2015 Outlook. Recommendation Change. Deutsche Bank Markets Research. Industry Solar. North America United States. Date 8 January 2015

Transcription

1 Deutsche Bank Markets Research North America United States Industrials 2015 Outlook Industry Date 8 January 2015 Recommendation Change Vishal Shah Research Analyst (+1) vish.shah@db.com Overall Thoughts We believe the recent volatility in solar stocks, driven largely by oil price weakness, presents an attractive entry point for investors as we expect 2015 to be a year of stable industry pricing and accelerating volume growth. We expect a balanced supply demand outlook as strong demand from the US and improving demand from China/other emerging solar mkts offsets any potential demand weakness in the UK/Japan. While weak oil prices could remain an overhang, our work suggests very little impact on solar demand fundamentals and expect several company specific positive catalysts in terms of execution of new/existing yieldcos. Our top picks include SUNE, SCTY, VSLR and SPWR. Q4 earnings season/q1 seasonality & Fundamentals: While solar stocks gave up nearly all of the outperformance during Q4'14, mostly due to declining oil prices, we believe Q4 earnings of most companies would generally be inline/ahead of expectations. China demand in 2014 could turn out to be 9-10GW vs expectations of 13GW, but we believe strong demand from markets such as UK/Japan would likely cover up any shortfall in China demand. We also expect companies to talk about incremental progress in the permitting and payment process in China and believe a number of companies could get closer to launching yieldcos of international assets during 1H15 timeframe. While normal seasonality could likely impact China demand in 1H15, we expect UK/Japan to act as primary drivers for strong 1H volumes/margins. Consensus estimates have generally come down and we believe Q1 guidance from most companies would be more or less inline with expectations. We expect other company specific catalysts such as announcements of yieldcos/project sales along with acquisition of new project pipelines to act as important catalysts for solar stocks over the next 3 months. Q1 seasonality typically affects module pricing, but we expect strong demand from higher priced markets such as the UK and Japan to drive pricing and margin improvement in Q1. We also see limited supply growth in the near term and as such expect relatively stable pricing environment. Oil price impact on demand: As explained in a few sections of the note, oil represents only about 5% of global electricity production and in some of the important solar markets such as US, China, oil based electricity generation is less than 5% of the total. Moreover, the fuel cost of oil based electricity generation even at $50 oil prices is in the 7-9c/kWh range and as shown in the note, the marginal electricity cost is higher than solar in many regions worldwide. Bottom line is that oil prices do not have a material impact on solar demand. Other important themes: 1) Strong, diverse demand drivers; US rooftop market will be the key highlight and utilities will also start competing in the solar market. Project pipelines and margins will continue to expand. Expect cost reduction to also drive module margins higher, 2) Yieldcos will continue to remain popular source of project capital funding, but companies with first mover advantage will be in a better position to build pipelines/acquire development assets, 3) Policy environment will continue to improve. Given that anti-dumping duties for Chinese modules are expected to be completely removed, we expect companies to see margin expansion. Jerimiah Booream-Phelps Research Associate (+1) jerimiah.booream-phelps@db.com Key Changes Company Target Price Rating YGE.N 5.00 to 3.00(USD) Buy to Hold Source: Deutsche Bank Top picks SunEdison (SUNE.N),USD18.16 Vivint (VSLR.N),USD8.16 SunPower (SPWR.OQ),USD23.85 City (SCTY.OQ),USD49.32 Source: Deutsche Bank YGE Downgrade / Sector Valuation/Risks We are downgrading Yingli Green Energy to a Hold based on ongoing balance sheet concerns and limited positive catalysts/financial flexibility. Please see page 33 for details. We value most solar companies using a mix of PE multiples and DCF s. Sector Risks include: 1) Grid prices change slower/faster than expected 2) Policy changes in key markets (China/Japan/US) 3) Input price volatility Deutsche Bank does and seeks to do business with companies covered in its research reports. Thus, 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. DISCLOSURES AND ANALYST CERTIFICATIONS ARE LOCATED IN APPENDIX 1. MCI (P) 148/04/2014. Buy Buy Buy Buy

2 Capacity Additions (GW) 8 January 2015 Key Themes 1) Expect more countries to reach grid parity in 2015: Unsubsidized rooftop solar electricity costs anywhere between $0.13 and $0.23/kWh today, well below retail price of electricity in many markets globally. The economics of solar have improved significantly due to the reduction in solar panel costs, financing costs and balance of system costs. We expect solar system costs to decrease 5-15% annually over the next 3+ years which could result in grid parity within ~50% of the target markets. If global electricity prices were to increase at 3% per year and cost reduction occurred at 5-15% CAGR, solar would achieve grid parity in an additional ~30% of target markets globally. We believe the cumulative incremental TAM for solar is currently around ~140GW/year and could potentially increase to ~260GW/year over the next 5 years as solar achieves grid parity in more markets globally and electric capacity needs increase. Figure 1: is Increasingly Important to the Global Generation Mix capacity adds (as % of global capacity adds) are well into the double digits 25% 20% % % % 0 0% Non Wind/ Capacity Wind Capacity % Source: EIA, GWEC, Deutsche Bank Estimates Note: 5% growth rate in total capacity additions from Wind Installs assumed flat at 5 year average of 39.5GW. Installs are DB ests 2) demand growth should be increasingly more diverse and a greater percentage of demand should be from sustainable markets. While some uncertainty about growth in markets such as Japan and the UK will remain an overhang on solar stocks in 1H15, we expect sector volatility to decrease in 2H15 timeframe as demand from the US, China and other emerging markets takes off. 3) Rooftop solar demand in the US should accelerate, especially as leasing companies expand in more states, get new sources of financing and customer adoption increases ahead of the 2016 expiration of the ITC. 4) 2015 will be a transformative year for the US utilities - we expect some utilities to enter the residential solar market and compete directly with city and Vivint whereas other utilities would most likely continue to lobby against the growth of distributed solar market. Page 2

3 5) We expect a relatively stable supply demand outlook in 2015 timeframe as most companies are planning only limited supply growth. Q1 seasonality typically affects module pricing, but we expect strong demand from higher priced markets such as the UK and Japan to drive pricing and margin improvement in Q1. We also see limited supply growth in the near term and as such expect relatively stable pricing environment. 6) Our constructive view on solar is largely dependent on improving cost curve of the underlying technology. Overall solar system costs have declined at ~15% CAGR over the past 8 years and we expect 40% cost reduction over the next 4-5 years as a solar module costs continue to decline, panel efficiencies gradually improve, balance of system costs decline due to scale and competition, global financing costs decline due to development of new business models and customer acquisition costs decline as a result of increasing customer awareness and more seamless technology adoption enabled by storage solutions. 7) YieldCos will continue to gain popularity among investors and solar companies looking to lower cost of capital will be an important year for YieldCos, both in terms of asset and geographic diversification. We expect solar only YieldCos to become more active in wind and other renewable assets; YieldCos with emerging market exposure to go public and more companies to spin off their solar assets into YieldCos; 8) Lower cost of capital will become a key growth enabler for companies with first mover advantage in the YieldCo space and also act as a significant catalyst in lowering the cost of solar power in emerging markets such as India. 9) Project development pipelines of most companies would increase at a much faster pace, both due to organic and inorganic growth. We expect developers with an off taker in the form of a YieldCo to take on increasingly riskier projects. because the demand for fully developed projects is expected to remain relatively high, we expect more development capital to flow into the sector. 10) Project margins will continue to increase in 2015 before reaching peak levels in 2016 timeframe. Low financing costs and strong demand from YieldCos will continue to drive prices of downstream projects higher. While regional mix may have some impact on project margins, we do not anticipate significant variations in different regions and expect margins closer to 20% levels in the near term. 11) 2015 earnings for most solar companies would become increasingly more volatile. As companies decide to hold more projects in balance sheet, we expect a near term negative impact both on income statement and balance sheet (in terms of increasing working capital requirements). We also expect earnings to be somewhat lumpy as timing of project completions and sales would be harder to predict. 12) Policy focus will remain front and center in many markets. Within the US, the extension of ITC (set to expire in 2016), extending the MLP status to renewables, net metering 2.0 in California, grid integration in Hawaii and grid access charges in several states would be some of the policy items impacting solar supply chain. Additionally, trade case development in China, US and other global markets would be an important theme to watch in Adverse Page 3

4 trade policies certainly pose the risk of slowing down growth in important solar markets, especially in light of the recent gas price weakness. That said, we believe a positive resolution of these trade disputes is likely and would set the stage for stronger growth in ) Q4 earnings season/q1 seasonality: While solar stocks gave up nearly all of the outperformance during Q4'14, mostly due to declining oil prices, we believe Q4 earnings of most companies would generally be inline/ahead of expectations. China demand in 2014 could turn out to be 9-10GW vs expectations of 13GW, but we believe strong demand from markets such as UK/Japan would likely cover up any shortfall in China demand. We also expect companies to talk about incremental progress in the permitting and payment process in China and believe a number of companies could get closer to launching yieldcos of international assets during 1H15 timeframe. While normal seasonality could likely impact China demand in 1H15, we expect UK/Japan to act as primary drivers for strong 1H volumes/margins. Consensus estimates have generally come down and we believe Q1 guidance from most companies would be more or less in line with expectations. We expect other company specific catalysts such as announcements of yieldcos/project sales along with acquisition of new project pipelines to act as important catalysts for solar stocks over the next 3 months. 14) Oil price impact on demand: As explained in a few sections of the note, oil represents only about 5% of global electricity production and in some of the important solar markets such as US, China, oil based electricity generation is less than 5% of the total. Moreover, the cost of oil based electricity generation even at $50 oil prices is the 7-9c/kWh range and as shown in the note, the marginal cost is higher than solar in many regions worldwide. Bottom line is that oil prices do not have a material impact on solar demand. 15) How to Make Hay While the Sun Shines? The solar sector has been generally under owned by institutional investors and expect greater institutional ownership to drive near term positive momentum for the sector. We expect a number of new business models focused on the downstream part of the value chain to emerge and expect innovative private companies to drive cost improvement/solar adoption. Both of these set of companies stand to generate significant shareholder value, in our view. We believe companies involved in financing/downstream part of the value chain stand to generate the most significant shareholder value in the near term. We expect these companies to be in a unique position to take advantage of the financing arbitrage offered by inefficient private markets and publicly trade "yield" vehicles. is achieving grid parity in a number of new markets globally and we expect companies involved in project development/financing to benefit the most from the significant volume growth over the next few years. As storage costs start to improve we expect companies with cost competitive storage solutions to create the most shareholder value. Page 4

5 Key Picks In the case of SPWR, weak 2015 outlook was the primary catalyst for share price reaction post analyst day, but the decision to hold back projects for a possible yieldco was the main reason for weak outlook. While capacity constraints could limit earnings upside through 2016, EBITDA and CAFD guide provided during the analyst day suggest that the shares could nearly double from current levels. We expect further clarity around yieldco plans during Q1 to act as the next catalyst for shares. For SCTY, we continue to see upside to 2016 volume targets and believe at current valuations, shares are discounting almost no value for the development business post 2016 timeframe. Finally, VSLR has guided to 100% yoy shipments growth in 2015 and '16 timeframe. Upside to these targets could act as a significant positive catalyst for shares. Seasonality will play a role in Q1 volumes, but expect strong growth from Q2. The company also has industry leading customer acquisition costs. We believe an increasing conversion of Vivint inc customers into solar customers along with greater penetration in commercial markets would result in further reduction of customer acquisition costs. Figure 2: Target Price and Downside Scenario Stock Target Price Closing Price (1/7/14) Downside Scenario SUNE $40 $18.16 $17.0 SCTY $90 $49.33 $29.5 VSLR $20 $8.16 $6.5 SPWR $43 $23.85 $22.0 TSL $15 $8.59 $8.3 Source: Deutsche Bank, Thomson Reuters SUNE For SUNE, we are keeping our $40PT unchanged and we see limited downside to SUNE shares. Even if we assume the development business is worth next to nothing, the value from TERP and the IDR s provide value as shown below. Figure 3: SUNE Bear Case Range Source: Deutsche Bank PT Value Downside Value Devco $20.5 $3 TERP Value to SUNE $8.1 $7.1 IDR Value $9.3 $5.1 SEMI $1.8 $1.8 Total $39.6 $17.1 Page 5

6 SUNE Valuation/Risks We value SUNE via SoTP utilizing multiple DCF's. We apply a 10% discount rate for the first ten years, while our TGR is 1.5% with a 15% discount rate. We value the Semiconductor business, the project development business, ownership in Terraform, and incentive distribution rights from the yieldco business separately to arrive at our SoTP Risks: 1) Pipeline/Backlog is not replenished 2) existing projects underperform expectations 3) input price volatility 4)negative changes to govt incentives 5) Acquisitions economics are less favorable than expected 6) Capital markets deals do not close. SPWR Our downside scenario assumes the multiple assigned to SPWR s 2016 earnings is cut in half (from 20 to 10x) and the target yield on their potential yieldco is 10%, rather than 6%. Figure 4: SPWR Scenario PT Value Downside Value 2016 EPS $1.96 $1.96 Multiple 20X 10x Value to SPWR $39.2 $19.6 Yieldco Target cashflow $75 $75 Target Yield 6% 10% Value to SPWR $7.9 $4.7 Source: Deutsche Bank Total ($/Share) $42.8 $22.1 *Discounted back at 10% SPWR Valuation/Risks Our price target based on 20x 2016E EPS of $1.96 plus $1.25B in target yieldco equity value (6% yield on $75M CAFD). This is discounted back 10% to arrive at our PT of $43. Downside risks include: Regulatory/execution issues for installations, ability to fill pipeline, financing costs/access, capacity expansion timing, cost reduction abilities, competitor panel pricing, pace of lease monetization/expansion VSLR/SCTY For our downside risk to SCTY/VSLR shares, we assume the businesses stop completely in 2016 (using our yearly build up, which is unchanged). Page 6

7 Figure 5: SCTY/VSLR Scenario Downside Value ($/Share) Inception Total VSLR $0.11 $0.49 $1.04 $1.71 $3.16 $6.5 SCTY $2.30 $3.21 $5.33 $8.56 $10.09 $29.5 *VSLR uses an 18% discount rate, SCTY uses a 12% discount rate Current PT Value DCF Value Including through Terminal Value VSLR $20 SCTY $90 Source: Deutsche Bank VSLR Valuation/Risks: We use a sum of the parts valuation with an 18% discount rate to value current and future leasing business cash flows and arrive at our $20 PT. We apply a higher discount than SCTY due to smaller platform and younger business. Risks include: 1) Adverse regulatory shifts on the state or federal level which could impact net metering, or other solar incentives 2) Changes in input prices 3) Headline risk from increased scrutiny of large utilities and lawmakers; 4) Inability to acquire project financing at attractive rates. 5) Competitive dynamics from new entrants or large incumbents 6) Widespread Customer defaults or bookings cancellations SCTY Valuation/Risks: We use a sum of the parts valuation with a 12% discount rate to value current/future leasing business cash flows and arrive at our $90 base case. Downside risks include changes to net metering, panel/labor prices, inability to attract financing, and headline risk from utilities/govt agencies. TSL In order to evaluate the downside valuation case for Trina, we looked at a sum of the parts / net asset value valuation where the company sells all of its manufacturing capacity, downstream solar projects, and satisfies debt obligations. Under the scenario below, we see valuation support at $8.3/sh. Page 7

8 Figure 6: TSL Liquidation Value Source: Deutsche Bank Manufacturing Business MW Value per watt Value ($M) Ingot to Module 1700 $0.40 $680 Cell to Module 1300 $0.20 $260 Module Only 800 $0.05 $40 Enterprise Value $980 Projects 2014 Projects (MW) 353 Liquidation Value ($/W) $1.00 Enterprise Value $353 Net Debt ($M) Cash $ Restricted Cash $ 97.7 S/T Debt $ L/T Debt $ Net Debt ($643.5) Total Equity Value $690 $/Share $8.3 TSL Valuation/Risks Our $15 PT is based on 11x 2016 EPS discounted back by 15%,. Downside risks incl.: Ability to execute on co. s downstream strategy. Polysilicon & other input price volatility. Degree of success in cost reduction efforts. Future capex funding requirements. Negative changes to Chinese govt incentives. Shifts in global solar demand. Page 8

9 Index of Brent Crude Oil Index of Brent Crude Oil 8 January 2015 Oil Concerns Unfounded Correlation between oil price and solar stock performance has increased significantly since oil broke below $100/barrel. This has not been the case in the recent past and we believe oil largely does not affect electricity prices and therefore should not affect solar installations. In this section, we highlight the following points: 1) Late 2014 negative price action in solar stocks is strongly correlated with the decline in the price of oil. 2) Actual oil fired generation to produce electricity is very expensive, even with low oil prices. 3) installations are not competing with oil fired generation in most instances 4) Unsubsidized solar competes with the price of electricity, which is unrelated to oil prices. 5) Companies with exposure to distributed generation are best positioned to capitalize on long term fundamentals. Correlations with Oil Have Increased Sharply Using the Guggenheim ETF (TAN) as a proxy for solar stocks generally, we examined the correlation between solar stocks and oil price over the last 2 years. We believe investors began irrationally selling solar stocks around the time oil broke $100/barrel. As shown below, solar stock prices have had no significant correlation with oil prices until 4Q 14. Brent crude first (recently) closed below $100/barrel on September 8, Figure 7: 2013: Oil is Not Correlated With Stocks 110 Figure 8: 1Q-3Q 2014: Still No Correlation R² = R² = In FY 2013, there was no correlation between the price of oil and the price of solar stocks and this was true for most of 2014 as well Index of TAN ( ETF) Index of TAN ( ETF) Source: Deutsche Bank, Thomson Reuters Note: TAN and Brent Crude prices are indexed to 100 on Jan 1, 2013 Note: Daily data runs from Jan 1, 2013 to Dec 31, 2013 Source: Deutsche Bank, Thomson Reuters Note: TAN and Brent Crude prices are indexed to 100 on Jan 1, 2014 Note: Daily data runs from Jan 1, 2014 to Sept 7, 2014 Page 9

10 Henry Hub Price ($/MMBtu) Index of Henry Hub Nat Gas Index of Brent Crude Oil 8 January 2015 Figure 9: 4Q 14: Stocks Are Suddenly Positively Correlated With Oil Price Movements but when oil started dropping in 4Q R² = correlations between oil price movement and solar stock price movement increased dramatically from nearly 0 to ~ Index of TAN ( ETF) Source: Source: Deutsche Bank, Thomson Reuters Note: TAN and Brent Crude prices are indexed to 100 on Jan 1, 2014 Note: Daily data runs from Sept 8, 2014 to Dec 26, 2014 On Sept 8, 2014, the S&P 500 closed just above 2,000 (At ). In the same timeframe as shown in the graph above, the S&P increased over 4% in value (to on Dec 26). Therefore, the decline in solar stocks cannot likely be attributed to market weakness. We also examined natural gas prices during the same time period shown above to determine if the weakness in solar stocks could be related. Figure 10: Nat Gas: Generally Stable in 4Q $5.00 $4.50 $4.00 $3.50 $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 For the majority of 4Q, nat gas prices were relatively stable or up... $0.00 9/8/ /8/ /8/ /8/2014 Figure 11: Nat Gas and Stocks: Uncorrelated R² = and correlations between nat gas price and solar stock price are insignificant Index of TAN ( ETF) Source: Deutsche Bank, Thomson Reuters Note: Daily data runs from Sept 8, 2014 to Dec 26, 2014 Source: Deutsche Bank, Thomson Reuters Note: Daily data runs from Sept 8, 2014 to Dec 26, 2014 Page 10

11 % Electricity Generation from Oil - Worldwide 8 January 2015 Based on our analysis and conversations with traders/investors, we continue to believe that much of the late 2014 weakness in solar stocks is related to declines in the oil price. This is not based on concerns rooted in fundamentals, in our view. Worldwide Oil Use in Electricity Generation Most Countries Generate Less than 5% of their Electricity From Oil In aggregate, the world generates ~3.9% of electricity from oil. However, the US generates only 0.9% of electricity from oil and China generates even less 0.17%. Figure 12: Oil Use for Electricity Generation has Declined Substantially 25% 20% Electricity production from Oil has decreased to less than 4% globally 15% 10% 5% 0% Source: World Bank, 2011 Data Furthermore, the majority of the countries which generate greater than 5% of their energy generation are not expected to be notable solar markets. The only truly notable solar market where oil accounts for double digit (10%) percent of total generation is Japan. The Japanese solar market is largely influenced by constructive government policy and generates significant portions of total installs in non-utility scale solar (which is the only segment that would compete with oil fired generation). Page 11

12 Figure 13: Countries Generating >5% of Electricity from Oil Count Country % Electricity from Oil Greater than 500MW Market? Count Country % Electricity from Oil Greater than 500MW Market? 1 Malta 99% 29 Morocco 26% 2 Eritrea 99% 30 Sudan 25% 3 Benin 99% 31 Togo 24% 4 Cyprus 96% 32 Indonesia 23% 5 Lebanon 95% 33 Gabon 21% 6 Jamaica 92% 34 Guatemala 19% 7 Cambodia 90% 35 Singapore 18% 8 Senegal 86% 36 Cameroon 18% 9 Haiti 79% 37 Oman 18% 10 Yemen, Rep. 78% 38 Mexico 16% Yes Jordan 73% 39 Nigeria 16% 12 Nicaragua 66% 40 Egypt, Arab Rep. 16% 13 Kuwait 62% 41 Argentina 15% 14 Honduras 55% 42 Venezuela, RB 14% 15 Sri Lanka 50% 43 Iraq 13% 16 Dominican Republic 48% 44 Japan 10% Yes - Current 17 Libya 44% 45 Greece 10% 18 Cuba 43% 46 Chile 10% 19 Panama 41% 47 Costa Rica 9% 20 Syrian Arab Republic 40% 48 Malaysia 8% 21 Pakistan 35% 49 Israel 7% 22 El Salvador 34% 50 Croatia 7% 23 Kenya 33% 51 Italy 7% Previously 24 Ecuador 33% 52 Peru 6% 25 Angola 29% 53 Algeria 5% 26 Iran, Islamic Rep. 28% 54 Portugal 5% 27 Uruguay 27% 55 Spain 5% 28 Saudi Arabia 26% Yes Source: Worldbank, 2011 Data World % 3.9% What is the cost? In order to examine our hypothesis that solar is not related to the price of oil, we ran an analysis to estimate the cost of electricity generated from oil. Electricity Generation Cost From Oil as a Feedstock While there are significant differences between all-in oil fired generation costs across the world, we believe that the actual cost of electricity from oil generation is significantly higher than electricity generated from solar in most instances. According to the EIA, there are ~5.86MBTU in a barrel of oil, and a typical oilfired electricity generating plant uses ~10,991 BTU to produce a single kilowatt hour of electricity. Therefore, at $50/barrel, the fuel cost alone to produce electricity is over 9 cents/kwh, and every $10 change affects the cost by ~2 cents/kwh. Page 12

13 Figure 14: Fuel Cost ONLY Per kwh generated from Oil Cost/Barrel Cost/kWh Cost/Barrel Cost/kWh $30 $0.06 $80 $0.15 $40 $0.08 $90 $0.17 $50 $0.09 $100 $0.19 $60 $0.11 $110 $0.21 $70 $0.13 $120 $0.23 Source: EIA, Deutsche Bank Actual Dispatch Curve in The US Furthermore, even the above estimates drastically understates the actual incremental cost of electricity generated from oil. Shown below is the actual estimated supply curve in New England (in the United States) for capacity available during summer Page 13

14 Figure 15: Variable Cost of Electricity in New England Marginal cost of electricity generated from Oil is well above the retail price of electricity in most instances. Source: SNL. Added text and emphasis from Deutsche Bank This regional breakdown will generally hold true for most other regions as well, in our view. Each data point in the chart above is an estimate of the incremental cost of wholesale power. As shown above, oil fired generation cost typically ranges from ~10 cents per kwh to ~50 cents per kwh. With solar PPA s being signed at levels in the mid single digits (cents/kwh), the value proposition versus oil fired generation is compelling. Page 14

15 Brent Crude ($/Barrel) Consumer Price Index of Electricity ($/MMBtu) 8 January 2015 Long Term Relationship between Oil and Electricity As shown below, the long term trend in the price of electricity is upwards, while the price of oil is much more volatile. If there were a fundamental basis for oil price changes affecting the price of electricity, the price of electricity should change much more. Figure 16: 40 Year Trend of Grid Sourced Electricity vs Oil Price $120 Since 1970, the retail price of electricity in the US has increased by almost ~600%... $35 $100 $30 $80 $25 $60 $40...while the price of oil has shown significantly more volatility $20 $15 $10 $20 $5 $0 $0 Retail Electricity ($/m BTU) Oil (Brent) Source: EIA, Deutsche Bank We view small scale, rooftop solar as one of the most attractive growth markets for solar installers, which also has clear a read-through for commercial and industrial segments (grid connected electricity prices for commercial and industrial customers is correlated with changes for residential customers). Page 15

16 Jan-90 May-91 Sep-92 Jan-94 May-95 Sep-96 Jan-98 May-99 Sep-00 Jan-02 May-03 Sep-04 Jan-06 May-07 Sep-08 Jan-10 May-11 Sep-12 Jan-14 Oil Price ($/Barrel) Electricity Price (Cents/kWh) 8 January 2015 Figure 17: Electricity Prices Across Consumer Segments Trend Together 160 Brent WTI Elec Price (Residential) Elec Price (Industrial) Elec Price (Commercial) Source: EIA, Deutsche Bank What Makes an Electricity Bill? While much of the previous analysis has centered around generation costs from oil and all in electric costs, we note that over 40% of the average electric bill in the US can be attributed to transmission and distribution (T&D) costs. This is because the structure of most mature electric markets allow utilities to recoup costs for large upfront capital expenditures from transmission and distribution. This system has developed over the last century as the modern electric cost-recovery method. Most investor owned regulated utilities are allowed to generate a regulated return over a multi-year timeframe. This system has facilitated grid build out across the US and other countries (although specific cost recovery mechanisms vary) as utilities are allowed to operate as a natural monopoly and are financially incentivized to build infrastructure (in the form of long term returns on upfront capital investment). Hence, the cost recovery for all necessary infrastructure including but not limited to electric generation assets necessitates the inclusion of T&D costs in the consumer s electric bill. Page 16

17 Figure 18: T&D Expense are a Significant Portion of Electric Bills 31% Generation Transmission 58% Distribution 11% Source: EIA, Deutsche Bank Therefore, even if feedstock for electricity generation were to decline substantially, transmission and distribution costs will continue to make up a large portion of an electric consumer s bill. This System Favors New Business Models Given that solar companies are not regulated utilities, any installer/leasing business that has exposure to distributed generation (DG) (Residential, Commercial, and Industrial installations) is inherently competing with the grid cost of electricity, which includes all 3 cost components (Generation, Transmission, and Distribution). However, DG installations do not require significant investments in transmission and distribution. In fact, most rooftop installations do not require any distribution infrastructure to operate behind the grid (behind the grid refers to the concept of using electricity at the generation source as a factory or residential customer would). Therefore, companies with significant exposure to DG installations are well positioned to compete with the grid price of electricity over the long term. city and Vivint both have little (SCTY) or no (VSLR) utility scale exposure, while companies like Sunedison and Sunpower are also well positioned for long term DG exposure. Page 17

18 Figure 19: Installer Exposure to Segments FSLR Residential Commercial & Industrial Power Plant DB Est 0% 5% 95% 100% 95% 5% SPWR 2015 Guide 30% 13% 57% 70% 87% 43% SCTY DB Est 80% 20% 0% 20% 80% 100% SUNE DB Est 5% 70% 25% 95% 30% 75% VSLR DB Est 100% 0% 0% 0% 100% 100% Source: Deutsche Bank, Company Reports Long Term Risk: Evolving Utility Business Models Based on ongoing disputes in Arizona, California, and Colorado, we see the beginnings of what could indicate a long term shift in how utilities and their regulatory commissions interact with solar. In 2015, we expect several key decisions from utility regulators to continue shaping this debate. Certain utilities are currently arguing that owners of solar installations are not paying enough to support the grid, because transmission and distribution charges are generally based on metered electricity use. When a solar installation connects to the grid, it generates a portion of the owners electricity use and effectively acts as a reduction in grid demand. In most cases, this leads to a proportional decrease in the dollar charge for grid-sourced electricity (which includes a proportional charge for T&D cost recovery). companies, individual users, and freedom-of-choice advocates believe this representation does not accurately account for the positive external contributions that solar installations provide. Theoretically, large scale distributed generation adoption should lower peak electricity demand, reduce strain on the grid, provide emissions-free electricity with no fuel cost, and Page 18

19 lower the amount of necessary future investment in the grid on all fronts. In a scenario where smart grids allow distributed solar resources to be dispatched as requested by the grid operator, the benefits from DG installations should increase. These debates are ongoing and unlikely to be finalized in the short term. On one hand, Arizona has already implemented a grid access charge of $0.70/kw (~$4-5 dollars per month for a typical residential system), which was still considerably lower than the ~$20/month grid charge requested by the utility. On the other hand, key states like California and New York continue to provide a constructive regulatory environment that favors solar. Long term, we believe the business models for solar and utility companies will necessarily shift as grid penetration rates increase (currently no more than 1-2% in even the high penetration states). Grid access charges could increase, utilities may start to compete more directly with solar installers, and cost recovery mechanisms generally will go through a rigorous analysis in most major solar markets. Arizona is generally considered one of the most contentious regions for debate in the US, yet solar leasing companies like SCTY have continued to ramp their installation rates despite this. Page 19

20 Solomon Islands Tonga Vanuatu USA Virgin Islands Australia Philippines Jordan USA Hawaii Brazil Italy Sweden Guyana Germany Japan Chile Papua New Guinea Hong Kong Denmark Belgium Spain United Kingdom Mexico Iran Netherlands France Jamaica Ireland New Zealand Serbia Pakistan South Africa Portugal India Hungary Israel Turkey United States Peru China 8 January 2015 Grid Parity is Here Over 50% of Countries under Review are Likely at Grid Parity Today Our analysis indicates that a wide range of countries throughout the world are at grid parity today in high electricity price and/or high sunlight regions. Using our levelized cost of energy (LCOE) model, we estimated typical cost per watt versus estimated electricity prices in the region to determine what countries are likely at grid parity. We have also examined key price drivers in the US market over the next several years and projected cost reductions from ~$2.90/w residential (including customer acquisition and other costs) to $~$1.80/w, which assumes cost reductions of ~10-15% each year. For our exercise, we gathered and estimated data points on average or high/low range residential electricity prices for over 60 countries worldwide. Incorporating a variety of assumptions on cost per watt, sunlight, inverter replacement, O&M cost, and other financial/operational decisions to estimate LCOE. In the figure below, we provide a high level overview of our estimates and demonstrate that we believe a wide range of countries are at grid parity today. Figure 20: Countries With Regions of Grid Parity $1.00 $0.90 $0.80 $0.70 $0.60 $0.50 $0.40 $0.30 $0.20 $0.10 $0.00 LCOE ($/KWh) Cost of Electricity ($/KWh) Source: Deutsche Bank Estimates Page 20

21 Figure 21: Countries With Regions of Grid Parity Data Country Grid Parity Insolation (kwh/m2/year) Cost of Electricity Comp Premium/ LCOE ($/kwh) Discount IRR (20 Year System) IRR (30 Year System) Australia Yes 1833 $0.49 $0.15 -$ % % Belgium Yes 867 $0.32 $0.24 -$ % 9.38% Brazil Yes 1667 $0.37 $0.18 -$ % 44.63% Chile Yes 1750 $0.25 $0.12 -$ % 29.40% Denmark Yes 813 $0.44 $0.35 -$ % 17.51% France Yes 1083 $0.21 $0.16 -$ % 7.58% Germany Yes 958 $0.33 $0.19 -$ % 16.55% Guyana Yes 1667 $0.28 $0.12 -$ % 35.49% Hungary Yes 1042 $0.26 $0.24 -$ % 8.67% Ireland Yes 750 $0.31 $0.27 -$ % 5.90% Israel Yes 1917 $0.16 $0.14 -$ % 12.00% Italy Yes 1292 $0.31 $0.14 -$ % 27.97% Japan Yes 1167 $0.28 $0.14 -$ % 19.11% Mexico Yes 1792 $0.20 $0.13 -$ % 15.09% Netherlands Yes 917 $0.32 $0.27 -$ % 10.59% New Zealand Yes 1167 $0.20 $0.18 -$ % 6.26% Papua New Guinea Yes 1417 $0.30 $0.17 -$ % 26.28% Peru Yes 1667 $0.13 $0.12 -$ % Philippines Yes 1583 $0.34 $0.10 -$ % 52.84% Portugal Yes 1458 $0.28 $0.25 -$ % 23.14% Spain Yes 1500 $0.24 $0.16 -$ % 14.69% Solomon Islands Yes 1417 $0.87 $0.14 -$ % 17.88% Sweden Yes 833 $0.30 $0.13 -$ % 9.15% Tonga Yes 1583 $0.63 $0.13 -$ Turkey Yes 1500 $0.14 $0.13 -$ United Kingdom Yes 792 $0.21 $0.14 -$ % USA Virgin Islands Yes 1667 $0.56 $0.19 -$ % 8.35% USA Hawaii Yes 1917 $0.39 $0.20 -$ Vanuatu Yes 1417 $0.60 $0.15 -$ China Yes vs High Electricity Price 1333 $0.11 $0.11 $ Hong Kong Yes vs High Electricity Price 1333 $0.25 $0.15 -$ % 14.38% India Yes vs High Electricity Price 1604 $0.12 $0.10 -$ Iran Yes vs High Electricity Price 1583 $0.21 $0.16 -$ % 14.29% Jamaica Yes vs High Electricity Price 1750 $0.18 $0.14 -$ % 13.55% Jordan Yes vs High Electricity Price 1917 $0.35 $0.13 -$ % % Pakistan Yes vs High Electricity Price 1833 $0.16 $0.13 -$ % 10.92% Serbia Yes vs High Electricity Price 917 $0.14 $0.11 -$ South Africa Yes vs High Electricity Price 1833 $0.17 $0.14 -$ United States Yes vs High Electricity Price 1458 $0.18 $0.17 -$ % 19.51% Total Count 39 Note: Calculations do not account for any subsidies current or future. Electricity Prices are estimated for residential consumers. Source: Deutsche Bank Estimates Page 21

22 Cost of Electricity/LCOE ($/kwh) Cost of Electricity/LCOE ($/kwh) 8 January 2015 What does the Future Hold for Grid Parity? While we believe systems can currently achieve grid parity within some regions of a wide range of countries, what will happen in the future? We conducted a scenario analysis where we assumed yearly average electricity price increases of 3% coupled with 5%, 10%, and 15% yearly overall system cost reduction through Our findings indicate that under a blue sky scenario, as much as 80% of the target markets could be at grid parity. Figure 22: 80% of our target markets could be at Grid Parity by 2017 $0.30 $0.25 These additional countries could reach grid parity by the end of $0.20 $ Which could equate to ~80% of the world at grid parity in a blue sky scenario $0.10 $0.05 Croatia Bulgaria Malaysia Malaysia Uruguay Singapore Finland Latvia Thailand Dubai Indonesia Canada Future Electricity Price (3% CAGR) GP at 5% Yearly Cost Reduction GP at 10% Yearly Cost Reduction GP at 15% Yearly Cost Reduction Source: Deutsche Bank Estimates While actual cost reduction may vary between these scenarios, we believe the trend is clear: grid parity without subsidies is already here, increasing parity will occur, and solar penetration rates are set to ramp worldwide. Figure 23: or 68% with flat electricity prices $0.25 $0.23 In a flat electricity price scenario, ~68% of countries under review could be at grid parity in $0.21 $0.19 $0.17 $0.15 $0.13 Croatia Bulgaria Malaysia Flat Electricity Price GP at 5% Yearly Cost Reduction Source: Deutsche Bank Estimates Page 22

23 Where Are We Today: Our Take One of the most prevalent metrics for direct cost comparison is cost per watt, which we have estimated for various regions throughout the world. While cost per watt is an appropriate measure to normalize cost comparisons, we note that economies of scale present in different segments/markets will skew cost per watt within regions. While some markets like the US and Japan will have a large portion of residential installations with less economies of scale, Utility scale or large DG markets like India and China will inherently achieve a lower cost per watt. Therefore, we have provided a starting point for analysis coupled with a scenario at different cost points in the previous section. Module Cost of Production Today Total module costs of leading Chinese solar companies have decreased from ~$1.31/W in 2011 to ~$0.50/W in 2014 primarily due to reduction in processing costs and polysilicon costs and improvement in conversion efficiencies. We see total costs coming down 30-40% over the next several years We think it is realistic to expect at least 30-40% reduction in cost per watt in key solar markets, while the greatest cost reductions are likely to come from the residential segments as scale and operating efficiencies improve. There is historical precedent for this in the oldest major solar market in the world Germany. In fact, costs today are well below costs in the United States and other less mature markets, and total installed costs have declined ~40%+ over the last ~3 years in the country. The exact drivers behind cost declines may vary between countries, but we believe the German example continues to prove that overall system costs have yet to reach a bottom even in comparatively mature markets. Total Cost Reduction Will Be Multi-Faceted: Mostly Not From Polysilicon While much of the cost reduction over the last 5-10 years has resulted from polysilicon price reductions, future cost reductions will necessarily come from non panel related balance of system costs. Polysilicon price reductions have accounted for significant portions of cost reductions, and were once the largest single cost component in panels, but this has changed drastically and rapidly over the last decade. In 2014, polysilicon represented no more than cents per watt so even if costs are halved, the effect on the total system cost would be incremental not revolutionary. However, there are significant other cost drivers that we believe the industry will leverage to drive down LCOE over the next several years. We have outlined our estimates of current and future cost trajectory in the US below, which we expect to mirror other regions cost roadmaps. Page 23

24 Figure 24: City 2017 Cost Targets Source: City Investor Presentation Panels: $0.75/w $0.50/w Panel prices in the US are already among the highest in the world today, so there would likely be price reductions simply through price arbitrage on a multi-year time horizon. However, we also believe there are fundamental reasons that panel prices worldwide are likely to trend lower over the next several years. While overhangs like trade cases or minimum price agreements could cloud the near term, we believe market inefficiencies will be worked out over the long term and the clearing price will reach $0.50 or lower within the next several years. Companies like SunEdison have publically targeted $0.40 cent per watt panels by the end of 2016, and many Tier 1 Chinese manufacturers are achieving sub $0.50/w already in Given that most manufacturers are improving 1-2 cents per quarter, less than ten cents improvement (to reach $0.40) over the next 12 quarters is likely conservative. If panels are sold at a 10 cent gross margin for a total cost of $0.50/w, manufacturers would achieve 20% gross margin well above recent historic averages. Furthermore, transportation costs and soft costs which inefficiently raise the price of panels should gradually improve as governments work through trade issues Inverter: $0.25/w $0.17/w Inverter prices typically decline 10-15% per year, and we expect this trend to continue into the future. Large solar installers are already achieving ~$0.25/w or lower on large supply deals, and we expect additional savings will be found over the next several years. Component cost reduction, next generation improvements, and incremental production efficiencies will drive savings on the manufacturing side, while new entrants and ongoing price competition Page 24