Competitive Power: The Recovery Continues, Why Some Fears are Mis- Placed, and What Concerns Us

Transcription

1 Competitive Power: The Recovery Continues, Why Some Fears are Mis- Placed, and What Concerns Us September 30, 2014 RESEARCH TEAM Dan Eggers, CFA (212) Thomas Murray (212) Matthew Davis (212) DISCLOSURE APPENDIX CONTAINS IMPORTANT DISCLOSURES, ANALYST CERTIFICATIONS, INFORMATION ON TRADE ALERTS, ANALYST MODEL PORTFOLIOS AND THE STATUS OF NON-U.S ANALYSTS. FOR OTHER IMPORTANT DISCLOSURES, visit or call +1 (877) US Disclosure: Credit Suisse does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the Firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. CREDIT SUISSE SECURITIES RESEARCH & ANALYTICS BEYOND INFORMATION Client-Driven Solutions, Insights and Access 1

2 Power Market Fundamentals Continue to Improve Positives Include Tightening with Coal Closures plus RPM Changes Our take: The underlying fundamentals for competitive power are continuing to improve, consistent in the measured direction we identified in the spring, but with the inevitable bumps along the way for a seasonally weather exposed, commodity cyclical business. We appreciate the market doubts how bad does gas basis turn, what about the weather, newbuild plants are coming, etc. but see the impact of retiring coal plants and push by ISOs to support improved system reliability collectively working to the benefit of power. Our favorite Integrated Power stock remains Exelon (EXC) where we see an attractive entry valuation, reliable 3.7% dividend yield, and leverage to a power and capacity market recovery plus additional unique levers including benefit from a potential carbon price in IL / US, accretion from the POM acquisition, and execution on operational goals. The positives we see for power are rooted in the tightening supply-demand fundamentals: 1) Power markets for (slides 10-17) should benefit from significant coal plant closures primarily in 2015/16 that will strip away lower cost mid-merit generation that is still needed at peak periods (was 5-6% of PJM load in the winter), forcing more expensive units to dispatch leading to higher market clearing prices and increased volatility during high demand periods as seen last winter. This part of our thesis continues to advance as expected. 2) Prospective changes to RPM capacity auction rules (slides 18-24) look constructive for incumbent generators as PJM works to enhance system reliability after last winter s wake-up call that should produce better compensation of physical generation with rule changes around the performance capacity standard, updated VRR demand curve and net CONE calculation, and prospective court mandated changes to demand response. >> We expect generators to benefit from a combination of higher capacity and energy prices with the mix evolving over time as rules develop with energy favored in the next few years and capacity later; overall the balance of positive drivers is constructive plus valuations look interesting from here (slides 51-54). 2

3 Power Market Fundamentals Continue to Improve Newbuilds will come but need to be put in context; natural gas price uncertainty still feels like biggest risk 3) Newbuild CCGTs are looking economic (slides 25-33) in most markets particularly with higher capacity prices, supporting more plant additions in Newbuild economics will benefit from further expansion in either spark spreads or capacity prices although returns sensitivities need to be watched. 4) Power prices have more downside support (slides 34-38) than we hear from bears. We see newbuild CCGTs having some impact on power markets by adding more efficient capacity (slide 35) although we think the downside is more limited than feared considering (a) the cash operating costs of other marginal generation relative to newbuild CCGTs that will limit the drop to the marginal cost of new CCGTs (slide 36) and (b) the reality that newbuild CCGTs need a combination of capacity revenues and reasonably healthy spark spreads to generate an economic return (slide 37). The assumption prices revert to heat rate times delivered gas price is incomplete in our mind since the newbuild gas plants have a lot of capital costs that need to be serviced. We appreciate the risk of eventual power price pressure (keeping in mind it is 3+ years away when the next potential wave of plants is added), but think the balance of higher capacity prices net of potential drag on power prices is still a net positive relative to current forward economics. 5) Lower gas prices and wider gas basis are a valid market concern and will be impacted by weather, the rate of production volume gains that have been a concern (slide 40), and ability to move gas to demand centers. We worry about gas prices given the linkage to power prices (slides 42-43) that in turn have had a reasonably high correlation to Integrated stock prices YTD (slides 44-45) although we think tightening power market conditions will limit some of the power price downside even if gas prices erode from here. Source: CS Estimates 3

4 A Few Trading Points to Watch We see a couple pivot points to consider over next 9 months Timing of investments in commodity cyclical investments is important for successful trading if not capable of buying and holding through volatility on the way to a structurally better market. We generally think of sell-siders acting like traders as too cute by half, but tactically today we think of a few discrete moves ahead for power: Now through February. We think investors should be moving more positive on power stocks sometime between now and end of 3Q earnings (which will be sloppy and could weigh on the stocks a bit). We think enthusiasm will pick up as we see more action on prospective RPM changes and go into winter where we can again see market volatility with winter gas deliveries and the impact of retiring coal plants providing power price support. Our obvious concern in mid-fall is the outlook for winter weather which if we don t have any would lead to natural gas price downside and some likely bleed out for the power stocks. March through May. We expect investor and sell-side attention to hone in on RPM auction expectations, creating the annual waves of uncertainty and speculation around results that leads to mini-cycles of excitement and dread. We think this period is the hardest to trade but will largely be an exercise in managing prompt market expectations against more structural views on power markets that will require better economics through energy / capacity to attract investment and support reliability goals. Beyond May. We expect some consternation over newbuild CCGT exposure coming out of RPM and the worries about the drag on power prices in with more lower cost generation. We think that concern is valid but also believe the combination of energy and capacity prices will support a better earnings outlook than assumed today. Until newbuild power plants are installed, look for power prices to be more constructive with the balance of economic power shifting from energy to capacity; either way the group looks better. 4

5 Updated Estimates and Target Prices CS Current Base Estimates vs Prior Estimates Price Target Price /29/2014 Current Up/Down Prior Current Prior % Change Current Prior % Change Current Prior % Change ETR % % % % EXC % % % % FE % % % % PEG % % % % CS Current Base Estimates vs Consensus CS Consens % Diff CS Consens % Diff CS Consens % Diff ETR % % % EXC % % % FE % % % PEG % % % CS Current MTM Estimates vs Consensus New Consens % Diff New Consens % Diff New Consens % Diff ETR % % % EXC % % % FE % % % PEG % % % We are updating our earnings estimates and target prices to incorporate power price forwards at September 26 th (basically end 3Q14) from our late July update where power prices are up modestly but still down from the June 30 mark (slide 57). Our updated target prices built from our sum-of-the-parts methodology show upside across the board except for FE where we still struggle with the high debt load in our build up to fair value. We discuss valuation further on starting on slide 51 where we see consolidated P/E and EV/EBITDA multiples looking interesting across the board particularly when coupled with good dividend yields at low payout ratios. Source: Thomson Reuters, Company data, CS Estimates 5

6 % of 2019 EPS % of 2019 EPS Positive Case 1: Higher Capacity, Energy or Both Integrateds Have Real Leverage to Either Revenue Stream 2019 Earnings Sensitivity to $2 Increase in Power Price 2019 Earnings Sensitivity to a $50 Move in Capacity Prices 10.0% 9.0% 8.0% 7.0% 6.0% 5.0% 9.2% 7.0% 4.6% 12.0% 10.0% 8.0% 6.0% 10.1% 8.3% 8.3% 4.0% 3.0% 3.5% 4.0% 4.0% 2.0% 1.0% 2.0% 0.0% EXC FE PEG ETR 0.0% EXC PEG FE ETR EPS Impact: $0.31 $0.23 $0.14 $0.23 EPS Impact: $0.33 $0.25 $0.27 $0.27 The Integrateds still offer considerable earnings leverage to increases in power and / or capacity prices even on relatively small moves. With markets looking to place a higher premium on reliability and availability, we see upward support for profits through some combination of these drivers. Looking forward the ultimate decision is revenue allocation for generators between capacity and energy revenue. Regardless, the impact to earnings from a $2/MWh (left) or $50/MW-day (right) has a similar effect on EPS. Source: PJM, Company data, CS Estimates 6

7 Capacity Price Change ($/ MW-d) Positive Case 1: Higher Capacity, Energy or Both Even if Pressuring One, the Other Likely More than Offsets Breakeven on Energy and Capacity $90 $70 $50 $30 $10 -$10 -$6 -$4 -$2 $0 $2 $4 -$30 -$50 Change in Power Price ($/ MWh) ETR EXC FE PEG The support of higher power and / or capacity prices provides an offset to the other if concerned about deterioration, while leaving considerable room for upside with improvement in either input (company specific on slide 48) We appreciate the source of earnings uplift for generators will likely shift between power and capacity prices depending on capacity rule evolution and newbuild plant response, but the improvement in one driver can absorb fairly healthy potential deterioration in the other. For the Integrateds, ETR, EXC and FE have similar earnings sensitivity shapes between capacity and power prices given the big baseload fleet each operates where capacity revenues are spread across more MWhs of production; PEG s more balanced fleet benefits comparably more from capacity. Source: Company Data, CS Estimates, SNL Financial 7

8 Cash Cost ($/MWh) Bear Argument 1: Power Prices Can Continue to Fall Cash costs of Other Plants Help to Limit Price Downside as Well Cash Cost of Different Generation Types CCGT (ADHub) CCGT (PJMW) Nuclear NAPP (Barge) CCGT (PJME) CCGT (NIHub) PRB 8800 NAPP (Rail) CAPP (Rail) All-In Fuel O&M Environ Costs Capex PJME PJMW ADHub NiHub Capex Environ Costs O&M All-In Fuel Capex costs reflect the ongoing maintenance capex required to sustain the plants but not captured in earnings Environmental costs include the consumables to run equipment, representing a real burden for coal plants with PRB most impacted from TRONA costs O&M costs represent the historically observed costs of operating power plants specific to fuel / plant type All-in-Fuel costs represent the plant specific costs; For CCGTs we use current forwards at local natural gas basis with AD-Hub using Dominion South gas, PJM-W using a blend of Dominion South and TETCO-M3, and PJM-E using TETCO-M3 Bears argue weak gas and newbuild CCGTs will drive power prices appreciably lower than forwards. We think downside power price cases will ultimately be limited with the worst case scenario at the cash operating costs of marginal plants that should limit the amount of absolute power price downside, even assuming deteriorating gas prices. Against 2016 forwards, we already see prices rubbing against cash costs especially in Ni-Hub and AD-Hub. Source: Credit Suisse Estimates, SNL Financial, Company Data 8

9 IRR (%) Bear Argument 2: Newbuild CCGTs squeeze power prices Without healthy spark spreads, newbuild can t work Capacity and Power Price (No Utilization Change) - PJMW 14% 12% 11.1% 10.1% 10% 7.8% 8% 6.8% 6% 3.9% 4% 2.4% 1.2% 2% 12.0% 8.9% 5.2% 2.9% 0% -2% -1.1% $140 $160 $180 $200 $220 $240 $260 Capacity Price ($/MW-Day) $5 Spark Spread $10 Spark Spread $20 Spark Spread $30 Spark Spread Discount Rate The risk to power price argument also holds that newbuild CCGTs will lead to erosion in power prices because highly efficient plants burning cheap natural gas will weigh down prices at the margin generally based on the argument that plant heat rate times gas price plus O&M sets the price (aka gas plants don t need to make money on selling power). We think this argument is short-sighted because when we look at newbuild CCGT economics, even with good capacity prices we still need to see healthy spark spreads to cover costs and earn a return. Spark spread discipline has existed over time, suggesting the downside to power prices isn t as bad as pitched. Source: PJM, Credit Suisse estimates 9

10 1) Power Markets Poised to Structurally Improve Since the spring we have argued that power markets will benefit from the long awaited coal plant closures due to EPA MATS rules (slide 11), particularly as we still see relatively high dependence on plants set to close during peak periods (slide 13). Coal plant closures and ongoing challenges around winter gas deliveries are leading to increased power market volatility (slide 15) that should push power price realizations and forwards higher with time. Future power prices can find real upside even if we just see volatility in a select number of hours; a full year can really be shaped at the margins so losing lower cost plants that run during high demand periods should put upward pressure on those high value hours. 10

11 PJM SERC MISO WECC SPP CAISO NEPOOL NYISO ERCOT 2,002 1,817 1,734 1, Retirement (MW) 6,908 6,841 18,322 Retirement (MW) 24,737 1) 2015 and 2016 are Big Coal Plant Retirement Years Retirements by RTO Retirements by Year 30,000 25,000 20,000 18,000 16,000 20,000 15,000 14,000 12,000 10,000 10,000 8,000 6,000 5, ,000 2, PJM SERC MISO WECC SPP CAISO NEPOOL NYISO ERCOT Long awaited coal plant retirements due to EPA s MATS rules are now here, changing the profile of the US power generation fleet with PJM and SERC most impacted; we still think the yet to be announced MISO plant closures (not in the numbers above) will add to the list of closures. Loss of mid-merit coal plants will impact regional dispatch outlooks as still important contributions to regional supplies will soon be lost (slide 13 for PJM), increasing dependence on higher cost marginal resources than we have previously seen. Source: Company Data, SNL, CS Estimates 11

12 Capacity (GWs) 1) PJM Will Lose Capacity, Net of Gas Plant Additions PJM Capacity Additions and Retirements Capacity Retired Added 2014 Capacity Retired Added 2015 Capacity Retired Added 2016 Capacity Retired Added 2017 Capacity Starting Capacity Added Retirements To date 12 GWs of coal capacity have been retired in PJM with another 13 GWs slated for retirement through Based on our analysis we estimate that another ~4GWs of capacity could be retired (not in chart) as companies are forced to show their hands with the MATS deadline. Notably, the plants set to close still run hard and are an important slice of total generation in higher demand periods (slide 13); the closures will impact the market. Currently ~9 GWs of new capacity is under construction (slide 26) that will help to dampen the impact of closures but not at the same rate, creating some support for power and capacity pricing. Note: Plant additions only include plants under construction. Slide 21 shows capacity at varying phases of development. Source: SNL Financial 12

13 Utilization (%) % of Total Generation 1) Retiring Plants Still Needed, Especially in Peak Months Utilization Rates of PJM Plants To Be Retired Market Share of PJM Coal Plants to be Retired 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 6% 5% 4% 3% 2% 1% 0% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Plants retiring in continue to run hard with utilization rates and market share higher in 2014 than we have seen in recent years. We appreciate these plants run less than in the past, but are still called regularly during higher demand summer and winter periods. These plants have marginal dispatch costs of no more than $50-60 / MWh, providing an insurance policy for PJM in high demand periods. Losing this insurance policy will push up the price of marginal electricity particularly in the more volatile, high priced hours (slide 15). Source: SNL Financial 13

14 Retirement (MW) Retirement (MW) 1) Coal Closures Impact Rest of RTO Area the Most Coal Retirements in PJM Coal Retirements by Zone 12,000 10,000 8,000 6,000 4,000 2, ,645 10,167 2,704 2,020 2, ,204 12,000 10,000 8,000 6,000 4,000 2,000 0 Rest of RTO SWMAAC EMAAC WMAAC Important to appreciating the regional impact of pending coal plant closures, the Rest of RTO zone of PJM will face considerable closures in 2015 and 2016 (right chart) accounting for the bulk of retirements. While capacity prices already incorporate the shift of capacity resource from physical power plants to demand response (see slide 22 for uptake of DR), we think energy markets will continue to see sustained higher volatility particularly as the coal plants serving as needed mid-merit resources are taken off-line. Losing these plants in lower power price RTO regions AD-Hub, Ni-Hub could represent a more meaningful opportunity to boost future power prices and generator earnings favoring EXC and FE. Source: SNL Financial, Company Data, CS Estimates 14

15 $/MWh (Real-Time) $/MWh (Real-Time) 1) PJM Volatility Is Already Producing Higher Actual Prices 1Q Hourly Power Price vs Load (Capped at $400/ MWh) ,000 70,000 90, , , ,000 YTD Hourly Power Price vs Load Demand (MWs) ,000 1,800 1,600 1,400 1,200 1, ,000 70,000 90, , , , ,000 Demand (MWs) Q14 saw record high demand and hourly power prices, reflecting the strain on the power system with high demand, outages and gas delivery challenges. The elevated volatility should be seen (we think) as a reminder of the delivery challenges ahead particularly with power plant closures and as growing market share for gas generation puts a bigger call on spot gas deliveries in the winter (boosting power prices). We would expect plants to be more ready to run this winter but mix and gas availability will still add to volatility. Even with considerably lower summer 2014 demand, prices in 2014 have consistently cleared at higher prices than in prior years at similar demand levels, highlighting the tightening market conditions. Source: Velocity Suite 15

16 Hours Considered Super Peak Pricing Increase ($/MWh) Hours Considered Super Peak 1) Volatility Matters: A Few Hours Can Swing Total Price A Few Hours of High Prices Changes the Year Year YTD % % % % % % Observed Price at the Top Hours Peak % Year YTD % % % % % % Wtd Avg Annual Price Power Price Volatility Impact to ATC Price at varying Percentages of Super Peak Hours YTD % 5.00% 10.00% We think an interesting, and underappreciated, impact on overall power prices is the amount of full-year effect from a small number of hours that price at peak levels. Demonstratively on the top left table, the top 0.5% hours of demand increased the full year price by $ / MWh and the top 5% of hours increased the full-year price by $5-15 / MWh (top left table). In a near-dated future where mid-merit coal plants are retired, the prospects for increased volatility at peak hours increases creating prospects for higher full-year power prices. 1% of Hrs 5% of Hrs 10% of Hrs Source: SNL Financial, CS Estimates 16

17 % of 2016 EPS % of 2019 EPS 1) Meaningful EPS Leverage to Higher Power Prices 2016 Earnings Sensitivity to $2 Increase in Power Price 2019 Earnings Sensitivity to $2 Increase in Power Price 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 6.5% 5.1% 3.5% 2.9% 10.0% 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 9.2% 7.0% 4.6% 3.5% 2.0% 2.0% 1.0% 0.0% EXC FE PEG ETR 1.0% 0.0% EXC FE PEG ETR EPS Impact: $0.17 $0.15 $0.09 $0.17 EPS Impact: $0.31 $0.23 $0.14 $0.23 Looking across the group, EXC with over 200 TWh of generation has the most leverage to a power price recovery across all years with high single digit EPS upside on a $2/MWh increase above our current assumptions. We show the 2019 EPS impact appreciating still far away to highlight the full leverage to power prices since we assume no meaningful hedges are in place at that point. Source: Company data, CS Estimates 17

18 2) Potential RPM Auction Changes are Encouraging PJM wants to fix RPM to improve reliability and accountability In the last 6 weeks a series of changes have been proposed for the RPM capacity auction that, while inevitably will change and evolve, are a clear directional push by PJM and stakeholders to promote a more reliable power system going forward after the vulnerabilities seen last winter that should ultimately support better generator economics. PJM s proposals around a performance capacity metric that better compensates and demands more of a core fleet of capacity (~85% would need to meet the new performance standard), updated VRR demand curve, and revised Net CONE methodology all fit the bill of helping to deliver a more robust and dependable system. Demand Response (DR) rule changes could create an unexpected layer of upward support for RPM prices depending on resolution following the DC Appeals Court ruling that DR is a retail product rather than wholesale and by extension (and another FERC decision) is not eligible to be treated as capacity in RPM auctions. While early, the impact of stripping DR from PJM capacity could have a real impact, since DR has been 40-60% of total recent reserve margins and on a static basis would have pushed the 2017/18 auction to $282 / MW-day from $120. While some DR will likely find life again as a state regulated product that lowers the demand forecast that goes into reserve margin math, we have doubts whether the states would be able to develop the rules and mechanisms in time for the 2018/19 auction. We worry an air pocket for DR could drive capacity prices above a reasonable equilibrium price, potentially offering a false signal to build more new capacity than needed and leaving investors hard pressed to find confidence in what is the right durable to justify capitalizing (this is akin to the often unsuccessful way of boiling a frog by throwing it into a hot pot where it will just jump out). We see a lot of work ahead for PJM and the states, but do think the mission of lifting capacity prices to support greater system reliability can be realized to the benefit of customers and generators. 18

19 07/08 08/09 09/10 10/11 11/12 12/13 13/14 14/15 15/16 16/17 17/18 $MW-Day 2) RPM Has Been Consistently Unpredictable PJM RPM Auction Pricing $300 $250 $200 $150 $100 $50 $0 RTO MAAC EMAAC PS PSNORTH Capacity revenues represent a meaningful revenue stream for power generators but the significant variability in auction results has reduced the efficacy of RPM largely due to dilution from nongeneration capacity resources and rule changes by PJM. Responding to recent and expected reliability concerns, PJM is pursuing a range of changes to RPM (compounded by court action around demand response) that could meaningfully increase auction result outcomes from what we have seen in the past. Source: PJM 19

20 RTO MAAC EMAAC SWMAAC RTO MAAC EMAAC SWMAAC RTO MAAC EMAAC SWMAAC Capacity Price ($/MW-d) 2) New VRR Curve Would Increase the Auction Price PJM s Recommended VRR Curve Impact of Recommended VRR Curve Change Brattle Recommendation PJM Recommendation $250 $200 $150 $100 14% 16% 16% 16% 31% 9% 9% 9% 22% 22% 22% 22% 35% 30% 25% 20% 15% $50 10% $0 5% 2015/16 BRA 2016/17 BRA 2017/18 BRA Actual New VRR Delta (%) PJM has proposed an updated VRR demand curve that effectively changes the intersection points between reserve margin and multiplier on Net CONE, boosting the clearing price for capacity at most reserve margin levels than what was used in the past. PJM s recommended VRR curve would have resulted in higher past auction prices: 15/16 BRA results would have been higher by $20-26 / MW-day (~14%-16% increase) 16/17 BRA results would have been higher by $11-19 / MW-day (~9%-31% increase) 17/18 BRA results would have been higher by $26 / MW-day (~22% increase) Source: Triennial Review of RPM: Draft Study Results 20

21 2) Capacity Performance Is Coming, Would Increase Price Category Capacity Performance Availability Expectations Limitations Penalties Penalty Window All hours of the year none Real-Time LMP Charge Year Round - Hot Weather Alert / Cold Weather Alert / Max Emergency Alerts Base Capacity All hours of the year none Real-Time LMP Charge for generators / Current DR penalties for DR Summer Extended DR Limited DR For Generators - Max Emergency Generation Loaded in summer or winter; Current rules for DR May through October 10 hours per day Current DR penalties Summer - DR activations June through September 10 x 6 Current DR penalties Summer - DR Activations The table above lays out the capacity products as detailed by the PJM Staff in their Aug 20 th proposal with the Capacity Performance product aimed at providing PJM with 1) fuel security, 2) enhanced operational performance in peak periods (winter and summer), 3) high generation availability, 4) flexible operation parameters and 5) operational diversity. While we believe that some form of a capacity performance product will have a positive and upward impact on capacity price, there are still a number of details to be worked out before putting numbers around that impact, with the main points of concern as highlighted by stakeholders comments: 1) penalty structure and terms for non-compliance, 2) targeted amount of capacity qualifying under the capacity performance product and 3) how to deal with the transitional period between 2015 and the next auction period of 2018/2019. Source: Energy Information Administration 21

22 Reserve Margin 2) DR Has Been ~40-60% of PJM Reserve Margin Total PJM RTO Reserve Margin (08/09-16/17) DR+EE % of Reserve Margin 0.0% 5.0% 10.0% 15.0% 25.0% Total Reserve Margin 22.5% 20.0% 17.5% 17.8% 17.5% 16.5% 18.1% 20.9% 20.2% 19.6% 20.2% 21.1% 19.7% 08/09 09/10 10/11 0.5% 0.8% 0.8% 15.0% DR + EE 11/12 1.2% 12.5% 10.0% 7.5% Traditional Generation 5.0% 2.5% 0.0% 08/09 09/10 10/11 11/12 12/13 13/14 14/15 15/16 16/17 17/18 12/13 13/14 14/15 15/16 16/17 17/18 6.8% 7.8% 11.9% 11.5% 9.7% 8.8% DR has accounted for 7-12% of the ~20% reserve margin or between 40-60% of the total reserve margin, leaving DR as an essential piece of resource from a reliability standpoint. Looking historically, if PJM were to remove DR from the auction, all else equal the price in the 17/18 auction would have cleared at $280 / MW-day instead of the $120 actuals although we think the impact is arguably less since newbuild capacity and generator bidding would likely have been different. Source: PJM 22

23 Reserve Margin Reserve Margin 2) RTO Zone Most Exposed to DR Rest of RTO Reserve Margin (incl. ATSI) MAAC Reserve Margin 25.0% 22.5% 20.9% 20.0% Total Reserve Margin 20.2% 19.6% 20.2% 21.1% 19.7% 25.0% 22.5% 20.9% 20.0% Total Reserve Margin 20.2% 19.6% 20.2% 21.1% 19.7% 17.5% DR + EE 17.5% DR + EE 15.0% 15.0% 12.5% 12.5% 10.0% 10.0% 7.5% 7.5% 5.0% 5.0% 2.5% Traditional Generation 2.5% Traditional Generation 0.0% 12/13 13/14 14/15 15/16 16/17 17/18 0.0% 12/13 13/14 14/15 15/16 16/17 17/18 DR resources contribute the greatest to the rest of RTO area (71% of total PJM capacity) where DR has been 45-75% of the total rest of RTO reserve margin. We would expect any changes to the RPM auction that reduce the share of DR resource because of the recent court decision or by setting more rigorous performance standards for capacity should support RTO prices and likely further link capacity prices between RTO and MAAC as seen in the last recent auction. Source: PJM 23

24 2) Locational Deliverability Areas (LDAs) of PJM PJM Zones Source: PJM 24

25 3) Newbuild CCGTs Approaching Economic Returns, Especially with a Higher RPM Outlook Newbuild CCGTs are at / near reasonable economic returns using current power forwards, particularly if capacity prices shift higher (slides 30-33). We still believe meaningful newbuild will require IRRs higher than mid-cycle (8% unlevered) given inconsistency of power markets and immediate market dilution that occurs when you add new capacity (the forwards are based on current market so adding more MWs will push both power and capacity prices lower), but improving economics will likely attract more investment. PJM currently has 9 GW under construction and 17 GW committed from prior auctions (slide 26), although not all are being built (slide 27). 25

26 3) Newbuilds are Expected to Come to PJM Already PJM Gas Plants Under Construction / In Development 6,000 5,000 4,000 3,000 2,000 1, Construction Begun Early Development Advanced Development Announced For plants to be online for 2016, they would need to be under construction today. In 2017 there are a lack of plants under construction (almost none), which worries us that the necessary capacity might not be available when needed. Plant Name Developer MW % of Total In Servi ce Status Warren County Power Station Dominion Resources, Inc. 1,329 4% 2014 CB Clayville Unit #1 Vineland City of - (NJ) 64 0% 2015 CB Garrison Energy Center Calpine Corporation 618 2% 2015 CB Nelson Energy Center Invenergy LLC 571 2% 2015 CB Newark Energy Center EIF Management, LLC 735 2% 2015 CB Perryman Exelon Corporation 120 0% 2015 CB Brunswick County Power Station Dominion Resources, Inc. 1,360 4% 2016 CB CPV St Charles Energy Center Marubeni Corporation, Warburg Pincus LLC, Co 2,900 9% 2016 CB Panda Liberty Generating Station (Moxie Liberty) Panda Energy International, Inc % 2016 CB Patriot Power Generation Plant (Moxie Patriot) Panda Energy International, Inc % 2016 CB Woodbridge Energy Center Competitive Power Ventures Holdings, LLC, To 2,100 7% 2016 CB Oxbow Creek Energy Bregal Energy 20 0% 2015 AD Roundtop Bregal Energy 21 0% 2015 AD Good Spring NGCC 1 & 2 ITOCHU Corporation, EmberClear Corp. 1,340 4% 2016 AD Oregon Clean Energy Project North America Project Development LLC 799 3% 2017 AD Stonewall Combined-Cycle Project Panda Energy International, Inc., Green Energy P 1,500 5% 2017 AD Wildcat Point Generation Facility Old Dominion Electric Cooperative 1,000 3% 2017 AD Mattawoman Power Plant Panda Energy International, Inc % 2018 AD Souderton Cogeneration Project Blue Earth Inc. 19 0% 2014 ED Bayles Energy Project Bregal Energy 21 0% 2015 ED NSF Indian Head Facility (Naval) US Navy 30 0% 2015 ED Red Glen Energy Plant Bregal Energy 21 0% 2015 ED West Deptford LS Power Group 1,500 5% 2015 ED Florey Knobb IC Plant Florey Knob Energy LLC 21 0% 2016 ED Hickory Run Energy Station LS Power Group 900 3% 2016 ED Rolling Hills Generating - CC Tenaska Energy Inc 1,274 4% 2016 ED Carroll County Energy Center Advanced Power AG 742 2% 2017 ED CPV Smyth Energy Center Competitive Power Ventures Holdings, LLC 714 2% 2017 ED Keys Energy Center Genesis Power LLC 755 2% 2017 ED Sunbury Repower CC Corona Power, LLC 1,065 3% 2017 ED York 2 Energy Center Calpine Corporation 760 2% 2017 ED Middletown Energy Center NTE Energy LLC 525 2% 2018 ED Moundsville Power Project Moundsville Power LLC 549 2% 2018 ED Oakwood Hills Energy Center Northland Power Incorporated, Enventure Partn 1,350 4% 2018 ED Tenaska Lebanon Valley Generating Station Tenaska Energy Inc 950 3% 2018 ED Westmoreland Generating Station Tenaska Energy Inc 950 3% 2018 ED Elkton IC Merck & Company, Inc. 1 0% 2015 A Triton East and West Cogeneration Triton College 0 0% 2015 A Deepwater Repowering Calpine Corporation 370 1% 2017 A Lordstown Generating Station Clean Energy Future, LLC 800 3% 2018 A Shell Chemical Appalachia CC Plant Shell Chemical Appalachia LLC 265 1% 2018 A TOTAL 30,682 Note: CB = Under Construction,; AD = Advanced Development; ES = Early Development; A = Announced Source: SNL Financial 26

27 Capacity (MWs) 3) But Not All Committed New Generation is Being Built Capacity Resource Cleared vs Capacity Additions (incl Planned) 5,000 4,000 3,879 3,000 2,000 1, ,000-2,000-3,000-2,404-4,000-5,000-6,000-4,857 '13/'14 '14/'15 '15/'16 '16/'17 '17/'18 Built vs. Comitted at RPM Difference RPM Incremental Resource RPM Incremental Resource Additions Auction Period '13/'14 '14/'15 '15/'16 '16/'17 '17/'18 Total Total Capacity 203, , , , ,964 - Capacity Offered 1,278 1,036 6,844 5,195 6,128 20,481 Capacity Cleared 1, ,899 4,282 5,927 16,802 Actual Additions / Planned Additions 455 4,295 4,334 1,878 1,071 12,032 Of the 17 GW of plants that have cleared the last 5 auctions, there is currently ~5 GW of capacity that has been committed but is not yet under construction (which it would need to be in order to hit the 2016/17 planning period). The failure to build committed capacity (even if economically repurchased in supplemental auctions) remains a contentious issue as PJM focuses on physical generation support for reliability, especially now that PJM is contemplating new penalties on not delivering. Source: PJM, SNL Financial, CS estimates *Note: Cleared vs Non-Clearing Gap for 16/ 17 auction period subject to change as plant additions are finalized/timelines released 27

28 Capacity Factor (%) 3) PJM is Already in a CCGT Renaissance of Sorts PJM Gas Additions (Actual and In Development) Capacity Factors for Operating Gas Plants (> 500 MW) 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 - Total Additions Construction Begun Advanced Development Early Development Announced PJM saw a big construction cycle in the early 2000s with the next several years showing a healthy planned newbuild plan. An interesting question to watch will be the construction capacity to add substantially more MWs with a tighter supply of trade labor due to demand for other industrial construction (LNG, chemical plants, pipelines, etc) Year Placed into Service Observed utilization rates for prior CCGT construction has generally been in the 60-80% level (some lower) which is important to consider when looking at newbuild plant economics (slides 34-38); how hard new capacity can run and how it impacts run-times for existing CCGTs is an important consideration for developers. Source: SNL 28

29 ) US Overall Expecting an Uptick in Newbuild Gas US Total Gas Additions (Actual and In Development) US New Build Capacity Development 50,000 30,000 45,000 40,000 35,000 25,000 20,000 30,000 15,000 25,000 20,000 15,000 10,000 5,000 10,000 5, Construction Begun Advanced Development Early Development Announced The US saw a huge gas build out in the early 2000s and is seeing an increase in planned newbuild generation in the wake of coal plant closures, bolstered by limits on what newbuild generation capacity can be built in the US given CO2 concerns and growing confidence in durably lower gas prices. Looking at projects under construction, there are a lot of proposed plants that are not currently under construction starting in As the end of 2014 approaches, we either need to see more plants under construction to deliver on proposed in-service dates or else look for slippage beyond 2017 (since construction is generally months for CCGTs). Source: SNL 29

30 IRR (%) 3) Capacity Prices Can Help Drive Improved Returns IRR Sensitivity from Capacity Price Change (65% Utilization, Base Spark Spread) 10.0% 9.5% 9.0% 8.5% 8.0% 7.5% 7.0% 6.5% 9.6% 9.2% 9.1% 8.8% 8.7% 8.6% 8.2% 8.0% 8.4% 7.8% PJMW PJME AD-Hub 8.4% 7.9% 7.5% 7.2% 7.0% Using current forward power prices and most recent capacity prices, newbuild CCGTs in PJM are hovering around the 8% after-tax, unlevered IRRs (dark blue bars) we have generally assumed are needed to justify investment. $0 $10 $25 $50 $75 Further improvement in capacity prices due to the design changes discussed in the preceding section would further support newbuild economics (the colored adders), although we should stress that significant investment when just clearing newbuild economics likely yields inadequate returns since forwards are based on what exists in the market today; adding capacity will in part weigh on economics that will hinder target returns. We still think developers spending $1 BN+ on a new CCGT need to be judicious in their returns triggers to justify investment. Note: Base assumptions in blue use September 26 forwards and flat RPM prices of $120/ MW-day for PJM-W, PJM-E and AD-Hub. See Appendix (slide 58) for more detail. Model is available on request. 30

31 Utilization Rate Utilization Rate Utilization Rate 3) Newbuild IRR Sensitivity (Capacity Price vs Utilization) New CCGTs look even better with higher capacity prices PJMW - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 50% 5.9% 6.1% 6.4% 6.9% 7.3% 60% 7.2% 7.4% 7.6% 8.1% 8.5% 65% 7.8% 8.0% 8.2% 8.7% 9.1% 70% 8.4% 8.5% 8.8% 9.2% 9.6% 80% 9.5% 9.7% 9.9% 10.3% 10.7% PJME - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 50% 6.5% 6.7% 7.0% 7.4% 7.8% 60% 7.8% 8.0% 8.2% 8.6% 9.0% 65% 8.4% 8.6% 8.8% 9.2% 9.6% 70% 9.0% 9.2% 9.4% 9.8% 10.2% 80% 10.2% 10.3% 10.5% 10.9% 11.2% AD-Hub - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 50% 5.2% 5.4% 5.8% 6.3% 6.7% 60% 6.4% 6.6% 6.9% 7.4% 7.8% 65% 7.0% 7.2% 7.5% 7.9% 8.4% 70% 7.6% 7.7% 8.0% 8.4% 8.9% 80% 8.6% 8.8% 9.0% 9.4% 9.8% Examining the returns on building a CCGT at various utilization rates and capacity factors, they are becoming more attractive today. However, if capacity, spark spreads or utilization rates were to inch higher the returns would become even more attractive. Note: Base assumptions in red use September 26 forwards and flat RPM prices of $120/ MW-day for PJM-W, PJM-E and AD-Hub. See Appendix (slide 58) for more detail. Model is available on request. 31

32 Chg in Spark Spread Chg in Spark Spread Chg in Spark Spread 3) IRR Sensitivity (Capacity Price vs Power Price) Spark spread expansion helps new plants meaningfully PJMW - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 (4.0) 6.5% 6.7% 7.0% 7.4% 7.9% (2.0) 7.2% 7.3% 7.6% 8.1% 8.5% % 8.0% 8.2% 8.7% 9.1% % 8.6% 8.8% 9.2% 9.6% % 9.1% 9.4% 9.8% 10.2% PJME - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 (4.0) 7.3% 7.5% 7.7% 8.1% 8.5% (2.0) 7.9% 8.0% 8.3% 8.7% 9.1% % 8.6% 8.8% 9.2% 9.6% % 9.1% 9.4% 9.7% 10.1% % 9.7% 9.9% 10.3% 10.6% AD-Hub - IRR on Newbuild CCGT Chg in Capacity Price ($/ MW-d) $0 $10 $25 $50 $75 (4.0) 5.6% 5.9% 6.2% 6.6% 7.1% (2.0) 6.3% 6.5% 6.8% 7.3% 7.7% % 7.2% 7.5% 7.9% 8.4% % 7.8% 8.1% 8.5% 8.9% % 8.4% 8.7% 9.1% 9.5% If spark spreads continue to move upward (slide 43), a $4/MWh increase in spark spreads would drive at least a 1% increase in returns. Note: Base assumptions in red use September 26 forwards and flat RPM prices of $120/ MW-day for PJM-W, PJM-E and AD-Hub. See Appendix (slide 58) for more detail. Model is available on request. 32

33 Utilization Rate Utilization Rate Utilization Rate 3) IRR Sensitivity (Power Price vs Utilization) Utilization rates will drive returns but not as flexible PJMW - IRR on Newbuild CCGT Chg in Spark Spread (4.0) (2.0) % 4.8% 5.3% 5.9% 6.4% 6.9% 60% 5.9% 6.6% 7.2% 7.8% 8.3% 65% 6.5% 7.2% 7.8% 8.4% 9.0% 70% 7.0% 7.7% 8.4% 9.0% 9.6% 80% 8.1% 8.8% 9.5% 10.2% 10.8% AD-Hub - IRR on Newbuild CCGT Chg in Spark Spread (4.0) (2.0) % 4.0% 4.6% 5.2% 5.8% 6.3% 60% 5.1% 5.8% 6.4% 7.0% 7.6% 65% 5.6% 6.3% 7.0% 7.6% 8.3% 70% 6.1% 6.9% 7.6% 8.2% 8.9% 80% 7.1% 7.9% 8.6% 9.3% 10.0% PJME - IRR on Newbuild CCGT Chg in Spark Spread (4.0) (2.0) % 5.5% 6.0% 6.5% 7.0% 7.5% 60% 6.7% 7.3% 7.8% 8.3% 8.9% 65% 7.3% 7.9% 8.4% 9.0% 9.5% 70% 7.8% 8.4% 9.0% 9.6% 10.2% 80% 8.9% 9.5% 10.2% 10.8% 11.4% Although a utilization rate of 65% is a reasonable assumption for a new CCGT, some of the CCGTs of the early 2000s build out are achieving utilization rates of ~80%. A new CCGT would be expected to be dispatched more frequently in the near team due to more efficient technology which enables a lower variable cost profile although overall fleet dispatch profiles could still act as a limiting factor on realized utilization rates. Note: Base assumptions in red use September 26 forwards and flat RPM prices of $120/ MW-day for PJM-W, PJM-E and AD-Hub. See Appendix (slide 58) for more detail. Model is available on request. 33

34 4) Newbuild CCGTs Will Weigh Modestly on Power Prices but Unlikely to Crush Prices (as some fear) The effect and really the goal of higher capacity prices in PJM is to attract new physical capacity that will inevitably lead to lower power prices as newer, more efficient generation lowers clearing prices and more physical capacity reduces some of the volatility that is likely to be seen in That said, cash operating costs for a range of power plants (slide 36) will provide price support at levels higher than some fear and close to where we are in the RTO regions since existing plants cannot survive on negative cash margins. Further, newbuild CCGTs need revenues from both capacity and positive energy margins to hit target IRRs (slide 37); those simply multiplying a CCGT heat rate by a gas price to calculate future power prices ignore the need for positive CCGT margins and the history of protecting spark spreads (slide 38). Looking at the Integrateds, slide 48 shows the earnings equilibrium between capacity and power price changes; under most scenarios economics still look better than current forwards. 34

35 Marginal Power Price ($/MWh) 4) New CCGTs Modestly Lower the Clearing Price Impact of New CCGT Plants on PJM Dispatch Curve with DR Removed $80 $70 $60 $50 $40 $30 $20 $ Cl eari ng Pri ce MWs of CCGTs ATC On Peak Of f Peak Status Quo , % Chg -0.5% -0.6% -1.3% $/MWh Chg , % Chg -1.5% -0.7% -7.1% $/MWh Chg , % Chg -5.8% -1.3% -12.1% $/MWh Chg $0 0 25,000 50,000 75, , , ,000 Capaity (MWs) Nuclear Gas Coal Oil Renewables & Others New Gas Net Import New Wind New Solar ATC On Peak Off Peak Running our PJM dispatch curve, we see the addition of 10 GW of newbuild CCGTs beyond those already in queue lowering our observed ATC (around the clock) power price by ~$ / MWh, which is disruptive but not as terminal as often feared; an extra 15 GW would be more painful. We appreciate actual dispatch behavior will often be different than an optimal model run but we think the magnitude of change is relevant to appreciating market impact. Source: PJM, Energy Velocity, CS Estimates 35

36 Cash Cost ($/MWh) 4) Power Plant Cash Costs Provide Price Support Cash Costs for Different Generation Types vs 2016 Power Price Forwards CCGT (ADHub) CCGT (PJMW) Nuclear NAPP (Barge) CCGT (PJME) CCGT (NIHub) PRB 8800 NAPP (Rail) CAPP (Rail) All-In Fuel O&M Environ Costs Capex PJME PJMW ADHub NiHub Capex Environ Costs O&M All-In Fuel Capex costs reflect the ongoing maintenance capex required to sustain the plants but not captured in earnings Environmental costs include the consumables to run equipment, representing a real burden for coal plants with PRB most impacted from TRONA costs O&M costs represent the historically observed costs of operating power plants specific to fuel / plant type All-in-Fuel costs represent the plant specific costs; For CCGTs we use current forwards at local natural gas basis with AD-Hub using Dominion South gas, PJM-W using a blend of Dominion South and TETCO-M3, and PJM-E using TETCO-M3 Above we show the fully loaded cash cost for the typical power plant based on fuel input. We think downside power price cases will ultimately be limited with the worst case scenario at the cash operating costs of marginal plants that should limit the amount of absolute power price downside, even assuming deteriorating gas prices. Against 2016 forwards, we already see prices rubbing against cash costs especially in Ni-Hub and AD-Hub. Source: Credit Suisse Estimates, SNL Financial, Company Data 36

37 IRR (%) 4) New Build CCGTs Need Energy Margin Too Capacity and Power Price (No Utilization Change) - PJMW 14% 12% 11.1% 10.1% 10% 7.8% 8% 6.8% 6% 3.9% 4% 2.4% 1.2% 2% 12.0% 8.9% 5.2% 2.9% 0% -2% -1.1% $140 $160 $180 $200 $220 $240 $260 Capacity Price ($/MW-Day) $5 Spark Spread $10 Spark Spread $20 Spark Spread $30 Spark Spread Discount Rate Part of the bear argument against power prices with the construction of newbuild CCGTs especially if also negative on natural gas prices is that power prices will fall to the cash cost of the CCGT using plant heat rate times gas price (and then sometimes adding O&M). The problem with this thesis is newbuild CCGTs need capacity revenues and positive spark spreads to support adequate IRRs. The chart above shows the IRRs for a generic CCGT at PJM-West using a constant spark spread, capacity price (x-axis) and 75% utilization rate. This argues that newbuild CCGTs need a price to succeed that is heat rate times gas price plus spark spread ($3.50 gas x 6.5 heat rate + $20 / MWh spark = $42.75 / MWh, well above the simple $3.50 x 6.5 heat rate + $5 O&M = $28 / MWh). Source: PJM, Credit Suisse estimates 37

38 01/12 05/12 09/12 01/13 05/13 09/13 01/14 05/14 09/14 Spark Spread ($/ MWh) 01/12 05/12 09/12 01/13 05/13 09/13 01/14 05/14 09/14 01/12 05/12 09/12 01/13 05/13 09/13 01/14 05/14 09/14 Spark Spread ($/ MWh) Spark Spread ($/ MWh) 4) Plus, Spark Spreads Have Been Durable PJMW Forward Spark Spreads Using Dominion South / M3 Gas PJM-E Forward Spark Spreads using TETCO M3 Gas $22 $20 $18 $16 $14 $12 $10 $8 $22 $20 $18 $16 $14 $12 $10 $21 $19 $17 $15 $13 $11 $9 $7 $5 Cal 14 Cal 15 Cal 16 AD-Hub Forward Spark Spreads Using Dominion South Gas Cal 14 Cal 15 Cal 16 Cal 14 Cal 15 Cal 16 Further thinking through the pressure on power prices with newbuild CCGTs and lower gas prices, we have seen durably healthy spark spreads in PJM forwards for extended periods of time operators protect spark spreads because they have bills they need to pay. We see this as further affirmation that power prices will find more downside protection than being argued (and is why we build our base power price forecasts off of spark spreads vs heat rates). Source: SNL 38

39 5) Natural Gas Remains the Wild Card, Creating Sentiment Risk Against Improving Power Market Fundamentals US natural gas production has continued to surprise to the upside with better than expected and accelerating volume growth (slide 40) that when paired with the mild summer has led to considerable refill in natural gas inventories (slide 41). While we remain optimistic around the fundamental improvement in the power markets, we would be remiss to ignore the correlation between gas and power prices (slides 42-43) and subsequently power and stock prices (slides 44-45) which could push the group around particularly in the near-term irrespective of fundamentals. 39

40 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 May-10 Jul-10 Sep-10 Nov-10 Jan-11 Mar-11 May-11 Jul-11 Sep-11 Nov-11 Jan-12 Mar-12 May-12 Jul-12 Sep-12 Nov-12 Jan-13 Mar-13 May-13 Jul-13 Sep-13 Nov-13 Jan-14 Mar-14 May-14 Jul-14 Sep-14 (3.2) (1.4) (1.5) (1.0) (0.3) (0.2) (0.2) ) Gas Production Growth Continues to Beat Year-on-Year Dry Gas Production (1) (2) (3) (4) After a fairly muted gas production ramp in 2013 and early 2014, we have seen year-on-year growth increasing materially as producers benefit from more productive wells than originally expected and increased gas take-away capacity has allowed volumes to beat on the upside. The issue to watch from here will be the rate of volume growth which could sustain at a high level with additional takeaway capacity coming on-line in 4Q14 and producers still realizing good returns even if assuming more stressed downside cases to gas prices. We admittedly see big production growth and mild weather representing the greatest risk to the near-term power story. Source: Bentek 40

41 5) Gas Storage Approaching More Normal Levels Contrary to our assumptions from earlier this year, natural gas inventories have rebounded sharply with the combination of better than expected natural gas production growth (slide 40) and the impact of mild summer weather that meaningfully hampered power demand and need for gas generation. Credit Suisse estimates still call for gas inventories to be below normal at the start of the winter gas season on November 1 but the rate of production growth and pick up in demand will be bigger determinants in how natural gas moves this winter. Source: Energy Information Administration 41