Two years ago, before it was even built, the Macho Springs Solar Facility in New Mexico made headlines thanks to a 20-year, 5.7-cents per kilowatt-hour power purchase agreement with El Paso Electric. The 30 percent federal investment tax credit and a generous New Mexico production tax credit helped make that price possible, but another factor was the likelihood that Macho Springs would be very productive.

And it’s proving to be just that – like many new utility-scale PV plants in the West, which are rewriting expectations for big solar performance in sunny regions.

Numbers on Macho Springs from the Federal Energy Regulatory Commission and Energy Information Administration differ slightly, but together they suggest the 50-megawatt plant operated in 2014 at a sizzling capacity factor around 33 percent after going online in the spring. (Capacity factor measures actual output against maximum possible output over a period of time.)

Macho Springs, now owned by Southern Company but built by First Solar using its thin-film panels on single-axis trackers, saw production fall off from October through December, when the solar resource is at its nadir in southwestern New Mexico, but still managed a 26 percent capacity factor in that period, according to FERC data.

The plant’s performance is no anomaly. As we calculated a few months ago, the 250-megawatt California Valley Solar Ranch in Central California, which uses SunPower’s crystalline silicon panels on single-axis trackers, operated at a 31.8 percent capacity factor in its first 12 months of operation.

These real-world numbers vary wildly from those often cited by solar skeptics. For instance, Charles Frank, in a widely publicized May 2014 Brookings Institution paper, argued that solar and other renewables suffer from “very low capacity factors and low reliability.” Frank assumed a PV capacity factor of 15.5 percent, based on old Energy Information Administration data, although he did allow that capacity factors can range “as high as 19 percent in the state of Arizona.”

In fact, even as far back as 2012, the 31.5-MW Long Island Solar Farm – in New York, not Arizona – operated at a 19 percent capacity factor in its first full year online. And that wasn’t a fluke; the plant, which uses fixed-tilt, crystalline silicon panels, produced almost exactly as much power in 2013 and then again in 2014.

Under recently updated methodology, the EIA now reports solar PV capacity factor data only as far back as August 2014, and the figure for the U.S. utility-scale fleet that month was 31.9 percent. Even in December, the EIA reported the capacity factor at 15.6 percent – higher than Frank’s overall figure.

Remarkably, government researchers say these figures don’t fully capture the improving productivity of utility-scale solar PV.

As module efficiency increases, cost-conscious developers tend to “either use fewer of them to reach a fixed amount of capacity (thereby saving on balance-of-system and land costs as well) or, alternatively, use the same number of them to boost the amount of capacity installed on a fixed amount of land,” Lawrence Berkeley National Laboratory scientists said in a study released last September. “In other words, for PV more than for other technologies like wind power, efficiency improvements over time show up primarily as cost savings rather than as higher capacity factors.”

Nevertheless, the researchers did report increasing capacity factors, with “average capacity factor among 2012 projects sharply higher (to nearly 30%) relative to the average capacity factors achieved by projects built in either 2011 or 2010 (just under 25% in either case).”

In addition to taking advantage of excellent solar resources, improved module efficiency and the use of tracking, high-productivity developers have capitalized on declining module costs to boost the size of their arrays relative to the inverters that convert DC power to AC. This “clips” some power off peak performance, but can be more economical and, by stretching out full output, increase the capacity factor.

So it was that LBNL scientists, taking into account the kinds of capacity factors we’re seeing today – between 30 and 35 percent – said that “although $50/MWh (5 cents/kWh) solar may not generate as much of a return as sponsors might like, financial modeling – not to mention the recent spate of market announcements – suggests that it is indeed feasible under the right conditions.”