Dear Mr. President,
Nearly five years has passed since you revealed plans to reinstall solar panels to the White House in October 2010.
In the meantime we have seen solar panel costs come down significantly (see Figure 1) and residential-scale systems today are about half the price they were back in 2010. Is the rumor that your staff had superior foresight and delayed the installation at your residency to save tax payers money true?
Following this price drop and federal and state level renewable energy initiatives, electricity generated from photovoltaics (PVs) seem to have become a rapidly growing source of carbon-free power in the United States and nowadays we can buy solar panels at local retailers. This is quite a different environment from a decade ago when solar panels were mostly available at limited specialty stores and maybe hobby shops. According to the Solar Energy Industries Association, there were over 20GW of installed solar facilities in the US with 6GW installed in 2014 alone.
Source: NREL Open PV Project monthly averages
PV systems are unique in that they can be installed in a relatively short time frame (weeks to months rather than years to decades required for other traditional power generation resources) and scalable and deployable in various sizes and configuration while still using the same core technology.
One extreme of the PV systems installed today is the centralized and large utility-scale systems that are usually larger than 5MW and often times extend to hundreds of MW. These free-field applications typically interconnect to the high-voltage transmission grid, supplying energy to the buyer, in many cases the local electric utility. The other extreme is the small and distributed residential installations that are mostly smaller than 10kW and average around 5kW. These installations on the more confined spaces of residential rooftops typically attach to the local electric utility’s distribution system, and generally send surplus power into that system and supply some of the on-site load requirements of the residential host.
Out of the aforementioned 20GW of solar systems installed today, utility-scale systems account for around 60% while residential-scale systems account for about 20%. The remaining are commercial and industrial installations. Yet, partially due to its popularity and visibility, residential-scale PV systems (also commonly known as rooftop PV systems) are often the main solar application considered in policy and regulatory circles when discussing solar options. We feel that utility-scale PV systems may often be overlooked.
Utility- vs residential-scale PV
While various studies compare the cost of solar to the cost of other generation resources, a comparison between different types of solar systems, i.e. utility- vs. residential-scale, was rarely properly quantified. Therefore we recently completed a study that compares the benefits of utility- vs. residential-scale PV systems by assessing the difference in generation cost to consumers. For simplicity, we compared equal amounts of solar installations (measured in DC capacity of the solar panels, or W-DC) for both residential and utility-scale PV systems in a representative electric utility setting.
The study provides a case study of comparing the per-megawatt hour (MWh) generation costs of adding 300MW-DC of PV panels either in the form of: (1) 60,000 distributed 5kW residential-scale (rooftop) systems owned or leased by retail customers, or (2) 300MW of utility-scale solar power plants (see Figure 2) that sell their entire output to the local utility under long-term power purchase agreements, using the Xcel Energy Colorado system as a representative electric utility.
Xcel Energy Colorado was selected for this case study because it is a reasonably representative of a midsize utility system in the Western US from a number of perspectives. These include, among others, the size of system, load profile, the current level of penetration of residential-scale PV systems in its service territory, and the mix of urban and rural load and distribution feeders. With Xcel Energy Colorado expecting to install similar magnitudes of solar over the next few years, the study looks at a near future year of 2019, making the comparison a realistic case.
We first thought about how the future world may look like. It was difficult with so many varying views and possibilities. We are in 2015 and according to the movie “Back to the Future” from the mid-1980s, we could have had hover boards and Delorean time machines. Forecasting the future, as I’m sure you are well aware of given your job and responsibility, was difficult so we decided to illustrate the potential 2019 world assuming things won’t change drastically.
Our Reference Case represents such a world and assumes reduction in investment tax credit (ITC) by the end of 2016 (down to 0% for residential system owners and 10% for businesses including utility-scale and residential-scale leases) under the current tax code, and various parameters consistent with what we see today including an inflation rate of 2%. We also added five scenarios with varying ITC, PV cost, inflation, and financing parameters.
The study finds that customer generation costs per solar MWh are estimated to be more than twice as high for residential-scale systems compared to the equivalent amount of utility-scale PV systems. Under the Reference Case, generation costs per solar MWh were found to be approximately 16.7c/kWh ($167/MWh) for residential-scale systems and 8.3c/kWh ($83/MWh) for the equivalent amount of utility-scale PV systems, resulting in a difference of 8.4c/kWh ($84/MWh). This generation cost difference between the utility- and residential-scale PV systems owned by the customer ranged from 6.7c/kWh ($67/MWh) to 9.2c/kWh ($92/MWh) across the other scenarios (See Figure 3). To put this in perspective, the cost difference observed here is approximately half to three quarters of the 2014 national average retail all-in residential electric rate of 12.5c/kWh.
Another way to look at this difference is that residential-scale PV systems cost $195 million more than the utility-scale systems under the Reference Case on an NPV basis over 25 years. Roughly 1,200MW of residential-scale PV systems were installed in 2014 alone. If the same amount of residential-scale PV systems (1,200MW) were installed in 2019, that would cost customers roughly $800 million more in NPV than a comparable purchase of utility-scale systems, under conditions assumed for the Reference Case.
The cost gap between utility and residential-scale PV
The large gap in per-MWh costs between utility- and residential-scale systems results principally from two factors.
The first factor is the lower total plant costs per installed kilowatt for larger facilities. Analyzing price trends for the last decade through mid-2014, the study estimated the installation cost for PV systems in 2019 to be $1.43/W-DC for utility-scale systems and $2.25/W-DC for residential-scale projects, indicating that utility-scale system costs are lower by nearly 40% compared to residential-scale systems. Recent price data indicate that prices could be even lower, however, no attempt to update prices to the latest information.
The second factor is the difference in power produced from the same PV capacity (300MW-DC) due to optimized panel placement, tracking and other economies of scale and efficiencies associated with utility-scale installations. For a representative year, assuming identical solar irradiance conditions in the greater Denver area, the utility-scale system is estimated to produce nearly 600GWh of power while residential-scale systems production would be slightly above 400GWh, resulting in a nearly 50% difference in power production.
The difference in production expands to other benefits that were not the focus of the study. For example, the fuel savings associated with replacing production from conventional thermal generation resources, and reduced air emission and avoided water usage of these conventional generation resources that are generally thought to be proportional to production, will be larger for utility-scale PV systems by an additional 50% than residential-scale systems.
On the longer term planning horizon, a larger production will likely lead to lower needs of capacity of other generation resources for resource adequacy (i.e. ensuring there is enough generators to serve the peak load). Literature research indicates that avoided transmission and distribution costs advantage by residential-scale systems (if any, which may depend on where and how densely they are located) are not large enough to significantly impact this gap in benefits of utility-scale PV system from generation quantity and overall costs discussed earlier. Comparative total (direct and indirect) job impacts of utility- vs. residential-scale cannot be determined without a much more complete analysis, but we believe there is likely no significant difference.
In summary, the overall benefits of utility-scale solar are likely to be substantially larger than the benefits of rooftop solar per dollar spent on solar PV capacity.
Implications for policy
While the results of our analysis solely apply to Xcel Energy Colorado system and should not be transferred to other regions without proper calibrations, the general relationship we find between costs is likely to hold true for many US utilities with significant solar potential.
However, it is also important to take into account the variations among power systems that largely depend on the geographical region. For example, installing utility-scale PV systems to serve load centers with limited land for installation or transmission capacity to deliver from utility-scale PV systems built remotely (such as New York City) may have different challenges. These challenges may overall favor smaller distributed installations represented in this study by the residential-scale PV systems. Smaller distributed systems may be more suitable for developing microgrids and system resiliency (although it may require changes from today’s normal residential-scale installation and set-up practice).
But what we wanted to communicate with you, Mr. President, is that the findings of the study illustrating the difference in utility- and residential-scale PV systems can have various policy implications.
First, utility-scale PV systems can be installed at a significantly lower cost – about half the cost - for achieving a given level of solar penetration, regardless of the ultimate goal.
Second, utility-scale PV systems can produce nearly 50% more power replacing generation from traditional generation resources and lead to larger environmental benefits. These benefits include reduction in air emission or water usage; reduction in fuel usage at the traditional generator resources; increase stability in power cost that is typically impacted by fluctuations in fuel costs; or avoiding the need for future generation resource additions to maintain resource adequacy. With the likely onset of new greenhouse gas savings targets from the newly released Environmental Protection Agency (EPA) rules, the options for reducing carbon emissions and the costs of achieving them will take on even greater importance.
Finally, utility-scale solar systems allow everyone to access to solar power. From the standpoint of cost, equity, and environmental benefits, large-scale solar is a crucial resource. This is a significant consideration for utilities and their regulators, and various policy makers starting with you and your staff, Mr. President, assuming that you are looking to maximize the benefits of procuring solar capacity at the lowest overall system costs.
Therefore, while both types of PV systems are going to be helpful and necessary in order to decarbonize our power system, at present, we hope that you recognize the difference and as of today, utility-scale PV systems may often offer power and environmental benefits more cheaply and equitably than distributed, residential-scale PV systems. It is certainly possible that in some locations, residential-scale PV will have cost and system advantages over utility-scale solar. The study looked only at likely costs, not the extremes of possible costs for either approach.
Finally, we sincerely hope that the PV panels you installed at your residency will not be taken down by your successor, as we observed back in the late 1970s and mid-1980s.
 Today about 70% of residential systems are leased from third party owners. Industry reports and our own calculations indicate that the cost of solar power to residential customers from leased systems is typically larger than the cost of solar power from otherwise-identical systems that are customer-owned. Therefore, the calculated per-MWh difference between utility- and residential-scale leased systems is even larger than the difference between utility- and residential-scale owned systems. In addition, industry reports indicate that customer ownership is likely to overtake leasing in the next several years. Because our target year is 2019, customer ownership is the more logical benchmark for comparison.