The U.S. Solar Market: Assessing the Potential

A February 2009 report by the Lawrence Berkeley National Laboratory found that the installed costs before tax incentives for residential and commercial photovoltaic systems had fallen to $7.60 per watt from $10.50 per watt in 2007 dollars. The report also found that the costs for silicon PV panels had stalled in 2005, that costs varied widely from state to state and that economies of scale improved for systems larger than 750 kW. All the while, installation costs in Japan and Germany situated themselves well below those in the U.S., introducing the possibility that U.S. installation costs could drop.

Certainly, the cost of installing U.S. PV systems is declining. Recent developments suggest a dramatic shift over the next few years in PV economics. After years of tight supplies, there is a surplus of PV modules. Strong European demand in the first half of 2008 increased industry capacity and production and prices have fallen since (see Figure 1, below). Silicon and panel producers have significantly increased their capacity and the surplus of panels on the market will continue for years to come. Additionally, the 2009 liberalization of the investment tax credit (ITC) will lower PV installation costs for commercial customers and some residential buyers.

The incentives that lower the costs of PV include state cash incentives, the federal ITC and accelerated depreciation. In analyzing the economics of solar PV systems, Standard & Poor's ignores potential benefits from renewable energy certificates (RECs), state ITCs and city and local government incentives. These benefits have generally been small and, if available, usually on a restricted basis.

Average pretax cash incentives declined from 2002 to 2007 by an average of $1.90 per watt for systems less than 100kW (see Figure 2, below), mainly because of declining incentives in California, which links incentives to system installation costs and which represents 75 to 80 percent of all systems in the Lawrence Berkeley study. The federal ITC rose 30 percent, but did not benefit residential systems much due to a $2,000 cap, which the Emergency Economic Stabilization Act of 2008 eliminated starting Jan. 1, 2009. This repeal now provides stronger incentives for residential systems.

The average pretax cash incentives have been reduced since 2002 to 2007 by an average $1.40 per watt for commercial systems with 100kW to 500kW. The uncapped ITC meant commercial PV systems received greater total financial incentives, at about $4.00 per watt in 2007, including state cash incentives and state and federal ITCs. Meanwhile, we expect commercial demand to grow as utilities can also now use the ITC.

Incentives Vary Widely

There is a wide difference in the incentives that various states offer for solar power, which means the growth of solar power will vary substantially by state, even given the same pre-incentive capital cost (see Table 1, below).

Cash incentives for PV in Northeastern states such as Connecticut, Massachusetts, New Jersey, New York and Pennsylvania stand out. These strong incentives, combined with generally high retail rates, make the Northeast one of the regions where solar will grow fastest–even faster than the Southwest, where state incentives are not as high despite the region's abundance of sunlight.

California has unique prospects, in that it has excellent exposure to sunlight, among the highest retail electric rates in the country and public sentiment and regulatory policies that favor renewables.

To help put numbers to solar economics, Standard & Poor's has developed a levelized cost of electricity (LCOE) model. For simplicity, we use three capacity factors for solar PV units: 19 percent (which represents the prime regions of the desert Southwest from West Texas to California); 15 percent (which represents much of the Southeast from Eastern Texas to Florida); and 13.5 percent, which represents much of the Northeast. (Site-specific capacity factors will vary somewhat around these averages.) These capacity factors reflect the combination of sunlight levels and operating performance for these projects that the National Renewable Energy Laboratory's Solar Advisor Model provides. These calculations assume polysilicon panels are fixed-tilt systems with no sunlight tracking. We used the "Modified Accelerated Cost Recovery System" (the method of accelerated asset depreciation the U.S. income tax code requires) to depreciate the assets. It provides for depreciation over five years for solar PV and 20 years for gas-fired plants.

At the wholesale level, government-mandated renewable portfolio standards (RPS) will be a major motivator for building utility-scale PV projects. Still, the LCOE of the various options will be critical in determining whether utilities will build or buy solar power. Indeed, with PV proponents aiming to achieve grid parity (match current electricity rates) in the long run, even without incentives, LCOE is an important tool. Given that PV tends to produce electricity roughly coincident with peak load (where gas-fired peakers are the current option of choice), it makes sense to gauge the viability of PV by comparing the two. We recognize that the hour-to-hour variability of solar power creates the need for back-up generation as solar installations rise as a percentage of the total system. However, we believe such a comparison is still valid and important (see Table 2, below).

When we make this comparison, we find that solar PV is close to wholesale grid parity in the Southwest. At $5.00 per watt, solar PV is competitive with a gas peaker plant running at a 20 percent capacity factor, assuming a long-term average natural gas price of $7 per million Btu (mmBtu). Given current pricing trends, costs for large-scale PV may come in under $5.00 per watt.If gas prices are higher, or RPS standards make RECs valuable or if carbon legislation requires gas peakers to buy carbon credits, PV can be a cheaper peaking alternative than gas.

Of course, the $7-per-mmBtu price for natural gas we have assumed for purposes of comparison is much higher than the current NYMEX spot price. Nevertheless, we believe this is a reasonable proxy for the long-term price of natural gas as economic conditions improve and if carbon legislation increases demand (and hence prices) for natural gas.

A combination of RECs and higher gas/carbon prices will be needed for PV to achieve parity with gas in other regions. If the amount of sunlight available only affords a 15 percent capacity factor, the installed cost will have to be $3.50 per watt at most. However, solar panel manufacturers and developers say that such a price will soon be attainable, given falling panel prices and the entry of aggressive Chinese manufacturers. So utility-scale solar PV will likely become a competitive peaking resource in many regions of the U.S., besides being useful to meet RPS standards. Capital costs required to attain parity with peakers will depend on various combinations of gas prices and carbon costs.

Comparing Costs

Comparing the LCOE of PV commercial and residential retail installations with the retail-delivered cost of electricity that includes transmission and distribution components is instructive. A comparison with average retail commercial and residential rates in various states reveals that it is not states with the best solar resources that reach grid parity fastest but those with the highest retail rates. State incentives for PV systems are also important because these sometimes can be as large as, if not larger than, federal ITC benefits. Not all customers pay the same rate, but statewide averages from the U.S. Energy Information Administration indicate the relative attractiveness of PV generation.

In general, current capital costs for commercial PV are about $5.50 per watt to $6.50 per watt, depending on installation size. Incentive levels are high in many Northeastern states, at about $4.00 per watt. At these levels, the LCOE of PV systems was below average commercial rates in these states as of February 2009. The Southwest, especially California, turns economical if we include state incentives. Arizona recently put in place a cash incentive payment linked to generation. The Southeast will clearly not be economical for a long time given the lack of state incentives and generally low rates. By contrast, if the cost to install commercial PV systems reaches about $5.00 per watt as panel prices fall, many states in the Northeast may reach parity even without incentives.

Residential grid parity is farther away but is closest in the Northeast. For states that have tiered rates based on consumption levels, the state average residential rate grossly understates the actual per-kilowatt hour rates customers in the highest consumption tiers pay, making solar PV very reliable.

Impeding Solar Adoption

Although some data suggests that solar is at or close to grid parity in several parts of the U.S. including incentives, a few factors, mostly short-term in nature, are impeding the growth of solar. It is important to realize that LCOE comparability, although critical, is only one of several factors that will influence solar PV growth.

The key factor impeding utility-scale solar PV projects is access to capital, both debt and equity. Most solar project developers need to attract outside capital for 100 percent of project costs. The tax equity market has all but dried up or is at best extremely expensive. Even though these projects benefit from a power purchase agreement with a utility, debt financing is yet to begin in a big way although considerable interest is building.

Indeed, there is a great interest in large-scale solar PV projects, although many are being held back by weak financial markets, the need to obtain a purchase power agreement with a utility and permitting issues. Certainly, residential and commercial penetration will be faster in states with higher retail rates and supportive state legislation, rather than in states with the best solar resources. Solar panel manufacturers and project developers are targeting an all-in installed cost of $2.00 per watt and a pan price of $1.00 per watt. This could lead to grid parity, even without incentives.

Swami Venkataraman is Director, Corporate and Government Ratings at Standard & Poor's.