LONDON — As the deployment of grid-connected solar photovoltaic (PV) systems has increased, so too has the desire to track the installed price of these systems over time and by location, customer type, and system characteristics.
A new report from the Lawrence Berkeley National Laboratory (LBNL) attempts to do just that by summarising trends in the installed price of grid-connected PV systems in the US from 1998-2011, with preliminary data for 2012. The analysis is based on project-level data for more than 150,000 individual residential, commercial, and utility-scale PV systems, totaling more than 3 GW and representing 76 percent of all grid-connected PV power installed.
Utility-scale PV Emerges
The majority of systems (64 percent) were installed within the last three years of the sample period, and the utility-scale systems in the sample, as might be expected, were even more heavily concentrated within the latter years, with 90 percent of these systems installed from 2009-2011.
California and New Jersey represent 54 percent and 18 percent of the sample, respectively, in terms of total installed capacity. Arizona, Pennsylvania, Massachusetts, North Carolina, and New York each represent 2 percent-7 percent of the sample capacity, with the remaining 20 states comprising 9 percent in total.
The U.S. PV market has diversified significantly in recent years. California, though still the largest market, represents a smaller share (40 percent), with correspondingly greater representation among the other leading state markets.
The 80 utility-scale PV systems studied are located in a total of 17 states, with almost 90 percent of the capacity distributed across eight states (California, New Jersey, New Mexico, Nevada, Colorado, Arizona, Florida, and Texas). The size of the utility-scale PV systems in LBNL’s data sample ranges widely, from 2 MW up to 58 MW. Most of the systems (66 percent) are smaller than 10 MW, though most of the sample capacity (72 percent) consists of systems larger than 10 MW.
Key findings are as follows:
The installed price of utility-scale systems varies significantly across projects. Among the 49 projects in the data sample completed in 2011, for example, installed prices ranged from $2.4/W to $6.3/W, reflecting the wide variation in project size (from 2 MW to 35 MW), differences in system configurations (e.g., fixed- tilt vs. tracking and thin-film vs. crystalline modules), and the unique characteristics of individual projects.
Discerning a time trend for the installed price of utility-scale PV is challenging, given the small and diverse sample of projects. As a rough measure, the capacity-weighted average installed price declined from $6.2/W for projects installed during 2004-2008, to $3.9/W for projects installed during 2009-2010, and to $3.4/W for projects installed in 2011.
Larger utility-scale systems have lower installed prices. In particular, among projects installed in 2011, the installed price of projects larger than 10 MW generally ranged from $2.8/W to $3.5/W, whereas projects smaller than 10 MW span a broader range, with most priced between $3.5/W and $5.0/W.
Installed price trends according to system configuration are less evident. Among <10 MW utility-scale projects installed in 2011, systems using thin-film modules are relatively low priced, compared to crystalline systems with and without tracking. Among projects >10 MW, however, no clear differences in installed prices are observable either between crystalline and thin-film systems or between systems with and without tracking.
Within the class of systems 2-10 MW in size, utility-scale systems (ground- mounted, by definition) generally have slightly lower installed prices than similarly sized commercial rooftop systems. Although median installed prices are similar between these two groups, the distribution is skewed lower for utility-scale systems, with one-third priced from $2.9/W to $3.5/W, whereas the lowest-priced third of the large commercial roof-mounted systems range from $3.6/W to $3.8/W.
Installed prices have declined by 5 percent-7 percent per year, on average, depending on system size. Price declines, however, have not occurred at a steady pace. In particular, installed prices declined markedly until 2005, but then stagnated through roughly 2009 while the PV supply chain struggled to keep pace with surging worldwide demand.
Since 2009, installed prices have fallen precipitously as upstream cost reductions – principally PV module cost reductions – worked their way through to end consumers, and as state and utility PV incentive programmes continued to ramp down their incentives. From 2010 to 2011, installed prices fell by $0.8/W (14 percent) for systems greater than 100 kW. Preliminary data for the first half of 2012 show that installed prices have continued to fall, and declines in global module prices over the first half of 2012 suggest that installed system prices will continue to decline as projects in the development pipeline, whose costs reflect current module pricing, are constructed.
Over the longer term, however, installed prices have fallen also as a result of reductions in non-module costs (which include such items as inverters, mounting hardware, labour, permitting and fees, overhead, taxes, and installer profit). During the first half of the historical period, the overall decline in total installed prices was primarily attributable to a reduction in non-module costs. Within the past several years, however, module prices have declined at a much faster pace than non-module costs, and as a result, non-module costs have grown in terms of their relative share of total system costs. This shift in the cost structure of PV systems has heightened the emphasis within the industry and among policymakers on reducing non-module costs and, particularly, business process (or “soft”) costs.
Prices vary across projects
A convergence of prices, with high-priced outliers becoming increasingly infrequent, is consistent with a maturing market characterised by increased competition among installers and module manufacturers and by better-informed consumers. A narrowing trend was most evident between 1998-2003 and 2004-2008. Since then, the spread in the installed price distribution has remained relatively stable, and a significant degree of variability in pricing across systems has persisted.
The installed price distribution for 10-100 kW and >100 kW systems exhibit considerable spread, though somewhat less so than for smaller systems.
Economies of Scale
Larger PV installations benefit from economies of scale by spreading fixed project and overhead costs over a larger number of installed watts and, depending on the installer, through price reductions on volume purchases of materials. The median installed price for systems >1000 kW in size ($4.5/W) is 42 percent lower than for systems <2 kW ($7.7/W).
To a limited extent, economies of scale help to explain the long-term decline in median installed prices. Median system sizes have risen over time, though by varying degrees. The growth in median system sizes was relatively modest and uneven, and is unlikely to have had any material influence on the observed price declines, either over the long-term or within the more recent past, especially given the declining returns to scale at larger system sizes.
Prices Differ Among States
Substantial differences in median installed prices due to the impact of state or local conditions were observed. For systems >100 kW, median installed prices range from $4.5/W in Pennsylvania to $6.2/W in Arizona. California is a relatively high-cost state, pulling installed price statistics for the entire country upward.
Differences in installed prices across states reflect an array of underlying drivers. Larger or more mature state and regional PV markets can facilitate lower prices through greater competition and efficiency, more extensive bulk purchasing, and better access to low-cost products. That said, a strong correlation is not always evident between state market size and installed system prices, demonstrating that other factors also clearly play an important role in determining state-level pricing. For example, states with higher incentives and/or higher electricity rates may have higher prices.
Variability in prices across states also likely derives from differences in administrative and regulatory compliance costs (e.g., incentive applications, permitting, and interconnection) as well as differences in labour wages. State-level price variation can also arise from differences in the characteristics of the systems installed in each state, such as typical system size, mounting structures and the prevalence of tracking equipment.
Higher than Other Markets
Notwithstanding the significant installed price reductions that have already occurred in the US, international experience suggests that greater near-term reductions may be possible.
LBNL compared median installed prices between the U.S. and Germany across a broader range of residential and commercial system sizes, excluding sales tax/VAT and focusing on systems installed in 2011. The results indicate that the sizable US-German PV price gap for small residential systems also extends to larger residential and commercial systems, with German system prices that are 44 percent-46 percent lower than US systems, across the three system size ranges.
Given that modules and other hardware items are effectively commodities, with only marginal price differences across countries, much of the pricing variation across countries can be attributed to differences in “soft costs”.
These differences may, in turn, be partly attributable to differences in the cumulative size of each market, where Germany and Italy had amassed roughly 25 GW and 13 GW of grid-connected PV capacity through 2011, far more than any other individual country, potentially allowing for learning-based cost reductions. That said, larger market size alone does not account for the entirety of the differences in average installed costs among countries.
While residential and commercial PV systems are primarily roof-mounted, LBNL’s data sample includes a modest number of ground-mounted residential and commercial systems, with either fixed-tilt or tracking. They compared installed prices across these three system configurations – roof-mounted, ground-mounted with fixed-tilt, and ground-mounted with tracking – focusing on residential and commercial systems only.
Comparing rooftop systems to fixed-tilt ground-mounted systems, LBNL found that installed prices have generally been greater for ground-mounted systems.
Among systems installed in 2011 in particular, the median installed price of fixed-tilt ground mounted systems was $0.4/W greater than rooftop systems within the >100 kW size range. Across all size ranges, fixed-tilt ground-mounted systems have generally had a higher median installed price than similarly sized rooftop systems.
Comparing ground-mounted systems with fixed-tilt to those with tracking, LBNL found that systems with tracking have generally had higher installed prices than ground-mounted systems without tracking. Among systems installed in 2011, the median installed price of ground-mounted systems with tracking was $1.5/W greater within the 10-100 kW size range. Curiously, within the >100 kW size range, systems with tracking had a slightly lower median installed price than ground-mounted systems without tracking ($4.9/W vs. $5.1/W), though this finding may simply be an idiosyncrasy of the small sample sizes, as for all system sizes and in all other years, systems with tracking had consistently higher installed prices than fixed-tilt ground mounted systems.
Installed Price Has Declined
The installed price of utility-scale PV systems varies widely. Among the 49 projects completed in 2011, for example, installed prices ranged from $2.4/W to $6.3/W, and an even wider spread is evident for the other two time periods shown. This wide distribution invariably reflects a combination of factors, including differences in project size (which range from 2 MW to over 35 MW for systems installed in 2011) and differences in module type (thin-film vs. crystalline) and system configuration (e.g., fixed-tilt vs. tracking systems). The wide price distribution is also attributable to the presence of systems with unique characteristics that may increase costs.
The capacity-weighted average installed price declined from $6.2/W for projects installed during 2004-2008, to $3.9/W for projects installed during 2009-2010, and to $3.4/W for projects installed in 2011. Of some note, the decline in capacity-weighted average price in 2011 occurred in spite of the fact that utility-scale system sizes were somewhat smaller in 2011 than in 2009-2010.
The wide range of prices is partially attributable to differences in project size and configuration. Clearly, larger systems tend to have lower prices.
These project size-based trends undoubtedly reflect underlying economies of scale. Other factors may also be at play, such as differences between the site characteristics typical of larger versus smaller projects and differences in the characteristics of the project developers (e.g., larger projects may be more likely to be developed by more experienced and/or vertically integrated companies).
The relationships between system configuration and installed price are somewhat less discernible (again, noting that a comparison of installed price ignores any performance-related differences associated with module efficiency and the use of tracking equipment). Among the class of utility-scale projects smaller than 10 MW, the thin-film projects (which include a group of five similarly configured and priced projects installed by a single southwestern utility) are all at the low end of the spectrum. Among the projects larger than 10 MW, however, no clear differences in installed prices are observable either between the crystalline and thin-film systems or between the systems with and without tracking. The absence of a visible trend of course does not mean that differences in system configuration have no impact on price; rather, within this small sample, the impact is lost within the noise of the myriad other factors that influence installed prices (e.g., regulatory compliance costs for projects built on public vs. private land, whether private land is leased or owned, design requirements associated with specific climatic conditions, etc).
Ground Mounted vs. Rooftop
A large contingent of utility-scale PV systems in the US consists of ground-mounted systems less than 10 MW in size. While the market for utility-scale systems in this size range has grown considerably in recent years, so too has the market for multi-MW commercial rooftop systems.
Among the lowest-priced third of ground-mounted systems, installed prices range from $2.9/W to $3.5/W, whereas the lowest-priced third of roof-mounted systems range from $3.6/W to $3.8/W. These results appear to suggest that, under the right conditions, ground-mounted systems within this size class may offer a price advantage relative to similarly sized rooftop systems.
Interestingly, this is the opposite of the trend observed previously among smaller residential and commercial system sizes, where installed prices were consistently higher for ground-mounted than for rooftop systems. This reversal of trends may, in part, reflect fixed costs and other economies of scale associated with site preparation for ground-mounted systems.
Available evidence confirms that the installed price of PV systems (i.e., the up-front cost borne by the PV system owner) has declined substantially since 1998, though both the pace and source of those cost reductions have varied over time. Prior to 2005, installed price reductions were associated primarily with a decline in non-module costs. Starting in 2005, however, installed price reductions began to stall, as the supply-chain and delivery infrastructure struggled to keep pace with rapidly expanding global demand. Starting in 2008, global module prices began a steep downward trajectory, driving installed price reductions of 25-35 percent among residential and commercial installations by 2011.
Non-module costs, in contrast, have remained relatively stagnant since 2005. Trends in non-module costs may be particularly relevant in gauging the impact of state and utility PV deployment programmes.
Unlike module prices, which are primarily established through global markets, nonmodule costs consist of a variety of cost components that may be more readily affected by local programmes – including deployment programmes aimed at increasing demand (and thereby increasing competition and efficiency among installers) as well as more-targeted efforts, such as training and education programmes. Historical non-module costs reductions from 1998-2005 suggest that PV deployment policies have, in the past, succeeded in spurring cost reductions; however, the fact that non-module costs have remained largely unchanged since 2005 highlights the potential need to identify new and innovative mechanisms to foster greater efficiency and competition within the delivery infrastructure.
Preliminary data for California systems installed in the first half of 2012 indicate that installed prices have continued to decline. Notwithstanding this success, further price reductions will be necessary if the U.S. PV industry is to continue its expansion, given the expectation that PV incentive programmes will also continue to ratchet down financial support. Lower installed prices in Germany and other major international markets suggest that deeper near-term cost reductions in US are, in fact, possible and may accompany increased market scale. It is also evident, however, that market size alone is insufficient to fully capture potential near-term cost reductions, as suggested by the fact that many of the U.S. states with the lowest installed prices have relatively small PV markets. Targeted policies aimed at specific cost barriers (for example, permitting and interconnection costs), in concert with basic and applied research and development, may therefore be required in order to sustain the pace of installed price reductions on a long-term basis.
Finally, installed prices vary substantially across system sizes, market segments, technology types, and applications. Policymakers may wish to evaluate whether differential levels of financial support are therefore warranted, for example to avoid over-subsidising more cost-competitive installations while providing sufficient support for promising but less mature technologies and applications.
However, the wide distribution of installed price for utility-scale PV reflects a combination of factors, including differences in project size (which range from 2 MW to over 35 MW for systems installed in 2011) and differences in module type and system configuration.
Discerning a time trend is challenging, given the small and diverse sample of projects. As a rough measure of this trend, the capacity-weighted average installed price declined from $6.2/W for projects installed during 2004-2008, to $3.9/W for projects installed during 2009-2010, and to $3.4/W for projects installed in 2011. Of some note, the decline in capacity-weighted average price in 2011 occurred in spite of the fact that utility-scale system sizes were somewhat smaller in 2011 than in 2009-2010.