The Future of Solar Energy in the us: a technological, Political and Financial Enquiry

• 
  Edison Electric Institute (EEI)
  

• 
  Americans for Prosperity (AFP)
  

There are two important investment components for financing large-scale solar projects—debt and equity. Equity capital is provided by the project developers as well as by equity partners (institutional investors, hedge funds, private investors, individual firms, etc.) and this contribution serves as the initial “seed capital” that allows the project to start. This equity capital is usually between 15-20% of the total capital investment, however, the contribution is not made until debt financing is secured (without lender/creditor commitment to finance the rest of the project, providing the initial equity capital can be highly risky).⁶⁰ Equity contributions are used to cover initial due-diligence expenses, which can broadly be categorized as legal, technical, and financial in nature; undertaking these due-diligence efforts is essential for project development. The cost of the preliminary legal assessment of necessary permits and contract terms (for engineering, construction, operations, and management) and technical due-diligence (ensuring the feasibility of the solar project and that permits and contracts accommodate the technical requirements of the project) will almost entirely be borne by the equity sponsors.⁶¹ Lender financing will only come into play once the equity stage due diligence is complete; debt capital will be used to support the more comprehensive legal, financial, and technical due-diligence required to begin project development. Given that loans are the most crucial component in solar project finance, it is important to understand how this capital will be deployed by private sector participants and under what conditions. The following sections will outline the current salient government incentive programs, partnerships, and financing structures that will allow solar infrastructure projects to progress from the equity stage to commercial viability.

Source:Department of Energy Loan Programs Office

Of the $41.5 billion remaining in direct loans and loan guarantees under the Loan Program Office, $16 billion is destined towards the Advanced Technology Vehicles Manufacturing program.⁶⁶ This leaves $25.5 billion for the innovative energy technologies, of which $21 billion is committed towards existing non-renewable energy sources—fossil and nuclear energy.⁶⁷ At the moment, only about 11% ($4.5B) of the $41.5 billion availability for government funding is destined towards renewable energy.

Source:Department of Energy Loan Programs Office

These loan guarantees can best be described as incentives for solar infrastructure projects, however, these are distinct from public-private partnerships which hold promise for future solar infrastructure development (given that the current level of loan guarantees and incentives is low relative to what is required to hit federal and state mandates for solar electricity generation).

Source:Department of Energy Loan Programs Office

Source:Department of Energy Loan Programs Office

The Energy Department report also suggested that another failure (due to poor government due diligence) on the scale of Solyndra “could seriously undermine the [Department of Energy Loan Guarantee program] and its goals, perhaps even leading to its termination.” ⁶⁹ Although the program appears to have stabilized and currently has a low default rate across its loan portfolio (2.28% in losses as a percentage of total commitments as of September 2014), future defaults could further reduce the limited availability of loan guarantees for solar photovoltaic infrastructure projects.⁷⁰ While the risk of future solar photovoltaic project defaults cannot be dismissed, it is also worth evaluating some of the current solar projects that have been successful within the Department of Energy’s Loan Guarantee portfolio—Desert Sunlight, Mojave, and Ivanpah. The following sections are case studies of these successful projects and detail the costs, benefits, and financing mechanisms used to bring the utility-scale projects online.

4.1.1. Desert Sunlight Photovoltaic Case Study

Desert Sunlight is the second largest solar PV project in the LPO portfolio with nearly $1.5 billion guaranteed by the DoE.⁷¹ The loan was guaranteed for issuance September 2011 for construction of a 550-MW PV solar generation plant in Riverside County, California.⁷² One of the remarkable aspects of this project is the amount of financial institutions involved—fourteen in total. The DoE notes in Desert Sunlight’s project summary that “the loan guarantees helped attract new lenders into the utility-scale photovoltaic market and provided them with experience financing utility-scale photovoltaic projects.”⁷³ Given that the project was financed as a syndicated loan with NextEra Energy, General Electric & Sumitomo of America as project owners, it provided substantial underwriting experience for all the lenders involved and set the stage for future syndicated loans of solar PV projects. The DoE also noted that an additional 17 PV projects with capacity greater than 100MW had been financed “without loan guarantees and many of them by banks that LPO had worked with through [Financial Institution Partnership Program].”⁷⁴

Source:Department of Energy Loan Programs Office

The Desert Sunlight can be deemed a success due to the fact that the project reached full commercial viability in January 2015 and is currently generating electricity. Given that the Department of Energy’s Loan Guarantee program is judged based on the economic impact of the projects in its portfolio, it is also worth noting that 550 temporary construction jobs were created for the DS project along with 15 permanent jobs. In terms of environmental impact, 614,000 metric tons of annual carbon dioxide emissions are projected to be reduced due to the electricity generated by Desert Sunlight (annually 1,060,000 MWh).

4.1.2. Mojave Concentrating Solar Power Plant (CSP) Case Study

Mojave is another successful solar project in the LPO portfolio. With 250-MW capacity, the parabolic trough concentrating solar power plant (CSP) in San Bernardino County, California reached commercial viability in December 2014, barely three years after the $1.2 billion loan guarantee necessary to begin construction was secured by the DoE ($1.2 billion was guaranteed of a total investment of $1.6 billion).⁷⁵ The project owner was spanish multinational Abengoa S.A. and its solar construction subsidiary, Abengoa Solar, LLC.^{76 77} The plant assets are currently owned by Abengoa’s YieldCo (a term to be defined later in the financing section of this paper); Mojave is expected to generate 617,000 megawatt-hours of clean energy annually and reduce 329,000 metric tons of carbon dioxide emissions per year. It is also important to note that the loan guarantee decision was aided by the fact that Abengoa had agreed to a 25 year power purchase agreement (PPA) with Pacific Gas & Electric (for the sale of energy produced by Mojave).⁷⁸ Despite the PPA and estimated economic impact (830 temporary construction jobs, 70 permanent jobs, local tax revenues from electricity sales), the project was received with much opposition from politically conservative groups and environmentalists in the wake of the Solyndra debacle.^{79 80} Unsubstantiated accusations of waste, corruption, and illegal activities were attributed to Abengoa, however, none of these accusations have been addressed by politicians or government officials. By all accounts, the project is currently a success, as it is at full operational capacity and generating electricity. ⁸¹

Our final case study of solar projects is Ivanpah, a 392 megawatt concentrating solar power (CSP) plant located in Ivanpah Dry Lake, California. The Department of Energy issued three loan guarantees in April 2011 covering a notional amount of $1.6 billion in loans; January 2014 marked the beginning of Ivanpah’s commercial operations.⁸² When completed, Ivanpah was the largest CSP plant in the world and according to the DOE, “Ivanpah nearly doubled the amount of solar thermal energy produced in the United States in previous years.”⁸³ The economic and environmental impact of Ivanpah in the region is staggering, with annual electricity production of 940,000 megawatt-hours, 500,000 metric tons of carbon dioxide emissions prevented annually, 1,000 temporary construction jobs created with 61 permanent jobs. In recognition for its technological innovation and economic impact, Ivanpah was named “Plant of the Year” by POWER Magazine (a publication that has covered the power generation industry for over 130 years), given its status as largest solar thermal plant in the world and being the first commercial-scale solar project to use power tower technology.^{84 85} As is the case with Desert Sunlight and Mojave, Ivanpah is currently a successful project within the Department of Energy’s loan guarantee portfolio. As aforementioned, the Department of Energy currently has around $4.5 billion of available loan guarantees earmarked for renewable energy (not specifically solar) under Section 1703 of Title XVII Innovative Clean Energy projects. As we have seen with the case studies discussed in this section, even partial loan guarantees of solar mega projects (>250MW capacity) can result in highly successful economic and environmental impacts to the community, exceeding the risk of default in guaranteeing loans (which for large-scale commercial power projects is much lower early stage unproven technology companies). Additionally, loan guarantees result in further financing transactions without loan guarantees, as investment banks and financial institutions gain experience and familiarity underwriting and syndicating solar power infrastructure projects.

Public-private partnerships are generally collaborative agreements between the government (usually federal and/or state) and private sector participants (investment banks and institutional investors) to develop long-term infrastructure projects. These projects are usually extremely costly to build and development takes several years to bring the project to full completion.⁸⁶ Given the high initial capital costs and delayed payback period (toll roads/utilities that only begin earning revenues upon completion—after multi-year construction period), the entire project poses significant risk for any one party to bear.⁸⁷ Through public-private partnerships, the private investor still bears a significant amount of risk (along with substantially all managerial responsibilities/costs), however, the participation by a government entity decreases the size of the burden.⁸⁸ The government entity can choose to either finance part of the project directly (assuming it has the funds available to do so) or it can provide the private party with an agreement to allow the private investors to receive the operating profits of the infrastructure project.^{89 90}This agreement can contain royalties to be received by the government or revenue-sharing models, both of which are increasing in proportion to the government entity with initial government capital contribution.

Specifically, when it comes to solar infrastructure projects that are expected to generate significant electricity, power purchase agreements can serve as the government contribution towards the partnership (while providing little to no direct capital upfront and the private investor(s) bearing the entirety of the initial capital cost).⁹¹ The importance of power purchase agreements for electricity-generating infrastructure is easy to see—investors undertake the task of financing a project with heavy initial capital costs and sustained cash outflows with a delayed payback period. The uncertainty of future revenues dissuades investors from committing their capital; the risk of realizing inadequate (low or negative) returns is high when there is no agreement in place for securing future demand. The power purchase agreement eliminates the uncertainty as project financiers can guarantee baseline future demand to cover their fixed costs (including costs of entry) as well as their variable costs of production.⁹² PPAs benefit both parties, as the government utility (party directly purchasing electrical output) secures future supply of electricity that conforms with government mandates while the private sector investors realize a return on their investment via the guaranteed purchases.

Although companies developing utility-scale solar projects have several financing alternatives to choose from, it is important to identify those that will be the most prominent for future projects. Given that the Department of Energy Loan Guarantees are nearly at their guarantee limit for renewable projects (and by extension, utility-scale solar) it is unlikely that this program will serve as a meaningful financing source for more than 2-3 utility-scale projects (>250 MW). With this in mind, Public-Private Partnerships and late-stage YieldCos. serve as more pragmatic sources of funds for companies developing utility-scale solar projects. Public-Private Partnerships can help finance projects early (assuming DOE Loan Guarantee limits are not increased) and the YieldCo. structure can provide early-stage investors with an exit for their investment. One indirect benefit of the Department of Energy’s Loan Guarantee program was the direct underwriting experience it provided to financial institutions (incentivized by the government’s loan guarantees) that would persist beyond the life of the loan guarantees. As the DOE program is phased out (assuming no extension of the loan guarantee limits—currently an uncertain prospect), the experienced financial institutions have continued participating in the financing of solar projects, which can be expected to continue in the future. The larger solar companies that are developing utility-scale solar projects can negotiate their purchase-power agreements and bundle the renewable assets attractively for sale to its YieldCo., distributing the funds received to early-stage equity investors and using the remaining balance towards new capital investments (utility-scale solar or otherwise).

We now use a basic electricity distributor model to illustrate a utility firm’s profit maximization decision. Our analysis is done in two stages. First, we aim to illustrate this representative firm’s choice of solar capacity up to 2040 under assumptions of projected electricity price and capital cost of increasing capacity. Then, we analyze the comparative statics of this choice of solar capacity under different ranges of assumptions on government subsidy, technological capability, and returns to scale of added capacity. By modeling these comparative statics, we examine the effects of the different possible outcomes of the technological, political, and financial situations detailed in this paper.

Our representative firm wants to maximize profits over a 25 year horizon, or in other words is making its decision based on information up to the year 2040. To model this consideration, we use a dynamic profit function:

Note that the profit each year is given by the revenue, electricity price (p_i) multiplied by electricity generation (y_i), minus the cost, a function of total solar capacity (s_i), plus government subsidy. The firm considers the net-present value of future profits, given by discounting by the risk-free rate (r).

We make the assumption that the electricity price will not remain constant. Looking at the historical prices over the past 10 years, prices grew 25% and exceeded the natural inflation rate of 21%.⁹⁹ Thus, we use the U.S. Energy Information Administration’s (EIA) projected electricity prices to capture this non-inflation price increase. We use the EIA’s reference case for oil price, economic growth, and oil and gas resources up to 2040; these assumptions formulate the EIA’s electricity price projection.¹⁰⁰

We also model the firm’s costs as a function of its total solar capacity. As described previously, the predominant cost is the capital cost of solar installations. We want to capture the relationship between these capital costs plus operations costs and the number of utility-solar installations. However, the privatized nature of these installations make it difficult to get data on the number of installations each year; instead, we use total solar capacity as a proxy for measuring the number of installations. Furthermore, it is not standard to measure costs per total capacity. Cost more commonly given by levelized cost of electricity (LCOE), which is measured per electricity generation (y_i) rather than total electricity capacity (s_i). Thus, it is necessary to define the relationship between capacity and generation:

This serves as our production function. A_i is a measure of technological efficiency, or in other words the capacity factor of the solar installations. According to literature, the current maximum capacity factor for solar is about 25%.¹⁰¹ Note also that in order to solve our maximization problem, the production function cannot have increasing returns to scale. We make the assumption that doubling total capacity will increase generation by less than double; intuitively, simply increasing the available capacity does not mean that all of that will be utilized in the presence of other sources of electricity. Thus, we have the following constraints:

Now that we can express generation in terms of total capacity, we can define our cost as a function of generation:

Note that since LCOE is measured as the cost per unit of generation, we can express the cost as LCOE multiplied by generation. LCOE represents the per-megawatthour cost (in real dollars) of building and operating a generating plant over an assumed financial life and duty cycle. Key inputs to calculating LCOE include capacity factor (A_i), capital costs of total capacity (K_i), and variable maintenance and operations costs (N_i).¹⁰² As we did for electricity prices, we use the EIA’s projections for LCOE up to 2040 for our model.¹⁰³

Next, we define the rate at which total solar capacity grows. Since our representative firm makes its choice of total capacity based on information up to 2040, we assume that the firm will plan its capacity expansion at present (i=0). Further, we assume that the firm will execute its expansion plan at a rate such that the growth rate is constant. Intuitively, we do not expect total capacity to increase by a constant lump-sum each year. We expect that as more solar installations are made, the lump-sum additions from year to year will rise. From literature, certain projections suggest that capacity growth will be exponential.¹⁰⁴ For the sake of accuracy, we make the more conservative assumption that the growth will be linear:

Finally, we define government subsidy as a function of generation. As detailed previously, the government has a number of policies supporting solar. For our model of the representative firm’s decision, we define government support as a per unit subsidy on generation:

This formulation is preferable to a lump-sum subsidy, which would not be distortionary on the firm’s decision. Further, we prefer using generation rather than total capacity since key government policies, such as RPS and net metering, focus on generation.

We rewrite our maximization problem as:



5.2.0 Profit Maximization

To solve the general form of our profit maximization, we first factor and substitute in the production function:

We also iteratively solve our capacity growth relationship to get:

We define the following for simplicity’s sake:

We now differentiate with respect to n to find the optimality condition:

Having found our optimality condition, we can now examine the firm’s choice of n under assumptions on A_i, g_i, r, and α, along with given p_i and LCOE_i. As a base case, we make the following assumptions:

Solving for optimal n, we get:

Intuitively, we can understand why the optimal growth rate is negative by comparing the projected prices and costs. As shown in the chart below, the projected price does not exceed the projected cost until 2040. Since the firm is considering the net present value its profits up to 2040, there is no growth rate that allows the firm to achieve a positive profit. Under these assumptions of no government subsidy, the firm will not choose to increase its solar capacity under optimality.

Having made this observation that the firm cannot profit without government subsidy, we proceed by analyzing the relationship between optimal n and different values of g. From our optimality condition, we know that:

This is easy to see intuitively: the more government subsidy, the more the firm will want to grow its solar capacity.

Next, we want to understand what level of government subsidy will incentivize the firm to choose a positive growth rate. We hold the same assumptions as before, except we vary g. We define that government subsidy is constant in all periods:

We vary our x, beginning at 0 and incrementing by .001, to build a piece-wise curve illustrating the relationship between n and g. We introduce the following constraint on n to eliminate negative and infinite growth:

Intuitively, this limits the firm’s growth rate such that it can at most double its capacity in one year. As shown below, the firm will choose not to grow for all values of g up to .951, after which it will always choose the maximum growth rate of 1. This is the smallest g at which the firm will be able to profit; in our further analysis, we will call this point the optimal g.