of participation in the day-ahead market. SPP

stakeholders then approved the removal of the day-ahead must-offer with no additional

physical withholding provisions, and SPP filed the tariff revision with FERC in the summer of



2017. FERC denied the removal of the limited must-offer requirement, as it did not include

physical withholding non-compliance penalties.45

The MMU continues to recommend updating the day-ahead must offer requirement and

addressing FERC’s concerns. However, given the status of other higher priority initiatives, the

MMU assigns a low priority to addressing the issue at this time. See further discussion in

Section 8.3.6.

3.4 VIRTUAL TRADING

Market participants in SPP’s Integrated Marketplace may submit virtual energy offers and bids

at any settlement location in the day-ahead market. Virtual offers represent energy sales to

the day-ahead market that the participant needs to buy back in the real-time market. These

are referred to as “increment offers”, which are like generation. Virtual bids represent energy

purchases in the day-ahead market that the participant needs to sell back in the real-time

market. These are referred to as “decrement bids”, which are like load. The value of virtual

trading lies in its potential to converge day-ahead and real-time market prices, and improve

day-ahead unit commitment decisions.

In order for virtual transactions to converge prices, there must be sufficient competition in

virtual trading; transparency in day-ahead market, reliability unit commitment, and real-time

market operating practices; and predictability of market events. Since the market began in

2014, there has been moderate, and increasing levels of virtual participation. Figure 3—26

displays the total volume of virtual transactions as a percentage of real-time market load

along with wind output levels.

45 See FERC ruling at https://www.ferc.gov/CalendarFiles/20171013130834-ER17-2312-000.pdf

https://www.ferc.gov/CalendarFiles/20171013130834-ER17-2312-000.pdf



Figure 3—26 Cleared virtual transactions as percent of real-time load

As shown in the figure, virtual transactions averaged 14.6 percent of real-time market load,

compared to 13.9 percent in 2016. Historically, the greatest increases in virtual transactions

as a percentage of load have been with cleared virtual offers. However, this trend stopped in

2018, as the percent of virtual offers to load was 8.1 percent, the same as 2017. This percent

was 5.4 percent in 2016 and 4.0 percent in 2015. Virtual cleared bids increased from 4.2

percent in 2016 to 5.8 percent in 2017 and further upward to 6.5 percent in 2018. Days with

high wind output see an increase in virtual offer activity. Virtual bids typically increase during

high load hours.

At just under 15 percent of load, the average hourly total volume of cleared virtuals ranged

from 1,947 MW of withdrawal, to 2,436 MW of injection. The net cleared virtual positions in

the market averaged about 489 MW of injection or supply each hour.

While the majority of virtual transactions occurred at wind resources in 2017, a trend that has

been increasing since mid-2015, fewer virtual transactions occurred at wind resources in

2018. Figure 3—27 illustrates the settlement location types where virtual offers clear.

0

2,000

4,000

6,000

8,000

0%

5%

10%

15%

20%


Win

d g

ener

atio

n (G

Wh)

Perc

ent

of l

oad

Cleared bids as a percent of load Cleared offers as a percent of load

Wind output (GWh)



Figure 3—27 Cleared virtual offers by settlement location type

This figure shows that over 43 percent of the virtual offers cleared at variable energy

resources during 2018.46 This is a decrease from 49 percent in 2017 and an increase from 36

percent in 2016. Even so, virtual offers at wind locations remain the largest volume of any

single location type. These large volumes highlight the possibility that participants of wind

resources may be missing financial opportunities by under-scheduling in the day-ahead

market.47

This is in contrast with the locational volumes of virtual bids. Cleared virtual bids were

primarily at resources other than variable energy resources, followed by load locations.

Figure 3—28, below, shows the cleared virtual bids by settlement location types.

46 This includes both dispatchable and non-dispatch variable energy locations. 47 Section 4.1.5 on price divergence discusses the effects of unscheduled wind in the SPP market.

0

500

1,000

1,500

2,000

2,500


Cle

ared

vir

tual

offe

rs (G

Wh)

Non-dispatchable variable energy resource Dispatchable variable energy resource Non-variable energy resource Hub Interchange Load



Figure 3—28 Cleared virtual bids by settlement location type

Forty-one percent of cleared virtual bids in 2018 were transacted at non-variable energy

resource locations, which is down from 43 percent in 2017 and up from 39 percent in 2016.

Another 31 percent of the total cleared virtual bid volume was transacted at load locations in

2018, which is up slightly when compared to 2017 (28 percent) and 2016 (26 percent). The

associated quantities also increased 26 percent from 2017 to 2018 and 50 percent from 2016

to 2017. Virtual bids at wind locations declined both as a percent of total virtual bids and in

volumes in 2018.

Figure 3—29 shows cleared demand bids that offered more than $30/MWh over the cleared

day-ahead price, and the supply offers offered at less than $30/MWh under the cleared day-

ahead price.

0

500

1,000

1,500

2,000

2,500


Cle

ared

vir

tual

bid

s (G

Wh)

Non-dispatchable variable energy resource Dispatchable variable energy resource Non-variable energy resource Hub Interchange Load



Figure 3—29 Virtual offers and bids, day-ahead market

These types of bids and offers are called “price-insensitive” and occurred more often with

bids up until 2017. Price-insensitive bids have remained steady with about 12 to 15 percent

of cleared bids since 2016, but that number has increased for offers from six percent in 2016

to 11 percent in 2018. Price-insensitive bids and offers are willing to buy/sell at a much

higher/lower price that could lead to price divergence rather than competitive, or price-

sensitive, bids and offers leading to price convergence between the day-ahead and real-time

markets. Price-insensitive bids and offers usually occur at locations with congestion and

arbitrage against the day-ahead and real-time price differences. Given that price-insensitive

bids and offers are likely to clear, these can be unprofitable if congestion around these

locations does not materialize, leading to divergence between the markets.

Financial information for virtual trades is shown monthly and on an annual basis for 2018 in

Figure 3—30.

-6,000

-4,000

-2,000

0

2,000

4,000

6,000


Ave

rag

e ho

urly

vo

lum

e (M

W)

Cleared bids < $30/MWh Cleared offers < $30/MWhCleared bids > $30/MWh Cleared offers > $30/MWhUncleared bids Uncleared offersNet cleared



Figure 3—30 Virtual profits with distribution charges

Month Raw profit Raw loss

Raw net profit

(prior to fees)

RNU charges/ credits

Day-ahead make-whole

payment charges

Real-time make-whole

payment charges

Virtual transaction

fee

Total net profit

January $ 22.3 $ (13.9) $ 8.4 $ 0.9 $ 0.3 $ 2.3 $ 0.0 $ 4.9 February 15.9 (14.3) 1.6 0.2 0.1 0.7 0.0 0.5 March 15.0 (12.5) 2.5 0.1 0.2 1.0 0.0 1.2 April 19.6 (15.2) 4.4 0.8 0.2 0.8 0.0 2.6 May 14.1 (13.9) 0.2 0.4 0.1 2.1 0.0 (2.4) June 15.5 (12.7) 2.8 0.3 0.1 1.2 0.0 1.1 July 9.4 (8.2) 1.2 0.2 0.1 1.9 0.0 (1.1) August 8.0 (6.0) 2.0 0.2 0.1 1.2 0.0 0.5 September 14.1 (10.0) 4.1 0.5 0.1 0.9 0.0 2.5 October 19.7 (16.0) 3.8 0.4 0.1 2.1 0.0 1.2 November 22.4 (15.6) 6.8 0.7 0.1 3.2 0.0 2.8 December 17.6 (11.1) 6.5 0.7 0.1 1.6 0.0 4.0 Total 193.5 (149.3) 44.2 5.4 1.7 18.7 0.5 17.8

All figures in $ millions

Virtual trades profited in aggregate for the year by about $44 million, a 19 percent decrease

year-over-year. Virtual bids can be charged distribution fees for day-ahead make-whole

payments and virtual offers are susceptible to real-time make-whole payment distribution

fees. In addition, both types of transactions can receive revenue neutrality uplift

charge/credits and are assessed a $0.05 per virtual bid or offer transaction fee for processing

virtual transactions. The average 2018 rates per megawatt for day-ahead make-whole

payments, real-time make-whole payments, and real-time revenue neutrality uplift

distributions are $0.04/MWh, $0.50/MWh, and $0.14/MWh, respectively. When factoring in

these charges and credits, the net virtual bidding profits for 2018 were $17.8 million, 60

percent lower after factoring in fees. Net profits in 2018 decreased 42 percent from $34.6

million in 2017.

Every month in 2018 was profitable in aggregate for virtual transactions before factoring in

transaction fees. However, after accounting for these fees, May and July were unprofitable in

aggregate. In the 58 months since the market began, only nine months have had a net loss

when factoring in fees. The highest payout months in 2018 happened in January and

December with net payouts just over $4.9 million and $4.0 million, respectively. These

months coincide with high wind periods and low load months when high price differences



can occur between day-ahead and real-time markets as a result of under-scheduled wind in

the day-ahead market.48

Net profits are typically small when assessed on a per megawatt basis. Figure 3—31 illustrates

the monthly average profit per megawatt for a cleared virtual in 2018.49

Figure 3—31 Profit and loss per cleared virtual, average

The chart shows that when factoring in all fees the average profit per megawatt for 2018 was

$0.46 per cleared megawatt, almost a 50 percent decrease from $0.91 in 2017, and a 28

percent decrease from the $0.59 in 2016.

There were 84 SPP participants who transacted virtuals in 2018. Figure 3—32 illustrates each

virtual participant’s virtual portfolio for the year by both net megawatts cleared and net

profits before adjusting for fees.

48 Section 4.1.4, price divergence, discusses the effects of unscheduled wind in the SPP market. 49 A change was made to the calculation of revenue neutrality uplift charges for virtuals. This correction resulted in a $0.06 per cleared megawatt reduction in the 2017 totals presented last year.

-$1.00

$0.00

$1.00

$2.00

$3.00


Cle

ared

vir

tual

pro

fit/l

oss

per

MW



Figure 3—32 Virtual portfolios by market participant

Five participants accounted for over 46 percent of the virtual profits before fees, which can

also be referred to as gross profits. These participants also account for roughly 50 percent of

the cleared transactional volume in the market. In aggregate, virtual trading generated gross

profits for most participants. However, nineteen virtual participants lost money on a gross

profit basis. The total losses before fees amounted to roughly $1.2 million, and five entities

accounted for over $900,000 of that loss. However, total gross trading profits exceeded total

gross trading losses by more than forty-fold.

Five participants also accounted for over 46 percent of the virtual profits after fees, which can

also be referred to as net profits. These participants account for roughly twenty-one percent

of the transactional volume in the market. In aggregate, virtual trading generated net profits

for most participants. However, twenty-eight virtual participants lost money on a net profit

basis. The total losses after fees amounted to roughly $3.2 million, and five entities

accounted for over $2.1 million of that loss. However, total net trading profits exceeded total

net trading losses by almost seven-fold.

Additionally, Figure 3-28 highlights the disparity in the trading fees paid by each market

participant. These fees totaled more than $26.2 million in 2018, they include: virtual

transaction fees (two percent), real-time revenue neutrality uplift fees (21 percent), day-ahead

make-whole payment fees (six percent), and real-time make-whole payment fees (71

percent). Virtual bids are subject to virtual transaction fees, real-time revenue neutrality fees,

and day-ahead make-whole payment fees. Whereas virtual offers are subject to virtual

transaction fees, real-time revenue neutrality fees, and real-time make-whole payment fees.

-1

0

1

2

3

4

5

6

7

-$1

$0

$1

$2

$3

$4

$5

$6

$7

Cle

ared

vir

tual

GW

h (m

illio

ns)

Vir

tual

+p

rofit

/-lo

ss ($

mill

ions

)

Cleared virtual megawatts Virtual +profit/-loss before fees Virtual +profit/-loss after fees



Nearly three quarters of the total fees assessed to virtual transactions are assessed only to

virtual offers.

The discrepancy in virtual fees relates to the quantity calculation associated with payers of

real-time make-whole payments – specifically, the real-time net settlement location deviation

hourly amount. This determinant accounted for 78 percent of the real-time make-whole

payments in 2018, or roughly $35 million. Virtual offers paid 53 percent of this amount,

nearly $19 million. As the name implies, the quantity applied to applicable non-virtual

transactions includes only the incremental deviations from day-ahead, however the quantities

assessed to virtual offers include the full virtual offer quantity.

This calculation methodology when combined with the larger make-whole payments

normally associated with real-time, generally leads to higher fees associated with offers when

compared to virtual bids. In 2018, the fees associated with virtual offers amounted to $1.11

per megawatt compared to $0.25 per megawatt for virtual bids. This calculation

methodology and associated incentives could be part of the reason why virtual trading

offsets only part of the under scheduling of wind resources in the day-ahead market. The

market monitor will continue to evaluate these trends going forward.

Cross-product market manipulation has been a concern in other RTO/ISO markets, and

extensive monitoring is in place to detect potential cases in the SPP market. For example, a

market participant may submit a virtual transaction intended to create congestion that

benefits a transmission congestion right position. Generally, this behavior shows up as a loss

in one market, such as a virtual position, and a substantial associated benefit in another

market, such as a transmission congestion right position. In the SPP market, twelve market

participants lost more than $10,000 in 2018, which is only slightly more than in 2017.



Southwest Power Pool, Inc. Prices Market Monitoring Unit


4 PRICES

This chapter covers prices in the SPP market, along with related metrics on negative prices,

make-whole payments, and long-run price signals for investment. Highlights of this chapter

include:

• Day-ahead market prices averaged $25.08/MWh in 2018, up about nine percent from

2017. The average real-time price for 2018 was $24.58/MWh, an increase of five

percent over 2017. Although gas prices declined from 2017 to 2018, higher loads in

2018 and declines in the frequency of negative price intervals contributed to higher

average prices in 2018.

• SPP hub prices converged in 2018, with only a $1/MWh difference between the North

and South hub prices, compared to a historical difference of around $5/MWh. This is

primarily attributed to reductions in congestion due to transmission expansion, as well

as higher prices in the northern portion of the SPP footprint.

• Price divergence (absolute) between the day-ahead and real-time markets has

increased by nearly 60 percent from 2016 to 2018, climbing from $5.56/MWh to

$9.42/MWh. Analysis has found that underscheduling of renewable resources and

short-term ramping limitations are drivers of this divergence.

• The frequency of negative price intervals declined in 2018 from 2017. However, the

MMU remains concerned about the frequency of negative price intervals. Negative

prices may not be a problem in and of themselves, however, they do indicate an

increase in surplus energy on the system.

• Sixteen resources received over $1 million in make-whole payments in 2018, up from

the twelve resources in 2017.

• One resource received over $3.6 million dollars in make-whole payments. The

resource was mainly used for transmission and voltage support.

• Roughly seventy-eight percent of real-time make-whole payments were paid for by

resources that withdrew more in real-time than their day-ahead market cleared

megawatts. Virtual offers were the largest source of payments, paying for forty-one

percent of the 2018 real-time make-whole payments.

• Revenues have been insufficient to support the cost of new entry of scrubbed coal,

advanced combined-cycle, and advanced combustion turbine generation since the



inception of the Integrated Marketplace, and 2018 was no exception. From 2015

through 2018, prices did support the going forward costs of combined-cycle and

combustion turbine units, though they did not support the cost of scrubbed coal units.

4.1 MARKET PRICES AND COSTS

This section reviews market prices and costs by focusing on the fuel prices, price volatility,

negative prices, operating reserve prices, and market settlement results including make-

whole payments. Overall, annual energy prices remained stable compared to previous years

with just a slight increase in both the day-ahead and real-time prices in 2018. However,

annual numbers may mask underlying issues related to market flexibility and efficiency. For

instance, we discuss increasing periods of price volatility and instances of negative prices

later in this chapter.

4.1.1 ENERGY MARKET PRICES AND FUEL PRICES

Figure 4—1 below compares day-ahead and real-time prices in SPP between 2007 and 201850

with natural gas prices.

Figure 4—1 Energy price versus natural gas price, annual

50 From 2007 to 2013, the average price from the Energy Imbalance Service market is shown. The 2014 real-time average includes two months of prices from the Energy Imbalance Service market and 10 months of prices from the Integrated Marketplace.

$0

$2

$4

$6

$8

$0

$20

$40

$60

$80

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

$/M

MB

tu

$/M

Wh

Real-time price Day-ahead price Panhandle Eastern Pipeline gas cost



Historically, electric market prices have followed the cost of natural gas. As natural gas prices

have remained low overall, so have SPP market prices. The average gas cost, using the price

at the Panhandle Eastern Pipeline (PEPL), decreased by two percent from 2017 to 2018.

Although electricity prices remained low in 2018, they were higher in comparison to previous

years. Day-ahead market prices averaged $25.08/MWh in 2018, up about nine percent from

2017. The average real-time price for 2018 was $24.58/MWh, an increase of five percent

over 2017. Although gas prices declined from 2017 to 2018, higher loads in 2018 driven by

weather, as discussed in Chapter 2, and declines in the frequency of negative prices

contributed to higher prices in 2018.

Figure 4—2 illustrates day-ahead energy prices, as well as gas costs, on a monthly basis for

2018, along with an annual comparison for the past three years.

Figure 4—2 Day-ahead energy price versus natural gas cost

On a monthly basis, natural gas prices have averaged around $2.50/MMBtu at the Panhandle

Eastern hub. The last three months of 2018 saw a steep climb in gas prices, peaking at

$3.67/MMBtu in November. Day-ahead prices were highest in November at around

$33/MWh, and real-time prices were highest in October at about $31/MWh. This is atypical

as prices in the SPP market tend to peak in the summer months, as temperatures increase

and loads peak. Gas prices were highest in November and contributed to the highest

electricity prices in November. Prices were lowest in March, with both day-ahead and real-

time electricity prices both below $19/MWh, as periods of high wind generation coincided

with low loads and low gas prices.

$0

$1

$2

$3

$4

$0

$10

$20

$30

$40


$/M

MB

tuDay-ahead price Panhandle Eastern Pipeline gas cost

$/M

Wh



Changes in gas prices have historically had the highest impact on electricity prices compared

to other fuels. This is because the short-run marginal costs of coal fired generation

historically have been cheaper than natural gas-fired generation. However, as natural gas

prices have fallen, the short-run marginal costs of natural gas fired generation have been

more competitive with coal fired generation, and in some instances displacing coal

generation. Figure 4—3 compares various fuel price indices with real-time prices.

Figure 4—3 Fuel price indices and wholesale power prices

This figure shows that regional natural gas prices trended up from 2016 to 2018, following

the national trend as represented by the Henry Hub, especially during the winter months

(November to December of 2018). The Henry Hub gas price averaged $3.17/MMBtu for

2018, while both Southern Star and Panhandle Eastern were around $2.60/MMBtu in 2018.51

The difference between Henry Hub prices and Panhandle Eastern and Southern Star prices

has grown over the last few years, with differences averaging $0.22/MMBtu in 2016,

$0.33/MMBtu in 2017, and $0.57/MMBtu in 2018. This difference is likely driven by pipeline

constraints in the Texas and Oklahoma area. Often, natural gas is a byproduct of oil drilling.

51 The relevant natural gas prices for the SPP market are those of the Henry Hub, the Panhandle Eastern Pipeline (PEPL), and Southern Star. These prices do not include transport costs.

$0

$10

$20

$30

$40

$0

$1

$2

$3

$4

$5


$/M

Wh

$/M

MB

tu

Real-time LMP Henry Hub Panhandle EasternSouthern Star Wyoming 8400 0.80 Rail Wyoming 8800 0.80 Rail



Natural gas production has continued to outpace takeaway capacity in this area, with

incremental production volumes quickly inundating any available space in the pipelines and

keeping supply-area prices at discounts compared to other trade hubs. Natural gas prices in

West Texas (Waha and El Paso) have been less than $1/MMBtu on a number of days, which

have helped high heat rate units be profitable.

Coal prices have remained relatively stable since 2016 with a slight increase overall from

2017 to 2018.52 Annual average coal prices at Powder River Basin for both types, 8,400

Btu/lb., and 8,800 Btu/lb., increased by about five percent. The price for 8,400 Btu/lb.

increased from $0.53/MMBtu in 2017 to $0.55/MMBtu in 2018, and the price for 8,800 Btu/lb.

increased from $0.67/MMBtu in 2017 to $0.71/MMBtu in 2018.

Controlling for changes in fuel prices helps to identify the underlying changes in electricity

prices from other factors.53 Figure 4—4 below adjusts the marginal energy cost for changes in

fuel costs.54

52 Platts’ coal prices are exclusive of transport costs. Transportation costs can have a significant impact on a coal resource’s short-run marginal costs, and may often exceed commodity costs 53 In addition to fuel, other variables also affect real-time prices. These variables include seasonal load levels, transmission congestion, scarcity pricing, and wind-powered generation. 54 The marginal energy component (MEC) indicates the system-wide marginal cost of energy (excluding congestion and losses). Fluctuations in marginal fuel prices can obscure the underlying trends and performance of the electricity markets. Fuel price-adjusted marginal energy costs is a metric to estimate the price effects of factors other than the change in fuel prices, such as changes in load or changes in supply. It is based on the marginal fuel in each real-time five-minute interval, when indexed to the three-year average of the price of the marginal fuel during the interval. If multiple fuels were marginal in an interval, weighted average marginal energy costs are based on the dispatched energy of different fuel types.



Figure 4—4 Fuel-adjusted marginal energy cost

As the figure shows, fuel-adjusted marginal energy costs were lower in 2018 compared to

nominal marginal energy costs55 both annually and in all months, except for May. While the

average nominal marginal energy cost in 2018 increased 15 percent from 2017, the fuel-

adjusted marginal energy cost also increased by 13 percent. Looking by month, the highest

nominal and fuel-adjusted marginal energy cost occurred in October. SPP experienced 26

price spikes in the $500/MWh to $1,000/MWh range in October. This was over three times

the historical monthly average for that price range. One of the main causes appears to be an

increase in mid-term and long-term wind forecast errors. When large wind dips are not

accurately forecasted, SPP operators may take manual action to bring more expensive

capacity on-line.

The largest differences between nominal and fuel-adjusted prices occurred in October,

November and December. During this period, there was a large jump in natural gas spot

prices. For instance, the spot index price for natural gas at the Panhandle hub increased 72

percent from September to November from $2.13/MMBtu to $3.67/MMBtu. The SPP market

normally experiences more units operating on oil and higher priced fuel during periods of

winter market conditions and limitations on natural gas pipelines.

55 Nominal marginal energy costs represent the non-fuel adjusted marginal energy costs.

$15

$20

$25

$30

$35


$/M

Wh

Average MEC Fuel-adjusted MEC



4.1.2 HUB PRICES

SPP has two hubs: the SPP North hub and the SPP South hub. The SPP North hub represents

pricing nodes in the northern part of the SPP footprint, generally in Nebraska. The SPP South

hub represents pricing nodes in the south-central portion of the footprint, generally in central

Oklahoma. Typically, the SPP South hub prices exceed the SPP North hub prices. This was

again true in 2018, but to a lesser degree than in previous years. The general pattern of

higher prices in the south and lower in the north is primarily due to fuel mix and congestion.

Coal, nuclear, and wind are the dominant fuels in the north and west. Gas generation

represents a much larger share of the fuel mix in the south and east.

Figure 4—5 shows the annual average day-ahead and real-time energy prices at the two SPP

market hubs.

Figure 4—5 Hub prices, annual average

On an annual basis, hub prices have converged with North hub prices averaging around

$24/MWh, and South hub prices averaging around $25/MWh for 2018. Historically, the

South hub price has been about $5/MWh higher than the North hub price. This reduction in

price spread can be attributed to reductions in congestion, which was primarily a result of the

addition of the second circuit of the Woodward to Mathewson 345kV line in early 2018 (see

Section 5.1.3 for further detail), as well as higher prices in the northern portion of the SPP

footprint.

Hub prices are shown on a monthly basis for 2018 in Figure 4—6 and Figure 4—7.

$0

$10

$20

$30

2016 2017 2018

$/M

Wh

North day-ahead North real-time South day-ahead South real-time



Figure 4—6 Hub prices, day-ahead, monthly

Figure 4—7 Hub prices, real-time, monthly

Since the beginning of the Integrated Marketplace in 2014, July 2017 was the first month

where the North hub real-time price exceeded the South hub real-time price. The North hub

again exceeded the South hub in the months of March, April, June, and August 2018. South

hub and North hub prices were nearly identical, particularly in the summer months, as

congestion that separated the regional prices declined.

Average on-peak, day-ahead prices for the SPP hubs, as well as other RTO hubs in the region

are shown in Figure 4—8.

$10

$20

$30

$40


$/M

Wh

North day-ahead South day-ahead

$10

$20

$30

$40


$/M

Wh

North real-time South real-time



Figure 4—8 Comparison of RTO/ISO average on-peak, day-ahead prices 2016 2017 2018

SPP North hub $ 24 $ 25 $ 30

SPP South hub $ 29 $ 31 $ 30

ERCOT North hub $ 26 $ 26 $ 42

ERCOT West hub $ 26 $ 26 $ 39

MISO Arkansas hub $ 27 $ 31 $ 34

MISO Louisiana hub $ 34 $ 34 $ 44

MISO Minnesota hub $ 25 $ 28 $ 32

MISO Texas hub $ 31 $ 38 $ 36

PJM West hub $ 35 $ 34 $ 42

Just as the average on-peak, day-ahead price at the SPP North hub increased from 2017 to

2018, all of the other hubs studied increased in price with the exception of the SPP South hub

and the MISO Texas hub. The average on-peak day-ahead SPP South and North hub prices

were similar as transmission expansion reduced congestion. The ERCOT hub price increases

were rather significant, increasing by over 50 percent from 2017 to 2018. These price

increases were likely driven by tight capacity conditions, particularly during the summer

months.

4.1.3 ENERGY PRICE VOLATILITY

Price volatility56,57 in the SPP market is shown in Figure 4—9 below. As expected, day-ahead

prices are much less volatile than those in real-time. The day-ahead market does not

experience the actual (unexpected) congestion and changes in load or generation found in

the real-time market. Real-time volatility tends to peak in the spring and fall, roughly

corresponding with times of higher wind and lower load, but can also peak during the

summer months because of peak load conditions.

56 Volatility is calculated as the standard deviation for load-serving entities in the SPP market. The standard deviation is calculated using hourly price in the day-ahead market and interval (five minute) price in the real-time market. 57 A measure of volatility is also shown earlier in this report at Figure 3—19. That volatility calculation is based on the interval-to-interval change in marginal energy cost. The volatility calculation in Figure 4-9 is based on the interval locational marginal prices for load-serving entities in the SPP market. Although the results are different, the magnitude of the annual change is very similar.



Figure 4—9 System price volatility

Volatility in the day-ahead market has climbed steadily from 2016 to 2018, while real-time

volatility decreased from 2017 to 2018. Even so, real-time price volatility in 2018 was almost

50 percent higher than in 2016. The high real-time volatility in 2017 can mostly be attributed

to high levels of wind generation in the spring and fall of that year. High real-time market

volatility during October and November 2018 can mostly be attributed to several short-lived

scarcity pricing events, as well as fluctuations in wind output.

Price volatility varies across the SPP market footprint for asset owners primarily because of

congestion on the system, which is based on the layout of the transmission system and the

distribution of the types of generation in the fleet. The volatility for the majority of asset

owners is grouped very closely to the SPP average in both the day-ahead and real-time

markets as shown in Figure 4—10.

0

20

40

60


Vo

latil

ity (s

tand

ard

dev

iatio

n)

real-time volatility day-ahead volatility



Figure 4—10 Price volatility by market participant

Similar to 2017, the area of the footprint that experienced the most volatility in 2018 was in

southwest Missouri. This area was impacted by external flows and congestion during the

year, mostly on the Neosho-Riverton flowgate, which is discussed in more detail in Section

5.1.4.3.

4.1.4 DAY-AHEAD AND REAL-TIME PRICE CONVERGENCE

Price convergence between day-ahead and real-time prices is important, because the more

day-ahead prices reflect real-time prices, the better unit commitment and positioning of

resources occurs for real-time operations. Figure 4—11 shows day-ahead and real-time prices

monthly for 2018 and annually for the past three years.

12.67

44.19

0

10

20

30

40

50

60

70V

ola

tility

(sta

ndar

d d

evia

tion)

Day-ahead volatility SPP day-ahead average Real-time volatility SPP real-time average



Figure 4—11 Day-ahead and real-time prices

Figure 4—12 below shows the monthly difference between day-ahead and real-time prices for

2018, as well as annually for the past three years.

Figure 4—12 Difference between day-ahead and real-time prices

Day-ahead prices exceeded real-time prices in February, July, September, and October.

February had the highest spread of real-time prices above day-ahead at $4.50/MWh, while

November had the highest spread of day-ahead prices over real-time at $3.00/MWh. On an

annual basis, the day-ahead premium has returned in 2018, with day-ahead average energy

$0

$10

$20

$30

$40


Pric

e ($

/MW

h)

Day-ahead price Real-time price

–$5

–$4

–$3

–$2

–$1

+$0

+$1

+$2

+$3

+$4


Day

-ahe

ad p

rice

–re

al-t

ime

pri

ce

($/M

Wh)



price fifty cents higher than real-time. Going back to the beginning of the Integrated

Marketplace in 2014, the only year that had an annual premium for real-time prices was 2017.

Real-time price volatility drove the annual real-time price premium in 2017.

While average prices in the day-ahead and real-time markets have been close over the past

several years, average prices can mask real-time volatility and underlying price differences.

In 2018, real-time system volatility was down from 2017, but was still up nearly 50 percent

from 2016. These metrics indicate that while average prices are similar, the underlying prices

in the day-ahead and real-time markets were different. The averaging of price spikes, and in

particular, high prices during periods of scarcity, drove real-time average prices up, closer to

day-ahead prices. We attribute these short-term, transient price spikes with limitations in

ramping capability.58

In this section, we highlight underlying differences in prices after controlling for scarcity

events. This analysis shows that a significant volume of generation, particularly from wind

resources not accounted for in the day-ahead market, drives down real-time prices.

There are many factors that cause prices to diverge between the day-ahead and real-time

markets. Some of those factors may include, but are not limited to:

• Day-ahead offers may include premiums to account for uncertainty in real-time fuel

prices.59

• Load and wind forecast errors can cause differences in the real-time market results.

• Participants may not offer in all of their load or generation in the day-ahead market.

• Modeling differences including transmission outages between the two markets.

• Generation outages or derates that were different in real-time than was anticipated in

the day-ahead.

58 For further information on ramping issues, see Section 3.3.1. 59 Additionally, revision request 239 allowed historic fuel cost uncertainty to be considered in the development of mitigated energy offers.



• Impacts from other RTOs, that were not anticipated, affect the SPP real-time market.

• Changes in imports and exports from other systems in the real-time markets.

• Unanticipated weather changes affect the real-time markets.

Price divergence60 between the day-ahead and real-time markets at the system level is shown

in Figure 4—13 below. Market participants may be willing to pay a premium for more

certainty in prices. This can result in higher prices in the day-ahead market. A large

divergence between day-ahead and real-time prices may also indicate that actual conditions

in the market do not match expected conditions. An extended period of a large variance

between day-ahead and real-time prices may indicate a structural or design deficiency in the

market.

Figure 4—13 Price divergence

The absolute price divergence has increased by nearly 60 percent from 2016 to 2018,

climbing from $5.56/MWh to $9.42/MWh. Analysis by the MMU has found that

underscheduling of renewable resources and short-term ramping limitations are drivers of

this divergence. The MMU made recommendations in the 2017 Annual State of the Market

report to address the underscheduling of renewable resources and to implement a ramping

60 Price divergence is calculated as the difference between day-ahead and real-time prices, using system prices for each five-minute (real-time) or hour (day-ahead) interval. The absolute divergence is calculated by taking the absolute value of the divergence for each interval.

-30%

-20%

-10%

0%

10%

20%

-$30

-$20

-$10

$0

$10

$20


Div

erg

ence

per

cent

age

Div

erg

ence

($/M

Wh)

Divergence (day-ahead – real-time) Divergence (absolute) Divergence percentage



product. The MMU anticipates that addressing these two recommendations would improve

absolute price divergence.

Figure 4—14, below, shows the marginal energy costs for both the day-ahead and real-time

markets during on-peak hours after controlling for scarcity events.61 Figure 4—15 shows the

same information, but for off-peak hours.

Figure 4—14 On-peak marginal energy prices, excluding scarcity hours

61 These numbers reflect only hours where scarcity demand curves where not applied for any interval during the hour. SPP uses scarcity demand curves for intervals when ramp or capacity requirements cannot be met through dispatch. Scarcity demand curves are discussed in detail in Section 4.2, below.

$0

$10

$20

$30

$40

$50


$/M

Wh

On-peak real-time marginal energy cost On-peak day-ahead marginal energy cost



Figure 4—15 Off-peak marginal energy prices, excluding scarcity hours

The marginal energy cost is one of three components that factor into location marginal prices

and represents the marginal cost to provide the next increment of dispatch absent losses and

congestion. Both charts clearly show that day-ahead prices are usually at a premium when

compared to real-time prices when controlling for scarcity, particularly in the off-peak hours.

In 2018, day-ahead marginal energy costs, for all hours, were just over three percent higher

than real-time prices.62 This is much lower than the ten percent price divergence in 2017 and

the twelve percent in 2016. The majority of the decrease can be attributed to the high

volume of price spikes seen in the final quarter of 2018.

The main contributors influencing the price differences are offered MW versus cleared MW of

wind resources in the day-ahead market, self-committing of units after the day-ahead market,

and economic reliability unit commitments. In fact, only 85 percent of the wind generation

was cleared in the 2018 day-ahead market, up three percent from 2017. This changes the

62 The MMU observed that 77 percent of the hours in 2017 had higher marginal energy cost in the day-ahead market than the real-time market. This is after removing any hours associated with scarcity pricing.

$0

$10

$20

$30

$40


$/M

Wh

Off-peak real-time marginal energy cost Off-peak day-ahead marginal energy cost



supply curve in real-time by shifting it outward and causes real-time prices to drop relative to

the day-ahead market. Furthermore, the market appropriately honors the minimum

submitted limits of all committed resources. With the unanticipated generation, many non-

wind units are dispatched down by the market to their minimum capacity limits, allowing

wind to set prices. When this happens prices often go negative as the energy offers for wind

units are typically negative to account for production tax credits.63

Figure 4—16 shows average hourly incremental differences in megawatts produced between

the real-time and day-ahead market in 2018.

Figure 4—16 Average hourly real-time generation incremental to day-ahead market

Wind generation had just under two-thirds of the 1,800 MW of incremental real-time

generation in 2018, with an hourly average of 1,118 MW of additional generation in real-time.

Self-committed generation accounted for an additional 255 megawatts and reliability unit

committed generation averaged about 275 MW. On average, SPP is a net exporter in both

the day-ahead and real-time markets and sees an average hourly decrease of 132 MW in the

real-time market compared to the day-ahead. This results in additional capacity committed in

day-ahead not needed in real-time. Averaging 490 MW, net virtual positions helped to offset

the additional generation, but only accounted for about 27 percent of the difference for the

year, which is down from about 650 MW (34 percent) in 2017.

63 Negative prices are discussed in detail in Section 4.1.5.

Wind

Net virtual supply (day-ahead)

RUC/manual committed

Self-committed after day-ahead market

Import increase

0

400

800

1,200

1,600

2,000

Additional energy Reductions

MW

h

Net additional real-time supply



4.1.5 NEGATIVE PRICES

With the prolific growth of wind generation in the SPP market, the incidence of intervals with

negative prices continues to be a concern. However, the frequency of negative price

intervals decreased from 2017 to 2018, as shown in Figure 4-17. This decline was primarily a

result of congestion reductions as a result of transmission expansion (see Section 5.1.3).

Higher loads were also a factor in lowering the frequency of negative priced intervals (see

2.2.3.

In 2018, less than one percent of all asset owner intervals64 in the day-ahead market had

prices below zero, as shown in Figure 4—17. This is down from three percent of all intervals in

2017.

Figure 4—17 Negative price intervals, day-ahead, monthly

While the same pattern holds in the real-time market (see Figure 4—18), negative price

intervals in the real-time market occurred almost five times more frequently than in the day-

ahead market.

64 Asset owner intervals are calculated as the number of asset owners serving load that are active in an interval. For example, if there 60 asset owners active in one five-minute interval throughout an entire 30 day month, the total asset owner intervals would be 518,400 for the month (60 asset owners * 288 intervals per day * 30 days).

0%

3%

6%

9%

12%


Perc

ent

neg

ativ

e p

rice

inte

rval

s

LMP −$0.01 to −$25.00 LMP −$25.01 to −$50.00 LMP less than −$50.00



Figure 4—18 Negative price intervals, real-time, monthly

The frequency of negative price intervals in the real-time market fell to just over three percent

of 2018 intervals, down from just over seven percent in 2017. Negative prices in the day-

ahead market were almost exclusively between −$0.01/MWh and −$25/MWh, with only two

percent of intervals with negative prices having prices lower than −$25/MW. However, in the

real-time market 15 percent of intervals with negative prices had prices lower than

−$25/MWh.

Additionally, occurrences of negative prices in the day-ahead market are most prevalent in

the overnight, low-load hours as shown in Figure 4—19.

0%

3%

6%

9%

12%


Perc

ent

neg

ativ

e p

rice

inte

rval

s

LMP −$0.01 to −$25.00 LMP −$25.01 to −$50.00 LMP less than −$50.00



Figure 4—19 Negative price intervals, day-ahead, by hour

This figure shows that the day-ahead negative price intervals in 2018 during overnight hours

are below both 2016 and 2017 levels. Also of note, during the on-peak hours, less than 0.2

percent of intervals in the day-ahead market were negative in 2018.

Negative price intervals in the real-time market (see Figure 4—20) follow the same pattern as

the day-ahead market with most negative price intervals occurring in the overnight, low-load

hours.

Figure 4—20 Negative price intervals, real time, by hour

0%

5%

10%

15%

20%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Perc

ent

neg

ativ

e p

rice

inte

rval

s

Hour ending2016 2017 2018

0%

5%

10%

15%

20%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Perc

ent

neg

ativ

e p

rice

inte

rval

s

Hour ending2016 2017 2018



These negative price intervals in the real-time market occur much more frequently than the

day-ahead market, with a 2018 peak of nine percent of intervals in real-time in the last hour of

the day, compared to a peak of just over 2 percent in day-ahead. During 2017, the first five

hours and last two hours of the day experienced negative prices in over 10 percent of all

intervals. However, in 2018 no intervals had negative prices over 10 percent of the time. In

2018, the real-time market had an average of about 3.5 percent of intervals with negative

prices during the on-peak hours.

At the asset owner level (for those serving load), the distribution of negative price intervals

during 2018 clustered around the footprint average, as shown in Figure 4—21.

Figure 4—21 Negative price intervals, real-time, by asset owner

In 2018, only one asset owner experienced negative prices in more than 10 percent of

intervals. This is in contrast to 2017, when four asset owners experienced negative prices in

excess of 15 percent of intervals. On the low end, all but six asset owners experienced

negative prices in fewer than five percent of all intervals. Contrasting to 2017, only 18 asset

owners experienced negative prices in fewer than five percent of intervals.

While the frequency of negative price intervals declined in 2018, the MMU remains

concerned about the frequency of negative price intervals. Negative prices may not be a

problem in and of themselves, however, they do indicate an increase in surplus energy on the

system. This may be exacerbated by the practice of self-committing after the day-ahead

market. In the SPP market where there is an abundance of capacity and significant levels of

0%

3%

6%

9%

12%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61

Perc

ent

neg

ativ

e p

rice

inte

rval

s

Market participants (high to low)

LMP -$0.01 to -$25.00 LMP -$25.01 to -$50.00 LMP less than -$50.00 average



renewable resources, negative prices can occur when renewable resources need to be

backed down in order for traditional resources to meet their committed generation.

Moreover, unit commitment differences, due to wind resources not being in the day-ahead

market and then coming on-line for the real-time market, can create differences in the

frequency of negative price intervals between the day-ahead and real-time markets. This

disparity between the markets negatively impacts the efficient commitment of resources.

As more wind generation is anticipated to be added over the next several years, the

frequency of negative prices has the potential to increase. Negative price intervals highlight

the need for changes in market rules to address self-committing of resources in the day-

ahead market and the systematic absence of some forecasted variable energy resources in

the day-ahead market to improve market efficiency. These issues are discussed further in

Chapter 8.

4.1.6 OPERATING RESERVE MARKET PRICES

Operating reserve is made up of four products: (1) regulation-up, (2) regulation-down, (3)

spinning reserve, and (4) supplemental reserve. The regulation products are used to ensure

the amount of generation matches load on a subinterval basis. Generators respond to

regulation instructions in seconds. Spinning and supplemental products are reserved for

contingency situations and respond to instructions within ten minutes.

Average monthly real-time prices for operating reserve products are presented in Figure 4—

22.

Figure 4—22 Operating reserve product prices, real-time

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

Regulation up Regulation down Spinning reserves Supplemental reserves



Generally speaking, regulation-up and regulation-down usually have the highest market

clearing prices. Supplemental reserves always have the lowest average prices of the

operating reserve products, with prices averaging less than two dollars on an annual basis.

Spinning reserve prices typically fall between regulation-up and supplemental reserve prices.

However, 2018 saw three months with spinning reserve prices above regulation-down prices.

Day-ahead and real-time price patterns vary across the operating reserve products, see

Figure 4—23 through Figure 4-26.

Figure 4—23 Regulation-up service prices

From 2017 to 2018, the average real-time market clearing price for regulation-up decreased

from $10/MW to $9/MW. Day-ahead regulation-up market clearing price was virtually

unchanged from 2017 to 2018 at $8.50/MW. Monthly prices for regulation-up were highest

in the peak wind months during the spring and especially in the fall. The high prices during

these periods can mostly be attributed to higher wind penetration levels during these

periods.

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

day-ahead real-time



Figure 4—24 Regulation-down service prices

Regulation-down market clearing price in the real-time market averaged about $6.40/MW in

2018 after peaking in 2017 at nearly $10/MW. The higher 2017 prices can generally be

attributed to the high levels of wind production in that year, particularly in high wind and low

load months. During these months, many thermal units were operating at their economic

minimums, which are lower than regulation minimums. Costs are higher to move these units

up to the regulation range. Day-ahead regulation-down market clearing prices also have

returned to 2016 levels at around $4.70/MW after peaking in 2017 at about $6.60/MW.

Figure 4—25 Spinning reserve prices

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

day-ahead real-time

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

day-ahead real-time



The market clearing price for real-time spinning reserves actually dropped slightly from 2017

to 2018 at just over $5/MW. However, October 2018 experienced the highest monthly

market clearing price for real-time spinning reserves, at just over $10/MW, since the market

began in 2014. This can be mostly attributed to the increase in scarcity events in that month.

Day-ahead spinning reserves averaged almost $5.50/MW for the year and peaking in the fall

months as natural-gas prices increased.

Figure 4—26 Supplemental reserve prices

Supplemental reserve market clearing prices remained low in both markets, with prices

averaging about $1/MW. On a monthly basis, the October average real-time market clearing

price for supplemental reserves was just over $3/MW. This is the highest price since May

2014.

In March 2015, SPP introduced a new product paying regulating units for mileage costs

incurred when moving from one set point instruction to another. These mileage payments

are paid directly through the operating reserve prices shown for regulation-up and

regulation-down, as shown in Figure 4—27. The market calculates a mileage factor for both

products each month that represents the percentage a unit is expected to be deployed

compared to what it cleared. If a unit is deployed more than the expected percentage, then

the unit is entitled to reimbursement for the excess at the regulation mileage marginal

clearing price. If the unit is deployed less, it must buy back its position.

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

day-ahead real-time



Figure 4—27 Regulation mileage prices, real-time

On an annual basis, average monthly regulation-up mileage prices for 2018 rose slightly from

2017 at around $10/MW. Average monthly regulation-down mileage prices in 2018 returned

to 2016 levels after spiking in 2017, almost exactly matching the pattern in regulation-down

prices in those years. The decline in 2018 mileage prices is a result of changes in bidding

patterns compared to fall 2017.

Overall, the cost of ancillary services in 2018 was $76 million, a decrease of three percent

from 2017. This can be primarily attributed to a slight decrease in ancillary service prices,

and a slight increase in reserve requirements.

The MMU analyzed regulation mileage prices in 2017 and found a design inefficiency.

Mileage prices are not set by the marginal resource’s cost like other products. Instead,

resources are cleared for regulation based on their service offers. These service offers are

derived by taking the competitive offer for regulation and adding the mileage offer to it after

discounting the mileage offer by the applicable mileage factor. For instance, the service offer

of a resource with a competitive regulation-down offer of $1 and a regulation-down mileage

offer of $36 would be $10 if the mileage factor is 25 percent.65 If the $10 service offer is

economic, then the resource will clear for regulation-down and the regulation-down mileage

price will be set at $36 if this is the highest mileage price that cleared in the market.

65 $1 + $36 * 0.25 percent = $10

$0

$4

$8

$12

$16


Mar

ket c

lear

ing

pri

ce ($

/MW

)

Regulation-up mileage Regulation-down mileage



The MMU has observed instances where resources cleared with regulation-down competitive

offers of $0 and mileage offers just under $50. These units consistently cleared with this offer

strategy because the service offer was near $10.50 (e.g. 21 percent*$50) which was lower

than the services offers of other resources offering in higher competitive offers. For instance,

another resource may offer in a $12 competitive offer and $0 mileage offer. This would make

that resource’s service offer $12 ($12+0*21 percent). In this circumstance, the resource with

the highest service offer will set the regulation-down price at $12, but the mileage offer will

be $50, set by the highest cleared mileage offer.

In addition, the MMU observed systematic overpayment of regulation mileage in the day-

ahead market, which appears to be the result of the mileage factor being set consistently too

high relative to actual mileage deployed. This occurred because the mileage factor is being

set on historical instructed regulation megawatts rather than deployed regulation. When

resources have to buy back their position, they typically have to buy back at the inflated

mileage offer. Using the example above, if a resource clears for 10 megawatts it will receive

the $12 clearing price for a total payment of $120, which was set using a $0 mileage offer.

However, if it does not get deployed for regulation it will have to buy back 2.1 megawatts at

the $50 mileage offer, because they performed less than expected. The unit was paid 2.1

megawatts at a $0 price for expected mileage at the clearing, but the buyback is now $105.

This makes the total payment to the resource for clearing regulation $15 or $1.50 per cleared

megawatt.

The instructed values for regulation are on average two and a half times what resources

perform. If the mileage factor was forecasted in the exact amount of what was performed

then the excess mileage payments should closely offset the unused mileage charges.

However, this is not the case. The reason for the difference is that regulation is deployed on

a 4 second basis but it is settled on a five minute basis. Resources can be directed to move

up 10 megawatts during at the beginning of the interval. However, 20 seconds later they

may be directed to hold off on providing that regulation. If their ramp rate is only 10

megawatts per minute, they will only have provided 3.3 megawatts of regulation. This is

generally what causes the instructed values to vary from the actual values.

Figure 4—28 below illustrates the differences between the unused mileage charges and the

excess payments.



Figure 4—28 Regulation mileage payments and charges

Negative values represent the payments made to resources that deployed for more

regulation megawatts than were expected and positive values represent charges made to

resources that deployed for less than what was expected. In 2018, roughly twice as much

was charged for mileage buyback as was paid out for excess mileage deployment, with $4.2

million being paid for excess mileage and $8 million being charged for unused mileage. This

is consistent with all prior years.

The MMU is concerned that participants with resources frequently deployed for regulation

will have an incentive to inflate the mileage prices by offering in $0 regulation offers and high

mileage offers and screens for manipulative behavior. The MMU also has concerns that the

inflated mileage factors are causing units to buy back megawatts at the inflated amounts

which can ultimately lead to higher uplift costs in the market. As such, the MMU has

recommended that SPP review and revise the regulation mileage pricing approach to send

more appropriate price signals.

4.1.7 MARKET SETTLEMENT RESULTS

The day-ahead market accounted for 98 percent of the energy consumed in the Integrated

Marketplace. This is consistent with the last two years. Figure 4—29 shows that

approximately 264 terawatt-hours of energy were purchased in the day-ahead market at load

settlement locations, of which only six terawatt hours were in excess of the real-time

consumption, requiring a sale back to the market.

-$8

-$4

$0

$4

$8

$12

2016 2017 2018

$ m

illio

ns

Regulation-down excess mileage payment Regulation-down unused mileage chargesRegulation-up excess mileage payment Regulation-up unused mileage charges



Figure 4—29 Energy settlements, load

Negative gigawatt hours denote withdrawals from the grid. Negative cash flows denote

charges to load-serving entities. Positive gigawatt hours represent sales of day-ahead

gigawatt hours back to the real-time market and negative cash flows represent payments to

load owners for those sales. As stated above, during several hours of 2019 load over

purchased in the day-ahead market causing a sale back in real-time. However, there were

also several hours where the load consumed in real-time was greater than the day-ahead

cleared quantities. Just over five terawatt hours of energy was purchased in the real-time

market because the real-time demand was higher than that of the day-ahead market. The

close relationship of day-ahead load consumption to real-time load consumption is a sign of

an efficient day-ahead market. Theses charts highlight the increase in load in 2018, noted in

Chapter 2, as well as the corresponding increase in market prices as described above.

Day-ahead generation accounted for 90 percent of generation settled in the market, a one

percent increase from last year and a two percent decrease from 2016. Figure 4—30 presents

the settlement of SPP generators.

-50,000

0

50,000

100,000

150,000

200,000

250,000

300,000

2016 '17 '18

Load

(GW

h)

-$2,000

$0

$2,000

$4,000

$6,000

$8,000

2016 '17 '18

Cas

h flo

w ($

mill

ions

)



Figure 4—30 Energy settlements, generation

Positive gigawatt hours denote injections into the grid. Positive cash flows denote payments

to generators. Negative gigawatt hours represent repurchases in the real-time market and

negative cash flows represent charges to generators for those repurchases. Ten percent of

the energy cleared in the day-ahead market was settled by purchasing energy in the real-

time market rather than generating the energy, which was about one percent more than what

was experienced in the prior three years.

SPP plays the role of the customer in the operating reserve market. At hour ending 6:00 AM

on the day before the operating day, SPP posts the forecasted amount of each operating

reserve product that is to be procured. This data sets the demand for the products for the

day-ahead market. SPP can change the demand levels after the clearing of the day-ahead

market, but this a rare occurrence. Even though the demand is essentially the same between

the day-ahead market and the real-time market, there is considerable activity with respect to

the operating reserve products in the real-time market. Figure 4—31 presents the settlements

data for operating reserves.

-50,000

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

2016 '17 '18

Gen

erat

ion

(GW

h)

-$2,000

$0

$2,000

$4,000

$6,000

$8,000

2016 '17 '18

Cas

h flo

w ($

mill

ions

)



Figure 4—31 Operating reserve product settlements

Positive numbers represent the cleared gigawatt hours for ancillary services. Purchases

represent the repurchase of day-ahead cleared ancillary service in the real-time market.

A large percentage of day-ahead sales (35 percent) are settled in the real-time market by

repurchasing the operating reserve product rather than supplying the service in the real-time

market. This is in contrast to 10 percent of the real-time energy generation in excess of day-

ahead cleared generation, which is settled at real-time prices. This 2018 repurchase rate is

relatively the same as the prior years. However, the repurchase is down five percent from the

40 percent that occurred in the first 12 months of the market.

Sixty-six percent of the 2018 real-time regulation-up service was settled at day-ahead prices,

exactly the same as the previous year. The corresponding percentages for regulation-down

service, spinning reserves, and supplemental reserves are 63 percent, 69 percent, and 61

percent respectively. These results were similar with the respective numbers in 2017 of 63

percent, 70 percent, and 62 percent. This essentially means that operating reserve products

are being moved around to different resources, due to their day-ahead and real-time

clearing, in about the same volumes as last year. This causes some resources to buy back

their day-ahead cleared megawatts at real-time prices, while other resources clear those

some quantities of megawatts and get paid real -time prices for their clearing.

0

3,000

6,000

9,000

12,000

2016 '17 '18 2016 '17 '18 2016 '17 '18 2016 '17 '18

Regulation-up Regulation-down Spinning reserves Supplementalreserves

GW

Day-ahead sales Real-time sales Real-time purchases



4.2 MAKE-WHOLE PAYMENTS

The Integrated Marketplace provides make-whole payments (MWPs) to generators to ensure

that the market provides sufficient revenue to cover the cleared offers providing energy and

operating reserves for a period in which the resource was committed. To preserve the

incentive for a resource to meet its market commitment and dispatch instruction, market

payments should cover the sum of the incremental energy cost, start-up cost, no-load cost,

transition cost, and cost of operating reserve products. The make-whole payment provides

additional market payments in cases where revenue is below a resource’s offer to make the

resource whole to its offers of operating reserve products, incremental energy, start-up,

transition, and no-load.

For the resources that are not combined-cycle, settlements separately evaluate: (1) day-

ahead market commitments based on day-ahead market prices, cleared offers and dispatch;

and (2) reliability unit commitments based on real-time market prices, cleared offers, and

dispatch. Combined-cycle resources can be cleared in both the day-ahead and real-time

markets at the same time, which is unique to combined-cycles. As a result, settlements must

evaluate the revenues and cost of both real-time and day-ahead commitments when

calculating real-time make-whole payments for combined-cycles.

For 2018, day-ahead market and reliability unit commitment make-whole payments totaled

approximately $72 million, up six percent from $68 million last year. Make-whole payments

averaged about $0.26/MWh for 2018, which is the same as 2017. In comparison to other

RTO/ISO markets, SPP’s make-whole payments were in the range of uplift costs, which varied

from $0.08/MWh to $0.29/MWh in 2017 and 2018.66

Figure 4—32 shows monthly and annual day-ahead make-whole payment totals by

technology type. Figure 4—33 shows the same make-whole payment information for

reliability unit commitment.

66 2017 ISO NE State of Market Report https://www.potomaceconomics.com/wp-content/uploads/2018/06/ISO-NE-2017-EMM-SOM-Report_6-17-2018_Final.pdf, page 39; 2018 State of the Market Report for PJM http://www.monitoringanalytics.com/reports/Presentations/2019/IMM_MC_SOM_20190321.pdf, page 9.



http://www.monitoringanalytics.com/reports/Presentations/2019/IMM_MC_SOM_20190321.pdf



Figure 4—32 Make-whole payments by fuel type, day-ahead

Figure 4—33 Make-whole payments by fuel type, reliability unit commitment

Day-ahead make-whole payments constituted 38 percent of the total make-whole payments

in 2018. Gas-fired resources represent about 81 percent of all make-whole payments, with

69 percent of all make-whole payments to simple-cycle gas resources through reliability unit

commitment make-whole payments.

$0

$2

$4

$6

$8


Mill

ions

Coal Gas, combined-cycle Gas, simple-cycle Hydro Wind Other

$0

$10

$20

$30

$40

$50

2016 '17 '18

Mill

ions

$0

$2

$4

$6

$8


Mill

ions

Coal Gas, combined-cycle Gas, simple-cycle Hydro Wind Other

$0

$10

$20

$30

$40

$50

2016 '17 '18

Mill

ions



Make-whole payments occur for several reasons, which include some of the following: local

reliability commitments, uncaptured congestion in the day-ahead market, inflexibility of

resources to move in economic ranges between on-peak and off-peak hours, and excessive

transmission congestion not being solved by the market. Make-whole payments to resources

in the “other” category primarily represent payments to oil-fired resources. The increase in

real-time make-whole payments in 2018 can primarily be attributed to more resources being

brought on from the reliability unit-commitment processes for capacity needs.

Figure 4—34 shows the share of each cause of make-whole payments in the real-time and

day-ahead markets.

Figure 4—34 Make-whole payments, commitment reasons Real-time commitment reason

2016 2017 2018

Intra-day RUC 24.4%% 35.7% 26.1% Manual, SPP capacity 2.6% 4.5% 24.9% Short-term RUC 12.9% 6.1% 13.7% Manual, voltage 8.0% 6.4% 4.4% Day-ahead RUC 20.1% 7.5% 1.8% Manual, stagger 0.0% 0.0% 0.9% Other 0.0% 0.8% 0.3% Manual, intra-day RUC 8.4% 6.1%

Manual, day-ahead RUC 0.3% 1.2%

Manual, off-line supplemental 0.1% 0.0% 0.0%

Day-ahead commitment reason 2016 2017 2018

Day-ahead market 73.2% 93.7% 96.9% Manual, voltage support 26.8% 6.3% 3.1%

Voltage support commitments decreased in 2018 when compared to prior years. There was

a large increase in percentage of real-time make-whole payments paid to resources running

on for capacity. In fact, just under 25 percent of the real-make-whole payments were paid to

resources committed manually for capacity needs, which occurred over five times more than

in 2017. Manual, intra-day and day-ahead, RUC were removed as an option in 2018, thus

make-whole payments that were attributed to those categories in prior years are now

captured in the Manual, SPP capacity category.



There was an initiative in late 2018 to clean up the categorizing of manual commitments.67

Prior to the initiative operators often categorized capacity commitments under the intra-day

RUC and day-ahead RUC manual codes. However, after the initiative the manual RUC codes

were removed, manual commitments for capacity are all coded under the SPP capacity code.

This explains about half of the increase in make-whole payments for resources committed for

capacity. The residual increase can most likely be attributed to the increase in wind volatility

(see Section 4.1.3) and resource outages in October and November (see Figure 3—14).

These led to the increased need for manual commitments as rampable capacity was often

scarce.

Make-whole payments associated with voltage support commitments do not follow the same

uplift process outlined in Section 4.3.1. Instead, the cost of these make-whole payments are

distributed to the settlement areas that benefited from the commitment by way of a load ratio

share. Figure 4—35 illustrates the level of make-whole payments associated with voltage

support commitments.

Figure 4—35 Make-whole payments for voltage support

The make-whole payments stemming from voltage support commitments are down 18

percent from 2017 and down 75 percent from 2016. The high make-whole payments in 2016

67 SPP MMU, Quarterly State of the Market report, Summer 2018 special issues section, page 49-6: https://www.spp.org/documents/58811/spp_mmu_qsom_summer_2018_final.pdf.

$0.0

$0.2

$0.4

$0.6

$0.8

$1.0


Mak

e-w

hole

pay

men

ts ($

/mill

ions

)

Day-ahead Real-time

$0

$3

$6

$9

$12

$15

2016 '17 '18

Mak

e-w

hole

pay

men

ts ($

/mill

ions

)

https://www.spp.org/documents/58811/spp_mmu_qsom_summer_2018_final.pdf



were directly attributed to some resource outages in the south central section of the SPP

footprint.

Most SPP resources received modest total annual make-whole payments as highlighted in

Figure 4—36.

Figure 4—36 Concentration of make-whole payments by resource

The majority of resources in SPP received less than $250,000 in make-whole payments in

2018. Seven resources received over $1 million, compared to eight resources with $1 million

make-whole payments in 2017. One resource received $3.7 million dollars in make-whole

payments, with $800 thousand of those dollars coming voltage support commitments. The

vast majority of the make-whole payments to this resource occurred because of manual

reliability commitments made to control a temporary flowgate in the area.

Figure 4—37 reveals some concentration in the market participants that received the highest

levels of make-whole payments. The overall numbers have not changed significantly over the

past few years.

$0

$1

$2

$3

$4

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340

Tota

l mak

e-w

hole

pay

men

ts($

mill

ions

)

Resources receiving make-whole payments



Figure 4—37 Market participants receiving make-whole payments 2016 2017 2018

> $1 million

> $5 million

> $10 million

> $1 million

> $5 million

> $10 million

> $1 million

> $5 million

> $10 million

Market participants receiving make-whole payments

12 4 2 12 5 1 16 5 1

% of make-whole payments by category

92% 61% 38% 90% 61% 18% 94% 57% 14%

In 2018, there were 16 market participants that each received annual make-whole payments

in amounts greater than $1 million. These 16 market participants accounted for 94 percent of

the total make- whole payments paid out in 2018. Twelve of those same 16 participants also

received over $1 million each in 2017. All but one of those 12 participants received over a $1

million in make-whole payments for each of the last three years. In 2018, there were five

participants that received over $5 million each in make-whole payments and out of that one

received $10 million. The participant with $10 million in total make-whole payments

accounted for 14 percent of all make-whole payments paid out in 2018. This was down from

18 percent in 2017.

4.2.1 MAKE-WHOLE PAYMENT ALLOCATION

The allocation of both day-ahead and real-time make-whole payments has important

consequences to the market. In principle, for market efficiency purposes uplift cost allocation

should be directed to those members that contributed to the need for the make-whole

payments (i.e., cost causation).

For the day-ahead market, make-whole payment costs are distributed to both physical and

virtual withdrawals on a per-MWh rate. The per-MWh rate is derived by dividing the sum of

all day-ahead make-whole payments for an operating day by the sum of all cleared day-

ahead market load megawatts, export megawatts, and virtual bids for the operating day. The

average per-MWh rate for withdrawing locations in the day-ahead market was just over

$0.09/MWh in 2018. This is approximately one and a half cents less than the 2016 and 2017

averages.



For the real-time market, make-whole payment costs are distributed through a per-MWh rate

that is assigned to all megawatt-hours of deviation in the real-time market. The average real-

time distribution rate was $0.98/MWh for 2018, up $0.08/MWh from $0.90/MWh in 2017.

There are eight categories of deviation and each category receives an equal amount per

megawatt when the cost of make-whole payments is applied. This can be seen in the

settlement location deviation charge, show in Figure 4—38.

Figure 4—38 Make-whole payments by market uplift allocation, real-time

Uplift type Deviation

MWs (thousands)

Uplift charge

(thousands)

Share of MWP

charges

Cost per MW of

deviation Settlement location deviation

37,685 $ 35,373 77.8% $ 0.94

Outage deviation 4,306 $ 5,064 11.1% $ 1.18

Status deviation 1,616 $ 1,721 3.8% $ 1.07

Maximum limit deviation 1,478 $ 1,598 3.5% $ 1.08 Reliability unit commitment self-commit deviation

590 $ 635 1.4% $ 1.08

Uninstructed resource deviation

566 $ 533 1.2% $ 0.94

Reliability unit commitment deviation 283 $ 266 0.6% $ 0.94

Minimum limit deviation 190 $ 231 0.5% $ 1.21

Even though each category of deviation is applied the same rate for deviation, approximately

78 percent of the real-time make-whole payment costs were paid by entities withdrawing

(physical or virtual) more megawatts in the real-time market than the day-ahead market.

Transactions susceptible to this charge are virtual offer megawatts, real-time load megawatts

in excess of the day-cleared megawatts for a unit, exporting megawatts in real-time in excess

of the export megawatts cleared in the day-ahead market, and units pulling substation power

in excess of any megawatts produced by the unit. However, virtual offers are the most

susceptible as 100 percent of their megawatts are considered incremental. Due to this fact

virtual offers alone paid for forty-one percent of all real-time make-whole payments.

4.2.2 REGULATION MILEAGE MAKE-WHOLE PAYMENTS

In March 2015, SPP introduced regulation compensation changes for units deployed for

regulation-up and regulation-down. One component of the regulation compensation

charges is regulation-up and regulation-down mileage make-whole payments for units that



are charged for unused regulation-up or regulation-down mileage at a rate that is in excess of

the regulation-up or regulation-down mileage offer.

SPP calculates mileage factors monthly for both regulation-up and regulation-down. These

mileage factors are ratios of historical averages of the percentage of each regulation product

deployed to the regulation product cleared in the prior month. The regulation-up mileage

factor and regulation-down mileage factors averaged 15 percent and 25 percent,

respectively, for 2018. The regulation-down mileage factor is up one percent from last year

and the regulation-up mileage factor is down three percent.

The mileage factor is a key component in the computation of mileage make-whole payments.

When the mileage factor is greater than the percentage of deployed regulating megawatts to

cleared regulating megawatts for each product, the resource must buy back the non-

deployed megawatts at the mileage marginal clearing price for the respective product. If the

mileage marginal clearing price used for the buyback is greater than the unit’s cost for the

product a make-whole payment may be granted.

Figure 4—39, below, illustrates the mileage make-whole payments for 2018 and the prior two

years.

Figure 4—39 Regulation mileage make-whole payments

The large increase in regulation-down make-whole payments for the second half of 2018 can

be directly attributed to the mileage clearing prices for regulation-down being nearly

$36/MW for some intervals in the latter half of 2018. The same trend also happened in 2017,

$0

$40

$80

$120

$160

$200


Mak

e-w

hole

pay

men

ts ($

/tho

usan

ds)

Regulation-up mileage MWP Regulation-down mileage MWP

$0

$200

$400

$600

$800

$1,000

2016 '17 '18



but the prices during that period got as high as $50/MW. The highest regulation-down

mileage prices happened in March 2016, when they got as high as $60/MW. The MMU

identified a design inefficiency in the regulation-mileage compensation process. Details of

this issue are described in section 4.1.6.

4.2.3 POTENTIAL FOR MANIPULATION OF MAKE-WHOLE PAYMENTS

The 2014 Annual State of the Market Report, the MMU highlighted specific vulnerabilities that

market participants could potentially manipulate in SPP’s make-whole payment provisions.

The below vulnerabilities were directly associated with the FERC order regarding the make-

whole payments and related bidding strategies of JP Morgan Ventures Energy Corp:68

1) make-whole payments for generators committed across the midnight hour,

2) make-whole payments for regulation deployment, and

3) make-whole payments for out-of-merit energy.

In each case, a market participant has the ability to position its resource to receive a make-

whole payment without economic evaluation of its offers by the market.

A revision request was passed by the SPP board during the July meeting.69 The revision

request corrects the issue one by limiting make-whole payments for any resource with multi-

day minimum run times to the lower of the market offer or the mitigated offer. This limitation

only applies to offers falling in hours in which the resource’s offers were not accessed by one

of the security constrained unit commitment (SCUC) processes and the resource bid at or

above their mitigated offer on the first day. Subsequent to board approval of the proposal,

SPP legal staff identified tariff language that requires modification as it conflicts with the

language passed in the proposal. This has required that additional tariff modifications go

through the stakeholder process.

The MMU strongly recommends that SPP and its stakeholders clean up the tariff language

necessary to implement the changes to address the gaming issue. Addressing this matter in

a timely manner is a high priority for the MMU.

68 See 144 FERC 61,068. 69 Revision request 306 (2014 ASOM MWP MMU Recommendation [3-Day Minimum Run Time])



Two revision requests were implemented in 2018 concerning regulation adjustments. One

adds an assessment of economics when deploying units for regulation and the other caps the

offers used for the adjustment. The MMU feels that these two revision requests adequately

close the gaming opportunities present in the regulation adjustment.70

4.2.4 JOINTLY-OWNED UNIT MAKE-WHOLE PAYMENTS

Another make-whole payment concern existed related to jointly-owned resources and the

combined resource option. At the time the MMU made their original recommendations, the

market committed jointly-owned units as one unit, dispatched each separate owner on a

percentage of ownership, and paid make-whole payments for energy based on the individual

owners’ energy offers. This allowed a shareowner to benefit from a higher energy offer than

its co-owners through high minimum energy costs in the make-whole payment.

In August 2017, SPP implemented changes based on two revision requests to eliminate the

gaming opportunities present in the original design. The new design eliminates the potential

gaming opportunity by taking all owners’ pricing points of a combined resource option for

jointly-owned resources and aggregates those price points into a single energy offer curve

for the unit. This revision request also enforced all shares under a combined resource option

to have a minimum capacity limit of zero megawatts when being assessed for dispatch. Start-

up and no-load costs were still reimbursed on the same percentage of ownership method

used prior to the changes.

With the new method, individual shares only received a dispatch instruction greater than zero

megawatts based on the new aggregated energy offer curve, which eliminated the original

gaming opportunity. However, while the change eliminated the gaming opportunity outlined

in the 2014 annual state of the market report, the design introduced new issues. There are

multiple issues:

• The first issue is related to how the new design allocated costs when these resources

were self-committed and uneconomic. During these uneconomic periods, the lowest

offers of the jointly-owned unit will be cleared to meet the unit’s minimum physical

capacity operating limit. The participants with the lowest offers would pay the costs of

running uneconomically.

70 The revision requests were 242 and 243.



• The second issue occurs when a jointly-owned unit clears in any process after the day-

ahead market—such as the reliability unit commitment process. Market participants

can inflate their offers at the time of dispatch and be made whole to that inflated offer,

even though the jointly-owned unit may be running at its physical minimum limit. This

is because each jointly-owned unit owners’ minimum operating limits are considered

zero megawatts for settlement even though the unit is only running above zero

megawatts because of the physical minimum limits of the unit.

• The third issue is with regards to real-time make-whole payment distributions. These

distributions are not accurately applied because the co-owners of jointly owned units

each have zero minimum capacity limits in settlements. They are allocated costs

incorrectly for real-time make-whole payments for the difference in the megawatts

between the minimum limit of the resource and the desired energy level.

• Additionally, there are other deviation calculations that are based on the difference of

a unit in manual status output versus where they would be desired to run if the

resource were dispatchable. When this happens there is a potential for the combined

unit’s desired energy to be lower than the physical minimum limit of the unit. This is

because the settlement process calculates the desired energy quantities with the

assumption that the unit can run at zero megawatts, which is below the physical

minimum of the unit. As a result, the jointly-owned unit asset owners could have

distribution charges unduly applied.

Both SPP stakeholders and FERC approved a new design in 2018.71 The new design is

expected to be implemented after implementation of the new settlement system, which is

currently expected to take place in early 2020. The new design requires that jointly-owned

units offer in as one unit. The market system will dispatch the resource as one unit, and then

the settlements process will allocate the cost and revenues out by percentage of ownership

of the resource. The new design will address the concerns outlined above.

71 [Revision Request 266-JOU Combined Single Resource Modeling post Settlement Share Allocation, https://www.spp.org/spp-documents-filings/?id=21069..]

https://www.spp.org/spp-documents-filings/?id=21069



4.3 TOTAL WHOLESALE MARKET COSTS AND PRODUCTION COSTS

The average annual all-in price, which includes the costs of energy, day-ahead and real-time

reliability unit commitment make-whole payments, operating reserves,72 reserve sharing

group costs, and payments to demand response resources, was $27.65/MWh in 2018. This

compares to the average price of energy at load pricing nodes in SPP’s real-time market for

2018 of $27.08/MWh.73 The all-in price was just over 14 percent higher than the 2017

average all-in price, which is partially attributed to the increase in load and decline in

negative prices in 2018.74 Figure 4—40 plots the average all-in price of energy and the cost of

natural gas, measured at the Panhandle Eastern (PEPL) hub.

Figure 4—40 All-in price of electricity and natural gas cost

72 Operating reserves are resource capacity held in reserve for resource contingencies and NERC control performance compliance, which includes the following products: regulation-up service, regulation-down service, spinning reserve and supplemental reserve. 73 The cost of energy includes all of the shortage pricing components. 74 The Reserve Sharing Group costs and payments to demand response resources were negligible for both years.

$0

$1

$2

$3

$4

$0

$10

$20

$30

$40


Nat

ural

gas

co

st ($

/mm

Btu

)

All-

in p

rice

($/M

Wh)

Energy (LMP) Reserves Day-ahead MWP Real-time MWP Gas PEPL



The figure shows that the vast majority of costs are from the day-ahead and real-time energy

markets.75 The graph also shows that the market cost of operating reserves and make-whole

payments constituted approximately two percent of the all-in price, with make-whole

payments and operating reserves amounting to $0.26/MWh and $0.31/MWh, respectively.

Production cost “is defined as the settlement cost for the market … for all resources.”76

Production cost, in this case, is the sum of four components:

• energy: cleared megawatts multiplied by locational marginal prices;

• ancillary service: cleared operating reserves multiplied by market clearing prices ;

• start-up: “…the out of pocket cost that a Market Participant incurs in starting up a

generating unit from an off-line state ….”;77 and

• no-load: “…the hourly fee for operating a synchronized Resource at zero … output.”78

Figure 4—41 shows the average daily production cost for the day-ahead market.

Figure 4—41 Production cost, daily average, day-ahead

75 Shortage pricing is included in the energy component and not easily separated out in the SPP settlement data. See Section 4.2 for a discussion of shortage pricing impacts. 76 Integrated Marketplace Protocols, Section 7.2.1 77 Integrated Marketplace Protocols, Start-Up Offer 78 Integrated Marketplace Protocols, No-Load Offer

$0

$10

$20

$30


Mill

ions

Energy Ancillary service No-load Start-up



The daily average day-ahead market production cost was essentially unchanged from 2016

to 2017. However, the daily average production cost increased more than eighteen percent

from 2017 to 2018. The increase in production cost was almost fully attributed to the energy

component. The energy component is sensitive to numerous inputs, which include fuel cost,

ancillary service scarcity, and load levels. The range of the 2018 daily day-ahead production

costs exceeded $50 million ranging from $2 million to $52 million. Additionally, more than

eighty-one percent of the daily production costs ranged between $10 million and $30

million.

Figure 4—42 Production cost, daily average, real-time

The average daily real-time market production cost was also essentially unchanged from

2016 to 2017. However, real-time production cost increased more than 16 percent from

2017 to 2018. The increase in production cost, similar to day-ahead, was also almost fully

attributed to the energy component. The range of the 2018 daily real-time production costs

was almost $60 million ranging from $4 million to $63 million. Additionally, more than 87

percent of the daily production costs ranged between $10 million and $30 million.

4.4 LONG-RUN PRICE SIGNALS FOR INVESTMENT

In the long term, efficient market prices provide signals for any needed investment in new

transmission, generation, and ongoing maintenance of existing generation to meet load.

Given the very high amount of capacity in the SPP system at peak, which the MMU estimates

was about 35 percent in 2018 (see Section 6.1.3), the MMU does not expect market prices to

$0

$10

$20

$30


Mill

ions

Energy Ancillary service No-load Start-up



support new entry of non-wind generation investments. While the SPP markets on their own

may offer low incentives for new growth in renewable generation and storage resources,

some additional reasons for these investments include federal and/or state incentives,

expansion of corporate goals, and emission reduction plans among others.

The MMU conducted an analysis to determine if the SPP market would support investments

in new thermal generation by analyzing the fixed costs, and annual fixed operating and

maintenance costs of three generation technologies relative to their potential net revenues79

at SPP market prices. The plants considered include a scrubbed coal plant, a natural gas

combined-cycle, and a combustion turbine. Figure 4—43 provides the cost assumptions and

Figure 4—44 shows the results of the net revenue analysis. The analysis assumes the market

dispatches the hypothetical resource when price exceeds the short-run marginal cost of

production. In addition to these assumptions a capital recovery factor of 12.6 percent was

used in the annual fixed operating and maintenance cost component.

Figure 4—43 Net revenue analysis assumptions

Scrubbed coal Advanced gas/oil combined-cycle

Advanced combustion Turbine

Size (MW) 650 429 237

Total overnight cost ($/kW-yr.)

$5,089 $1,108 $680

Variable overhead and maintenance ($/MWh)

$7.17 $2.02 $3.54

Fixed overhead and maintenance ($/kW-yr.)

$70.70 $10.10 $6.87

Heat rate (Btu/kWh) 9,750 6,300 9,600 Source: EIA Updated Capital Cost Estimates for Utility Scale Electricity Generating Plants, April 2018

79 Net revenue is equal to revenues minus estimated marginal cost.



Figure 4—44 Net revenue analysis results

Technology

Average marginal

cost ($/MWh)

Net revenue from SPP market

($/MW yr.)

Annual revenue

requirement ($/MW yr.)

Able to recover

new entry cost

Annual fixed O&M cost

($/MW yr.)

Able to recover

avoidable cost

Scrubbed coal $ 29.60 $ 35,845 $ 709,754 NO $ 70,700 NO

Advanced gas/oil combined-cycle

$ 18.27 $ 69,751 $ 149,238 NO $ 10,100 YES

Advanced combustion turbine

$ 28.31 $ 36,880 $ 92,261 NO $ 9,780 YES

Revenues have been insufficient to support the cost of new entry of generation for all three

technologies since the inception of the Integrated Marketplace, and 2018 was no exception.

In 2015, 2016 and 2017, prices did support the ongoing maintenance cost of combined-

cycle and combustion turbine units, though it did not support the cost of scrubbed coal units.

This is consistent with the 2018 results shown above. Declining gas prices, excess capacity,

and negative prices are the leading contributors to the decline in the profitability of coal

plants.

Figure 4—45 provides results by SPP resource zone, as indicated by the dominant utility in the

area.

Figure 4—45 Net revenue analysis by zone Scrubbed coal Gas/oil combined-cycle Combustion turbine

Resource Zone


($/MW yr.)

Able to recover

net entry costs

Able to recover

avoidable cost


($/MW yr.)

Able to recover

net entry costs

Able to recover

avoidable cost


($/MW yr.)

Able to recover

net entry costs

Able to recover

avoidable cost

AEP $ 24,723 NO NO $98,566 NO YES $55,047 NO YES

KCPL $ 25,781 NO NO $72,891 NO YES $43,327 NO YES

NPPD $17,865 NO NO $45,537 NO YES $30,635 NO YES

OGE $23,438 NO NO $71,113 NO YES $40,395 NO YES

SPS $17,884 NO NO $76,809 NO YES $44,781 NO YES

WAUE $22,846 NO NO $53,536 NO YES $14,801 NO YES

WR $22,108 NO NO $67,004 NO YES $39,832 NO YES

This shows that the conclusions do not vary geographically, even with differing energy prices

and fuel costs. Other RTO/ISO markets have experienced a “missing money problem” in

their markets, where net revenues do not support needed new investments. The MMU

expects the market to signal the retirement of inefficient generation. Aging of the fleet will



eventually change the peak available capacity such that price signals for higher net revenue

become increasingly important. The ability of market forces to provide these incentives and

long-run price signals is a strong benefit of the Integrated Marketplace.

Southwest Power Pool, Inc. Congestion and transmission congestion rights market Market Monitoring Unit


5 CONGESTION AND TRANSMISSION CONGESTION RIGHTS MARKET

This chapter reviews transmission congestion in the SPP market, as well as the transmission

congestion rights market in SPP. Key points from this chapter include:

• The areas that experienced the highest congestion costs in 2018 were in central

Kansas, southeastern Oklahoma, and southwest Missouri. Western Nebraska

experienced the lowest congestion costs for the year.

• Overall, congestion is down across the SPP footprint, with intervals having no

congestion increasing steadily from 2016 to 2018, and intervals having a breached

constraint steadily decreasing in the same period.

• Frequently constrained areas identified in the 2018 study include central Kansas and

southwest Missouri. The Texas panhandle frequently constrained area was removed

as a result of the 2018 study.

• In aggregate, market participants were effective in hedging congestion in the SPP

market using SPP’s congestion hedging products. Load-serving entities covered

more than 120 percent and non-load-serving entities covered more than 125 percent

of their total congestion cost.

• Individual market participants hedged congestion with varying degrees of

effectiveness — more than 30 percent of load-serving-entities did not fully recover

their congestion cost.

• Transmission congestion right funding fell within the target range. The annual

funding percentage exceeded 94 percent, and the annual shortfall improved by more

than $6 million. Auction revenue right funding decreased from 160 percent to 145

percent; however, the ARR surplus increased more than $20 million and exceeded

$120 million.

• Participants can transfer congestion rights through use of a bulletin board or sell back

positions in the auction. However, most congestion rights are not transferred or sold.

No bulletin board trades cleared during 2018 and have averaged less than one-fourth

of one percent of the total auction quantity since the start of the Integrated



Marketplace. Intra-auction sales80 increased in volume, but continue to average only

about five percent of the total auction volume.

5.1 TRANSMISSION CONGESTION

The locational marginal price (LMP) for the over 950 settlement locations in the SPP market

reflects the sum of three components:

1) marginal energy component (MEC) - system-wide marginal cost of the energy

required to serve the market,

2) marginal congestion component (MCC) - the marginal cost of any increase or

decrease in energy at a location with respect to transmission constraints, and

3) marginal loss component (MLC) - the marginal cost of any increase or decrease in

energy to minimize system transmission losses.

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝐿𝐿𝑀𝑀𝑀𝑀 + 𝐿𝐿𝑀𝑀𝑀𝑀 + 𝐿𝐿𝐿𝐿𝑀𝑀

LMPs are a key feature of electricity markets that ensure the efficient scheduling,

commitment, and dispatch of generation given the system load and reliability constraints.

LMPs also provide price signals for efficient incentives for future generation and transmission

investment and help guide retirement decisions.

This section focuses on the congestion and loss components of price and related items

including:

• geographic pattern of congestion and losses,

• changes in the transmission system that alter congestion patterns,

• congestion impacts on local market power,

• load-serving entities hedging congestion costs in the transmission congestion

rights market, and

• distribution of marginal congestion and loss amounts.

80 The sale of a previously acquired position in a subsequent auction.



5.1.1 PRICING PATTERNS AND CONGESTION

Figure 5—1 shows price contour maps representing the day-ahead and real-time average

prices in 2018.

Figure 5—1 Price map, day-ahead and real-time market

Annual average day-ahead market prices ranged from around $16/MWh on the western

edge of Nebraska, to around $36/MWh in the south-central section of Oklahoma. About 75

percent of this price variation can be attributed to congestion and 25 percent to marginal

losses, which is consistent with prior years. Because congestion is more volatile in the real-

time market, the average geographic price range is slightly larger, from $14/MWh to

$38/MWh.

In May 2017, upgrades were made at Woodward by adding a 138kV phase-shifting

transformer. This area was the most frequently constrained area in the SPP market before the

upgrade. However, since the phase-shifting transformer has been in service, the frequency of

Western Nebraska

New Mexico/ west Texas

Southwest Missouri

Central Kansas

Southeast Oklahoma

North Dakota/ Montana border

Woodward phaseshifter

2018 day-ahead locational marginal price

2018 real-time locational marginal price



congestion at Woodward has fallen significantly such that the MMU recommended no longer

designating Woodward as a frequently constrained area.81 Further upgrades in the area

completed in February 2018 included a second 345kV circuit of Woodward-Tatonga-

Mathewson. Buildout in this area has allowed higher levels of low-cost wind generation in the

western parts of the SPP footprint to serve load centers located on the eastern portion of SPP.

In addition, congestion has shifted eastward. The congestion pocket normally seen in the

Woodward area has shifted down towards southeastern Oklahoma.

As in prior years, the Hays, Kansas area was frequently congested with prices averaging

around $31/MWh in 2018. Lastly, while prices in the North Dakota averaged around

$22/MWh there is a pocket on the North Dakota and Montana border were prices averaged

around $28/MWh, which is consistent with prior years. The higher prices in this area can be

attributed to load associated with oil and gas exploration.

5.1.2 CONGESTION BY GEOGRAPHIC LOCATION

The major drivers of the congestion pattern in SPP are the physical characteristics of the

transmission grid and associated transfer capability, the geographic distribution of load, and

the geographic differences in fuel costs. The eastern side of the SPP footprint, with a higher

concentration of load, also has a higher concentration of high-voltage (345 kV) transmission

lines. Historically, high-voltage connections between the west and east have been limited, as

have high-voltage connections into the Texas Panhandle area.

The costs of coal-fired generation increases as transportation costs rise. For example,

transportation cost increases with distance from the Wyoming Powder River Basin near the

northwest corner of SPP’s footprint. This is important because coal is SPP’s predominant fuel

for energy generation at 42 percent in 2018.

Natural gas-fired generation, SPP’s largest fuel type by installed capacity (41 percent in

2018), resides predominantly in the southern portions of SPP. Wind-powered generation

generally lies in the western half of the footprint, and nuclear generation resides near the

center, while the majority of hydro is located in the north.

81 Frequently constrained areas are discussed further in Section 5.1.7.



These factors combine to create a general northwest to southeast split in prices. The

exception is slightly higher prices in the northern area of North Dakota resulting from the

growth of, and associated demand from, oil and gas exploration and production facilities.

Outside of the extreme northern part of North Dakota, the Integrated System typically sees

lower prices compared to the rest of the footprint.

Figure 5—2 depicts the average marginal congestion component for the day-ahead market

across the SPP footprint.

Figure 5—2 Marginal congestion cost map, day-ahead market

The lowest average marginal congestion costs occur near the Stegall tie-line in the western

Nebraska region, at −$7/MWh, and the highest marginal congestion costs lie in central

Kansas and southeast Oklahoma at around $5/MWh, and southwest Missouri at around

$7/MWh.

Hays

Neosho -Riverton

Woodward

Stegall tie-line

marginal congestion cost day-ahead market



The congestion in the Woodward area was persistent in 2017 until the phase-shifting

transformer upgrade went into place in May 2017. With the addition of the phase-shifting

transformer at Woodward, the congestion has shifted east to southeastern Oklahoma, but is

not as evident as the prior Woodward constraint. The Texas Panhandle congestion still

remains but has moved further south towards the Lubbock, Texas area.

5.1.3 TRANSMISSION EXPANSION

Several major transmission projects were completed during 2018 (shown on Figure 5—3

below) that will support the efficient transmission of energy across the SPP footprint.82

• Woodward-Tatonga-Mathewson 345 kV (circuit 2) o location: western Oklahoma o energized: February 2018

• Chisholm-Gracemont 345kV (circuit 1) o location: western Oklahoma o energized: December 2017

• Iatan-Stranger Creek 345kV voltage upgrade (circuit 1) o location: Kansas City area o energized: June 2018

• China Draw-North Loving-Kiowa-Hobbs 345k and terminal upgrades (circuit 1) o location: southeastern New Mexico o energized: June 2018

• Plant X-Tolk 230kV rebuild (circuit 1 and 2) o location: west Texas o energized: March 2018

82 See the 2019 SPP Transmission Expansion Plan Report at https://www.spp.org/documents/56611/2019_spp_transmission_expansion_plan_report.pdf.

https://www.spp.org/documents/56611/2019%20spp%20transmission%20expansion%20plan%20report.pdf



Figure 5—3 SPP transmission expansion in service, 2018

The lines depicted on the map in Figure 5—4 below are projects that will further enhance the

SPP transmission grid in future years.



Figure 5—4 SPP transmission expansion plan

The Integrated Transmission Plan (ITP) projects shown are recommended upgrades to the

extra-high-voltage backbone (345kV and above) for a 20-year horizon. The ITP process seeks

to target a reasonable balance between long-term transmission investments and congestion

costs to customers. The notification to construct (NTC) and ITP projects shown have received

a written notice from SPP to construct a transmission project that was approved by the SPP

board of directors. Planned projects that may provide relief for the most congested areas in

SPP are listed in Figure 5—10.

5.1.4 TRANSMISSION CONSTRAINTS

Market congestion reflects the economic dispatch cost of honoring transmission constraints.

SPP uses these constraints to manage the flow of energy across the physical bottlenecks of

the grid in the least costly manner while ensuring reliability. In doing so, SPP calculates a

shadow price for each constraint, which indicates the potential reduction in the total market

production costs if the constraint limit could be increased by one megawatt for one hour.

Figure 5—5 provides the top ten flowgate constraints by shadow price for 2018.



Figure 5—5 Congestion by shadow price, top ten flowgates

Flowgate name Region Flowgate location

VINHAYPOSKNO Central Kansas Vine Tap-North Hays 115kV ftlo Post Rock-Knoll 230kV (MIDW)

NEORIVNEOBLC* SW Missouri Neosho-Riverton 161kV (WR-EDE) ftlo Neosho-Blackberry 345kV (WR-AECI)

TMP109_22593* Eastern Oklahoma Stonewall Tap-Tupelo Tap 138kV (WFEC) ftlo Seminole-Pittsburg 345kV (CSWS-OGE)

TMP270_23432 Eastern Oklahoma Cleveland-Cleveland AECI 138kV (GRDA-AECI) ftlo Cleveland-Tulsa North 345kV (GRDA-CSWS)

TMP151_23193 SW Missouri Oakland East Switch-Joplin Atlas Junction 161kV ftlo Asbury Plant-Purcell Southwest 161kV (EDE)

TAHH59MUSFTS* West Arkansas/East Oklahoma Tahlequah-Highway 59 161kV ftlo Muskogee-Fort Smith 345kV (GRDA-OKGE)

HALTUCHALSWI^ West Texas (Lubbock area) Hale County-Tuco 115kV ftlo Swisher County-Tuco 230kV (SPS)

TMP127_23359 Western Nebraska Scotsbluff-Victory Hill 115kV (NPPD) ftlo Stegall Xfmr 345kV/1 (WAUE)

TEMP72_22893 Eastern Kansas Wolf Creek 345kv/1 (WR) ftlo Wolf Creek-Waverly 345kV (WR-KCPL)

TMP322_23590 Eastern Oklahoma Stonewall Tap-Tupelo Tap 138kV (WFEC) ftlo Sunnyside-Terry 345kV (CSWS-OKGE)

* SPP market-to-market flowgate

^ also includes congestion from TMP228_22916, which became HALTUCHALSWI on July 2, 2018

Overall, the level of congestion in the SPP footprint has declined from 2017 to 2018. In 2017,

the most congested flowgate (Woodward-FPL Switch for the loss of Tatonga-Mathewson) had

a real-time shadow price of nearly $60/MWh, and was congested in almost 25 percent of

real-time intervals. But in 2018, the flowgate with the highest congestion (Vine Tap-North

Hays for the loss of Post Rock-Knoll) had an annual real-time shadow price of just $20/MWh,

and was congested just over eight percent of all real-time intervals. In fact, in 2017, three

flowgates were congested in over 15 percent of all real-time intervals, while in 2018 no

flowgate was congested in over nine percent of all real-time intervals.

0%

10%

20%

30%

40%

$0

$10

$20

$30

$40

Perc

ent

cong

este

d

Shad

ow

pri

ce ($

/MW

h)

Real-time average shadow price Day-ahead average shadow priceReal-time percent intervals congested Day-ahead percent intervals congested



Most of the highest congested corridors on the SPP system are significantly impacted by

inexpensive wind generation. Of the ten most congested flowgates, those affected the most

by wind generation are the west-to-east flows through the Hays, Kansas area, and west-to-

east flows in eastern Oklahoma. The southwest Missouri area is also impacted by wind and

external flows. Projects are planned throughout the SPP footprint that provide for more

transfer of wind generation from west to east and are listed in Figure 5—10.

5.1.4.1 Central Kansas constraints

The central Kansas area is also impacted by wind generation and several constraints have

appeared consistently since prior to the start of the SPP market. Figure 5—6 compares

congestion over the years for day-ahead and real-time in the central Kansas area since start of

the SPP market.

Figure 5—6 Central Kansas congestion83

The central Kansas area continued to see congestion in 2018. The Vine – Hays 115kV for the

loss of Post Rock – Knoll 230kV (VINHAYPOSKNO) constraint experienced congestion in 30

percent of all intervals in the day-ahead market and eight percent of all intervals in the real-

time market. The yearly average shadow price for the real-time market was $20/MWh for

83 Values combined for SHAHAYPOSKNO (South Hays – Hays 115kV) and VINHAYPOSKNO (Vine – Hays 115kV) constraints for 2017. South Hays – Hays 115kV terminated May 2017, and Vine – Hays 115kV activated September 2017.

0%

20%

40%

60%

$0

$10

$20

$30

2015 2016 2017 2018

Perc

ent

cong

este

d

Shad

ow

pri

ce ($

/MW

h)




2018 compared to $27/MWh in 2017. A second Post Rock – Knoll 230kV circuit was

energized in December 2018. This may provide relief to this area.

5.1.4.2 Southwest Missouri

The Neosho – Riverton 161kV constraint is a market-to-market flowgate that is impacted by

SPP and MISO wind, as well as flows from neighboring non-market areas.84 Congestion in

this area dates back to prior to the start of the Integrated Marketplace. However, continued

addition of wind in SPP and neighboring areas have contributed to the increased congestion.

Figure 5—7 compares congestion on the Neosho – Riverton 161kV constraint since 2015.

Figure 5—7 Southwest Missouri congestion

Congestion increased substantially on this constraint in 2017 when compared to previous

years continuing into 2018. This constraint has been a focus of seams discussions on the

amount of market-to-market payments, power swings or volatility, and transmission upgrades

that may benefit SPP, MISO, and other neighboring non-market entities. The market-to-

market process has settled over $27 million in payments for the Neosho-Riverton constraint

from MISO to SPP, and is discussed in Section 2.6.2.

84 Neighboring non-markets include; Tennessee Valley Authority, Associated Electric Cooperative Inc., and Southwestern Power Administration.

0%

10%

20%

30%

40%

$0

$10

$20

$30

$40

2015 2016 2017 2018

Perc

ent

cong

este

d

Shad

ow

pri

ce ($

/MW

h)




5.1.5 PLANNED TRANSMISSION PROJECTS

Figure 5—8 provides a list of projects that may alleviate congestion on the ten most

congested flowgates in the SPP system.

Figure 5—8 Top ten congested flowgates with projects Flowgate name Region Flowgate location Projects that may provide relief

VINHAYPOSKNO Western Kansas

Vine-Hays 115kV (WR) ftlo Post Rock-Knoll 230kV (WR)

Post Rock-Knoll 230kV, circuit 2 (energized December 2018, 2017 ITP10)

NEORIVNEOBLC * SW Missouri/ SE Kansas

Neosho-Riverton 161kv (EDE-WR) ftlo Neosho-Blackberry 345kV (WR-AECI)

Neosho – Riverton 161kV Terminal Upgrades (June 2018, 2017 ITP10) [Identified as an economic need in the 2019 ITP Assessment to be completed October 2019]

TMP109_22593* Eastern Oklahoma

Stonewall Tap-Tupelo Tap 138kV (WFEC) ftlo Seminole-Pittsburg 345kV (CSWS-OGE)

Stonewall Tap-Tupelo Tap 138kV Terminal Upgrades (July 2021, 2017 ITP10)

TMP270_23432 Eastern Oklahoma

Cleveland-Cleveland AECI 138kV (GRDA-AECI) ftlo Cleveland-Tulsa North 345kV (GRDA-CSWS)

None identified at this time. [Identified as an economic need in the 2019 ITP Assessment to be completed October 2019]

TMP151_23193 SW Missouri/ SE Kansas

Oakland East Switch-Joplin Atlas Junction 161kV ftlo Asbury Plant-Purcell Southwest 161kV (EDE)

None identified at this time. [Related to economic needs in the 2019 ITP Assessment to be completed October 2019]

TAHH59MUSFTS* West Arkansas/ East Oklahoma

Tahlequah-Highway 59 161kV ftlo Muskogee-Fort Smith 345kV (GRDA-OKGE)

None identified at this time

HALTUCHALSWI West Texas (Lubbock)

Hale County-Tuco 115kV ftlo Swisher County-Tuco 230kV (SPS)

None identified at this time. [Identified as an economic need in the 2019 ITP Assessment to be completed October 2019]

TMP127_23359 Western Nebraska

Scotsbluff-Victory Hill 115kV (NPPD) ftlo Stegall Xfmr 345kV/1 (WAUE)

None identified at this time

TEMP72_22893 Eastern Kansas Wolf Creek 345kv/1 (WR) ftlo Wolf Creek-Waverly 345kV (WR-KCPL)

None identified at this time. [Related to an economic and reliability need in the 2019 ITP Assessment to be completed October 2019]

TMP322_23590 Eastern Oklahoma

Stonewall Tap-Tupelo Tap 138kV (WFEC) ftlo Sunnyside-Terry 345kV (CSWS-OKGE)

Stonewall Tap-Tupelo Tap 138kV Terminal Upgrades (July 2021, 2017 ITP10)

* SPP market-to-market flowgate

5.1.6 GEOGRAPHY AND MARGINAL LOSSES

Variable transmission line losses decrease with increased line voltage or decreased line

length for the same amount of power moved. In the SPP footprint, much of the low-cost

generation resides at a distance from the load and with limited high-voltage interconnection.

The average variable losses on the SPP system for 2018 were 2.8 percent in the day-ahead

market. This is up slightly from 2.7 percent in 2017 and 2.3 percent in 2016. The marginal



loss component of the price captures the change in the total system cost of losses with an

additional increment of load at a particular location relative to the reference bus.

Figure 5—9 maps the annual average day-ahead market marginal loss components.

Figure 5—9 Marginal loss component map, day-ahead

The average day-ahead marginal loss component ranges from about −$4.45/MWh at the

Laramie River Station in eastern Wyoming, to −$3.32/MWh near North Platte, Nebraska, to

−$0.10/MWh in the Kansas City area, to $0.90/MWh in the Hobbs, New Mexico area, and up

to $2.67/MWh in the corners of Texas/Oklahoma/Arkansas. Negative values reduce prices

through the marginal loss component relative to the marginal energy cost. Positive values

increase prices as generation from these locations are more beneficial from a marginal loss

perspective. The $7.12/MWh spread between geographic prices is larger than the

$5.25/MWh spread in 2017.

day-ahead market marginal loss component



5.1.7 FREQUENTLY CONSTRAINED AREAS AND LOCAL MARKET POWER

Congestion in the market creates local areas where only a limited number of suppliers can

provide the energy to serve local load without overloading a constrained transmission

element. Under these circumstances, the pivotal suppliers have local market power and the

ability to raise prices above competitive levels thereby extracting higher than normal profits

from the market. SPP’s tariff provides provisions for mitigating the impact of local market

power on prices, and the effectiveness of market power mitigation is described in Section

7.2.2. Local market power can be either transitory, as is frequently the case with an outage,

or persistent, when a particular load pocket is frequently import-constrained.

Because the SPP tariff calls for more stringent market power mitigation for frequently

constrained areas, the MMU analyzes market data at least annually to assess the

appropriateness of the frequently constrained area designations. The 2017 frequently

constrained area study85 recommended the removal of the Woodward area and reduction of

the Texas Panhandle area, and these became effective on April 1, 2018.

The following year’s study (2018) recommended the removal of the reduced Texas

Panhandle (Lubbock) area and the addition of two new areas; southwest Missouri and central

Kansas. These recommendations began through the stakeholder process with a target of

being approved by the SPP board of directors in January 2019 followed by a tariff filing.

In the meantime, stakeholders improved this process by approving revision request 304

(Accelerate Frequently Constrained Area Process.) Under this improved process, the MMU is

to notify market participants of identified changes upon completion of the analysis and allow

a 14 day window for questions or concerns from stakeholders regarding the proposed

changes. After addressing any needed concerns or comment during this period, the

proposed changes will become effective. This improved process, approved by SPP

stakeholders, was accepted by FERC on December 22, 2018.86 After discussion with

stakeholders and implementation of the new process, these 2018 study changes became

effective on February 22, 2019.

85 The Frequently Constrained Area reports can be found on the SPP website: https://www.spp.org/spp-documents-filings/?id=25496. 86 http://elibrary.FERC.gov/idmws/file_list.asp?accession_num=20181220-3042


http://elibrary.ferc.gov/idmws/file_list.asp?accession_num=20181220-3042



5.1.8 MARKET CONGESTION MANAGEMENT

In optimizing the flow of energy to serve the load at the least cost, the SPP market makes

extensive use of the available transmission up to constraint limits. When constraints reach

their limits, they are considered binding. The market occasionally allows transmission lines to

exceed their rating if the price to correct the overload becomes too high. This is considered

a breached constraint. Figure 5—10 highlights day-ahead market binding, breached, and

uncongested intervals.

Figure 5—10 Breached and binding intervals, day-ahead market

The figure shows that uncongested intervals and breached intervals are rare in day-ahead.

Historically in the Integrated Marketplace, less than one percent of day-ahead market

intervals incur a breached condition compared to nearly 25 percent for the real-time

market.87

In the more dynamic environment of the real-time market, uncongested intervals and

breached intervals occur much more frequently than in the day-ahead market. Real-time

congestion is shown in Figure 5—11.

87 SPP uses hourly intervals in the day-ahead Market and five-minute intervals in the real-time market for scheduling, dispatch, and settlement purposes.

0%

20%

40%

60%

80%

100%


Uncongested intervals Intervals with binding only Intervals with a breach



Figure 5—11 Breached and binding intervals, real-time

As shown above, uncongested intervals have been steadily increasing on an annual basis,

with 18 percent of intervals with no congestion in 2018, up from nine percent in 2016.

Conversely, real-time intervals with a breached constraint in steadily decreasing, with 27

percent of intervals with a breach in 2018, compared to 37 percent in 2016. This increase in

uncongested intervals, and accompanying decrease in breached intervals, indicates less

overall congestion in the SPP market footprint. This can primarily be attributed to the

transmission expansion that has occurred over the past few years.

Market-to-market coordination with MISO, as discussed in Section 2.6.2, was implemented in

March 2015, and the integration of the Integrated System occurred in October of that year.

Both of these events increased the number of constraint breaches. A market-to-market

breach of a MISO constraint could be an indicator that MISO has more efficient generation

than SPP to alleviate congestion on that constraint. Of the 27 percent of the real-time

intervals with a breached constraint, over half of these had a breached market-to-market

constraint.

5.1.9 CONGESTION PAYMENTS AND UPLIFTS

Market participants in the energy market incur congestion costs and receive congestion

payments based on their marginal impact on total market congestion cost through the

marginal congestion component of price. Most SPP market participants owning physical

assets are vertically integrated, so their net congestion cost depends on two things. The first

0%

20%

40%

60%

80%

100%


Uncongested intervals Intervals with binding only Intervals with a breach



is whether they are a net buyer or seller of energy. The second is the relative marginal cost

component at their generation and load. For financial market participants, congestion costs

reflect the impact of virtual positions on a binding or breached constraint in the day-ahead

and real-time markets.

Figure 5—12 shows the annual day-ahead and real-time market congestion payments for

load-serving market participants during 2018.

Figure 5—12 Annual congestion payment by load-serving entity

Most load-serving entities face congestion costs, depicted as negative payments (charges) in

the graph, because they are part of vertically-integrated entities with higher marginal

congestion components at load than at resources. Day-ahead congestion payments by

ranked load-serving entities ranged from about $73 million in charges to almost $4 million in

payments.88

Market participants also receive payments and incur costs for real-time market congestion,

which are charged and paid based on deviations between day-ahead and real-time market

positions. At an aggregate level, absent the additional revenue neutrality uplift costs, 97

percent of the SPP load-serving entities’ net congestion costs were hedged with day-ahead

88 Day-ahead congestion collections funds transmission congestion rights. These rights are described in greater detail in Section 5.2.

-$100

-$80

-$60

-$40

-$20

$0

$20

Net

co

nges

tion

pay

men

t ($

mill

ions

)

Day-ahead congestion payment Real-time congestion payment Net payment



prices. Figure 5—13 provides the aggregate congestion costs and hedging totals for load-

serving entities, non-load-serving entities, and financial only entities.

Figure 5—13 Total congestion payments

(in $ millions) Load-serving entities

Non-load-serving and financial only entities

2016 2017 2018 2016 2017 2018

DA congestion $ 259.6 $ 364.6 $ 390.8 $ 81.2 $ 219.0 $ 169.7

RT congestion 14.0 40.6 (10.0) (76.5) (157.2) (97.8)

Net congestion 273.7 405.3 380.9 4.6 61.8 71.9 RT congestion uplift (39.1) (68.7) (74.8) (4.3) (10.2) (11.9)

Real-time market congestion ranged from over $6 million in payments to over $12 million in

costs for load-serving entities. Real-time market congestion ranged from $36 million in

payments to $10 million in costs for non-load-serving entities. Many of the non-load-serving

entities incurring costs represent wind farms, which may sell at negative prices or buy back

day-ahead market positions. The real-time market congestion payments result in a net

benefit of $10 million for load-serving entities. Total real-time market congestion payments

for non-load-serving and financial only entities also resulted in a net benefit and amounted to

$98 million.

Unlike day-ahead congestion, which funds transmission congestion rights, real-time market

congestion costs are allocated to market participants through revenue neutrality uplift (RNU)

charges. In 2017, SPP allocated about 86 percent of revenue neutrality uplift charges to load-

serving entities, resulting in an additional $75 million in congestion-related charges for load-

serving entities.89

89 Real-time congestion uplift is not allocated in the same proportion in which it is collected.



5.1.10 DISTRIBUTION OF MARGINAL LOSS REVENUES (OVER-COLLECTED LOSSES)

Both the congestion and loss components of prices create excess revenues for SPP that must

be distributed to market participants in an economically efficient manner. In the case of

marginal loss revenues, this requires that the distribution does not alter market incentives.

This was not the case during the first year of SPP’s market, and SPP took steps that largely

corrected the incentive issues. SPP proposed changes to the method for distributing over-

collected losses in FERC Docket No. ER15-763.90 The Commission accepted these changes,

which went into effect in April 2015.

The enhanced design consolidates the distributions of day-ahead and over-collected loss

rebates into one distribution. Under this design, both day-ahead and real-time over-

collected loss rebates are distributed on just real-time withdrawing megawatts. This includes

loads, substation power, exports, wheel-throughs, pseudo-ties, and bilateral settlement

schedules (BSS). The only exception is that both day-ahead and real-time bilateral settlement

schedules are entitled to the rebate. In addition to consolidating the distributions to only

real-time withdrawing megawatts, changes were made to loss pool allocations. Under the

old method, virtual transactions drove up the SPP loss pool allocations of over-collected

losses rebates, even though virtual activity was not eligible for rebates. This caused real-time

exporters to get a large percentage of the over-collected losses rebates which, during that

time, were typically a charge. Virtual transactions are no longer considered in the loss pool

distributions. These design enhancements better allocate the over-collected losses to the

transactions that contributed to the over-collection while removing some of the adverse

incentives present under the former design.

Over-collected losses for the past three years are shown in Figure 5—14.

90 https://www.ferc.gov/CalendarFiles/20150331172533-ER15-763-000.pdf

https://www.ferc.gov/CalendarFiles/20150331172533-ER15-763-000.pdf



Figure 5—14 Over-collected losses, real-time

A total of $145 million was paid out in over-collected losses rebates during 2018, with $133

million (89 percent) going to load. This is up from the $120 million in over-collected losses

rebates paid out in 2017 and $90 million paid in 2016.

The use of bilateral settlement schedules changes the distribution of over-collected losses.

The bilateral settlement schedules enable market participants to transfer energy from one

entity to another at a particular settlement location. It creates a financial withdrawal at the

settlement location for the seller and a financial injection at the settlement location for the

buyer. As long as the bilateral settlement schedules do not change the net withdrawal at the

location, the charges and credits for losses simply change hands between the entities owning

the bilateral settlement schedules. The MMU identified a design issue with the distribution of

loss rebates to bilateral settlement schedules. The issue was corrected in May 2018 with the

implementation of revision request 200.91

Where the bilateral settlement schedules create a net withdrawal that would not otherwise

exist, it creates credits and charges that would not otherwise exist. For example, if a bilateral

settlement schedule amount at a resource settlement location exceeds the cleared output of

the resource, it creates a net withdrawal, and the generation owner receives a loss

91 RR 200 (Design change for bilateral settlement schedule and over-collected losses distribution)

-$30

$0

$30

$60

$90

$120

$150

2016 2017 2018

Ove

r-co

llect

ed lo

ss p

aym

ents

($ m

illio

ns)

Load Resource Interface Hub Net payment



distribution credit for the excess megawatts of the bilateral settlement schedule. The same

occurs with a bilateral settlement schedule at hubs, where no energy is withdrawn, other than

a bilateral settlement schedule.

Prior to revision request 200, bilateral settlement schedules allowed participants to receive

loss rebates for withdrawing megawatts where losses were not paid into the market. Also,

the fact that bilateral settlement schedules can be entered subsequent to the initial

settlement statements provided information to participants on when to transact the bilateral

settlement schedules based on available loss rebates. This revision request addressed these

concerns.

A recommendation in the 2014 Annual State of the Market report was to remove bilateral

settlement schedule transactions from the over-collected losses distribution calculation.

Revision request 200 still allows for loss rebates to be paid to bilateral settlement schedules.

However, those rebates are now capped at the underlying energy flows of the bilateral

settlement schedules. The MMU believes that the adverse incentives given to bilateral

settlement schedules to transact in amounts that vary from the underlying flows of the

transaction have been removed.

5.2 CONGESTION HEDGING MARKET

In the Integrated Marketplace, the locational marginal prices assessed to load are generally

higher than the locational marginal prices assessed to generators. The largest portion of this

price difference is almost always attributed to congestion. This is an expected outcome and

central to the design of nodal electricity markets. The difference between what generators

are paid and what loads pay is often referred to as congestion rent. SPP remains revenue

neutral in all Integrated Marketplace transactions and therefore must allocate the congestion

rent back to the market participants. The congestion hedging market is the mechanism used

to allocate congestion overcollections.92

5.2.1 MARKET DESIGN

Market participants participate in the congestion hedging market by obtaining auction

revenue rights and/or transmission congestion rights. Auction revenue rights begin as

92 With respect to day-ahead congestion rent only.



entitlements associated with long-term, firm transmission service reservations. These

transmission service reservations are a revenue source for transmission owners and an

expense for transmission customers. More specifically, transmission owners receive revenues

from transmission customers for building and maintaining the transmission lines, and

transmission customers pay the transmission owners for the use of the lines by way of the

charges associated with transmission service reservations.93

Auction revenue rights link the transmission service, which provides physical rights, to the

Integrated Marketplace by converting these to financial rights. SPP verifies transmission

service entitlements, which become candidate auction revenue rights. To obtain auction

revenue rights, market participants nominate candidate auction revenue rights. If the

nominations pass the allocation’s simultaneous feasibility test, the candidate auction revenue

rights become auction revenue rights. The simultaneous feasibility test ensures that the

market’s aggregate nomination of auction revenue rights does not violate thermal constraint

limits under a single contingency.94 The test incorporates information from the network

model, which aids SPP in aligning the supply of auction revenue rights with the capacity of

the underlying transmission system. The simultaneous feasibility test aims to ensure the

revenues generated from the congestion right auction will sufficiently fund the quantity of

auction revenue rights nominated. If a candidate auction revenue right nomination fails the

simultaneous feasibility test, this reduces the quantity of auction revenue rights successfully

converted from candidate rights.95

Once market participants have successfully nominated their candidates into auction revenue

rights, they must choose to either hold their auction revenue right or convert it into a

transmission congestion right through a process known as self-conversion.96 If a market

participant holds their auction revenue right, they will receive, or pay, a stream of unchanged

cash flows over the life of the product. The size and direction of the cash flow depends on

the market’s collective assessment of the congestion rent along the auction revenue right

path as assessed by prices during the transmission congestion right auction. If a market

participant believes that the auction prices will undervalue the congestion rent associated

93 These charges are assessed through transmission settlements and include various tariff schedules. 94 SPP tariff, 5.3.3, Simultaneous Feasibility 95 2018 SPP MMU Spring Quarterly State of the Market Report, https://www.spp.org/documents/58275/spp_mmu_quarterly_spring_2018.pdf, page 50. 96 SPP tariff 5.4.1 (2)

https://www.spp.org/documents/58275/spp_mmu_quarterly_spring_2018.pdf



with their auction revenue right, the market participant will likely self-convert. When a market

participant self-converts, their auction revenue right becomes a transmission congestion

right, which means their cash flow is subject to the fluctuations in day-ahead market

congestion rent.

Financial-only97 market participants participate alongside traditional utilities in the

transmission congestion right auctions to provide additional liquidity and price discovery. All

participants compete for the residual network capacity, on price. The auction software

attempts to maximize auction revenue without violating the simultaneous feasibility test. This

test helps align the supply of transmission congestion rights with the residual capacity of the

underlying transmission system. If a transmission congestion right bid fails the simultaneous

feasibility test, the quantity of transmission congestion rights successfully obtained will be

reduced to the point where the bid no longer violates the test. Once a market participant

obtains a transmission congestion right, they can hold it through to settlement, offer it for sale

in a subsequent auction, or transact on the bulletin board outside of the auction cycle. The

overwhelming majority of positions are held through to settlement.

5.2.2 MARKET TRANSPARENCY 5.2.2.1 Hedging effectiveness by classification

The transmission congestion right and auction revenue right net payments paid to entities in

the SPP market are shown in Figure 5—15.

97 Finacial-only market participants do not generate or serve load in the SPP footprint.



Figure 5—15 Total congestion and hedges

(in $ millions) Load-serving entities Non-load-serving and

financial only entities 2016 2017 2018 2016 2017 2018

DA congestion $ 259.6 $ 364.6 $ 390.8 $ 81.2 $ 219.0 $ 169.7

RT congestion 14.0 40.6 (10.0) (76.5) (157.2) (97.8)

Net congestion 273.7 405.3 380.9 4.6 61.8 71.9

TCR charges 51.0 122.5 214.8 63.5 138.0 174.8

TCR payments (212.5) (308.8) (331.5) (158.7) (314.3) (262.8)

TCR uplift 21.1 24.3 26.1 18.1 31.8 32.2

TCR surplus * (4.2) (4.5) (5.4) (4.1) (5.9) (7.5)

ARR payments (74.8) (147.2) (249.1) (6.5) (11.7) (17.8)

ARR surplus (29.2) (94.1) (113.2) (1.9) (7.5) (9.6)

Net TCR/ARR (248.6) (407.9) (458.2) (89.6) (169.5) (90.7)

* remaining at year end

Payments to load-serving entities of $458 million exceeded their day-ahead congestion costs

of $391 million in 2018. Additionally, real-time congestion costs aided load-serving entities

by $10 million, thereby reducing total congestion cost to $381 million. This shows that

overall, load-serving entities did a good job at hedging congestion through the transmission

congestion right market. In aggregate for 2018, non-load-serving and financial only entities

collected transmission congestion right and auction revenue right net revenues of $91

million, which exceeded their day-ahead and real-time market congestions costs of $72

million. Overall, day-ahead congestion cost decreased four percent year-over-year, but

increased more than 40 percent over 2016.

5.2.2.2 Bidding behaviors

The SPP working groups continued dialogue over the past year with respect to obtaining

auction revenue rights, and by extension self-converted transmission congestion rights. The

market monitor reported on this topic in the 2018 spring quarterly state of the market

report,98 and is actively engaged in efforts to improve this issue. As noted above, in

aggregate, load-serving entities received more revenue from their congestion hedges than

98 2018 SPP MMU Spring Quarterly State of the Market Report, https://www.spp.org/documents/58275/spp_mmu_quarterly_spring_2018.pdf, page 50.

https://www.spp.org/documents/58275/spp_mmu_quarterly_spring_2018.pdf



they paid in day-ahead and real-time congestion cost. However, on the individual participant

level, some load-serving entities were under-hedged while others were over-hedged.

Figure 5—16 shows, by load-serving entity, the day-ahead congestion exposure along with

the value of auction revenue rights and transmission congestion rights as well as the net

overall position.

Figure 5—16 Net congestion revenue by market participant

The figure highlights that the majority of participants received positive net revenues, while a

handful of participants held portfolios that did not cover their day-ahead congestion costs.

For instance, the bottom five participants paid over $26 million more in congestion costs than

was offset by their auction revenue right and transmission congestion right positions. This is

substantially less than the $125 million for the bottom five participants in the 2017 calendar

year.

The range of participant outcomes is influenced by three main factors: hedging need,

individual participant bidding behavior, and the market’s collective bidding behavior.

With respect to hedging need, each participant experiences varying levels of congestion

exposure mostly related to geographic location and type of physical interconnection. The

various levels of congestion exposure lead to different hedging needs among market

participants.

-$100

-$75

-$50

-$25

$0

$25

$50

$75

$100

Mill

ions

Day-ahead congestion rent ARR + TCR settlement Net

Net gain

Net loss



The bidding behavior of the individual participant affects the auction revenue rights they

receive through the allocation. Participants can, and do, employ numerous strategies with

varying degrees of success.

The bidding behavior of the other market participants, as a whole, affects the ability of each

and every other market participant to obtain hedges through the auction revenue right

allocation. More specifically, if a transaction is physically feasible with respect to transmission

service, and by extension the day-ahead market, it does not necessarily mean the transaction

will also be feasible in the auction revenue right allocation. The issue arises because the

transmission system’s capacity, represented by transmission service requests,99 includes both

prevailing flow and counter-flow transactions. However, participants often choose not to

nominate their counter-flow candidate auction revenue rights in the allocation process, in

part, because these positions tend to carry negative cash flows. These incentives motivate

individual participants to abstain from counter-flow positions.100 By not nominating all

candidate auction revenue rights, the capacity in the allocation will not match the capacity of

the physical system. Because the basis of the auction revenue right is transmission service, if

the two capacities do not align, a participant’s auction revenue right may not perfectly hedge

their transmission service and their related day-ahead market activity.

Differences between outages modeled in the auction processes and day-ahead market can

also affect market participants’ ability to obtain hedges. Details on outage modeling are

discussed below.

5.2.2.3 Transmission outage modeling

When outages appear in the day-ahead market that were not in the transmission congestion

rights auction, they reduce system capacity and can cause underfunding. Figure 5—17 shows

transmission outages by reported lead time.

99 Long-term, firm, transmission service requests 100 SPP MWG Meeting Materials, 1/22/2019, https://www.spp.org/spp-documents-filings/?id=18437, 11a.SPP MMU_ARR Observation.pdf




Figure 5—17 Transmission outages by reporting lead time

SPP models only transmission outages that were reported at least 45 days prior to the first of

the month in the transmission congestion rights auction.101 However, SPP only requires

transmission owners to submit planned outages 14 days in advance.102 The above figure

shows that the majority of outages are not considered in the transmission congestion rights

markets solely due to the submission lead time. Roughly 86 percent of outages are ruled out

of the transmission congestion right model by this phase alone. This is a similar level as prior

years.

Figure 5—18 shows the duration in days for the different types of outages.

101 SPP Integrated Marketplace Protocols, Section 6.6 102 SPP Operating Criteria Appendix OP-2

0

500

1,000

1,500

2,000

2,500

Less than 7 7 to 15 15 to 30 30 to 45 45 to 60 60 to 75 More than75

Num

ber

of o

utag

es

Reporting lead time (days)

Planned Opportunity Forced Discretionary Operational Emergency Urgent

included in auction

not included in auction



Figure 5—18 Transmission outages by duration

Outages shorter than five days are excluded from auction revenue right/transmission

congestion right processes. This means the vast majority of outages (84 percent) are

excluded from the transmission congestion rights models because they are less than five

days.

Figure 5—19 shows outages by real-time, day-ahead, and transmission congestion right

markets.

Figure 5—19 Transmission outages by market

0

1,000

2,000

3,000

4,000

Less than 1 1 to 3 3 to 5 5 to 7 More than 7

Num

ber

of O

utag

es

Duration (days)


included in auction

not included in auction

0

1,000

2,000

3,000

4,000


Num

ber

of o

utag

es

Real-time Day-ahead Transmission congestion right



While the number of outages in the day-ahead and real-time are similar, the outages

represented in the transmission congestion rights market are only a fraction of the total

number of outages. The transmission congestion market only includes outages that are

longer than five days, and are submitted at least 45 days in advance of the first of the month.

This represented only about five percent of the total outages in the day-ahead market. These

differences in outages can create underfunding of transmission congestion rights.

Ideally, outages in the transmission congestion rights markets would be perfectly aligned

with the day-ahead market. However, the MMU understands the challenges associated with

accounting for outages in the transmission congestion rights market and recognizes that

there can never be an exact match among the markets. Even so, improving how outages are

handled and accounted for in the auction processes could help to improve underfunding.

We encourage stakeholders to consider improving how outages are accounted for in the

transmission congestion right auction process to improve the congestion hedging market

results.

5.2.3 FUNDING

Funding for transmission congestion rights increased slightly in 2018. Conversely, funding

percentage for auction revenue rights decreased in 2018. While the auction revenue right

funding percentage decreased, the auction revenue right surplus increased. As mentioned

in previous reports,103 the overfunding of auction revenue rights could be cause for concern.

The market monitor continues to encourage SPP to review and address the reasons for this

overfunding.

In the 2014 and 2015 Annual State of the Market reports, the MMU discussed a

recommendation and corresponding revision request104 that was intended to reduce over-

allocation of auction revenue rights and the resultant over-selling of transmission congestion

rights by reducing the transmission capability available in the annual auction revenue right

allocation. This revision request made its way through the SPP stakeholder process in 2015

and 2016, and FERC gave final approval in time for the 2016 annual auction revenue

right/transmission congestion right processes. The “TCR year” runs from June 1 through May

31 of each year, so it is not in-sync with the annual report year. Looking at calendar year data,

103 2017 SPP Annual State of the Market Report 104 Revision request 91



2016 included the revisions for half the year, 2017 and 2018 included the revisions for the

entire year.

Figure 5—20 Transmission congestion right funding levels, annual

The 2016 calendar year, which included the revisions for about half of the year, yielded 92

percent transmission congestion right funding. The 2017 calendar year, which included the

revisions for the entire year, produced 94 percent transmission congestion right funding.

The 2018 calendar year, which also included the revisions for the entire year, also produced

94 percent transmission congestion right funding. While the funding percentage very

marginally improved year-over-year, the shortfall improved by more than $6 million during

2018.

Monthly transmission congestion right funding levels and revenue are shown in Figure 5—21.

0%

20%

40%

60%

80%

100%

-$200

$0

$200

$400

$600

$800

2016 2017 2018

Fund

ing

per

cent

Mill

ions

Day-ahead revenue Transmission congestion right fundingSurplus/shortfall Funding percent



Figure 5—21 Transmission congestion right funding levels, monthly

The monthly averages appear fairly consistent, with the majority of the months hovering near

the 90 to 100 percent target range. The lowest funding percentage was 86 percent in

February and the highest was at 112 percent in July.105

Daily observations of transmission congestion right funding for the past three years are

shown in Figure 5—22.

105 Target range is specified in the Protocols section 5.3.3. “In the event the cumulative funding is at or below 90 percent or above 100 percent, MWG may approve an additional adjustment…”

0%

20%

40%

60%

80%

100%

120%

-$20

$0

$20

$40

$60

$80

$100


Fund

ing

per

cent

Mill

ions

Day-ahead revenue Transmission congestion right fundingSurplus/shortfall Funding percent



Figure 5—22 Transmission congestion right funding, daily

Most daily observations of transmission congestion right funding percent fall between 80

percent and 120 percent, as seen in Figure 5—22. While variation in funding can be expected

as a result of factors including transmission outages and derates, the fact that the majority of

funding falls in this range indicates that the overall process is generally effective. It is

important to note that there was an increase in funding events in excess of 155 percent

increased from one observation in both 2016 and 2017 to seven observations in 2018. This

data suggests that while the annual funding percentage is relatively stable year-over-year, the

upside volatility in the daily observations increased in 2018.

Another way to look at funding is through the funding level and not just percentage. Most

daily observations of transmission congestion right funding fell between plus and minus

$250,000 in 2018. However, the frequency of events outside this range continues to

increase. This data also suggests that while the annual funding shortfall has been relatively

consistent year-over-year, the volatility in the daily surpluses and shortfalls has increased.

The market monitor will continue to evaluate these trends going forward.

Figure 5—23 shows that the auction revenue right funding percentage has increased against

2016 levels in both 2017 and 2018.

0

10

20

30

40

50

60

Day

s

2016 2017 2018



Figure 5—23 Auction revenue right funding levels, annual

In 2016, auction revenue rights were 138 percent funded. In 2017, auction revenue rights

were 160 percent funded. In 2018, auction revenue rights were 145 percent funded.

Auction revenue right surpluses increased from $27 million in 2016 to $101 million in 2017

and increased again in 2018 to almost $123 million. While the decrease in funding

percentage is encouraging, the substantial and growing surplus presents a potential concern,

as it could be an indicator of inefficiency. The market monitor urges SPP, along with the

stakeholders, to determine the root cause of the overfunding, and analyze the surplus

distribution methodology to ensure it is equitably allocated.

Figure 5—24 shows the 2018 monthly funding levels and revenues for auction revenue rights.

100%

120%

140%

160%

180%

200%

$0

$100

$200

$300

$400

$500

2016 2017 2018

Fund

ing

per

cent

Mill

ions

Transmission congestion right revenue Auction revenue right funding Surplus/shortfall Funding percent



Figure 5—24 Auction revenue right funding levels, monthly

The shift in auction revenue rights funding beginning in June reflects the change in the TCR

year. The figure also shows the auction revenue right funding was elevated during the last

five months of the 2017 TCR year when compared to the first seven months of the 2018 TCR

year. This general trend was also observed in 2017.

5.2.4 INTRA-AUCTION SALES AND BULLETIN BOARD TRANSACTIONS

Intra-auction sales refer to the sale of a previously acquired transmission congestion right

position in a subsequent auction. Bulletin board transactions refer to trades where a market

participant buys or sells a position outside of the auction cycle through the SPP bulletin

board. Overall, both inter-auction sales and bulletin board transactions remain low.

Figure 5—25 shows the transaction volume by type as a percentage of all transmission

congestion right purchase volume.

0%

50%

100%

150%

200%

250%

$-

$10

$20

$30

$40

$50


Fund

ing

per

cent

Mill

ions

Transmission congestion right revenue Auction revenue right funding Surplus/shortfall Funding percent



Figure 5—25 Intra-auction sales and bulletin board transactions

The increase in bulletin board transactions from 2016 to 2017 was due, in large part, to one

market participant selling all of their positions to another market participant. The acquirer

also subsequently sold some of their newly acquired positions to third party market

participants in 2017. However this was unusual, and zero bulletin board transactions cleared

in 2018. Additionally, intra-auction sale volume remained relatively stable around six percent

of the total transmission congestion right volume.

Outside their relationship to the auction cycle, these transactions also differ from each other

in another material way. Bulletin board transactions are similar to the secondary equity

market, where a share of stock is offered for sale and that same share is later purchased by

another market participant. As such, the bulletin board transactions do not affect total

supply; they only affect ownership of the existing supply.

However, intra-auction sales can affect supply in addition to ownership. When market

participants offer their prevailing flow positions for sale intra-auction, the capacity of those

positions once used is available to the market again. But, the newly available system capacity

can be taken up by any path, not just the path sold intra-auction. Furthermore, to sell

counter-flow positions intra-auction, unclaimed prevailing flow capacity must exist for the

transaction to clear.106 This is because counter-flow intra-auction sales reduce the total

106 The additional capacity could also be provided by another counter-flow intra-auction bid.

0%

1%

2%

3%

4%

5%

6%

Intra-auction prevailingflow sales

Intra-auction counterflow sales

Bulletin boardprevailing flow

transactions

Bulletin board counterflow transactions

Perc

enta

ge

of a

ll tr

ansm

issi

on

cong

estio

n ri

ght

tran

sact

ions

2016 2017 2018



capacity available. The counter-flow now offered for sale, previously facilitated other

prevailing flow positions. If the counter-flow sale were to clear without considering supply,

the prevailing flow once facilitated by this counter-flow would no longer be feasible. So in

order for these existing prevailing flow transactions to remain feasible, additional prevailing

flow capacity must exist. Practically, the additional prevailing flow capacity required plays the

same role once played by the counter-flow being offered for sale. These circumstances likely

also incent market participant abstentions from counter-flow. Generally, if a market

participant holds a counter-flow position, it could be very difficult to sell the position intra-

auction, which is the main source of intra-marketplace liquidity.107

5.2.5 ISSUES, PROGRESS, AND NEXT STEPS

While the congestion hedging market is workably effective overall, the market monitor

highlights the following five areas where continued progress could bring about improved

market outcomes, risk reduction, and efficiency gains.

1. Obtaining auction revenue rights

As previously mentioned market participants experience varying levels of success in

obtaining auction revenue rights and by extension self-converted transmission

congestion rights. SPP and its stakeholders have presented various proposals

regarding the auction revenue rights allocation. The various working groups have yet

to reach consensus, and await the Holistic Integrated Tariff Team’s recommendations,

which are expected in April 2019. The market monitor encourages the continued

development of solutions that have the potential to reduce the magnitude of this

issue.

2. Credit policy

The SPP credit policy’s financial security requirements, associated with congestion

hedges, may not fully reflect the risks inherent in these products. The Credit Practices

Working Group is reviewing SPP’s credit policy and working toward a solution, which

more fully addresses the various risks associated with these products. Given the

gravity and magnitude of the GreenHat Energy default in PJM, the market monitor

107 Intra-market refers to inside the SPP Integrated Marketplace.



encourages the Credit Practices Working Group to make this particular issue its

highest priority going forward.

3. Allocation of the auction revenue right surplus

The auction revenue right surplus continues to increase and exceeded $120 million

dollars in 2018. SPP should determine the root cause of the surplus and further

determine if the current allocation methodology could be enhanced. The market

monitor views the enhancement of the allocation methodology to be a meaningful

potential source of efficiency gains.

4. Secondary intra-market liquidity

Bulletin board transaction volume ceased in 2018 and intra-auction sales, while stable,

continue to represent modest transaction volume. The limited liquidity associated

with these products is not unique to the SPP markets; however, improved liquidity

would likely prove beneficial for all market participants and enhance efficient auction

price formation. The market monitor encourages SPP to adopt policies and

procedures that deepen liquidity by incenting market participants to transact in SPP’s

secondary market.

5. Modeling inconsistencies

Modeling inconsistencies and outage discrepancies between the congestion hedging

and day-ahead models remain significant. The process and rules surrounding the

modeling of congestion hedging outages should be reviewed, and during the review

SPP and its stakeholders should determine if the current practice is appropriate. The

market monitor recognizes the significant challenges associated with model

convergence and encourages SPP’s continued focus in this area.



Southwest Power Pool, Inc. Planning process Market Monitoring Unit


6 PLANNING PROCESS

The scope of market monitoring work covers all aspects of the SPP market including activities

that could impact short and long-term operation of the market. The planning process and its

outcomes, in particular, affect congestion patterns, operational effectiveness, and costs, as

well as reliability.

The MMU’s capability to possess and analyze comprehensive market information positions it

to be able to provide valuable input in the planning process. For this reason, starting in 2018

the MMU, in its advisory capacity, has been involved in SPP’s planning process, primarily with

meetings and the discussions of the Economic Studies Working Group (ESWG).108 The MMU

believes its feedback in the planning process on relevant topics will improve planning

outcomes, and benefit the market as a whole.

Key highlights from this chapter include:

• Modified the Future 1 and Future 2 assumptions of the 2020 ITP—carried forward from

the 2019 ITP—on coal and gas generators’ retirement age from 60 years to 56 years for

coal and 50 years for gas.

• The proposed battery accreditation was set to 100 percent as opposed to 48 percent

following a mutual agreement with the SPP staff.

• In addition, the MMU supported the SPP staff’s recommendation for lowering the

modeling assumption of variable operations and maintenance cost (VOM) for

renewable generation in the ITP manual from $8/MWh for wind and solar units to

$0/MWh, consistent with generally observed VOM levels in mitigated offers. The

proposal was approved by the SPP stakeholder process.

• The MMU’s peak available capacity metric was 35 percent in 2018, up slightly from 33

percent in 2017.

108 The ESWG advises and assists SPP staff, various working groups, and task forces in the development and evaluation principles for economic studies. The group provides technical support for the development and application of economic studies. The group also reviews the economic planning processes for adherence to sound economic metrics methods and provides recommendations for improvement of the economic evaluations (see https://www.spp.org/organizational-groups/board-of-directorsmembers-committee/markets-and-operations-policy-committee/economic-studies-working-group/).

https://www.spp.org/organizational-groups/board-of-directorsmembers-committee/markets-and-operations-policy-committee/economic-studies-working-group/

https://www.spp.org/organizational-groups/board-of-directorsmembers-committee/markets-and-operations-policy-committee/economic-studies-working-group/



Addressing congestion related issues requires consideration of both transmission and

generation investment options from an overall market view. For instance, the current excess

generation capacity in the SPP market and its implication on market outcomes is a topic that

needs to be evaluated in the planning process. For the last several years, the persistently

high amount of available generation capacity has been a contributing factor to the relatively

low market prices in the SPP market. This affects the financial viability of generators as low

prices that are not supportive of the existing generation capacity makes future retirements

more likely.

The MMU outlined its first recommendations to improve planning processes in the 2017

annual report.109 In the 2017 report, the MMU pointed out that enhancing the accuracy of

planning processes with operational realities would enable SPP and its members to more

effectively plan for future system needs and conditions. The MMU observed that while SPP

and stakeholders have progressed in improving and aligning the planning and operational

processes,110 there are additional areas that could be improved upon. Specifically, we noted

that the economic studies and the resource adequacy processes would benefit from further

alignment with operational conditions.

The MMU was involved in helping stakeholders develop key inputs to the 2020 Integrated

Transmission Planning (ITP) scope and futures case studies. In particular, MMU feedback on

the following items helped shape planning assumptions:

• Coal and natural gas generators’ retirement age were modified from 60 years to 56

years for coal and 50 years for natural gas resources. These changes were based on

MMU analysis of information posted by the Energy Information Administration (EIA)

and were included in both the Future 1 and Future 2 assumptions of the 2020 ITP.

• The MMU assisted SPP staff in assessing how to consider the capacity value of battery

storage resources. The MMU recommended that the capacity of batteries be counted

at 100 percent of capacity as opposed to 48 percent. SPP staff presented and

stakeholders approved this value.

109 SPP Market Monitoring Unit Annual State of the Market 2017, published May 8, 2018 at 198-199. The following is based on that recommendation. 110 The report cited the introduction of an assessment for persistent operational needs and the criteria for identifying these needs in SPP Integrated Transmission Planning Manual (including sections 4.4, 5.3.4, and 6.1.4).



Section 6.1 of this chapter covers the resource adequacy process. The remainder of this

chapter covers in more detail on the transmission planning process and the input provided

by the MMU.

6.1 RESOURCE ADEQUACY

6.1.1 CAPACITY ADDITIONS AND RETIREMENTS

Nearly 48 percent of SPP’s fleet is more than 30 years old. In particular, nearly 87 percent of

coal capacity and just under 38 percent of gas capacity is older than 30 years. According to

the U.S. Energy Information Administration (EIA), the national average retirement age of coal-

fired generation that retired between 2008 and 2017 was 52 years.111 Aside from the

resources that joined SPP from Nebraska in 2009 and the Integrated System in 2015, the

greatest majority of significant new capacity in the SPP footprint over the last 10 years has

been wind capacity.

Figure 6—1 illustrates that certain segments of the SPP generation fleet are aging. The chart

highlights that almost 30 GW of generation is over 40 years old. Almost half of which is gas,

and about a third is coal.

Figure 6—1 Capacity by age of resource

111 See https://www.eia.gov/todayinenergy/detail.php?id=34452. The EIA analysis includes retirements from not just the electricity sector.

0

10

20

30

40

Coal Gas Hydro Wind Other

Nam

epla

te c

apac

ity (G

W)

0 to 30 years 30 to 40 years 41 to 45 years 46 to 50 years Over 50 years

0

20

40

60

80

100

Total

https://www.eia.gov/todayinenergy/detail.php?id=34452



Figure 6—2 shows the annual trend of capacity additions and retirements over the past three

years.

Figure 6—2 Capacity additions and retirements by year

Almost all of the coal and gas capacity retired since 2016 has been 1950s era plants. The

nuclear unit retired in 2016 was commissioned in the early 1970s. The wind units that retired

in 2017 were first-generation wind resources with very low capacity.

For capacity additions, wind generation has accounted for 85 percent of the additions over

the last three years and 97 percent of the additions in 2018, totaling over 2,300 MW of

nameplate capacity additions in 2018. Even with the increased amount of solar generation in

the generation interconnection queue, the last solar generation added to the SPP market was

in 2016 at 165 MW.

6.1.2 GENERATION CAPACITY COMPARED TO PEAK LOAD

In the 2017 annual report, the MMU introduced a new peak available capacity metric to

replace the previous reserve margin metric112 used in prior reports. The new metric uses a

percentage of the average maximum capacity for each resource during July and August, and

divides that figure by the nameplate capacity for the resource. This method essentially

112 The previous reserve margin metric used unit registration ratings (i.e. nameplate capacity) to determine system capacity, while wind counted at only five percent of registered capacity.

0

1,200

2,400

3,600

2016 2017 2018 2016 2017 2018

Nam

epla

te c

apac

ity (M

W)

Nuclear Coal Wind Solar Hydro Gas Oilretirements additions



creates a derated capacity value due to ambient temperatures for each resource and is a

more conservative measure of capacity when compared to nameplate, or even summer rated

capacity. A percentage is then calculated for each fuel type of resource and that percentage

is applied to the total nameplate capacity. Based on capacity factors during the afternoon

hours in July and August, wind and solar resources are derated accordingly. Solar resources

are derated to 50 percent of nameplate capacity. For wind resources, 24 percent of

nameplate capacity is included in the calculation.

The peak available capacity percent is the amount of extra system capacity available after

serving system peak load, and is shown in Figure 6—3.

Figure 6—3 Peak available capacity percent

Year Peak available capacity (MW) Peak load (MWh)

Peak available capacity percent

2014 62,332 45,301 38% 2015113 62,958 45,279 39%

2016 68,839 50,622 36% 2017 67,950 51,181 33% 2018 67,475 49,926 35%

For 2018, the peak available capacity percent was 35 percent, up from 33 percent in 2017.

At 35 percent, peak available capacity is still nearly three times higher than SPP’s minimum

required planning reserve margin of 12 percent.114 Also note that the peak availability

capacity metric will differ from the reserve margin calculated by SPP because of differences in

methodology. Most notably, the SPP methodology only includes capacity with firm

transmission, whereas our metric includes all system resources interconnected with the SPP

grid using derate factors.

A relatively high peak available capacity percentage such as this has positive implications for

both reliability and mitigation of the potential exercise of market power within the market.

However, it also contributes to downward pressure on market prices, negatively affects

113 2015 uses the total capacity on September 30, prior to the addition of the Integrated System. 114 SPP Planning Criteria 4.1.9.



revenue adequacy, and can burden ratepayers with additional and potentially unnecessary

costs.115

6.2 INTEGRATED TRANSMISSION PLAN

The SPP tariff116 requires SPP to conduct an annual Integrated Transmission Planning

Assessment to evaluate the transmission system upgrades for a ten-year planning horizon.

The planning assessment serves as a regional planning process. This process involves many

aspects of transmission planning including the considerations for reliability, public policy,

operational, and economic needs, and generator interconnection to develop a cost-effective

transmission portfolio for a ten-year planning horizon.117,118 The assessment employs a set of

modeling assumptions that are used as the starting point for planning studies and identifies

system needs from interconnection and transmission service requests within the limits of the

process timelines established. The process coordinates the evaluation of transmission

service needs and associated projects with those identified in the ITP assessment. This way

the current approach targets to facilitate continuity in the SPP’s overall transmission

expansion plan.119

Each integrated transmission plan has a study scope that receives input from stakeholders in

conjunction with the assumptions and parameters that are not standardized in the integrated

transmission process manual.120, The SPP transmission planning studies coordinate the

evaluation of transmission service needs—resulting from generator interconnection and

transmission service requests—and associated projects with those identified in the ITP

assessment to conduct an overall transmission expansion plan.

115 We recognize that grid resiliency is a topic of concern and discussion. However, we feel that resiliency must be properly measured and evaluated in the SPP market to ensure that ratepayers are not burdened by unnecessary costs of excess capacity. 116 See SPP OATT Sixth Revised Vol. No. 1, Attachment O Transmission Planning Process, Section III. 117 The latest version of the ITP manual published on October 17, 2018 is available at the SPP website. 118 SPP also performs a 20-year assessment. We have not covered this assessment in this report as this study is currently not under development. 119 Ibid. at 1. 120 See SPP Tariff Attachment O at Section III.1.c.



SPP, along with the ESWG and the Transmission Working Group (TWG), develops an

assessment scope for each ITP assessment for items that require SPP stakeholder review and

approval with each new study.121 The study scope document122 describes the assumptions

and methodologies that need to be updated at each study cycle so that the future

performance of the existing transmission system and any needed improvements could be

assessed.123

In 2018, the MMU participated in the development discussions of the 2020 ITP scope. The

MMU’s input covered an array of topics including the retirement age assumptions for coal

and gas-fired generators and accreditation level assumptions for new wind, solar, and battery

resources.

6.2.1 2020 ITP GENERATOR RETIREMENT ASSUMPTIONS

The initial 2020 ITP scope proposal carried forward the 2019 ITP assumptions which assumed

60 years as the retirement age of for coal and natural gas-fired generators. Given the

importance of such assumptions, the MMU recommended updating the 2020 ITP retirement

assumptions to no more than the Energy Information Administration (EIA) average for the last

five years (2013 through 2017) as the existing assumptions appeared to underestimate the

actual retirement trend in the SPP market. The averages were 56 years for coal and 50 years

for natural gas-fired generators.124 The MMU based its analysis on the historic retirement

data compiled by the EIA.125

121 Ibid. at 1-2. The assessment report requires approvals from the MOPC and the board. 122 The most recent scope document is the “2020 Integrated Transmission Planning Assessment Scope,” published on January 18, 2019. The document was approved by the SPP stakeholder process. 123 The ITP manual at 5. 124 The last five year averages through November 2018 (2014-2018) are even lower, 55 years for coal and 49 years for gas-fired resources. 125 The MMU used the EIA’s monthly generator retirement data for the entire U.S. to calculate historic averages for each year. The average retirement years were calculated for the three five-year sub periods to see historical trends for coal and gas-fired generators, and the most recent sub period was used for the recommendation. Only the electric utility generators were included in this calculation (See EIA-860M Preliminary Monthly Electric Generator Inventory (based on Form EIA-860M as a supplement to Form EIA-860 available at https://www.eia.gov/electricity/data/eia860M/).

https://www.eia.gov/electricity/data/eia860M/



The MMU further argued that its net revenue analysis indicates a new coal resource cannot

recover fixed operating and maintenance (O&M) costs at current market prices (see Section

4.4 for 2018 results). In addition, the trend for coal and natural gas resources’ retirements in

2018 was below the five-year EIA averages. As such, even lower retirement ages for the next

10 years could be expected. The MMU’s argument is also supported by the recent

announcements of several market participants redirecting their investment policies towards

more renewable resources and less carbon intensity.126

Figure 6—4 shows the data for average age of retirements for coal and natural gas-fired

resources for the U.S.127

Figure 6—4 Average age of retirements for coal and gas resources 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Coal 53 48 54 64 54 54 56 58 56 56 49

Gas 43 52 43 49 48 51 52 48 46 52 46

2002-17 2008-17 2013-17

Coal 55 55 56

Gas 47 49 50

Source: EIA-860 Electric Utility Retirements (data through December 2018).

126 For instance, Omaha Public Power District (OPPD) set a long-term goal of providing at least 50 percent of its retail sales from renewable sources and reducing its carbon intensity by 20 percent from 2010 to 2030 (S&P Global Market Intelligence published on Nov., 5, 2018). Xcel Energy Inc. announced that it plans to completely decarbonize its power supply portfolio serving eight states by midcentury, first cutting carbon emission by 80 percent by 2030, from 2005 levels in its utility service territories in Colorado, Michigan, Minnesota, New Mexico, North Dakota, South Dakota, Texas and Wisconsin (S&P Global Market Intelligence published on Dec., 4, 2018). AEP announced in February 6, 2018 that it will pursue a goal to reduce its CO2 emissions of its generating units 60 percent by 2030 and 80 percent by 2050 from 2000 levels (https://www.aep.com/news/releases/read/1503/AEPs-Clean-Energy-Strategy-Will-Achieve-Significant-Future-Carbon-Dioxide-Reductions-). 127 SPP staff did an analysis that compared the SPP region to the total U.S. and identified that the U.S. results were similar to the SPP regional results. See September 25, 2018 meeting materials, Item 3 – 2020 ITP Futures: https://spp.org/Documents/58652/eswg%20twg%20agenda%20&%20background%20materials%2020180925.zip.

https://www.aep.com/news/releases/read/1503/AEPs-Clean-Energy-Strategy-Will-Achieve-Significant-Future-Carbon-Dioxide-Reductions-

https://www.aep.com/news/releases/read/1503/AEPs-Clean-Energy-Strategy-Will-Achieve-Significant-Future-Carbon-Dioxide-Reductions-

https://spp.org/Documents/58652/eswg%20twg%20agenda%20&%20background%20materials%2020180925.zip

https://spp.org/Documents/58652/eswg%20twg%20agenda%20&%20background%20materials%2020180925.zip



Figure 6—5 below presents the age of all existing resources along with the original proposed

retirements and with the change associated with using EIA averages.

Figure 6—5 Average age of generators in the SPP market

The ESWG modified generator retirement assumptions used in the 2020 ITP to 56 years for

coal and 50 year for natural gas-fired resources. However, participants included an exception

to add back resources that repowered, even if the age limit was reached. This has resulted in

over 4,000 MW of resources that were set to retire based on age back into the study.128

6.2.2 2020 ITP RENEWABLE CAPACITY ACCREDITATION ASSUMPTIONS

In addition to generator retirements assumptions, the 2020 ITP scope includes assumptions

on how to value—or provide credit for—new wind and solar generation capacity. The value of

128 This was approved in the February 2019 ESWG meeting.

0

10

20

30

40

Coal Gas Hydro Wind Other

Nam

epla

te c

apac

ity (G

W)

0 to 30 years 30 to 40 years 41 to 45 years 46 to 50 years Over 50 years

set to retire under original assumption

additional retirement using EIA average



capacity is called accreditation.129 Each RTO/ISO market has its own assessment of the value

of new wind and solar resources in its region. The original 2020 ITP proposal was to carry

over the assumptions from the 2019 ITP, which valued new wind at 20 percent and new solar

at 70 percent.130,131

The MMU reviewed various accreditation calculations made by other RTOs/ISOs, and no

consistent method was found across the markets. In fact, while some value new wind and

solar capacity during peak or super-peak hours, others use performance during a small set of

the highest load hours. In some cases capacity is differentiated by location (such as coastal

versus non-costal), in other cases capacity is differentiated by technology (such as fixed

versus tracking solar).

As stated in Section 6.1.2, the MMU modified the reserve margin calculation methodology in

the 2017 annual report to make the reserve margin estimation more representative of

operational conditions. Using the new methodology for calculating the peak available

capacity metric, shown in Figure 6—3 above, the MMU developed a set of wind and solar

129 These assumptions for capacity contributions are used as inputs in development of the market economic models and reflect resources’ expected supply availability given a particular future. The final target in the process is to develop and recommend a final portfolio after considering solutions and portfolios addressing varying goals such as mitigating economic needs (cost effectiveness), reliability needs, public policy needs, and persistent operational needs. 130 These numbers were approved by the Supply Adequacy Working Group (SAWG) and used in the ITP. 131 The resource adequacy requirement in SPP is for load to procure generation equal to 112 percent of expected load. The accredited renewable capacity is capped at 12 percent of load-serving entity’s total load only for non-designated resources. The 12 percent cap for renewable capacity does not apply to designated resources. Designated resources are those that can be called upon at any time for the purpose of serving network load on a non-interruptible basis. The designated generation resources must be owned, purchased or leased by the owner of the network load. (See SPP Planning Criteria, Revision 1.6, January 31, 2019).



capacity factors to aid in decision-making on capacity accreditation assumptions.132 The

results are shown in Figure 6—6 below.

Figure 6—6 MMU calculation for wind and solar capacity factors133 Peak type Capacity factor Wind, peak hour 24%

Wind, super-peak hour 23%

Solar, peak hour 38-45%

Solar, super-peak hour 53-64%

Comparing the SPP accreditation numbers with historical performance, the MMU observed

that the accreditation numbers for new wind resources may be understated and accreditation

numbers for new solar resources may be overstated in the 2020 ITP scope. Consequently,

the MMU recommended that SPP stakeholders modify their accreditation numbers for new

wind and solar resources consistent with historic performance. SPP stakeholders, however,

decided to leave the existing accreditation values unchanged.134

132 The MMU used average historic peak hour performance in July and August over three years, and defined two sets of peak hours: peak hours and super-peak hours. Peak hours reflect the 16 hour peak load period (7 am to 11 pm, non-holiday weekdays) and the super-peak hours reflect the hours beginning 3 pm through 6 pm (four-hour period on all days). Weights were assigned to capacity factors observed in three years giving more weight to a more recent year assuming that a more recent capacity factor would better represent the current situation. As a result, wind peak hour and wind super-peak hours rounded to the same value for the 2015-2017 and 2016-2018 periods; solar peak hour was 38 percent for the 2015-2017 period and 45 percent for the 2016-2018 period; and solar super-peak hour was 53 percent for the 2015-2017 period and 64 percent for the 2016-2018 period. 133 These numbers are calculated from peak and super-peak hours in July and August. Peak hours are 16-hour blocks from 7am to 11pm on non-holiday weekdays. Super-peak hours are calculated from 3pm through 6pm. 134 The ESWG expressed that the Supply Adequacy Working Group (SAWG) has the expert knowledge and opinion on the subject and decided to keep the existing accreditation numbers.



Finally, the MMU recommended SPP develop a documented methodological approach

going forward for calculating renewable accreditation levels similar to other RTOs/ISOs and

stakeholders to provide input on that approach.135

6.2.3 OTHER TOPICS

In revision request 276, SPP staff recommended updating the pricing of variable operations

and maintenance cost (VOM) for renewable generation in the ITP manual. Specifically, SPP

staff recommended lowering the renewable modeling assumption of $8/MWh for wind and

solar units to $0/MWh in production cost calculations as an interim measure. The MMU

concurred with SPP staff’s assessment as $8/MWh for renewable VOM level was not reflective

of generally observed variable operations and maintenance cost levels in mitigated offers.

Thus, relative to the $8/MWh level, the MMU supported the SPP staff’s $0/MWh proposal

because it more closely reflects operational realities observed in the SPP market, and filed

comments accordingly.136

SPP staff also recommended 48 percent accreditation for battery storage to be used in the

2020 ITP scope. The MMU conducted an analysis by reviewing technical and market

information, and recommended a 100 percent accreditation assumption for this study. The

MMU’s recommendation was based primarily on the following: the general market practice

of capacity contracts for batteries enhances 100 percent accreditation; the physical capability

of the resource should be the main factor for such calculation; and there is no justification for

an accreditation number other than the 100 percent level at this time. After further

135 The Supply Adequacy Working Group (SAWG) is currently considering to adopt a specific methodology for determining accreditation figures as used in other markets. In its March 27, 2019 meeting, the SAWG approved a motion to use Effective Load Carrying Capability (ELCC) method as the guiding principle for the accreditation of solar, wind and storage resources to replace the current accreditation methodology found in section 7.1.5.3 (7) of the SPP Planning Criteria once new criteria language is approved. 136 In the March 2018 ESWG meeting, the stakeholders approved the $8/MWh number and rejected the $0/MWh recommendation. The Market Operations and Policy Committee overruled the ESWG decision and approved renewable VOM costs at $0/MWh at its April 2018 meeting.



consideration, SPP staff concurred with the MMU arguments and the ESWG approved the

accreditation for battery storage at 100 percent in December.137

In addition, the ESWG evaluated the potential of adding a third study case as part of the 2020

ITP. The third future would have considered a shift in environmental regulations—such as a

carbon tax adder—and a change in technologies/market trends including accelerated

deployment of storage devices or electric vehicles. The MMU supported evaluation of this

third scenario as it would help SPP and stakeholders plan for a wider range of future

possibilities than was captured in the other study cases. Moreover, the third study case was

similar to a study case in the MISO transmission plan.138 SPP stakeholders rejected the third

study case as they considered it too costly to include another case study as part of the

evaluation process.

6.3 ASSESSMENT

The MMU has a unique position to evaluate SPP data and market outcomes and provide

feedback in the planning process to improve efficiencies and outcomes. This involvement

has already proven beneficial as stakeholders approved some MMU recommendations for

the 2020 ITP.

The MMU believes the accuracy of (planning) projections for needed upgrades in

transmission infrastructure heavily depends on market data and analysis. When underlying

assumptions (or inputs) for economic modeling are not reflective of operational realities or

historical trends—such as the retirement ages of generators, the role of renewables, or

emerging new technologies—both the location and magnitude of need for new transmission

might be inappropriate. Consequently, the market signals for investment in transmission and

generation could be distorted resulting in unexpected congestion patterns and suboptimal

levels of transmission. This in turn would inevitably translate into suboptimal costs and

market prices.

137 The MMU raised the issue in ESWG’s November 2018 meeting, and the stakeholders approved the 100 percent recommendation in their December 3-4, 2018 meeting. 138 MTEP19 Futures: Summary of definitions, uncertainty variables, resource forecasts, siting process and siting results, undated (available at https://cdn.misoenergy.org/MTEP19%20Futures%20Summary291183.pdf).




For instance, planning for generation and resulting congestion patterns that would not likely

exist given prevailing trends would impose an unnecessary cost of transmission on market

participants in particular and ratepayers in general. Alternatively, a robust planning

projection could serve as a preemptive measure for eliminating or diverting congestion. A

particularly useful example is the case of addition of the phase-shifting transformer at the

Woodward substation in 2017 to address congestion. This upgrade substantially reduced

the frequency of congestion at Woodward such that the MMU recommended no longer

designating Woodward as a frequently constrained area. Furthermore, as highlighted earlier

in this report (see Section 5.1.3), further expansions at Woodward in 2018 were a significant

factor in reducing the incidence of negative prices. The MMU believes that SPP can further

align planning processes with market and operational realities to achieve full benefits of the

process.

Finally, the MMU has the most up-to-date cost data for all resources and can provide valuable

feedback in the planning process when financial assumptions are made. As in the case of

renewable variable operations and maintenance modeling assumptions, the MMU is

prepared to make directional guidance on the level of such costs without compromising

confidential market participant data.

6.4 RECOMMENDATIONS GOING FORWARD

Based on its experience with the planning process in 2018, the MMU recommends the

planning process continue to use data gathering and analysis methods to determine

economic modeling inputs so that a more accurate assessment can be made for the SPP’s

system needs and conditions. One way to accomplish this is to review best practices across

other organized markets.

For instance, the accreditation calculations for new solar and wind resources in other markets

encompass varying methods that differ by peak and super-peak hours, location (coastal

versus non-coastal) or type (fixed versus tracking). In some markets, the incremental effect of

the loss of load without the (renewable) resource is used in calculating accreditation levels.

Accreditation calculations for battery storage also require equally careful approaches. A

better way to accomplish this going forward is to develop a documented and dynamic

approach along with input from stakeholders.



Another important issue is the generator retirement age assumptions used in economic

modeling. In addition to market participant input, substantial consideration needs to be

given to the operational realities and historical trends when the assumptions for retirement

ages are made.

The futures used in the planning process allow for the assessment of likely deviations from

the (base) reference case where the impact could be significant. When including scenarios,

the ITP scope should consider developments occurring in the power sector in general with

due considerations given for probable shifts in environmental regulations, market trends, or

technological advancements. In addition, the planning processes should factor in

announcements with regard to major shifts in corporate policies given the evolving market

trends and technologies towards more renewable generation and reduced emissions.

The planning process should have the ability to assess a range of potential outcomes in

transmission planning. Following market trends helped shape the MMU’s thinking in terms of

how to consider the future generation mix and the potential need for an additional scenario.

The MMU believed that a third scenario would provide a bookend scenario for the 2020

analysis but was rejected by stakeholders as it was considered too costly to study.

While the transmission planning process theoretically can include a third scenario, in practice

it does not have the flexibility to include one given that cost of including a third scenario was

not unique to the 2020 planning process. This appears to be a significant shortcoming in the

planning process as the study process is limited to only two potential scenarios in its 5- and

10-year look ahead. This limits the range of potential outcomes that SPP could study139 or

alternatively requires the cases to be very broad and different from each other.

We recommend that SPP enhance their study process to allow the ability to study a range of

potential outcomes. If such range of potential outcomes are not captured in a third case

study, the MMU recommends to factor them into either an existing case study or potentially

as part of the 20 year assessment. The wider the range of possibilities studied, the more

robust the results will be.

139 The MISO transmission planning process studies four cases, see https://cdn.misoenergy.org/MTEP19%20Futures%20Summary291183.pdf.




Southwest Power Pool, Inc. Competitive assessment Market Monitoring Unit


7 COMPETITIVE ASSESSMENT

Chapter 7 of this report provides a competitive assessment of the SPP market. Key points

from this chapter include:

• Structural and behavioral metrics indicate that the SPP markets have been very

competitive over the last several years. The market share, HHI, and pivotal supplier

analyses all indicate minimal to moderate potential structural market power in SPP

markets outside of areas that are frequently congested.

• There were 35 percent of the hours in 2018 having the market share exceeding the 20

percent threshold, which was a significant increase from 2017. The data shows that

the increase in highest market share hours occurred following May 2018, which

coincides with the merger between Great Plains and Westar.

• The HHI index analysis show that 12 percent of hours were considered moderately

concentrated in 2018, with all of those hours occurring after the merger, while the

market was unconcentrated in all hours of 2016 and 2017.

• While the Great Plains and Westar merger has increased market concentration

compared to the last two years, the market was even more concentrated before 2016.

The MMU does not have any concerns with the market concentration increase with the

merger.

• The results of pivotal supplier analyses indicate that the percent of hours with pivotal

supplier is the highest (98 to 100 percent) in the New Mexico and West Texas region,

increasing with demand level. This is where one of the SPP’s frequently constrained

areas in 2018 was located.

• The monthly offer price markups of 2018 ranged from −$5.17 to $2.64/MWh for off-

peak periods and from −$6.01 to $3.07/MWh for on-peak periods. The lowest

markups occurred in spring and winter in off-peak and on-peak hours, when wind

generation was generally the highest.

• Both off-peak and on-peak average markups were at the lowest levels since

implementation of the Integrated Marketplace at around −$2.05/MWh and

−$2.41/MWh, respectively.

• The incremental energy mitigation was extremely low, at less than 0.01 percent in

both the day-ahead real-time markets in 2018. The overall mitigation frequency of



start-up offers in 2018 were at the lowest levels since the market started in 2014 as the

combined frequency of mitigation of start-up offers for day-ahead, reliability unit

commitment, and manual commitments decreased to 2.8 percent in 2018 from 3.2

percent in 2017.

• Similar to 2017, economic withholding mitigation remained infrequent. The results of

the output gap analysis indicate still a very low levels of (economic) output withheld in

the SPP footprint, less than 0.01 percent on annual average level. These outcomes are

consistent with the expectations of competitive market conduct.

• The unoffered economic capacity results show that on an annual average basis the

total unoffered capacity equaled 2.0 percent in 2016, 1.9 percent in 2017, and 3.2

percent in 2018. The majority of the outages were long-term, and were primarily the

result of maintenance in the spring and fall shoulder months.

• The start-up mitigation only occurred in less than three percent of intervals in 2018

and day-ahead mitigation accounted for 84 percent of the total start-up cost

mitigation. These figures were similar to results in 2017.

• Major maintenance costs will begin being included in mitigated start-up and no-load

offers in April 2019.

• Starting from 2017, the MMU has been involved with the SPP’s initiative to develop a

new process for evaluating generator retirement applications. Given the expected

market developments and increased rate of retirements, a streamlined process of

collecting generator specific (financial) data would greatly enhance the MMU’s review

of retirements and would provide sufficient time to recommend and develop

appropriate mitigation measures.

The SPP Integrated Marketplace provides sufficient market incentives to produce competitive

market outcomes in regions and periods when there are no concerns with regard to local

market power. The MMU’s competitive assessment provides evidence that in 2018 market

outcomes were workably competitive and that the market required mitigation of local market

power infrequently to achieve those outcomes.

The market power analysis in this report considers both the structural and behavioral aspects

of market power concerns. The structural aspects can be detected by various techniques

such as market share analysis, (market-wide) concentration indices, and pivotal supplier

analysis. The structural indicators are used to look for the potential for market power without



regard to the actual exercise of market power. Behavioral aspects, on the other hand, assess

the actual offer or bid behavior (i.e., conduct) of the market participants, and the impact of

such behavior on market prices by looking for the exercise of market power. These

behavioral indicators include offer price markup;140 economic withholding analysis,

addressed through automated mitigation; output gap analysis; and physical withholding.141

This chapter evaluates the SPP market’s competitive environment by establishing the level of

structural market power and then examining market prices for indications of market power

impact. The level of structural marker power is assessed both at the general level through

concentration indices and at the local—transmission constraint—level through pivotal supplier

analysis. Mitigation of economic withholding is accomplished ex-ante through automatic

market power mitigation processes that limit the ability of generators with local market power

to raise prices above competitive levels. The mitigation program is monitored and evaluated

to ensure it is efficient and effective. Accordingly, the following subsections examine the

significance of market power and the effectiveness of local market power mitigation in the

SPP markets.

7.1 STRUCTURAL ASPECTS OF THE MARKET

Three core metrics of structural market power are the market share analysis, the Herfindahl-

Hirschman Index (HHI), and pivotal supplier analysis. The first two of these indicators

measure concentration in the market and are of a static nature. Pivotal supplier analysis, on

the other hand, takes into account the dynamic nature of power markets and considers

changing demand conditions and locational transmission constraints in assessing potential

market power.

Figure 7—1 displays the market share of the largest on-line supplier (i.e., market participant)

in terms of energy output in the real-time market by hour for 2018.

140 While the SPP MMU uses offer price markup, other market monitors may use price cost markup as a behavioral indicator. 141 The uneconomic production analysis, as another behavioral indicator, is addressed through FERC referrals.



Figure 7—1 Market share of largest supplier

The market share ranged from 9.1 percent to 28.4 percent, exceeding the 20 percent

threshold142 in 3,093 of the hours (35 percent) for the year. This represents a significant

increase from the 9.3 percent to 16.9 percent range in 2017. The data show that the above

mentioned increase in highest market share hours occurred beginning June 2018 which

coincides with the merger between Great Plains (the parent company of KCP&L) and Westar

Energy that formed Evergy, Inc. Furthermore, these 3,093 hours covers both summer peak

and the following shoulder months indicating a consistent pattern. Note that although a

mere increase in market share in itself does not pose a threat to the structural

competitiveness of the SPP market, other relevant market data including pivotal supplier

hours and local market power mitigation must also be evaluated for competitive assessment

(see below).

142 The 20 percent threshold is one of the generally accepted metrics that would indicate structural market power. Note, however, that neither market share nor the HHI metric alone would be sufficient for the assessment of market power particularly in today’s spot electricity markets where load pockets formed by transmission congestion may lead to market power with much smaller market shares and/or HHI values.

0

5

10

15

20

25

30

Perc

ent

mar

ket s

hare

of

larg

est m

arke

t p

artic

ipan

t

20% threshold




The Herfindahl-Hirschman Index (HHI) is another general measure of structural market power,

analyzing overall supplier concentration in the market. It is calculated by using the sum of the

squares of the market shares of all suppliers in a market as follows:

𝐻𝐻𝐻𝐻𝐻𝐻 = ��𝐿𝐿𝑀𝑀𝑖𝑖∑ 𝐿𝐿𝑀𝑀𝑖𝑖𝑖𝑖

∗ 100�2

𝑖𝑖

According to FERC’s “Merger Policy Statement,”143 which is similar to Department of Justice

merger guidelines, an HHI less than 1,000 is an indication of an unconcentrated market, an

HHI of 1,000 to 1,800 indicates a moderately concentrated market, and an HHI over 1,800

indicates a highly concentrated market.

Figure 7—2 provides the number of hours for each concentration category in terms of actual

generation over the last three years.144

Figure 7—2 Market concentration level, real-time

Concentration HHI Level

2016 2017 2018

Hours % of Hours Hours % of

Hours Hours % of Hours

Unconcentrated Below 1,000 8,784 100% 8,760 100% 7,692 88%

Moderately concentrated

1,000 to 1,800 0 0% 0 0% 1,067 12%

Highly concentrated

Above 1,800 0 0% 0 0% 0 0%

The SPP market was unconcentrated in all hours of 2016 and 2017. However, 12 percent of

hours were considered moderately concentrated in 2018. The SPP market has never risen

above the highly concentrated threshold of 1,800 since the start of the Integrated

Marketplace in 2014. Figure 7—3 depicts the hourly real-time market HHI in terms of

generation for 2018.

143 Inquiry Concerning the Commission’s Merger Policy Under the Federal Power Act: Policy Statement, Order No. 592, Issued December 18, 1996 (Docket No. RM96-6-000). 144 The SPP MMU calculates HHI by actual sales. The FERC merger guidelines uses capacity owned.



Figure 7—3 Herfindahl-Hirschman Index

Hourly HHI values ranged from 502 to 1,283 during 2018, pointing to a higher trend relative

to 2017.

Figure 7—4 shows a graphical breakdown of the HHI for all hours since the start of the

Integrated Marketplace in March 2014. For the years with significant events impacting the

make-up of the market – 2015, for the addition of the Integrated System; and 2018, with the

creation of Evergy – the years are divided on the chart showing the HHI before and after the

event.

0

500

1,000

1,500

2,000

Her

find

ahl-H

irsc

hman

Ind

ex


Unconcentrated

Moderately concentrated

Highly concentrated



Figure 7—4 Herfindahl-Hirschman Index, annual

As shown above, the market remained unconcentrated from the date of the addition of the

Integrated System to the date of the KCPL/Westar merger creating Evergy.145 Even though

moderately concentrated hours increased following the creation of Evergy, just under 20

percent of hours were considered moderately concentrated. In contrast, in 2014, just over 50

percent of hours were considered moderately concentrated, and in 2015, prior to the

addition of the Integrated System, nearly 40 percent of hours were moderately concentrated.

While moderately concentrated hours increased following the creation of Evergy, it is not

necessarily a cause for alarm. Given the HHI results in conjunction with the elevated market

shares, a closer monitoring of market behavior by market participants in terms of the exercise

of local market power may be required.146

145 Note that the HHI analysis is performed at the market participant level. There may be asset owners under market participants on a contractual basis, where bidding control is not under the purview of the market participant. 146 Section 7.2 analyzes behavioral aspect of market power.

0%

20%

40%

60%

80%

100%

2014 2015pre-IS

2015post-IS

2016 2017 2018pre-merger

2018post-merger

Unconcentrated Moderately concentrated Highly concentrated



SPP market participants with generation spanning all supply segments,147 reflecting the

technology and fuel mix of resources, have the greatest ability to benefit from structural

market power. These market participants may frequently set prices regardless of the

technology type on the margin. Figure 7—5 provides the percent of real-time market intervals

that each market participant had a resource on the margin.148

Figure 7—5 Market participants with a marginal resource, real-time

The chart shows that the top three market participants each set price in more than 10

percent, but less than 15 percent, of all real-time market time intervals. Conversely, well over

half of all participants set price in less than one percent of all intervals.149

The MMU’s market share analysis and calculated HHI both indicate a moderate potential for

general structural market power in SPP markets outside of areas that are frequently

congested. Structural market power is also assessed at a more localized level and in the

context of locational transmission constraints by reevaluating frequently constrained areas

periodically and (re)defining them accordingly as was discussed in Section 5.1.7.

147 Supply segments include baseload, intermediate, and peak resources. 148 Kansas City Power & Light and Westar Energy were treated as separate entities for the purpose of this analysis. 149 The percentages on this chart are not additive because multiple market participants may have a resource on the margin during any given interval.

0%

5%

10%

15%

20%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63

Perc

ent

of r

eal-t

ime

inte

rval

s w

ith a

m

arg

inal

res

our

ce

Market participants



Pivotal supplier analysis takes into account the dynamic nature of the power market,

particularly demand conditions, and evaluates the potential for market power in the presence

of “pivotal” suppliers. A supplier is pivotal when its resources are needed to meet demand.

There may be one or more pivotal suppliers in a particular market defined by transmission

constraints and load conditions, and a supplier’s status of being pivotal may vary between

time periods irrespective of its size. In the market clearing process, market power is

evaluated locally through a local market power evaluation because the exercise of local

market power is relevant for determining prices.

The following analysis identifies the frequency with which at least one supplier was pivotal in

the five different reserve zones (regions) of the SPP footprint in 2018.150 One condition for a

supplier to have an ability to raise prices above competitive levels, is the frequency it

becomes pivotal. Another market condition is during times of shortage or high demand.

The mere size of a supplier has no link to being pivotal; however, suppliers with a high

frequency of being pivotal in tight supply periods have an even greater ability to exercise

market power. For this reason, the frequency of being a pivotal supplier is also analyzed at

various levels of demand across these five regions.

Figure 7—6 shows how frequently a supplier is pivotal at varying load levels in the five reserve

zones in the SPP market footprint.

150 SPP divides market resources (generation) into five reserve zones. For the purpose of this report, these reserve zones are named as “Nebraska”, “Western Kansas and Panhandles”, “New Mexico and West Texas”, “Kan., Mo., Okla., Ark., East Texas, La.”, and “Iowa, Dakotas, Montana”. Thus, each generation resource is mapped to one of these reserve zones. To define a load zone to match with a resource zone, each load settlement location was mapped to a reserve zone to approximate demand within a particular zone. Additionally, import limits are approximated by the average of the reserve zone limits for the times they were activated in 2018.



Figure 7—6 Hours with at least one pivotal supplier

The results indicate that the percent of hours with pivotal supplier is the highest (98 to 100

percent) in the New Mexico and West Texas region, increasing with demand level. This is

where one of the SPP’s frequently constrained areas in 2018 was located. This region is

followed by the Iowa, Dakotas, Montana region where, depending on the level of load, one

to 52 percent of the hours exhibit at least one pivotal supplier. Compared to previous years,

the Nebraska region increased from seven to 12 percent of high demand hours having a

pivotal supplier. The remaining regions experience pivotal supplier conditions for only

negligible periods.

7.2 BEHAVIORAL ASPECTS OF THE MARKET

7.2.1 OFFER PRICE MARKUP

In a competitive market, prices should reflect the short-run marginal cost of production of the

marginal unit. In SPP’s Integrated Marketplace, market participants submit hourly mitigated

energy offer curves that represent their short-run marginal cost of energy. Market

participants also submit their market-based offers, which may differ from their mitigated

offers. To assess market performance, a comparison is made between the market offer and

the mitigated offer for the marginal resources for each real-time market interval. Figure 7—7

0%

20%

40%

60%

80%

100%

Iowa, Dakotas,Montana

Nebraska New Mexico andWest Texas

Kan., Mo., Okla.,Ark., East Texas, La.

Western Kansasand Panhandles

Perc

ent

of h

our

s

< 20th between 20-80th > 80thDemand levels



provides the average marginal resource offer price markups151 by month for on-peak and off-

peak periods. Figure 7—8 shows the average on-peak offer price markup by fuel type of

resource. The 2018 data show that we see a growing trend of negative markups in both on-

and off-peak periods, but also in most months as well (see Figure 7—9 for annual ranges).

This implies that there is significant competition in the SPP market.

Figure 7—7 Average offer price markup of marginal resource, monthly

In 2018 the average monthly markups ranged from −$5.17 to $2.64/MWh for off-peak

periods and from −$6.01 to $3.07/MWh for on-peak periods. The lowest markups occurred

in spring and winter in off-peak and on-peak hours, when wind generation was generally the

highest.

151 Offer price markup is calculated as the difference between market-based offer and the mitigated offer where the market-based offer may or may not be equal to the mitigated offer. The MMU calculates a simple average over all marginal resources for an interval. The markups are weighted to reflect each marginal resource’s proportional impact on price.

-$8

-$6

-$4

-$2

$0

$2

$4

Jan 18 Feb 18 Mar 18 Apr 18 May 18 Jun 18 Jul 18 Aug 18 Sep 18 Oct 18 Nov 18 Dec 18

Offe

r p

rice

mar

gin

al c

ost

mar

kup

($

/MW

h)

Off-peak On-peak



Figure 7—8 Average on-peak offer price markup by fuel type of resource, monthly

The observed levels of negative markups indicate that many market participants’ real-time

market offers were below their mitigated offers.152 This could occur where generators decide

to offer below their marginal cost to maintain commitments in a very competitive

marketplace. For instance, wind resources at the margin increased from 10.4 percent of all

resource hours in 2017 to 12.2 percent in 2018. Figure 7—9 below points to a continued

declining annual trend of off-peak and on-peak average markups.

152 Wind units may have negative mitigated offers because of subsidies related to production tax credits.

-$30

-$20

-$10

$0

$10

Jan 18 Feb 18 Mar 18 Apr 18 May 18 Jun 18 Jul 18 Aug 18 Sep 18 Oct 18 Nov 18 Dec 18

On-peak

Offe

r p

rice

mar

gin

al c

ost

mar

kup

($

/MW

h)

Coal Gas Hydro Wind



Figure 7—9 Average offer price markup, annual

Both off-peak and on-peak average markups were at the lowest levels since implementation

of the Integrated Marketplace at around −$2.05/MWh and −$2.41/MWh, respectively,

showing lower levels compared to 2017 figures. Although a lower offer price markup level in

itself would indicate competitive pressure on suppliers in the SPP market, the observed

continuous and deepening downward trend may raise questions about the commercial

viability of generating units and the possibility of generation retirements. However, even

though these pressures have continued to mount over the last several years, SPP continues to

add new generation (mostly wind), and generation capacity relative to peak loads remains

high (see Section 6.1.2). More renewable generation is anticipated to be added over the

next few years, which is expected to outweigh planned retirements. As such the trend in

negative offer-price mark ups will likely continue, further putting pressures on the economic

viability of many generating resources.

7.2.2 MITIGATION PERFORMANCE AND FREQUENCY

SPP employs an automated conduct and impact mitigation process to address potential

market power abuse through economic withholding. The mitigation applies to resources that

exercise local market power in areas of transmission congestion, reserve zone shortages, and

manual commitments.

-$8

-$4

$0

$4

2016 2017 2018 2016 2017 2018

Off-peak On-peak

Offe

r p

rice

mar

gin

al c

ost

mar

kup

($

/MW

h)

High Average Low



SPP resources’ incremental energy, start-up, no-load, and operating reserve offers are subject

to mitigation for economic withholding when the following three circumstances occur

simultaneously in a market solution:

1) The resource has local market power;

2) The offer has failed the conduct test. Resources submit two offers for each product: a

mitigated offer representing the competitive baseline costs that must adhere to the

mitigated offer development guidelines153 and a second offer generally referred to as

a market offer. An offer fails the conduct test when the market offer exceeds the

mitigated offer by more than the allowed threshold; and

3) The resource either:

a) Is manually committed by SPP for capacity, transmission constraint, or voltage

support; or by a local transmission operator for local transmission problems or

voltage support; or

b) The application of mitigation impacts market prices or make-whole payments by

more than the allowed $25/MWh threshold.

Mitigation was very low overall, with some variation across products and markets. Figure 7—

10 shows that the mitigation of incremental energy, operating reserves, and no-load was

generally infrequent in the day-ahead market in 2018.

153 As indicated in Appendix G of the SPP’s Market Protocols.



Figure 7—10 Mitigation frequency, day-ahead market

Overall patterns for mitigation frequency for the three products in the day-ahead market

were similar to those in 2017. Mitigation frequency increased slightly in September and

October, particularly for no-load offers. The mitigation frequency for operating reserves

increased slightly starting in October through December. The application of mitigation in the

day-ahead market occurred at levels of 0.05 percent for operating reserves, 0.16 percent for

no-load, less than 0.01 percent for incremental energy, and about 2.4 percent of start-up

offers.

Mitigation of incremental energy in the real-time market is shown in Figure 7—11 below.

0.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%


Perc

ent

of r

eso

urce

ho

urs

miti

gat

ed

Energy Operating reserves No-load



Figure 7—11 Mitigation frequency, real-time market

For the real-time market, the mitigation of incremental energy has been at very low levels

since the start of the market with an annual average around 0.01 percent in 2018, which is

about the same as 2017 levels.

Figure 7—12 shows the mitigation frequency for start-up offers for the various commitment

types.

Figure 7—12 Mitigation frequency, start-up offers

0.00%

0.02%

0.04%

0.06%

0.08%

0.10%


Perc

ent

of r

eso

urce

ho

urs

miti

gat

ed

0%

2%

4%

6%

8%

10%


Perc

ent

of s

tart

-up

offe

rs m

itig

ated

Day-ahead Reliability unit commitment Manual



The overall mitigation frequency of start-up offers was the lowest since the market began in

2014, as it decreased in 2018 relative to 2017 levels, to under three percent. While the

frequency of manual and real-time mitigation dropped in 2018 compared to 2016 and 2017

levels, day-ahead mitigation remained less than 3 percent from 2016 to 2018. Day-ahead

mitigation accounted for 84 percent of the total start-up cost mitigation. The highest level of

mitigation of start-up offers was at 6.5 percent in June and fell to around 1.5 percent in

December.

7.2.3 OUTPUT GAP (MEASURE FOR POTENTIAL ECONOMIC WITHHOLDING)

Economic withholding by a resource is defined as submitting an offer that is unjustifiably high

such that either the resource will not be scheduled or dispatched, or if scheduled or

dispatched the offer will set a higher than competitive market clearing price. Accordingly,

the output gap metric aims to measure the economic (or competitive) amount of output that

was withheld from the market by submitting offers in excess of competitive levels. The

output gap is the amount of generation not produced as a result of offers exceeding the

mitigated offer above an appropriate conduct threshold. The conduct threshold is employed

to compensate for any inaccuracies or uncertainties in estimating the cost, similarly to the one

used in economic withholding mitigation. In this report, the output gap is calculated as the

difference between the resource’s economic level of output at the prevailing market clearing

price and the actual amount of production. The economic level of output is produced by a

generator between its minimum and maximum economic capacity.154

The output gap calculation used the same method as last year.155,156 We

implemented a multi-stage process to determine the economic output level for a unit

for output gap evaluation.

In the first stage, we determined if the unit would have recovered its start-up, no-load,

and incremental costs if ran for its minimum run time at the dispatch point dictated by

154 The MMU calculates this metric by including all resources’ total (reference level) capacity when calculating output gap percentages. 155 The metric is based on the approach used by Potomac Economics. 156 Under this method, units were grouped into two categories: economic units that are not committed, and committed units that are dispatched at lower levels than their economic level.



the prevailing day-ahead energy price. In the second stage, if a unit was economic

for commitment, we then identified the economic level of incremental output during

hours when it was economic to produce energy based on market prices. To reflect

the timeframe in which commitment decisions are actually made, we used day-ahead

prices for non-quick-start resources and real-time prices for quick-start resources in

assessing the output gap.

This year the MMU still considered the 17.5 percent conduct threshold for the

frequently constrained areas and the 25 percent conduct threshold for the rest of the

footprint to reflect the actual thresholds used in economic withholding mitigation.157

In order to account for the discrepancy between a resource’s offered capacity and the

dispatched amount (because of possible limitations in real-time market conditions

such as transmission constraints, operator actions or ramp limitations, virtual

participants), an upward adjustment is made by taking the greater of the day-ahead

scheduled or the real-time dispatched amount to reflect the actual amount of

production.

Note that certain market conditions such as congestion (supplier location), supplier size, or

high demand can create market power and facilitate economic withholding behavior. For

this reason, the output gap is calculated as percentages of total economic output withheld

compared to total reference capacity for the SPP footprint and for the two frequently

constrained areas. In addition, the output gap is calculated for the largest three suppliers

(market participant portfolios) in each area comparing the levels to those of the remaining

suppliers. Similar to the last year’s report, the annual calculations were run at varying levels of

demand as a potential market condition that can affect the withholding outcome.

The results in Figure 7—13 below show the SPP footprint-wide monthly levels of the output

gap from 2016 to 2018.

157 The two frequently constrained areas Texas Panhandle and Woodward stayed as such until April 1st after which Lubbock was designated as the only frequently constrained area in the SPP footprint. The following charts reflect this situation.



Figure 7—13 Output gap, monthly

Though slightly higher in 2018 compared to previous years, the output gap remained low

overall averaging less than 0.03 percent in all months, reflecting a high level of participation

in the market, overall. The increase in May was due to two resources including opportunity

cost adders only in their market offers without reflecting them in their mitigated offers. This

was also the reason for the slight increases in September and October. These impacts are

shown as the light purple bars in Figure 7—12. When controlling for these impacts, the

output gap remained low overall averaging less than 0.01 percent in all months.158

Figure 7—14 displays the output gap calculated by demand level and participant size for the

entire SPP market footprint, while Figure 7—15 shows the output gap for the Lubbock

frequently constrained area.159 The Woodward and Texas Panhandle frequently constrained

areas were only active for the first three months of 2018, with results of those frequently

constrained areas being insignificant, they are not included as a chart in this section. In

general more output is expected to be withheld at higher demand levels or by larger

suppliers. However, at times output may also be withheld in low load periods, as prices are

often negative during the lowest 20 percent of load hours.

158 These impacts have been excluded from the following charts in this section. 159 The Lubbock frequently constrained area was activated on April 1, 2018, therefore data shown only covers April through December 2018.

0.00%

0.01%

0.02%

0.03%


Out

put

with

held

Actual output gap OCC impacted gap



Figure 7—14 Output gap, SPP footprint

Figure 7—15 Output gap, Lubbock frequently constrained area

There was no output gap in the new Lubbock frequently constrained area for any demand

and supplier levels. In general, the results indicate a very low level of (economic) output

withheld in the SPP footprint, less than 0.01 percent on annual average level. These

outcomes are consistent with expectations of competitive market conduct.

0.00%

0.05%

0.10%

0.15%

0.20%

Top 3 suppliers Other suppliers Top 3 suppliers Other suppliers Top 3 suppliers Other suppliers

< 20th percentile between 20-80th percentiles > 80th percentile

Cap

acity

with

held

2016 2017 2018

0.00%

0.05%

0.10%

0.15%

0.20%

Other suppliers Top 3 suppliers Other suppliers Top 3 suppliers Other suppliers Top 3 suppliers

< 20th percentile between 20-80th percentiles > 80th percentile

Cap

acity

with

held

Total2018



7.2.4 UNOFFERED GENERATION CAPACITY (MEASURE FOR POTENTIAL PHYSICAL WITHHOLDING)

As part of the competitive assessment, we also looked into the potential physical withholding

behavior by generators throughout the 2016 to 2018 period. Physical withholding refers to a

conduct where a supplier derates a resource or otherwise does not offer it into the market.

Physical withholding may include intentionally not following dispatch instructions, declaring

false derates or outages, refusing to provide offers, or providing inaccurate capability

limitations. Any economic generation capacity that is not made available to the market

through a derate, outage, or otherwise not offered to the market is considered for this

analysis.160,161

We classified total economic capacity that was derated from respective reference levels by

reason and duration. Deratings can take the form of planned outages approved in SPP’s

outage scheduling system, forced outages, or any undesignated unoffered capacity.162 Any

deratings from reference levels including partial deratings are considered in this analysis.

Derates were divided into short-term and long-term. Those with less than seven days

duration were classified as short-term and the rest as long-term. This is because the

economic capacity that was not offered short-term has more potential for physical

withholding relative to long-term derates as it would be less costly—because of loss of sales—

for a supplier to withhold capacity for a short duration of time.

As in the case for economic withholding, potential for physical withholding is also affected by

various market conditions at the time offers are made including location (congestion),

supplier size, or demand levels. Larger suppliers would be in a more advantageous position

160 This analysis, in part, draws on “Assessment of the Market Monitoring Metrics for the SPP Energy Imbalance Service (EIS) Market,” Potomac Economics, December 2010 and “2016 State of the Market Report For the New York ISO Markets,” Potomac Economics, May 2017. 161 Economic capacity is determined in a similar way as in the output gap analysis in Section 7.2.3 by comparing resource’s (cost-based) mitigated offer to the prevailing locational price. 162 The planned maintenance outages by nuclear generation and unoffered capacity by hydro, wind and solar is excluded in this analysis.



to exercise market power. During tight market conditions, suppliers have more incentive and

opportunity to physically withhold capacity for strategic reasons. In addition, scheduling

maintenance outages in high demand periods may indicate a strategic behavior to create

artificial shortages.

In the assessment, we considered derated and unoffered economic capacity both in day-

ahead and real-time. Similar to the output gap analysis, the commitment decisions were

made based on day-ahead market outcomes for non-quick-start units and real-time market

outcomes for quick-start units.163 The unoffered capacity is calculated as the difference

between the unit’s economic capacity164 and its offered maximum economic capacity

operating limit during intervals when the unit was deemed economic (i.e., covering its costs

given the clearing price).

The following figures show unoffered economic capacity as percent of total resource

reference levels by month for the SPP footprint, for the Lubbock frequently constrained area,

and by supplier (participant) size against varying load levels.165

Figure 7—16 Unoffered economic capacity

163 See Section 7.2.3 for explanation of this method. 164 Bounded by a resource’s reference level. 165 Unoffered capacity percentages are calculated out of the total reference levels of the corresponding area (i.e., SPP footprint or each of the frequently constrained areas).

0%

1%

2%

3%

4%

5%

6%

7%

8%


Perc

ent

of c

apac

ity

Short-term outage Long-term outage All other unoffered Gas, internal combustion unoffered



Figure 7—16 shows that on an annual average basis the total unoffered capacity equaled 2.0

percent in 2016, 1.9 percent in 2017, and 3.2 percent in 2018.

The figure also shows that the majority of the outages were long-term and concentrated

during the fall and spring shoulder months. When short and long-term outages were

excluded from the averages, the remaining unoffered capacity amounts to 0.22 percent, 0.23

percent, and 0.39 percent for 2016 through 2018, respectively. From a competitive market

perspective, the results generally indicate reasonable levels of total unoffered economic

capacity. The latter (net of outages) results, which are very low, could also be interpreted to

indicate pressure on market participants, particularly on coal-fired resources, to offer—and

maintain commitments—given their long-term coal contracts.166 The general high levels of

self-scheduling of supply offers in the SPP market could be another contributing factor in this

outcome.

Figure 7—17 shows that short-term outages by either large suppliers or others do not rise with

increasing load across the SPP footprint.

Figure 7—17 Unoffered economic capacity at various load levels, SPP footprint

On the other hand, unoffered economic capacity of gas (peaker) units by large suppliers rises

with increased load albeit at very low levels of 0.03 percent, 0.04 percent, and 0.06 percent

166 Coal-fired resources may prefer to offer at times even below their mitigated offers to guarantee to be scheduled or dispatched. See also the offer price markup analysis earlier in Section 7.2.1 and the negative markup results reported therein.

0%

1%

2%

3%

4%

5%


< 20th percentile between 20-80th percentiles >80th percentile

Uno

ffere

d e

cono

mic

cap

acity




of respective load levels. The unoffered economic capacity of gas units by other suppliers

increases even at lower levels against load. Unoffered economic capacity compared to load

is more apparent for the remaining resource types, however, it does not exceed 0.22 percent

for either of the supplier groups which was the same as in 2017.

Another take away from the results is that while long-term outages constitute the majority of

total outages in the SPP footprint at higher load levels, short-term outages play a larger role

in the Lubbock frequently constrained area167, as shown in Figure 7—18. As with the output

gap metrics, the Woodward and Texas Panhandle frequently constrained areas were only

active for the first three months of 2018, with results of those frequently constrained areas

being insignificant, they are not included as a chart in this section.

Figure 7—18 Unoffered economic capacity at various load levels, Lubbock area

The Lubbock frequently constrained area shows the dominance of larger participants in

declaring long-term and short-term outages.

The SPP-wide generation outage data168 (see Section 3.1.4) show that most long-term

outages were for maintenance (70 percent). Of those outages, 77 percent were scheduled

during shoulder months. Out of the short-term outages, approximately 33 percent were

167 The Lubbock frequently constrained area was activated on April 1, 2018, therefore data shown only covers April through December 2018. 168 Covering all resources in the SPP market including nuclear, hydro, wind and solar generation.

0%

5%

10%

15%

20%

25%


< 20th percentile between 20-80th percentiles >80th percentile

Uno

ffere

d e

cono

mic

cap

acity




forced outages, with close to half of them in shoulder months. The results are generally

consistent with competitive market conduct.

7.3 OFFER BEHAVIOR DUE TO MITIGATION THRESHOLD

As discussed in the 2016 State of the Market report, the MMU observed inefficient market

behavior with regard to the mitigated threshold. The MMU recommended that mitigation

measures for resources committed for a local reliability issue be treated separately from the

mitigation measures for economic withholding.169 Resources that fall into this category are

not subject to the three tests associated with economic withholding, which is appropriate.

The MMU submitted revision request 231170 to the Market Working Group in May 2017 to

address this issue. In that revision request, the MMU proposed converting the 10 percent

threshold for local reliability mitigation to a 10 percent cap. The change was implemented in

December 2018 and caps offers at the mitigated offer plus ten percent for local

commitments.

7.4 START-UP AND NO-LOAD BEHAVIOR

Analysis of no-load and start-up offers showed that many market participants made start-up

and no-load offers considerably above their mitigated offer levels (see Figure 7—19).

Nonetheless, start-up mitigation only occurred in less than three percent of intervals in 2018

and day-ahead mitigation accounted for 84 percent of the total start-up cost mitigation.

These figures were similar to results in 2017.

169 In its 2016 report, the MMU recommended converting the 10 percent mitigation threshold for local reliability commitments to a 10 percent cap. 170 Revision request 231 mitigation of locally committed resources.



Figure 7—19 Mitigation start-up offer mark-up by fuel category

MMU’s further analysis and discussion starting in 2017 indicated that start-up mitigation was

most frequently the result of major maintenance costs inherent in the repeated heating and

cooling process of starting-up and shutting down units that were included in market based

offers but not mitigated offers. As the purpose of the mitigated offer is to prevent undue

costs not directly tied to a commitment or dispatch decision from being imposed on the

market, revision request 245 was developed and submitted in late 2017.171 This revision

request allows for the inclusion of major maintenance costs that can be directly tied to the

number of run hours or starts to be included in the mitigated offers for start-up and no-load

and properly evaluated by the market clearing engine in determining commitments.

Revision request 245 was approved by the FERC in October 2018.172 Major maintenance

costs can be included in mitigated start-up and no-load offers in April 2019.

7.5 GENERATOR RETIREMENTS

As discussed earlier in this report, the SPP market has currently a significant amount of excess

generation capacity. Furthermore, the substantial incoming—and projected—new capacity,

primarily wind, is likely to exacerbate this situation absent significant amount of generator

retirements. While the SPP tariff provides for a 12 percent capacity (reserve) margin for the

171 RR245 Mitigated Start-Up and No-Load Offer Maintenance Cost. 172 FERC Order Accepting Tariff Revisions Docket No. ER18-1632-001, October 18, 2018.

0%

100%

200%

300%

400%

500%

600%

700%

800%

900%

1000%

Coal Gas, CC Gas, steam Gas, turbine/engine Oil



footprint, the MMU’s calculated peak available capacity metric discussed in Section 6.1.3

point to levels well above the required level, at 35 percent. Calculating reserve margin

requirements is based on long established standards and provide sufficient cushion for long

term reliability whereas significant excess capacity impose additional costs on the market.

When such excess capacity exists, functioning competitive markets send signals mainly

through (low) energy prices173 indicating a need for the exit of that excess capacity.

A recent wave of generator retirements, particularly of coal-fired generation, has been widely

observed throughout the country. The SPP market should be expected to follow this trend

because of excess capacity, aging fleet, and cost disadvantages of certain types of

generation technologies vis-à-vis the prevailing market prices.

Considering these developments, and starting from 2017, the MMU has been involved with

the SPP’s initiative to develop a new process for evaluating generator retirement applications.

The MMU has approached this primarily from market economics and market power

perspectives. In fact, a strategically motivated generator retirement particularly in a

congested area may create structural and behavioral market power issues that may

potentially amount to physical withholding. Creating a shortage through a retirement may

lead to sustained price spikes and/or may benefit generators affiliated with the retiring

generator(s). The MMU is already evaluating generator retirements from reasonableness,174

market impact, and physical withholding perspectives. However, given the expected market

developments and increased rate of retirements, a streamlined process of collecting

generator specific (financial) data would greatly enhance the MMU’s review of retirements

and would provide sufficient time to recommend and develop appropriate mitigation

measures. This way, based on the MMU’s impact assessment, measures can be identified to

alleviate or preempt any detrimental impact such as transmission upgrades, frequently

constrained area designation, or stricter mitigation rules.

The MMU has already informed SPP staff and the market participants about its proposed

review process through internal meetings and presentations at the Market Working Group,

and will continue to provide feedback as necessary.

173 As highlighted earlier in Section 7.2.1, market offers are less than mitigated offers, which contributes to price formation. 174 Through both technical and economic justifications.



7.6 COMPETITIVE ASSESSMENT SUMMARY

Overall, structural and behavioral metrics indicate that the SPP markets have been very

competitive over the last several years. The market share, HHI, and pivotal supplier analyses

all indicate minimal to moderate potential structural market power in SPP markets outside of

areas that are frequently congested.

Market share indicators increased from minimal to lower moderate levels after Great Plains

and Westar merged to form Evergy, Inc. in June 2018. Furthermore, the percent of the hours

where the SPP market was unconcentrated declined from 100 percent to 88 percent. These

analyses both indicate a moderate potential for general structural market power in SPP

markets outside of areas that are frequently congested. The MMU will continue to evaluate

structural market power concerns going forward.

Meanwhile, there was one frequently constrained area in 2018 where the potential concerns

of local market power were the highest. 175 Ongoing analysis shows existing mitigation

measures have been an effective deterrent in preventing pivotal suppliers from unilaterally

raising prices.

Behavioral indicators were also assessed through the analysis of actual offer or bid behavior

(i.e., conduct) of the market participants and the impact of such behavior on market prices to

look for the exercise of market power. One such indicator, the negative offer price mark-ups

in 2018 continued to show increased negative offer levels similar to those in 2017. This also

points to a continued declining annual trend of off-peak and on-peak average offer price

markups. What is striking for 2018 is that not only the relative and absolute amounts with

respect to 2017 have increased but the pattern has changed to include winter and summer

peak period months as well. This could occur where—particularly coal—generators decide to

offer below their marginal cost to maintain commitments or when wind units—that may have

negative mitigated offers because of subsidies related to production tax credits—constitute a

sizeable share in total resource portfolio. As stated earlier, a lower offer price markup level in

itself would indicate a competitive pressure on suppliers in the SPP market, the observed

175 Texas Panhandle and Woodward maintained their frequently constrained area status until April 1, 2018 after which Lubbock was designated as the only such area in the SPP footprint.



continuous and deepening downward trend may raise questions about the commercial

viability of generating units and the possibility of generation retirements.

Similar to 2017, economic withholding mitigation remained infrequent. In particular, the

incremental energy mitigation was extremely low in both markets, at less than 0.01 percent in

both the day-ahead real-time markets in 2018. The overall mitigation frequency of start-up

offers in 2018 were at the lowest levels since market started in 2014 as the combined

frequency of mitigation of start-up offers for day-ahead, reliability unit commitment, and

manual commitments decreased to 2.8 percent in 2018 from 3.2 percent in 2017.

Meanwhile, the combined frequency of no-load, operating reserve, and incremental energy

mitigation decreased from their in 2017 levels and remained at negligible levels.

The system wide monthly output gap results show a very low-level—less than 0.03 percent—of

economic withholding in all months in 2016 to 2018 across the SPP footprint. These low

overall levels of economic output withheld are consistent with competitive market conduct.

The new metric introduced last year, the average unoffered economic capacity was around

two percent in the 2016 to 2017 period and was just over three percent in 2018. The majority

of the outages were long-term, and were primarily the result of maintenance in the spring

and fall shoulder months. Furthermore, the short-term outages by either large suppliers or

others do not rise with increasing load across the SPP footprint. The results are generally

consistent with the workings of a typical competitive market.

Meanwhile, the very low level—less than one percent—of unoffered capacity net of outages

could indicate a pressure on market participants to offer—and maintain commitments—given

their longer term agreements. The general high levels of self-committing supply in the

market could be another factor in the low levels of unoffered capacity.

Overall, the SPP Integrated Marketplace provides effective market incentives and mitigation

measures to produce competitive market outcomes particularly during market intervals

where exercise of local market power is a concern. The competitive assessment in this report

provides evidence that market results in 2018 were workably competitive and that the market

required mitigation of local market power infrequently to achieve those outcomes.

Nonetheless, mitigation remains an essential tool in ensuring that market results are

competitive during periods when such market conditions offer suppliers the potential to

abuse local market power.



Southwest Power Pool, Inc. Recommendations Market Monitoring Unit


8 RECOMMENDATIONS

One of the core functions of a market monitor is “to advise the Commission, the RTO or ISO,

and other interested entities of its views regarding any needed rule and tariff changes.”176

The MMU accomplishes this responsibility through many forums, including but not limited to

active participation in the SPP stakeholder meetings process, commenting on FERC notices

of proposed rulemakings, submitting comments at FERC on SPP filings, and making

recommendations in the Annual State of the Market report. This section outlines the MMU

recommendations to SPP and stakeholders to address our concerns with the current design,

rules, and processes.

This section highlights new recommendations, clarifies existing recommendations, and

outlines recommendations made in prior reports. For each recommendation, we present the

reasons for our recommendation and assign a priority. We also identify the current status of

previous recommendations. Overall, SPP and its stakeholders have made significant

progress on most outstanding MMU recommendations. Section 8.4 lists the status of past

and current annual report recommendations.

8.1 NEW RECOMMENDATIONS

The following recommendations are new for 2018. We have five new recommendations

including recommendations on parameter values and market power, on how SPP calculates

credit requirements in light of a credit situation that occurred in the PJM markets in 2018,

pricing capacity for uncertainties, improving results in the transmission planning processes,

and improving mileage price formation

8.1.1 LIMIT THE EXERCISE OF MARKET POWER BY CREATING A BACKSTOP FOR PARAMETER CHANGES

The market is currently vulnerable to the exercise of market power by the manipulation of

their resources’ non-dollar based parameters. Where there is no market power, a seller

cannot control price because other sellers are competing for revenue. Market power is a

market participant’s ability to manipulate price by manipulating either supply, demand, or

176 As defined by FERC in Order No. 719.



some amount of both through either dollar or non-dollar based offers. Where a seller has

market power, the seller can control price.

Market participants can exercise market power by manipulating a resource’s non-dollar-

based parameters. For instance, a resource can manipulate energy price by reducing its

maximum operating limit from its actual limitation so that it reduces the supply of energy and

raises the price. Non-dollar-based parameters should not be manipulated to affect market

clearing. Although SPP’s tariff and protocols have well-defined expectations and precise

limitations for the basis of dollar-based offer components in the presence of local market

power, the expectations for the basis of non-dollar-based offer components are much less

clearly defined and the limitations are much less precise.

The MMU recommends that SPP strengthen the language regarding non-dollar-based

parameters so that the expectation for the basis of these values is clear and the potential

exercise of market power is much more limited. The expectations for the basis of the

parameters should be clear and well-defined. Changes to these parameters should be

limited to actual capability and should be applied, at a minimum, in the presence of market

power. One option would be to require parameters to reflect actual limitations always.

Actual limitations could include physical and environmental limitations and potentially other

true and verifiable limitations. Another option could be to automatically apply parameter

mitigation in the presence of market power similar to the automatic mitigation of dollar-

based offer components.

8.1.2 ENHANCE CREDIT RULES TO ACCOUNT FOR KNOWN INFORMATION IN ASSESSMENTS

In 2018, GreenHat Energy, a financial-only market participant in PJM, defaulted on its

portfolio of congestion hedging products in the PJM markets. Current estimates of the

default exposure exceed $400 million.177 GreenHat Energy acquired its portfolio in

compliance with the PJM credit policy, which due to its design, required GreenHat Energy to

post less than $1 million in financial security. The disconnect between the projected loss and

the required financial security stems largely from PJM’s credit policy’s assessment of the

historic congestion patterns associated with GreenHat Energy’s positions. Planned

177 https://pjm.com/-/media/committees-groups/committees/mic/20190206/20190206-item-01a-informational-update-ferc-order-denying-waiver-request.ashx, page 5.





transmission expansion changed the historic congestion patterns in the PJM markets, which

caused GreenHat Energy’s congestion hedging portfolio to become unprofitable. Ultimately,

GreenHat Energy defaulted and has forced the rest of the PJM market to absorb the costs.

While the SPP market is different from the PJM market, SPP’s credit policy is similar to PJM’s

in some respects. For instance, SPP’s financial security requirements for transmission

congestion rights are based only on historic congestion patterns even when significant

transmission upgrades are planned or have occurred. For instance, as noted in the 2017

Annual State of the Markets report, congestion patterns in SPP shifted significantly after

installation of the phase shifting transformer at the Woodward substation, which changed

congestion patterns throughout the footprint.178 This known event was not factored into

SPP’s financial security requirements for transmission congestion rights even though it was

expected to have a significant effect on outcomes. Furthermore, the congestion pattern was

so noticeably shifted that after only a few months the MMU recommended eliminating a

frequently constrained area as a result of the expansion.179 However, SPP’s financial security

requirements only factored in the historic data.180

The MMU has engaged SPP’s Credit Practice Working Group and has contributed to the

dialogue about next steps. SPP and its stakeholders generally agree that updating the SPP

credit policy to protect from exposure such as that experienced in PJM is a priority. The

MMU concurs with their assessment and recommends that changes to the credit policy be

developed to protect stakeholders from risks of default, particularly with respect to

information that is known such as transmission expansion.

8.1.3 DEVELOP COMPENSATION MECHANISM TO PAY FOR CAPACITY TO COVER UNCERTAINTIES

Because of unexpected variations in wind output, SPP operators often needs to manually

commit resources (often in excess of 50 units) in order to meet instantaneous load capacity

178 https://www.spp.org/documents/57928/spp_mmu_asom_2017.pdf, page 141. 179 https://www.spp.org/documents/56330/fca%202017%20report%20-%20final.pdf 180 Per SPP Tariff Attachment X, Section 5A.2.


https://www.spp.org/documents/56330/fca%202017%20report%20-%20final.pdf



requirements. Often, however, these resources do not earn enough revenue to cover their

offered costs181 which contributed to an increase in real-time make-whole payments.182

Moreover, resources providing this reliability service are not compensated specifically for the

need. Even when resources are needed so much that they are committed for capacity, the

supplemental reserve price is still very low. While the MMU recognizes that SPP operators

may need to commit units to account for unforeseen circumstances, manual capacity

commitments occur often enough that systematic solutions should be developed. An ideal

design of a ramping product should allow for most of the capacity and stagger needs to be

met with the compensated ramping product instead of the uncompensated headroom. The

large number of manual capacity commitments may indicate that there is a need for a new

uncertainty product beyond just a short-term ramping product that might better address

problems in the 30-minute to 3-hour time frame.183 Developing products to reduce the need

for the large number of manual capacity commitments would help ensure appropriate

compensation is being provided for the reliability services provided.

The MMU has engaged SPP’s Market Working Group and has participated in a high level

discussion on the potential need for an uncertainty or standby reserve product. SPP and

stakeholders are generally interested in exploring this topic further. The MMU concurs and

recommends that further work be done to develop a compensation mechanism or product to

compensate capacity used to cover uncertainty of generation and load.

8.1.4 ENHANCE ABILITY TO ASSESS A RANGE OF POTENTIAL OUTCOMES IN TRANSMISSION PLANNING

SPP’s transmission planning process develops an annual look-ahead plan. This plan

evaluates transmission needs over a 5- and 10-year time horizon. The plan typically evaluates

181 Offered costs included their incremental energy, no load, start-up and reserve offers. When mitigated, the resource offers are replaced with the mitigated offers. 182 See Section 4.2. 183 The exact time periods of potential uncertainty products should be determined through the technical expertise of the stakeholder process.



two scenarios.184 The first case is a base case, and the second case is an emerging

technology case. The process could consider a third scenario. In 2018, stakeholders

requested that SPP staff consider an additional case. The third case would have considered a

shift in environmental regulations—such as a carbon tax or adder—and a change in

technologies/market trends including accelerated deployment of storage devices or electric

vehicles, higher levels of generation retirement, and higher penetration of renewables.185

MMU staff follows market trends to keep abreast of developments, including corporate

announcements on carbon reduction and renewable energy targets. Following these trends

helped shape the MMU’s thinking in terms of how to consider the future generation mix and

the potential need for an additional scenario. The MMU believed that a third scenario would

provide a bookend scenario for the 2020 analysis and supported inclusion of the scenario as

part of the 2020 Integrated Transmission Plan. Ultimately, stakeholders rejected the third

scenario at the Economic Studies Working Group, at the Market Operations and Policy

Committee, and at the Strategic Planning Committee. The primary reason cited was the cost

to perform the study.

While the transmission planning process theoretically can include a third scenario, in practice

it does not have the flexibility to include one given that cost of including a third scenario was

not unique to the 2020 planning process. This appears to be a significant shortcoming in the

planning process as the study process is only limited to two potential scenarios in its 10-year

look ahead. This limits the range of potential outcomes that SPP could study.186 We

recommend that SPP enhance their study process to allow the ability to study a range of

potential outcomes. If such range of potential outcomes are not captured in a third case

study, the MMU recommends to factor them into either an existing case study or potentially

as part of the 20-year assessment. The wider the range of possibilities studied, the more

robust the results will be.

8.1.5 IMPROVE REGULATION MILEAGE PRICE FORMATION

In addition to regulation capacity payments, resources that are deployed for regulation also

receive payments for costs incurred when moving from one set point instruction to another.

184 These scenarios are also known as futures cases. 185 This scenario was similar to a scenario used in the MISO transmission plan. https://cdn.misoenergy.org/MTEP19%20Futures%20Summary291183.pdf. 186 The MISO transmission planning process studies four cases.




These mileage payments are paid directly through regulation-up and regulation down

payments in the day-ahead market. The market calculates a mileage factor for both products

each month that represents the percentage a unit is expected to be deployed compared to

what it cleared. If a unit is deployed more than the expected percentage, then the unit is

entitled to reimbursement for the excess at the regulation mileage marginal clearing price. If

the unit is deployed less, the position must be bought back.

The MMU has identified a mileage price inefficiency. Mileage prices are not set by the

marginal cost of mileage like other products. Instead, units are cleared for regulation based

on what are known as regulation service offers. These service offers are calculated by taking

the competitive offer for regulation and adding the mileage offer to it after discounting the

mileage offer by the mileage factor.

The MMU has observed instances where resources cleared with regulation-down competitive

offers of $0 and mileage offers just under $50. These units consistently cleared with this offer

strategy because the service offer was near $10.50 (e.g. 21 percent*$50) which was lower

than the services offers of other resources offering in higher competitive offers. For instance,

another resource may offer in a $12 competitive offer and $0 mileage offer. This would make

that resource’s service offer $12 ($12+0*21 percent). In this circumstance, the resource with

the highest service offer will set the regulation-down price at $12, but the mileage offer will

be $50, which would set the mileage clearing price.

The MMU is also concerned that the mileage clearing price does not correctly reflect price

formation of mileage given that this price does not represent the expected or realized

mileage deployment. Furthermore, the MMU is concerned that participants with resources

frequently deployed for regulation will have an incentive to inflate the mileage prices by

offering in $0 regulation offers and high mileage offers. The MMU also identified systematic

overpayment of regulation mileage in the day-ahead market, which appears to be the result

of the mileage factor being set consistently too high relative to actual mileage deployed.

The MMU discussed our concerns with the Market Working Group at its August meeting.

While an action item was developed requesting SPP staff and the MMU to review the

effectiveness of the regulation mileage pricing process and present further options, no

additional work has been done since that time. We recommend that SPP staff review the

performance of regulation mileage, and develop potential approaches to improve regulation

mileage price formation. Furthermore, we recommend that SPP staff consider adjusting the



mileage factor. We believe that SPP staff and stakeholders should include these items as part

of its analysis and change development processes for moving forward.

8.2 CLARIFICATION OF PREVIOUS RECOMMENDATIONS

8.2.1 ADDRESS INEFFICIENCY CAUSED BY SELF-COMMITTED RESOURCES

The MMU and market participants have identified several reasons why they self-commit

resources in the market. Some of these reasons include contract terms for coal plants, low

gas prices that reduce the opportunity for coal units to be economically cleared in the day-

ahead market, long startup times, overtime costs, increased major maintenance costs, and a

risk-averse business practice approach. However, it is imperative to minimize the need to

self-commit resources to realize the full benefits of SPP’s market. While there may not be a

single reason causing market participants to self-commit resources, there can be ways that

SPP and its stakeholders can work to minimize the need to self-commit.

For instance, long lead-time and long run-time resources are often self-committed and

contribute to depressing prices in the SPP market. These resources are not appropriately

evaluated in the current market structure and can be committed by market participants

during uneconomic periods. The current clearing engine logic does not provide

commitments beyond the 24-hour period of the next operating day. The creation of a market

process that economically evaluates resources over a longer period will allow for more

efficient market solutions, as well as decreased production costs.187

In the current design, a resource that is required to run for multiple days is not evaluated by

the day-ahead market to see if the resource is economic over its minimum run-time. The

clearing engine may see that it is economic on the first day and issue the commitment, and

then in future days the resource will stay on until its minimum run-time is met even if it is

uneconomic. As such, many resources that have multi-day minimum run times avoid the

market clearing process and instead self-commit in the market based not on an evaluation by

the market, but on their own evaluation of market conditions. This is not the optimal solution

187 This would be different from the current multi-day reliability unit commitment process.



for the SPP market as it removes the ability for the SPP market software to evaluate and

commit the resources economically relative to all other resources in the market.

Adding multi-day unit commitment logic is at the top of the current stakeholder market

design initiative list and has been discussed in 2018 and 2019 in the stakeholder process.

SPP staff has proposed a multi-part approach to address multi-day unit commitment. First,

they have indicated that they prefer to provide a multi-day forecast of prices or schedules as

it would be quicker, easier, and less expensive to implement. The multi-day forecast would

serve to provide information to aid market participants that self-schedule to do so in periods

that would be more favorable for their resources and for the market. Second, SPP staff has

indicated that after the multi-day forecast was made available, they would consider

developing a multi-day unit commitment process.

The MMU is currently in the process of reviewing the SPP staff proposal to provide multi-day

forecast information. At this time, it is not clear if the benefits of this approach outweigh costs

and concerns. For instance, the accuracy of the forecast significantly declines the further out

the forecast goes. Thus, it is not clear how useful the information will be to market

participants if the accuracy of the multi-day forecast is low. Furthermore, there have been

several recent statements by multiple SPP market participants indicating corporate goals to

increase renewable generation targets or to set carbon reduction goals over the coming

years.188 This has the potential to change the resource mix such that many of the longer run

resources may no longer participate in the market in future years.

Given these concerns, the MMU clarifies its recommendation. Progress has been made in

addressing self-commitment reasons by allowing the inclusion of major maintenance in

mitigated offers effective April 18, 2019 and changing local mitigation to a threshold of 10

percent effective December 2018. Broadly, the MMU recommends that SPP and its

stakeholders continue to explore and develop ways to reduce the incidence of self-

commitment of resources outside of the market solution. With regards to the development

188 For instance, Omaha Public Power District (OPPD) set a long-term goal of providing at least 50 percent of its retail sales from renewable sources and reducing its carbon intensity by 20 percent from 2010 to 2030. Xcel Energy Inc. announced that it plans to completely decarbonize its power supply portfolio serving eight states by midcentury, first cutting carbon emission by 80 percent by 2030, from 2005 levels in its utility service territories in Colorado, Michigan, Minnesota, New Mexico, North Dakota, South Dakota, Texas and Wisconsin. AEP announced in February 6, 2018 that it will pursue a goal to reduce its CO2 emissions of its generating units 60 percent by 2030 and 80 percent by 2050 from 2000 levels. (See Section 6.2.1 for discussion).



of multi-day forecasting of prices or schedules or a multi-day unit commitment process, the

MMU supports these as attempts to reduce self-commitment of generation, but we will

carefully weigh the benefits and costs before we support any specific proposal especially in

light of the accuracy of forecasted information and the potential changing generation mix.

We continue, however, to view reducing self-commitment of generation as a high priority for

SPP and its stakeholders as this will enhance market efficiency and improve price signals.

8.2.2 ENHANCE COMMITMENT OF RESOURCES TO INCREASE RAMPING FLEXIBILITY

As noted in last year’s annual report, increasing ramping flexibility is becoming increasingly

important to integrate higher levels of renewable generation.189 For instance, one way to

address ramping flexibility is to ensure sufficient ramping resources are on-line to meet

changing conditions (see Section 8.3.1 below). Over-commitment of resources in real-time

can reduce ramping flexibility, suppresses prices, and leads to increased make-whole

payments. This can be caused by changing conditions between the time a resource is locked

into a commitment by the market software and the time the resource actually comes on-line.

The MMU recommends that SPP and its stakeholders address this issue by enhancing its

markets rules to enhance the commitment of resources to increase rampoing flexibility. Last

year, the MMU described this as enhancing decommitment of resources. However, having

explored this issue further, the issue is not just about decommitment of resources, but is also

about improving how resources are committed.

For instance, the market software frequently commits resources well in advance of when they

are actually required to start. This commitment is based on the known assumptions and

information available at the time the market engine clears the market. However, conditions

change over time. For example, load forecasts, wind forecasts, and outages change, and

resources trip off-line. Resources are committed because the market software evaluates it to

be profitable over that study period. When conditions change, the resource may no longer

be profitable; however, resources may continue to start-up or to run as the commitment is

189 https://www.spp.org/documents/57928/spp_mmu_asom_2017.pdf, page 193.




not changed or the start-up is not cancelled. This is due, in part, to the current tariff language

that only allows the decommitment of day-ahead market committed resources to prevent

anticipated excess supply conditions or other emergency conditions.

The MMU evaluated a day in June where wind output increased by several thousand

megawatts.190 We identified several quick-start and other short lead time resources that had

day-ahead schedules that ran in real-time.191 Real-time system marginal energy costs were

about $35/MWh less than day-ahead costs during the peak hours. Other markets have

different commitment rules that would help increase market flexibility. For instance, PJM and

California ISO delay the commitment of short start resources and quick-start resources until

the real-time market, making the day-ahead commitment instruction advisory. Furthermore,

the California ISO can also decommit resources after a resource’s minimum run time has

been met, ignoring the day-ahead commitment period. These commitment rules can help

increase the flexibility of resources in SPPs markets and help improve pricing outcomes.

The MMU recommends that SPP and stakeholders explore options, such as those noted

above, to enhance commitment of resources to increase flexibility. We view this as a high

priority item.

8.2.3 FURTHER ENHANCE ALIGNMENT OF PLANNING PROCESSES WITH OPERATIONAL CONDITIONS

Enhancing the accuracy of planning processes with operational realities enables SPP and its

members to more effectively plan for future system needs and conditions. Many of the

challenges outlined in this report—including congestion, negative prices, and low generator

net revenues—as well as improvements—such as transmission additions—are, in part, a

reflection of planning decisions. The more the planning process can learn from and

incorporate operational information, the more planning can identify and address concerns in

190 The MMU evaluated June 12, 2018. Wind was approximately 4,000 MW higher in real-time than expected in the day-ahead market. 191 The day-ahead market committed 25 quick start resources that could be started within 10 minutes or less and an additional 18 resources with start-up times less than four hours.



advance of market operations. SPP and stakeholders over the past few years have worked to

improve and align the planning and operational processes. For instance, the Integrated

Transmission Planning manual includes an assessment for persistent operational needs and

the criteria for identifying these needs.192 In 2017, the MMU identified two areas where

further alignment would be beneficial. We update the status of these recommends and

clarify this recommendation by adding a new item below.

First, in early 2018, SPP proposed to set the interim default variable operations and

maintenance costs for wind resources to $0/MWh in revision request 276. While the

Economic Studies Working Group rejected this revision request and narrowly passed an

amended version of this request, the Market Operations and Policy Committee reversed

direction and accepted the original SPP proposal. Based on the MMU’s review of operational

information, variable operations and maintenance costs for wind resources are closer to

$0/MWh. The 2019 Integrated Transmission Plan used the $0/MWh value. After reviewing

the results, stakeholders have agreed to continue to use the $0/MWh value going forward.

As such, the MMU believes that this recommendation is complete.

Second, it is important that the full set of generation that exits be considered at peak load

conditions. SPP’s deliverability study evaluates deliverable capacity purchases by a load

responsible entity and does not require firm transmission service to support the

capacity.193,194 Thus, we consider SPP to be accounting for this capacity and not

recommending any further action on this item on the resource adequacy

process. Accordingly, we withdraw this recommendation on this basis.

Third, after engaging in the development of the 2020 Integrated Transmission Plan scope in

2018, one of the key takeaways from our involvement in this process was that some of the

proposed economic modeling assumptions, particularly around generator retirement age

and renewable resource accreditation, differed from prevailing data trends. With regards to

retirement age, the MMU performed an analysis of public Energy Information Administration

192 See SPP Integrated Transmission Planning Manual (dated October 17, 2018), including sections 4.4, 5.3.4, and 6.1.4 (available at https://www.spp.org/Documents/22887/ITP%20Manual%20version%202.3.docx). 193 See SPP Open Access Transmission Tariff, Sixth Revised Volume, Attachment AA - Resource Adequacy, Section 10. 194 Hence, based on the information brought to our attention with regard to SPP’s deliverability study we clarify our earlier point made in last year’s report such that SPP excludes resources without firm transmission for resource adequacy calculation.

https://www.spp.org/Documents/22887/ITP%20Manual%20version%202.3.docx



(EIA) data and presented on prevailing trends. With regards to renewable accreditation, we

reviewed other market practices, and SPP’s actual wind and solar production data. This

information can be accessed by SPP staff and the MMU. Notably, SPP staff requested access

to the retirement information used by the MMU and performed an even more detailed

analysis of the EIA data, breaking down natural gas retirements by technology. The MMU

supported this SPP analysis.

Based on our experience in this process, we recommend that SPP staff continue to

collaborate with the MMU to build upon the work in 2018, by reviewing data and best

practices in other markets, factoring in market conditions and trends in decision-making

process, and performing analysis to support and inform assumptions. Doing so will help to

improve assumptions and ultimately study results.

8.3 ADDRESS PREVIOUS RECOMMENDATIONS

The MMU has provided recommendations to improve market design in each of our previous

Annual State of the Market reports since the launch of the integrated marketplace in 2014.

Overall, SPP and its stakeholders have found ways to effectively address many of our

concerns. However, there remain outstanding recommendations. A description of each of

these recommendations, their currents status, and our assessment of their priority are

outlined below.

8.3.1 DEVELOP A RAMPING PRODUCT

A ramping product that incents actual, deliverable flexibility can send appropriate price

signals to value resource flexibility. This resource flexibility can help prepare the system for

fluctuations in both demand and supply that result in transient short-term positive and

negative price spikes.

Today, the SPP real-time dispatch engine solves for only the current interval and has no look-

ahead logic to ensure that there is enough rampable capability to meet the needs of future

intervals. This can cause quick-ramping resources to be dispatched to their maximum limits

in one interval, and then, in the next interval, there is a shortage of ramp because the only

resources able to move have slower ramp rates. In these cases, SPP can have plenty of

capacity on-line, but not enough rampable capacity, which can result in scarcity pricing.



SPP’s scarcity events are very short-term, transient events that are frequently a result of a

shortage of rampable capacity. A ramping product will compensate resources for holding

back capability in one interval so it can be used as energy in a future interval. This will reduce

the frequency of scarcity events and provide value to the resources providing ramping

capability.

Both the California ISO and Midcontinent ISO have designed and successfully implemented

ramping products. In 2018 and into 2019, SPP, stakeholders, and the MMU have discussed

design elements of a ramping product. SPP staff hopes to have a design approach approved

by mid-year.

We agree with the assessment of SPP and its stakeholders that development of a ramping

product is a high priority initiative and continue to recommend that a ramping product

design be completed in 2019.

8.3.2 ENHANCE MARKET RULES FOR ENERGY STORAGE RESOURCES

With the increase in wind penetration in the SPP market, there is not only a need for resource

flexibility, but also for storage where the frequency of negative prices has increased, as

discussed in Section 4.1.5. Stored energy resources have the potential to address both the

need for flexibility and reduce the incidence of negative prices. However, SPP’s current tariff

does not easily allow these resources to integrate in our market. In order to capture the

benefits of these new technologies, a new market design needs to be developed.

FERC issued Order No. 841 in February 2018 to reduce barriers to participation and to

develop a participation model for electric storage resources.195 Over the course of 2018,

SPP, stakeholders, and the MMU worked on changes that would comply with FERC Order No.

841. These changes passed the October SPP Board meeting and SPP filed the changes with

FERC in December. The MMU filed supportive comments shortly after. In the MMU

195 https://www.ferc.gov/whats-new/comm-meet/2018/021518/E-1.pdf




comments, we identified multiple areas of further work. These areas include further

enhancements to electric storage integration include addressing the potential for storage

resources to exercise downward market power, the potential for market storage resources

(MSR) to manipulate the transmission market, possible market design gaps regarding major

maintenance and quick-start resource requirements, and the inefficient commitment of non-

continuously dispatchable resource requirements in relation to market storage resources.196

The MMU views integration of storage resources in the SPP markets as an ongoing high

priority as several outstanding items beyond compliance with FERC Order No. 841 need to

be addressed in order to fully integrate electric storage resources in the SPP markets.

8.3.3 ADDRESS INEFFICIENCY WHEN FORECASTED RESOURCES UNDER-SCHEDULE DAY-AHEAD

Our analysis shows that, on average, 85 percent of forecasted wind generation was cleared in

the day-ahead market in 2018, compared to 82 percent in 2017, and 89 percent in 2016. On

average for the year, over 1,000 MW of real-time wind generation was not included in the

day-ahead market in each hour.197 When this happens, we frequently observe day-ahead

prices exceeding real-time prices. While we also observe participants virtual transactions at

wind locations during these times, we find that price convergence in absolute terms is not

improving.198 As a result, under-scheduling of wind reduces the efficiency of the day-ahead

unit commitment process. In this case, other non-wind resources may be overcommitted in

196 Some market storage resources have a non-dispatchable range between their charging range and discharging range. The dispatch calculation for this type of non-continuous dispatch range is much more complicated than the typical linear dispatch calculation. SPP’s current proposal is to commit this type of market storage resource for either charging or discharging. This type of commitment is inefficient because it does not make the whole dispatch range available. For more detail, see Motion to Intervene and Comments of the Southwest Power Pool Market Monitoring Unit, Section I.B.5, Docket No. ER19-460, December 7, 2018. 197 From a reliability standpoint, the reliability unit commitment assesses wind resources at forecasted levels. However, the reliability unit commitment process cannot economically decommit resources scheduled by the day-ahead market. 198 In many cases, we find that virtual transactions are setting prices at wind locations with positive offer prices. In most cases, market participants do not offer wind resources at positive values. As a result, while the virtual transactions may be reflecting missing quantity values at a resource location, they are not reflecting the offer a wind resource would likely have. Moreover, virtual participants may pair virtual offers at resource locations with offsetting virtual bids at other locations. This effectively works to capture congestion differences between the day-ahead and real-time markets, but does not work to converge the energy price component.



the day-ahead market, which results in real-time prices significantly lower than day-ahead

prices when wind generation increases in the real-time market.

Systematic under-scheduling of wind resources in the day-ahead market can contribute to

distorting market price signals, suppressing real-time prices, and affecting revenue adequacy

for all resources. Therefore, the MMU continues to recommend that SPP and its stakeholders

address this issue through market rules changes that address the consequences of the under-

scheduling of forecasted supply resources in the day-ahead market. We consider this a high

priority as it helps to enhance market efficiency and improve price signals.

8.3.4 CONVERT NON-DISPATCHABLE VARIABLE ENERGY RESOURCES TO DISPATCHABLE

In the 2015 Annual State of the Market report, the MMU identified non-dispatchable variable

energy resources as a concern because of their adverse impact on market price and system

operations. These resources exacerbate congestion, reduce prices for other resources,

increase the magnitude of negative prices, cause the need for market-to-market payments,

and force manual commitments of resources that can increase uplifts. Going forward,

resource flexibility is essential to integrate an increasing volume of wind generation in the

SPP market. FERC demonstrated strong support for the elimination of most instances of non-

dispatchable resources with the approval of a rule change for the New England market in

December 2016.199 Furthermore, FERC also rejected California ISO’s proposal to extend the

transition period for protective measures related to non-dispatchable variable energy, thus

requiring dispatchability.200

The Market Working Group passed a proposal by SPP at the February 2018 meeting that

required full conversion of non-dispatchable variable energy resources. The proposal failed

to pass the Market Operations and Policy Committee meeting in April and was appealed to

the SPP board. The SPP board tabled discussion in April. In July, the Market Worked Group

passed a modified version of the proposal that allowed exemptions for resources exercising

rights under PURPA201 and for run-of-river hydro that could not be dispatchable. This

199 See FERC Docket Nos. ER17-68-000 and ER17-68-001. 200 See FERC Docket No. ER17-1337-000. 201 Public Utilities Regulatory Policies Act of 1978



subsequently passed the July Market Operations and Policy Committee meeting and SPP

board.

While the MMU preferred the original version of the proposal that was appealed to the April

SPP Board meeting, the MMU found the modified proposal acceptable and supported it at

the July SPP Board meeting and in the MMU comments with FERC.202 FERC approved the

filing in April 2019.

8.3.5 ADDRESS GAMING OPPORTUNITY FOR MULTI-DAY MINIMUM RUN TIME RESOURCES

Resources with minimum run times greater than two days have the opportunity to game the

market. The current implementation of the market rules limit make-whole payments to the

as-committed market offers for the first two days of a resource’s minimum run time. However,

after the second day, no rule exists to limit make-whole payments for a resource that

increases its offers from the third day onward until the resource’s minimum run time is

satisfied. For resources with minimum run times greater than two days, the market

participant knows that the resource is required to run and can increase their market offers

after the second day to increase make-whole payments.

The SPP board passed a proposal203 at the July meeting that would limit make-whole

payments for any resource with multi-day minimum run times to the lower of the market offer

or the mitigated offer. This limitation only applies for offers falling in hours not accessed by

one of the security constrained unit commitment (SCUC) processes and the resource bid at or

above their mitigated offer on the first day. The MMU supported the proposal. Subsequent

to board approval of the proposal, SPP legal staff identified internally inconsistent tariff

language that the revisions revealed by but did not address. This has required that

additional tariff modifications go through the stakeholder process.

The MMU strongly recommends that SPP and its stakeholders clean up the tariff language

necessary to implement the changes to address the gaming issue and provide an anticipated

202 SPP Market Monitoring Unit comments on ER19-356, https://elibrary.ferc.gov/idmws/common/opennat.asp?fileID=15111404. 203 Revision request 306 (2014 ASOM MWP MMU Recommendation [3-Day Minimum Run Time])

https://elibrary.ferc.gov/idmws/common/opennat.asp?fileID=15111404



implementation date for the system changes. Addressing this matter in a timely manner is a

high priority for the MMU.

8.3.6 REPLACE DAY-AHEAD MUST OFFER, ADD PHYSICAL WITHHOLDING PROVISION

FERC rejected in 2017 SPP’s proposal to remove the day-ahead must offer requirement and

indicated that it would consider removal of the requirement if it were paired with additional

physical withholding provisions.

While the MMU remains concerned with the current day-ahead must offer requirement, we

recommend that further consideration of this issue be a low priority. The MMU will continue

to track market performance concerns related to this provision and will consider raising the

priority on this matter if further issues are identified or current issues are exacerbated.

Otherwise, we regard other matters as having higher impacts to the market and priority for

development at this time.

8.4 RECOMMENDATIONS UPDATE

The table below lists the status of Annual State of the Market recommendations included in

the 2017 report and those that are new to this report. Recommendations closed prior to the

completion of the previous year’s report do not appear in this table. To review closed

recommendations that are not covered in this report, please review earlier reports. All

previous annual reports can be found at https://www.spp.org/spp-documents-

filings/?id=18512.

Figure 8—1 Annual State of the Market recommendations update

Recommendation Report

year Current status

1. Limit market power by backstopping parameter changes

2018 Awaiting stakeholder engagement

2. Enhance credit process to account for known information


3. Develop compensation or product for capacity used for uncertainties


4. Enhance ability for transmission planning to cover range of outcomes






5. Improve regulation mileage price formation

2018 Outstanding Market Working Group action item

6. Reduce self-scheduling in market

2017 Multi-day commitment process – under stakeholder consideration

7. Address under-scheduling of wind

2017 Under consideration by Holistic Integrated Tariff Team

8. Develop ramping product 2017 Under stakeholder consideration 9. Enhance unit commitment

logic 2017 Under stakeholder consideration

10. Develop energy storage design

2017 Awaiting FERC order and follow-up

11. Continue alignment of planning processes with operational conditions (multiple items)

2017

Wind VOM – procedure updated with MMU recommended value, recommendation complete Counting generator capacity – Deliverability study addresses MMU’s concern

12. Local reliability commitment mitigation threshold conversion to a cap

2016 Complete - implemented in December 2018

13. Non-dispatchable variable energy resource transition to dispatchable variable energy resource status

2015 FERC accepted filing in April 2019

14. Improved quick-start logic 2014 Awaiting FERC order

15. Manipulation of make-whole payment provisions (multiple items)

2014

Across midnight hour – awaiting FERC filing Out-of-merit payments – withdrawn by MMU Jointly-owned units – awaiting implementation Regulation – implemented in 2018

16. Day-ahead must offer requirement and physical withholding

2014 Assigned low priority, awaiting stakeholder engagement

17. Allocation of over-collected losses

2014 Awaiting implementation



COMMON ACRONYMS

1073 City of Malden (Mo.) – Board of Public Works

AECC Arkansas Electric Cooperative Corporation

AECI Associated Electric Cooperative, Inc.

AEP/AEPM American Electric Power

ARR auction revenue rights

BEPM Basin Electric Power Cooperative

BRPS Big Rivers Electric Corporation

BSS bilateral settlement schedules

BTU British thermal unit

CC combined-cycle

CDD cooling degree days

CHAN City of Chanute (Kan.)

COSN City of Superior (Nebr.)

COWP City of West Plains (Mo.) – Board of Public Works

CT combustion turbine

CWEP Carthage (Mo.) Water and Electric Plant

DA day-ahead

DAMKT day-ahead market

DA RUC day-ahead reliability unit commitment

DISIS definitive interconnection system impact study

EDE/EDEP Empire District Electric Co.

EHV extra high voltage

EIA Energy Information Administration

EIS energy imbalance service

ERCOT Electric Reliability Council of Texas

ETEC East Texas Electric Cooperative, Inc.

EUDO_X City of Eudora (Kan.)

FCA frequently constrained area

FCU_X Falls City (Nebr.) Utilities

FERC Federal Energy Regulatory Commission

FREM City of Fremont (Nebr.)

GATE_X Gateway Plant

GI generation interconnection

Southwest Power Pool, Inc. Common acronyms Market Monitoring Unit


GLDF generator to load distribution factor

GMOC/UCU Greater Missouri Operations Company (KCPL)

GRDA/GRDX Grand River Dam Authority

GSEC Golden Spread Electric Cooperative, Inc.

GW gigawatt

GWh gigawatt hour

HDD heating degree days

HHI Herfindahl-Hirschman Index

HMMU Harlan (Iowa) Municipal Utilities

HVDC high-voltage direct current

IA interconnection agreement

ID RUC intra-day reliability unit commitment

IDC interchange distribution calculator

INDN City of Independence (Mo.)

IOU investor owned utility

IPP independent power producer

IS Integrated System

ISO independent system operator

ITP Integrated Transmission Plan

JOU jointly-owned unit

KBPU Kansas City (Kan.) Board of Public Utilities

KCPL/KCPS Kansas City Power & Light

KMEA Kansas Municipal Energy Agency

KN01 Kennett (Mo.) Board of Public Works

KPP Kansas Power Pool

kV kilovolt (1,000 volts)

LES/LESM Lincoln (Nebr.) Electric System

LIP locational imbalance price

LMP locational marginal price

MCC marginal congestion component

MEAN Municipal Energy Agency of Nebraska

MEC/MECB MidAmerican Energy Company

MEUC Missouri Joint Municipal Electric Utility Commission

MIDW Midwest Energy Inc.



MISO Midcontinent Independent Transmission System Operator

MLC marginal loss component

MM million

MMBtu million British thermal units (1,000,000 Btu)

MMPA_X Minnesota Municipal Power Agency

MMU Market Monitoring Unit

MW megawatt (1,000,000 watts)

MWh megawatt hour

MWP make-whole payment

MRES Missouri River Energy Services

MUMZ_X Missouri River Energy Services

NCU_X Nebraska City (Nebr.) Utilities

NDVER non-dispatchable variable energy resource

NELI_X City of Neligh (Nebr.)

NERC North American Electric Reliability Corporation

NOAA National Oceanic and Atmospheric Administration

NPPD/NPPM Nebraska Public Power District

NSP/NSPP Northern States Power Energy

NWMT_X Northwestern Energy

NWPS Northwestern Energy

O&M operation and maintenance

OGE Oklahoma Gas & Electric

OMPA Oklahoma Municipal Power Authority

OOME out-of-merit energy

OPPD/OPPM Omaha Public Power District

OTPW Otter Tail Power Company

OTP_X Otter Tail Power Company

OTPR_X Otter Tail Power Company

PARL City of Piggott (Ark.) Municipal Light, Water, and Sewer

PBEL City of Poplar Bluff (Mo.) Municipal Utilities

PEOP_X Peoples Electric Cooperative

PEPL Panhandle Eastern Pipe Line

PISIS preliminary interconnection system impact study

PJM PJM Interconnection, LLC



PLWC Paragould (Ark.) Light & Water Commission

RC reliability coordinator

REMC Rainbow Energy Marketing Corporation

RNU revenue neutrality uplift

RR revision request

RSG reserve sharing group

RT real-time

RTBM real-time balancing market

RTO regional transmission organization

RUC reliability unit commitment

SC simple-cycle

SECI/SEPC Sunflower Electric Power Corporation

SPA Southwestern Power Administration

SPP Southwest Power Pool, Inc.

SPRM City Utilities of Springfield (Mo.)

SPS Southwestern Public Service Company

ST steam turbine

ST RUC short-term reliability unit commitment

TCR transmission congestion right

TEA The Energy Authority

TNGI_X City of Grand Island (Nebr.)

TNHP_X Heartland Consumers Power District

TNHU_X City of Hastings (Nebr.) Utilities

TNSK Tenaska Power Service Company

UGPM Western Area Power Administration, Upper Great Plains

WAPA Western Area Power Administration

WECC Western Electricity Coordinating Council

WFEC/WFES Western Farmers Electric Cooperative

WR/WRGS Westar Energy, Incorporated