A new study by The Brattle Group has found that utility-scale solar PV systems in the U.S. are a more cost-effective carbon-free power option than residential-scale PV systems -- at roughly half the cost per kilowatt-hour.

The analysis, prepared for First Solar, with support from the Edison Electric Institute, used data from Xcel Energy Colorado, a mid-sized investor-owned utility. Forecasting to 2019, the report compared the per-megawatt-hour customer supply costs of adding 300 megawatts DC of PV capacity in the form of 60,000 5-kilowatt residential-scale systems owned or leased by retail customers, versus 300 megawatts of utility-scale solar power plants that sell their generation to Xcel Energy under long-term power-purchase agreements.

The analysis projected that in 2019, utility-scale PV power costs in Xcel Energy Colorado's service territory will range from $66 per megawatt-hour to $117 per megawatt-hour across a range of scenarios. The projected power costs for a typical, customer-owned PV system will range from $123 per megawatt-hour to $193 megawatt-hour -- roughly double the cost.

The costs are even larger for leased residential-scale systems, between $140 per megawatt-hour and $237 per megawatt-hour.

For comparison, the national average electricity rate in 2014 was 12.5 cents per kilowatt-hour. The average retail rate for electricity in Colorado as of March was 11.74 cents per kilowatt-hour.

“There is no question that distributed solar is a popular and beneficial resource that many customers want and many customers are installing. But large-scale solar power plants are going to be substantially less expensive per kilowatt-hour than rooftop and allow more universal access to solar,” said Peter Fox-Penner, Brattle principal and co-author of the study. “So from the standpoint of cost, equity and environmental benefits, large-scale solar is a very, very important resource going forward.”

The report comes as utility regulators gather to discuss state policy strategies at the National Association of Regulatory Utility Commissioners' summer committee meetings in New York.

Past studies have examined the costs and benefits of residential solar compared to non-solar generation, and have also looked at the costs and benefits of existing solar tariff arrangements for residential customers (net metering in particular) for non-solar customers. The study, “Comparative Generation Costs of Utility-Scale and Residential-Scale PV in Xcel Energy Colorado's Service Area,” is one of the first to compare “solar to solar” in equal amounts, according to The Brattle Group.

“As somebody who does nothing but crunch numbers in this space, I was surprised, not that [utility-scale] was cheaper, but how much cheaper it was,” said Fox-Penner. “Solar panels are the ultimate modular product...and so you would expect if you buy 20,000 panels for one location, or 20,000 panels for 5,000 locations, they would all cost the same. So I wasn’t expecting as much of a cost difference, and I would not be surprised if regulators also have that impression.”

The Brattle Group based its study on historical PV installation data from the National Renewable Energy Laboratories to derive the cost decline rates from 2014 to 2019. The five forecast scenarios were based on with varying PV cost, inflation, financing parameters and different levels of federal Investment Tax Credit.

Declining manufacturing costs and the falling ITC are not the main reasons for the cost gap, since these factors affect both types of systems. Rather, the cost differential stems primarily from lower future plant costs for utility-scale solar facilities resulting from economies of scale, and from better tracking and positioning that allows for greater electric output than from an equal amount of residential PV. A higher capacity factor means that utility-scale solar also reduces more carbon dioxide.

The costs calculated are specifically “customer paid costs for generation,” in the case where Xcel Colorado buys the same amount of PV (300 megawatts) from two solar configurations and transfers those costs to the rate base. The study also evaluated the transmission and distribution costs of the two types of PV plants, as well as externalities, and concluded that they are unlikely to change the report’s conclusion.

In addition, the authors note that the findings are specific to Xcel Energy Colorado, and would need to be reconfigured to account for different cost drivers in other regions of the country. However, the report finds that the utility-scale solar will likely be cheaper than rooftop solar all over the U.S.