Transmission Metrics: Initial Results I tem No. A-4 March 1 7 , 2 - - PDF document

Page 1 of 22 S lide 1

I tem No. A-4

March 1 7 , 2 0 1 6

Transmission Metrics: Initial Results

Good morning Mr. Chairman and Commissioners. Office of Energy Policy and Innovation st aff has att empted to develop obj ect ive and st andardized measures of various charact erist ics of t he electric syst em and it s performance to help assess t he effect iveness of t he Commission’ s policies regarding t ransmission invest ment and t o inform potential policy revisions going forward. As t he team described in it s present ation at the April 2015 open meet ing, st aff considered a range of potentially relevant metrics in t hree broad categories: (1) met rics designed t o evaluate key goals

• f Order No. 1000; (2) met rics designed t o indicate whet her appropriat e levels of

t ransmission infrast ruct ure exist in a part icular region; and (3) metrics designed to permit analysis of t he impact of Commission policy changes by comparing key values before and after changes t ake place. In t he st aff report being released t oday, st aff describes our met hodology for calculat ing each of the t hree categories of met rics and t he results of t hat analysis. We will now provide a brief overview of t he report , which will be available t hrough t he www.ferc.gov websit e. To begin, my colleague Ben Fost er will discuss t he first met ric, whose development he led, which is int ended t o help assess a key goal of Order No. 1000 – nonincumbent participat ion in regional t ransmission planning processes.

AD15-12-000

Page 2 of 22 S lide 2

Key Goal Order No. 1000 – Nonincumbent Developer Reform

 Purpose: measure nonincumbent participation at

the bid/proposal stage of the Order No. 1000 regional transmission planning process

 Analysis focused on PJM and CAISO because

bid/proposal data from these two regions was available at the time of analysis

This met ric measures t he percent age of bids or proposals for new t ransmission proj ect s in t he Order No. 1000 regional t ransmission planning processes t hat nonincumbent transmission developers submit ted. At t he t ime t hat st aff was preparing t he report , relevant dat a was only available for CAIS O and PJM. As explained in more detail in t he report, st aff gat hered dat a from public document s posted on CAIS O’ s and PJM’ s websit es and elsewhere. S t aff applied Order No. 1000’ s definition of nonincumbent transmission developer, which turns on whet her a t ransmission developer has a ret ail dist ribut ion service t errit ory or footprint and, if so, whet her t he proj ect is located there. To det ermine t he incumbency st atus of developers submitting proposals, which was generally not available on the regions’ websit es, st aff compared t he zone in which each proposed proj ect would be built wit h t he developer’ s ret ail dist ribution service t erritory or foot print , where applicable.

Page 3 of 22 S lide 3

Competitive Proposals by I ncumbents vs. Nonincumbents

22% 35% 25% 33% 63% 38% 78% 65% 75% 67% 37% 62%

0% 10% 20% 30% 40% 50% 60% 70% 80% 2013 2014 2015 2013 2014 2015 CAISO PJa

n=9 n=20 n=4 n=43 n=304 n=105 Incumbent bonincumbent Incumbent bonincumbent

Sources: CAISO 2012-13 and 2013-14 transmission plans; PJa Transmission Expansion Advisory Committee and RTEP Proposal websites

Percent of annual proposals in CAI SO and PJM ( 2 0 1 3 -2 0 1 5 )

S lide 3 summarizes t he result s of st aff’ s analysis of t he bids and proposals t hat developers submitt ed from 2013 t o t he period in 2015 when t his report was being

• prepared. The figure shows t he percent age of proposals in each RTO t hat came from

incumbent and nonincumbent t ransmission developers during t he st udied period, wit h t he associat ed number of proposals received in each region and year. Overall, of t he 485 proposals submit ted in t he CAIS O and PJM regions, 53 percent were from incumbent s and 47 percent from nonincumbent s. On a regional basis, the percent age of proposals from nonincumbent s account ed for t wo-t hirds t o three-quart ers of proposals in each of the three years in CAIS

• O. In PJM,

t he percent age of proposals from nonincumbent s account ed for more t han 60 percent

• f all proposals in 2013 and t he st udied portion of 2015, but less t han 40 percent of

proposals in 2014, t he year in which PJM received t he maj orit y of it s proposals.

Page 4 of 22 S lide 4

Metrics to Help Assess Need for Transmission I nvestment

 Assumption: Persistent and costly congestion may

indicate need for additional transmission capacity

 Different approaches for bilateral markets and RTO/ISO

markets

• Bilateral Markets – based on number of interchange-curtailing

Transmission Loading Relief (TLR) events

• RTO/ISO markets – based on Locational Marginal Price data

Thank you, Ben. Next we will t urn t o met rics designed to help indicate whether appropriate levels of t ransmission infrast ructure exist in a region. Here, st aff relied on the assumption t hat persist ent cost ly congestion in an area may indicat e insufficient t ransmission invest ment because it may suggest t hat there is not enough available transfer capabilit y on t he transmission syst em t o support the delivery of less cost ly energy. Ideally, persist ent cost ly congest ion would be ident ified direct ly from hist orical energy price informat ion by looking for significant ly large price different ials t hat persist for ext ended periods of time. RTO/ ISO market s generat e pricing data direct ly applicable t o t his purpose, and as such, staff used this data t o calculat e the met ric for RTO/ IS O market regions. For non-RTO/ IS O market regions, st aff used a more indirect metric based on hist orical NERC Transmission Loading Relief (TLR) dat a.

Page 5 of 22 S lide 5

Load-Weighted Curtailment Frequency Metric

For non-RTO/ IS O market regions, my colleague Abdur Masood led st aff’ s invest igat ion

• f whet her NERC TLR procedures used t o manage congestion can serve as an indirect

measure of t he level of t ransmission infrast ructure in t he region. S pecifically, all

• t her t hings being equal, more TLR event s might indicat e a need for more

t ransmission infrast ruct ure and fewer event s might indicat e less need for addit ional t ransmission infrast ruct ure. In practice, st aff assumed t hat such a TLR-based met ric would need t o be used in conj unct ion wit h publicly available sources of pricing data, such as price indices or ret ail rate informat ion, in order t o incorporat e t he concept of cost ly congestion. In ot her words, even if a region experiences large numbers of TLR event s, in the absence of any significant and persist ent price differentials in t hat region, t he TLR event s might not indicat e a need for addit ional t ransmission infrast ruct ure. At t his point , I need t o not e t hat instead of TLRs, t he West ern Int erconnect ion manages unscheduled flows using a coordinat ed combinat ion of cont rollable devices, such as phase shifting t ransformers, and schedule curtailment s t hat st aff believes are similar t o TLRs but are not recorded in the NERC TLR logs. Thus, st aff did not calculat e this met ric for t he West ern Interconnect ion.

Page 6 of 22 For t he East ern Interconnection, TLR dat a is publicly available from NERC, but reliable price information for non-RTO/ IS O market areas is less readily available for t he t ypes of price indices or ret ail rat e data t hat st aff initially hoped t o use. However, in t he future st aff int ends t o explore whet her it could use FERC Elect ric Quart erly Report (EQR) wholesale pricing dat a t o calculat e t his met ric for non- RTO/ IS O market s. All j urisdictional and some non-j urisdictional wholesale sellers of elect ricit y submit EQR pricing dat a t o FERC, and st aff believes t hat t he approximat e location of associat ed t ransactions can be gleaned from t he data. Accordingly, EQR data may provide a comprehensive view of pricing t rends in bilateral market regions comparable t o what RTO/ IS O pricing dat a provides for organized market s. For t his report , t he basis of t his met ric is the number of interchange-curt ailing TLRs t hat t he t ransmission operat ors of the region report ed t o NERC. In order t o provide a basis for comparing bet ween regions of different sizes, st aff normalized t his metric based on t he retail load associat ed wit h the region in question.

Page 7 of 22 S lide 6

Load-weighted TLRs

(Number of TLRs/ GWh retail)

Sources: NERC TLR data

This slide shows t he load-weight ed TLR metric for S

• uthwest Power Pool,

Midcontinent Independent S yst em Operator, Inc., and Tennessee Valley Aut horit y, which were t he areas wit h t he highest numbers of TLRs. While MIS O and S PP operat e

• rganized market s t hat optimize dispatch based on congest ion, great ly reducing their

internal use of TLRs, it is st ill possible for RTOs t o require TLRs t o address unscheduled loop flow originating from out side t heir footprint s. Bot h MIS O and S PP have ext ensive borders wit h non-organized market areas, which may help explain t heir continuing use of TLRs. Overall, it appears t hat S PP consist ent ly experienced more TLR event s per gigawatt - hour of retail load than ot her regions during t he analyzed period. However, it should be not ed t hat S PP formed it s Consolidat ed Balancing Aut horit y and launched it s Int egrated Market place in March of 2014. Prior t o t hat , S PP was act ing as t he reliabilit y coordinat or for mult iple Balancing Aut horit y Areas and operat ed an imbalance market that was more limited in scope and capabilit y than the Int egrated Market place. The TLR logs show a significant decrease in the rat e of S PP’ s TLR use aft er t he consolidation and market st art -up t ook place. While correlat ion is not necessarily causat ion, t his is what we would expect t o happen; consolidat ing Balancing Aut horit y Areas and moving to a more comprehensive market st ruct ure

Page 8 of 22 should lead t o more efficient use of the associat ed existing t ransmission facilities, which should result in a decrease in the need for TLRs. The report notes certain pot ent ial concerns wit h reliance on a TLR-based met ric, such as t he fact that TLRs only represent t ransmission limitations bet ween Balancing Aut horit y Areas, and the fact that it is theoret ically possible for a syst em t o experience cost ly congest ion but not have a significant number of TLRs. However, on balance, st aff believes t hat a TLR-based metric can provide one useful data point in analyzing non-RTO/ IS O bilat eral market s. James Nachbaur will now discuss t he price different ial met ric he developed for RTO/ IS O market regions.

Page 9 of 22 S lide 7

RTO/ISO Market Price Differential Metric

S t aff developed a t ransmission invest ment met ric t hat reflect s persist ent differences in RTO/ IS O market nodal prices. This metric is expressed in years and it capt ures how long RTO/ ISO market nodal price differentials have occurred persist ent ly, t hough not necessarily at all t imes t hroughout a year. S t aff reasons t hat consecut ive years of significant price differentials could indicat e insufficient t ransmission infrast ructure because, for example, lower cost energy at lower-priced nodes is not being delivered t o t he node with higher prices. S t aff, however, not es t hat available t ransfer capabilit y bet ween places— and t he t ransmission invest ment t hat maint ains t hat capabilit y— may not be t he only variables relevant t o persist ent price differences. To calculat e this metric, st aff used real-t ime prices at load and generat or point s. S t aff gathered these prices from ABB Velocity S uit e. To avoid placing excessive weight on highly unusual prices, st aff used the 95t h and 5th percentiles of prices, rat her t han maximum or minimum prices, at each load and generat or point in each

• year. S

t aff t hen calculat ed market -wide average high and low generator and load prices in each year. Using t his informat ion, st aff identified point s whose high or low prices were at least one st andard deviation higher or lower t han the market -wide averages in each year.

Page 10 of 22 S t aff ident ified high-priced and low-priced point s in 2012, 2013, and 2014 t o determine where price separat ion occurred persist ent ly and had not yet been resolved, based on data available as of the time t his report was being prepared. To focus on t he persist ence of price separat ions, st aff t hen calculated how long ago t he current run of high or low prices began. There are many high-priced and low-priced point s.

Page 11 of 22 S lide 8

Source: Staff analysis of ABB Velocity Suite price data

As shown on t his slide st aff identified areas wit hin each RTO t hat cont ain mult iple point s wit h persist ent price separations in t he same direct ion. Finally, st aff identified for each region t he longest period of price separation experienced by a point in t hat

• region. That number is t he RTO/ IS

O Price Different ial met ric for that region.

Page 12 of 22 S lide 9

RTO/ I SO Market Price Differential Metric for Select Regions

Region identified by staff Baltimore, aaryland Upper teninsula Delmarva teninsula Long Island, New York Northwestern New Jersey North-central aISO Western Stt Dreater /hicago North Dakota- South Dakota- ainnesota border New York- /anada border Northern New York West-central North Dakota The Deysers, /alifornia Direction Iigh Iigh Iigh Iigh Iigh Iigh Low Iigh Low Iigh Low Iigh Low Low Low Low Low Start of price differential ‘05 ‘05 ‘06 ‘07 ‘07 ‘05 ‘05 ‘08 ‘07 ‘10 ‘06 ‘12 ‘10 ‘06 ‘06 ‘10 ‘11 Years of persistence through 2014 10 10 9 8 8 10 10 7 8 5 9 3 5 9 9 5 4

This slide summarizes t hese result s. As you can see, t here are several regions t hat experienced significant price different ials for up t o 10 years, at least t hrough 2014. At t his point , I would like t o emphasize a few import ant caveat s. By t hemselves, t hese met ric result s do not prove t hat t ransmission capacit y should necessarily be added in any of these areas. These dat a merely provide one indicat ion that it could be useful t o explore the economics of adding new t ransmission capacit y in t hese

• regions. Furt hermore, significant changes in underlying fundamental input s t o

elect ricit y prices, like t he t ypes of large-scale changes in relative fuel prices t hat we’ ve seen in recent years, could great ly impact price trends going forward. Accordingly, it would likely be very useful t o continue updat ing t his t ype of analysis as more recent dat a become available.

Page 13 of 22 S lide 10

Baseline Transmission I nvestment Metrics

 Basic measures of transmission investment  Investment dollars and circuit-miles were

normalized based on associated retail load

 Main use is to permit before-and-after comparison

following relevant changes in policy

Thank you, James. The t hird category of met rics is designed t o permit analysis of t he impact of Commission policy changes by allowing t he comparison of key values before and aft er policy changes t ake place. This category includes t hree int errelated met rics: (1) Load-weight ed Transmission Invest ment; (2) Load-weight ed Circuit - miles; and (3) Circuit -miles per Million Dollars of Invest ment . In combination, t hese t hree metrics allow for a comparison of how much t ransmission infrast ruct ure has been developed in each region and t he relat ive cost of that investment . Ben, who also led t he development of t hese met rics, will now discuss each met ric in t urn.

Page 14 of 22 S lide 11

Load-Weighted Transmission Investment Metric

This met ric describes t he load-weight ed dollar value of t ransmission facilities t hat went int o operat ion each year from 2008-2014 in t he eight NERC regions of the cont iguous U.S . Weight ing t ransmission invest ment dollars by associated ret ail load allows for comparisons bet ween regions of different sizes. While more load-weight ed invest ment may not always be bet ter t han less invest ment, t racking how t hese values change following changes in Commission policy may be informat ive. Transmission proj ect dat a are from t he CThree Group’ s Nort h American Elect ric Transmission Proj ect s database, and load dat a are from NERC’ s 2014 Elect ricit y S upply & Demand database. S t aff convert ed nominal cost or budget figures t o 2014 dollars using t he annual average of t he consumer price index for all urban consumers. To calculat e the final, load-weight ed met ric, st aff divided t he normalized invest ment figures for each NERC region for each year by t he ret ail load in each year.

Page 15 of 22 S lide 12

U.S. I ncremental Transmission I nvestment 2008-2014

12

$19.70 $7.28 $5.31 $5.05 $3.93 $4.25 $5.00 $2.42 $0.88

TRE 2013 w/ o CREZ $2.56

$4.72 $3.17 $2.80 $2.61 $2.34 $2.19 $1.88 $0.88 $0.51

$0 $4 $8 $12 $16 $20 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 TRE SPP NPCC WECC RFC All Regions MRO SERC FRCC

All New and Upgraded Projects in Operation, $/ MWh

Sources: C Three Group, bERC, .LS

S lide 12 shows load-weighted incremental t ransmission invest ment in dollars per MWh in t he eight NERC regions of t he cont iguous U.S . from 2008 t o 2014. The figures in red represent the load-weight ed invest ment across all seven years, while figures in black refer to t he highest load-weight ed dollar figure in each region. Overall, t he average load-weighted t ransmission invest ment for all regions for all years is over t wo dollars per MWh of load, alt hough invest ment s are “ lumpy” for most regions, as is t ypical for large infrast ructure proj ect s. Due to a maj or spike in t ransmission invest ment in 2013, t he average load-weighted investment for TRE (t he Texas Regional Entit y) over all years exceeds four dollars per MWh. Five of t he eight NERC regions (S PP, NPCC, WECC, RFC, and MRO) are in t he range of approximat ely $1- 3/ MWh on average over t he period, while two regions (S ERC and FRCC) fall below one dollar per MWh on average over t he period. The metric shows a generally increasing t rend of load-weight ed invest ment over t he period, wit h all regions except FRCC and MRO report ing t he great est load-weight ed invest ment in 2013 or 2014. The highest all-year average invest ment over t he period, of $4.72/ MWh, and highest single-year met ric ($19.70/ MWh in 2013) was in TRE. This was due t o the approximat ely $6.5 billion of proj ect s— t he largest single-year invest ment of any region— t hat went into operat ion in 2013, of which approximat ely $5.7 billion was

Page 16 of 22 under Texas’ Competit ive Renewable Energy Zone (CREZ) initiat ive, which aimed t o alleviate congest ion and int egrat e wind capacit y int o t he elect ric grid. Excluding t his large CREZ invest ment in 2013, invest ment in t hat year would be $2.56/ MWh and t he TRE regional average invest ment would be $2.22/ MWh, much closer t o the all-region

• average. Thus the changes in t his met ric over t he period perfect ly illust rat e the

powerful impact of one part icular policy initiative – Texas’ CREZ init iat ive.

Page 17 of 22 S lide 13

Load-Weighted Circuit- Miles Metric

The next met ric describes t he load-weight ed circuit -miles of t ransmission line added from 2008 t o 2014. As wit h t he previous metric, weighting t ransmission circuit -miles by associated ret ail load allows for comparisons bet ween regions of different sizes. For t his met ric, st aff filt ered t he dat a in t he C Three Group dat abase, removing t he data associat ed wit h those proj ect s t hat do not include a line component and a limit ed number of proj ect s wit hout a NERC region designat ion, or wit h mult iple designat ions. To determine the number of circuit miles for each proj ect , st aff mult iplied reported line miles by t he number of report ed circuits. In cases where the number of circuit s was not report ed, st aff conservat ively assumed t hat the line has only one circuit . To arrive at t he final met ric of load-weighted circuit miles, st aff divided t he circuit - mile figure for each NERC region for each year by t hat region’ s retail load.

Page 18 of 22 S lide 14

Circuit-miles of Transmission Added in U.S. 2008-2014

18.3 8.5 5.1 3.4 2.5 2.3 1.7 1.4 1.0

TRE 2013 w/ o CREZ 4.0

5.7 3.4 2.6 1.9 1.8 1.2 1.1 1.0 0.7

4 8 12 16 20 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 TRE SPP MRO All Regions WECC NPCC RFC SERC FRCC

New and Upgraded Lines in Operation, Circuit-miles/ TWh

Sources: C Three Group, bERC

S lide 14 shows load-weighted transmission line addit ions in circuit -miles/ TWh from 2008 t o 2014. Overall, t he result s for t his met ric are similar t o those for t he previous metric. TRE and S PP lead, and S ERC and FRCC lag, t he other regions in t erms of load-weight ed circuit -miles added, wit h five regions (WECC, NPCC, RFC, S ERC and FRCC) below t he all-region all-year average of approximat ely t wo circuit -miles/ TWh. TRE added t he most circuit -miles on a load-weight ed basis. As not ed above, t his is mainly due t o the CREZ proj ect s, most of which included a relatively long line component . Only WECC built longer lines on average than TRE, but it added fewer circuit -miles on an absolut e basis and, because it s load is almost twice that of TRE, on a load-weighted basis as well.

Page 19 of 22 S lide 15

Circuit-Miles Added Per Million Dollars Metric

This last met ric is designed t o provide a basis for assessing t he cost impact of different policy choices or fact ual circumst ances on t ransmission invest ment . S pecifically, t his met ric divides t he circuit -miles of t ransmission line added in t he cont iguous U.S . from 2008-2014 by t he amount of money invest ed over the same period in million dollars of invest ment . Data for t his met ric are also t aken from the C Three Group’ s t ransmission database. S t aff filt ered t he dat a as described earlier.

Page 20 of 22 S lide 16

Circuit-miles Added per Million Dollars 2008-2014

2.9 5.7 2.4 9.2 1.7 1.4 1.9 2.0 1.8 1.7 1.6 1.4 1.3 1.3 1.1 0.8 0.8 0.7

2 4 6 8 10 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 08 11 14 MRO FRCC SERC TRE SPP All Regions NPCC WECC RFC

New and Upgraded Lines in Operation, Circuit-miles/ $Million

Sources: C Three Group, .LS

S lide 16 shows circuit-miles per million dollars of t ransmission invest ment from 2008 t o 2014. Regions wit h higher figures represent a great er number of circuit-miles added per million dollars invested. By t his measure, MRO built t he most circuit -miles per million dollars on average across all years (1.7), compared t o a t otal of 1.1 for all regions. RFC, NPCC, and WECC built t he fewest circuit -miles per million dollars across all years, of less t han one. The difference in circuit -miles per million dollars invest ed may be due t o a range of factors, including terrain, populat ion densit y, and st at e policy choices, among ot hers. TRE and FRCC appear t o have t he most variabilit y in t heir result s, alt hough several proj ect s t hat went into operation in TRE in 2008, and FRCC in 2010 and 2012, have circuit -mile data but no associat ed dollar figure, which causes t hose years t o appear as out liers in t he figure above. S PP appears t o have t he least variabilit y in t his met ric across t he years. From a developers’ perspective, less variabilit y in cost s would likely be desirable, but more research is necessary t o determine what may be driving differences in t he number of circuit -miles built per million dollars among these regions.

Page 21 of 22 I would like t o emphasize t hat care should be t aken in at tempting t o use t he result s of t his met ric t o gauge the “ cost -effect iveness” of different regions’ t ransmission invest ment s because much of t he cost of a proj ect is driven by t he highly variable physical and regulat ory challenges part icular t o each region, proj ect , or developer. Nevert heless, st aff believes that t his met ric, in combination wit h the other t wo I’ ve j ust discussed, can provide useful insight s int o t he impact of Commission policy changes, part icularly when considered over t ime.

Page 22 of 22 S lide 17

I tem No. A-4

March 1 7 , 2 0 1 6

Transmission Metrics: Initial Results

Thank you, Mr. Chairman and Commissioners. This concludes our presentation and we welcome any quest ions you may have.