1. Next Steps to Reduce the RTM Settlement Timeline August 29, 2013 COPS Workshop Update to TAC Harika Basaran COPS Chair September 5, 2013

2. 2 August 29, 2013 COPS Workshop Highlights Well attended by diverse stake holders Detailed presentations by ERCOT & TDSPs All materials posted http://www.ercot.com/calendar/2013/08/20130829-COPS Goal was more fact finding and educating on the processes All participants agree that the faster cash clears, the better it is for the whole market Disagreement is on the tradeoffs between reduced time line and availability of accurate data & cost to system changes Another disagreement is on the impact on the credit calculations vs. reduced risk of uplift in case of market default Developed 3 options to achieve the reduction in settlement timeline Depending on the option, it may require significant protocol changes from EPS meter to dispute processes to credit calculations

3. 3 Options Proposed at the Workshop- 1.Add a secondary Initial Settlement Statement 1.OD + X, OD + Y where X + 7 = Y 2.Values of X and Y will need to be determined 3.It has been implemented in other ISOs (see slide #21) 4.Possibilities are: {1,8}, {2,9}, {3,10}, {4,11}, {5,12}, {6,13}, 5.Days between X and Y proposed at 7 but could be something different, must be not too short but also not too long 6.Consider also moving Finals and True Ups from the way they are now OD + 57, and OD + 180 7.What is the minimum value of X? 8.Can X be less than 3 if the prices are not final after 2 business days per protocols? 9.Can X be less than 3 if the MPs still can enter trades after the adjustment period per protocols? 10.Can X be less than 2 so that RTM settles before DAM? 11.Additional settlement run also means additional extract data to be posted 12.MPs with load & gen have to pull, save and maintain bigger databases

4. 4 Options Proposed at the Workshop- 2.Settle DAM & RTM at the same time but include only certain charge types for RTM on the first run, then run another complete RTM 1.Settle RTM at OD + X with DAM for all charge types that have validated data available 2.QSE to QSE trades, EPS metered generation and load 3.Run another complete RTM Initial settlement after Y days and rerun all the RTM charge types again 4.Which specific RTM charge types need to be settled more quickly and have complete and validated data to be settled in the first run? 5.Values of X and Y need also to be determined 6.Consider moving Final and True up as well

5. 5 Options Proposed at the Workshop- 3.Set target OD + 5 for June 1, 2014 and OD + 3 for June 1, 2015 1.Tie targets in reduction to offer cap increases 2.Move the days for Final settlement statement to OD + 45 or less depending on TDSP processes 3.Keep True up at OD + 180 for complex disputes 4.Give MPs and ERCOT time to upgrade, improve their systems to capture accurate data and to comply with faster times 5.Can be phased as going in stages to OD + 6, then OD + 5, OD + 4 and OD + 3 6.Require an additional settlement run only if necessary

6. 6 Issues with settling in very short times OD + 1, 2, 3 GENERATION DATA: Generation data is the most significant one & determines the ERCOT load Most generators have EPS (ERCOT polled settlement meter) requirement per protocols (Section 10.2.3) Currently no proxy or estimation process exist such as profiles for generation data or NOIE bi-directional points Can we use real time SCADA (telemetry) data for the first run? Generators have been always paid based on TDSP data Need new protocols and system changes to include SCADA (telemetry) data as proxy for EPS by ERCOT and MPs TDSPs may have to increase their wire charge to recover Tradeoff is ongoing cost to stakeholders vs. benefits from reduced default uplift risk that may or may not happen at some unpredictable frequency & magnitude

7. 7 Issues with settling very in short times OD + 1, 2, 3 cont. PRICES & TRADES: RTM prices are not final until after 2 business days per protocols (6.3.6) Do we want to settle RTM with unfinalized prices ? MPs can submit trades per protocols till next day 14:30 (4.4.2 (3)) Do we want to settle with incomplete trades? AVALIABILITY OF LOAD DATA FOR THE INITIAL RUN Total load volume ERCOT wide (# of ESIIDs as of July 2013) Total % Available by Initial Settlement OD+7 –AMS 45.60% 45.60 % –COMP_IDR 24.60 % 4.3 % –NOIE 26.80 % 24.3 % –NIDR 3:00 % 0.4 % –Impact of shortening time line will affect each of the meter types differently –Although majority AMS is available next day, ERCOT & TDSPs need to communicate to resolve any issues if time line is very short

8. 8 Issues with settling very in short times OD + 1, 2, 3 cont. GENERAL CONCERNS RAISED WITH SETTLING TOO FAST: ERCOT and TDSPs have no time to check or do any reasonable validation Possibility of charging the incorrect MP and paying to the incorrect MP Credit calculations will also use that first cut of initial settlement At what point does the risk of generation data availability become a “showstopper” or require a solution? At what point do we anticipate transitioning from settlement on predominantly TDSP- submitted usage data to predominantly ERCOT-estimated usage data? At what point is an additional settlement iteration needed in order to mitigate risk associated with waiting until Final Settlement to resolve data availability and accuracy issues? At X=6, 5, 4, 3? At what point do we anticipate introduction of credit risk that may outweigh credit benefit? Cost of to ERCOT and all stakeholders to improve accuracy of the data & complying with additional settlement run Tradeoff is cash clearing based on incomplete and not validated data until the second initial settlement run and cost of upgrades & staffing vs. benefits from reduced default uplift risk that may or may not happen at some unpredictable frequency & magnitude

9. 9 Next Steps

10. 10 Next Steps - Scheduled a second workshop in lieu of September COPS meeting –September 10, 2013 Tuesday WEBEX and Met Center Homework –Pros and cons of the each option for each MP –Impact on systems & staffing of each option –Is there a floor of OD + X that market should not go below that is determined by posting of the final RTM prices and availability of good generation data & trade data? –When does the cost of reducing time line and adding another settlement run becomes more than benefits achieved with reduced risk of default uplift ? Objectives of the second workshop –Is there an option all MPs can agree on? –Is there a hybrid of all 3 options most of the MPs can agree on? –If not define alternatives with pros and cons for each –Draft NPRR language and time line for implementation of the next steps

11. 11 The following slides are provided for informational purposes only and were developed by ERCOT Staff for the workshop

12. 12 Settlement Cycle Effect on ERCOT Credit Exposure

13. 13 Settlement cycle effect on credit exposure ERCOT’s credit exposure calculation addresses a number of credit risk components. Exposure that must be secured Exposure that may be unsecured Future exposure from CRRs Day Ahead Market exposure Real Time Market exposures Mass transition risk component Unbilled/unpaid amounts Estimates for new Counter- Parties Minimum collateral amounts Current exposure for QSEs Current exposure for CRRAHs ERCOT Public Unsettled Real Time component Settled Real Time component Unsettled Real Time component

14. 14 Settlement cycle revisions affect the indicated components. Exposure that must be secured Exposure that may be unsecured Future exposure from CRRs Day Ahead Market exposure Real Time Market exposures Mass transition risk component Unsettled Real Time component Estimates for new Counter- Parties Minimum collateral amounts Current exposure for QSEs Current exposure for CRRAHs Unsettled Real Time component Settlement cycle effect on credit exposure ERCOT Public Settled Real Time component Unsettled Real Time component Unbilled/unpaid amounts

15. 15 Estimated impact on total ERCOT credit exposures for different Real-Time Market settlement cycles. ERCOT Public Settlement cycle effect on credit exposure Notes: 1.M2 factor reflects unbilled historical activity. It reduces to 10 days in September to reflect the settlement cycle reduction from 9 to 7 days. The M1 factor reflects forward risk. 2.Price spikes will increase exposure under any scenario. 3.URTA adjusted by changing M2 factor. 4.Pro rata adjustments to RTLCNS.

16. 16 ERCOT Public Settlement cycle effect on credit exposure Estimated impact on credit exposures for different Real-Time Market settlement cycles.

17. 17 Estimated exposure does not decrease pro rata to RT settlement days. ERCOT Public Settlement cycle effect on credit exposure Offsetting factors Minimum collateral requirements and lookbacks / extrapolation dampen the impact of shortening the settlement cycle Market credits decrease at a rate approximately equal to market debits No adjustments made for increased resettlement risk Offsetting factors Minimum collateral requirements and lookbacks / extrapolation dampen the impact of shortening the settlement cycle Market credits decrease at a rate approximately equal to market debits No adjustments made for increased resettlement risk

18. 18 Current State

19. 19 Evolution of ERCOT’s RTM Settlement Timeline PeriodStatementInvoice 8/2000 –2/2002OD+3Issued Weekly, due in 5 BDs 3/2002 – 12/2005OD+17Issued Weekly, due in 5 BDs 1/2006 – 11/2011OD+10Issued Weekly, due in 5 BDs 12/2011 – 3/2013OD+9Issued Weekly, due in 5 BDs Daily, OD+9, due in 2 BDs 4/2013 – 7/2013OD+8Daily, OD+8, due in 2 BDs 8/2013 to currentOD+7Daily, OD+7, due in 2 BDs Under DiscussionX < OD+7Daily, OD+X, due in 2 BDs ERCOT Public

20. 20 Cash Clearing Stats on the Current OD+7 Schedule OD + "X" $ to ERCOT: # of ODs $ to ERCOT: % of ODs $ to Market: # of ODs $ to Market: % of ODs 914740.27%00.00% 105214.25%9526.03% 1113837.81%5113.97% 12246.58%17949.04% 1341.10%328.77% 1400.00%61.64% 1500.00%10.27% 1600.00%10.27% Total365100%365100% On an OD+7 timeline, the vast majority of the Operating Days have cash cleared through ERCOT within 13 calendar days from the OD. For each day removed from the timeline, a corresponding day is removed from the cash clearing timeline ERCOT Public

21. 21 Benchmarking with Other ISOs/RTOs Statement T+3, T+12, T+55,… Cash Clears ~OD+12* Statement T+7, T+59, T+180 Cash Clears ~OD+12* Statement T+7, T+47, T+77, T+137,… Cash Clears ~OD+22* Statement T+4 (credit only), T+7, T+14, T+55, T+105 Cash Clears ~OD+21* Statement T+1, EOM+5, ~T+120, ~T+240 Cash Clears ~OD+12* Statement T+4B, ~EOM+120, ~EOM+300 Cash Clears ~OD+12* T+10B, T+20B Cash Clears ~OD+43* Statement T+7B, T+2M Cash Clears ~OD+49* Statement Wkly (T+3 to T+9) Cash Clears ~OD+11* *Cash clearing timeline based on an example day of 6/6/2013. However, non-business days have the effect of creating a range of dates during which this activity can occur. Cash clearing is from OD to payment to the Market Participant. The statement and cash clearing timelines added to this map are approximations for an “apples to apples” comparison

22. 22 Advanced Meter Settlement Impacts – July 2013 July 2013: At month end, settling 6.3M ESIIDs using Advanced Meter data. July 2013: 97.0% of the load in ERCOT is settled with 15- min interval data (AMS, Competitive IDR, and NOIE IDR). ERCOT Public

23. 23 ERCOT Wide Load Volumes by Meter Type – INITIAL Settlement – July 2013 ERCOT Public

24. 24 ESI ID Data Availability “Look-Back” Analysis – August 2013 ERCOT Public

25. 25 ESI ID Data Availability “Look-Back” Analysis – August 2013 ERCOT Public

26. 26 ESI ID Data Availability “Look-Back” Analysis – August 2013 ERCOT Public

27. 27 ESI ID Data Availability “Look-Back” Analysis – August 2013 ERCOT Public

28. 28 QUESTIONS & COMMENTS