This article appeared in the Summer 2016 issue of The American Prospect magazine. Subscribe here.
The United States does not like to engage in explicit economic planning. Direct government pursuit of industrial objectives violates both our professed belief in free markets and our global commitment to liberal trade devoid of national favoritism. Nonetheless, the U.S. has been willing to use something close to economic planning when it comes to transitioning to both photovoltaic (PV) power installation and production of solar cells—a transition that markets won’t make on their own because of the current pricing advantage enjoyed by carbon-based fuels.
States and cities have subsidized solar startups. The federal government has used tax credits to subsidize solar production. The Department of Energy has underwritten research and even operated as a venture capitalist. However, because of the uncoordinated and slightly guilty nature of solar planning U.S.-style, the American effort on solar has been fragmented, bordering on incoherent—and is in grave danger of being displaced by a player that has no compunctions about using as much subsidy and as much government direction as necessary to capture the world’s solar production industry. That, of course, would be China. The U.S. has lost the first round. The question is whether we can lead again in next-generation solar—and what that means: PV production, installation, advanced R&D, or all three? It also requires U.S. policy to be a lot clearer about the linkages between these three aspects of a solar transition.
THE CURRENT ERA of solar includes lots of false starts, whose lessons should inform policy for the next era. In my 2010 book Emerald Cities (and in my 2009 article for the Prospect), I wrote enthusiastically about how states and cities were building a new solar production industry in the United States. An unlikely story I told was of two solar manufacturers in Toledo, Ohio. Toledo was long a center of glass production. Research at the University of Toledo used expertise in glass to create advantage in solar cells. When First Solar opened its first factory in nearby Perrysburg in 2001, it was the country’s largest PV solar panel producer.
Another manufacturing success out of Toledo was Xunlight, using flexible thin-film solar technology. The company was started by a University of Toledo physics professor, Xunming Deng, and his wife in 2002. With $3 million in startup funds from the university’s Innovation Enterprises, $2 million in state and local tax credits, and a $4 million state loan, the company began producing its flexible stainless steel solar cells. Xunlight also received $34.5 million in tax credits from the stimulus in 2010, in addition to a $3 million grant from the National Institute of Standards and Technology, a typical case of how solar entrepreneurs in the U.S. cobble together diverse subsidies.
In 2007, Xunlight raised $7 million in financing followed by another $22 million from major technology investment firms. It seemed on track to becoming an international leader in thin-film flexible solar modules. In 2010, Xunlight invested $2 million to open a production facility in Kunshan, China. The idea, or so CEO Deng said, was to assemble panels at a lower cost while expanding the Toledo plant. Two years later, the plant was sold. By the end of 2014, Xunlight filed for Chapter 7 bankruptcy and closed all operations, a victim of China’s pricing advantage.
In my home state of Massachusetts, Evergreen Solar was the darling of the solar industry. Started in 1994, Evergreen expanded in 2008 into a $450 million facility in the former Fort Devens, outside Boston. With more than $50 million in various state and local subsidies and land deals, Evergreen’s claim to fame was its patented string ribbon technology that used considerably less silicon than other panels. But silicon prices dropped, undermining Evergreen’s market advantage. By August 2011, Evergreen too had filed for Chapter 11 bankruptcy and sold its assets to pay off hundreds of millions of dollars to its creditors. Evergreen, along with Solyndra, has become a favorite whipping boy of conservatives warning against the perils of industrial policy.
WHAT HAPPENED? How could these promising companies go bankrupt when worldwide solar installations are growing? Last year, 30 percent of the new electricity generation in the country came from solar, and the industry projects about 100 gigawatts of annual deployment by 2018.
The main explanation is China. To establish a solar industry, China offered huge subsidies and free land to U.S. solar companies to locate there. The only stipulation was that they couldn’t sell their products in China. This happens to violate free trade principles and arguably trade law, but many U.S. companies took the deal and the U.S. government did not make an issue of it. Then China started developing its own solar industry and dumping products on the world market at below-cost prices—something that also violates trade law. By 2011, prices began falling dramatically and many U.S. and German producers couldn’t compete.
In 2012, U.S. manufacturers sought relief from Chinese dumping, and the Department of Commerce imposed tariffs averaging 31 percent. In 2014, Commerce closed loopholes and imposed tariffs averaging 52 percent (ranging from 26 to 165 percent) on PV panels from both China and Taiwan (because China had been using Taiwanese inputs to avoid tariffs). After review in January 2015, the department set the tariffs at an average of about 20 percent.
But the tariffs came too late to save much of the U.S. industry. At least 14 U.S. solar manufacturers went out of business by 2012, and almost 60 worldwide. With its explicit industrial policy of state-subsidized production and discrimination against imports, China is playing by its own rules and demolishing other nations’ efforts, even as it depends on their imports of its products. By 2015, seven of the top ten PV producers were Chinese. Only two U.S. companies, SunPower and First Solar, made the list. And although First Solar is building solar farms in the U.S. and throughout the world, almost all of its highly efficient panels are made in Malaysia.
And now China has created a huge internal market. China’s first big push in solar was an investment of $221 billion in renewable energy development as part of a stimulus package in 2009. This investment would help China meet a goal of producing 11.4 percent of its primary energy from renewable sources by 2015, and 15 percent by 2020. The five-year plan that started in 2011 originally had a goal of producing 5 gigawatts of solar energy, which was increased several times. A 2015 goal of adding 17.8 gigawatts was exceeded by 20 percent. The plan covering 2016 through 2020 calls for 100 gigawatts. China’s solar producers are building huge solar farms to meet the renewable goals of the five-year plans.
The good news is that solar PV in the U.S. is coming back—after dropping from the peak 2010 production of 1,205 megawatts to 517 in 2012, production was back up to 1,288 megawatts in 2015. While much of the growth has been in utility scale and rooftop residential installations, GTM Research predicts that the commercial sector of the market will reach $3.8 billion by 2020, due to the extension of the federal tax credit and increasing corporate adoption of solar.
But the comeback is largely based on states subsidizing a new round of incrementally better, tried-and-true crystalline solar PV panels and not cutting-edge technologies considered to be “next generation” solar. With this strategy, in the words of Varun Sivaram, who follows the solar industry at the Council on Foreign Relations, “solar is headed down a path of profitless prosperity.”
The issue of where production is located is thus related to the question of the solar industry’s rapidly evolving technology. Will the U.S. lead in developing next-generation technology, and if so, how? A related question is whether U.S. manufacturers truly need to be producing PV products using current technologies in order to keep moving to more advanced technologies.
MIT’s 2015 “Future of Solar” report maintains that considerable R&D needs to focus on achieving large-scale expansion of thin-film solar, integrating increasing amounts of solar power into the grid, and developing technology for large-scale storage. Further, reducing costs requires innovation in manufacturing technologies for producing solar, such as the elimination of expensive vacuum processing.
Why thin-film? Because the monocrystalline silicon cell technology that is currently being installed will not get much higher than 25 percent efficiency (the percentage of radiation hitting the cell that is converted to electricity). Many experts following the field argue that we need to double the long-run performance. Presumably, we need to incubate those companies that will be able to produce high-value-added products at low cost. If ultra-light, flexible solar panels could be produced at considerably lower cost per watt, companies could actually make real margins, not profitless technologies.
The Obama administration has been ramping up funding on renewable energy. The stimulus package invested $90 billion in clean tech, of which about $20 billion was on tax incentives for solar and wind, and about $50 billion for energy technology programs, mostly focused on the grid and energy efficiency. Although it is difficult to estimate total investment, as it’s spread across many departments and programs, annual federal funding on solar and other renewable energy is about $6.4 billion and will increase by 10 percent in the 2016 budget. And if it ever gets through Congress, Obama and 19 other world leaders launched the Mission Innovation initiative at the Paris Climate Summit, committing to doubling investment in clean tech research. If the U.S. is to be a solar leader, we need to direct more investment to developing more advanced panel and storage technologies.
THE GLOBALIZED NATURE of the industry makes it challenging to target domestic production as well as R&D and installation. A typical company may have its headquarters and some research in one country and its production in various locations. Even with the lure of subsidies, these corporate decisions are often beyond the reach of national policy.
In 2012, SolarWorld (headquartered in Germany), the largest crystalline PV solar producer in the U.S., closed its California operation and shut down lines in Hillsboro, Oregon. Then in 2014, when market demand had increased, the company announced a $10 million new line in Hillsboro that will increase production from 380 to 530 megawatts annually and increase employment from 700 to 900. Demand for the company’s 72-cell bifacial solar panel technology, which is up to 25 percent more efficient than other models, is growing, with an increase in shipments of 62 percent in 2015. But SolarWorld still imports panels it produces in Germany for installation in the U.S.
Georgia-based Suniva also struggled, but is coming back after a controlling stake was bought by China-based Shunfeng International Clean Energy. It moved production from overseas to a new domestic plant, which opened in 2014 in Saginaw Township, Michigan. The company was given a five-year $15 million tax credit from the Michigan Economic Growth Authority, and pledged to create 350 jobs and produce 170 megawatts of monocrystalline modules a year by 2017, which are Buy America–compliant—containing 80 percent U.S. content.
The nation’s biggest new solar production facility was originally scheduled to open later this year in Buffalo. In exchange for New York Governor Andrew Cuomo’s “Buffalo Billion” economic development program, which subsidizes construction of SolarCity’s “gigafactory,” the company committed to creating almost 1,500 production jobs to produce one gigawatt annually. This will be the installation company’s first foray into manufacturing. In fact, SolarCity argued against putting tariffs on Chinese panels—which is not surprising given that the company has received close to $1 billion in domestic tax subsidies and grants for installing Chinese panels throughout the U.S.
New York state is paying $750 million of the $900 million cost to build the factory, in exchange for SolarCity’s promise to spend $5 billion on the operation over a decade. The investment is part of a $1 billion commitment to create jobs and investment in targeted growth industries.
The significance of this plant is that it combines elements of crystalline with thin-film technology to make highly efficient panels. This is not next-generation solar technology, but it will make the most efficient panels on the market if the panels produced at scale perform as well as the pilot did. And the production process is supposedly highly efficient as well—taking considerably few steps than other producers. Despite its efficiency and a robust distribution and supply chain, its success as a manufacturer is not guaranteed. Its opening has been pushed back several times due to delays in obtaining equipment and is now scheduled for the summer of 2017.
The project, a signature initiative of Governor Andrew Cuomo, has been marred by scandal—exactly the sort of conflicts of interest that lead conservatives to conclude government should avoid industrial policy. An ongoing investigation is revealing that politically connected developers who were donors to Cuomo’s campaign got contracts for SolarCity and other major upstate economic-development initiatives.
Instead of opening the plant in 2016, the company reduced its projection of 1,400 factory employees to 500, bringing into question whether it was necessary for the state to build a 1.2-million-square-feet facility. And there is reason to question SolarCity’s employment commitment. When provided with an $11.8 million subsidy for a $27 million project to install solar panels at Oregon State University and the Oregon Institute of Technology, SolarCity originally partnered with SolarWorld to produce the panels in Hillsboro, Oregon, which would have generated $10 million in wages. Instead, the panels are being produced by convict labor. SolarCity gave the contract to Suniva under a subcontract with Norcross, which employed workers from Oregon’s Federal Correctional Institute. These inmates were paid 93 cents an hour, compared with the $11 starting wage at SolarWorld. This is not exactly the job creation envisioned by the Oregon Department of Energy’s Business Energy Tax Credit program, which has since been shut down.
THE SAGA RAISES the bigger question of whether states should be subsidizing solar production facilities to compete with China, especially when the domestic production represents only modest technical innovation. There is widespread agreement that we need to expand solar installations, which still produce less than 1 percent of our electricity. Our national policy to stimulate installation of solar is more coherent than our policy on innovation and production, but only marginally so.
One of the most important stimulants for private investment in installation is the federal tax credit for solar and wind energy. The tax credit has been in place since 2005, but only for short periods and with no guarantee of renewal. Investment slows when the credit’s future is uncertain.
The credit is essential to the business model of installers such as SolarCity, which leases its panels to homeowners. As the owner, the company gets the federal 30 percent investment tax credit. Fortunately, in 2015, Congress extended the tax through 2021. GTM Research estimates that this extension alone will result in more than 50 percent net growth in U.S. solar installations from 2016 to 2020. The credit will be eliminated for residential installations permanently in 2021, and reduced to 10 percent for commercial installations.
There is a crazy quilt of state policies to subsidize solar installations. A total of 29 states have renewable portfolio standards, which require utilities to purchase a defined percentage of renewable energy. And 41 states have mandatory net metering, in which consumers sell solar or wind power back to the utility. Connecticut, with low sunlight and high electricity rates, has heavy subsidies for solar. Nevada has a lot of sunlight and used to be third in the nation in solar installations. But in 2015, the Public Utility Commission increased fees to solar customers and reduced the price utilities pay to buy solar power back from rooftop panels—essentially negating net metering for retail customers, which allowed many to install solar to begin with. SolarCity, Vivint Solar, and Sunrun left the state as their business models cannot work without net metering.
THE DEBATE ABOUT solar parallels arguments about other industrial policies. The U.S. has given up much of its traditional manufacturing in such industries as machine tools, textiles, steel, and consumer electronics. Those losses have consequences both for jobs and for our balance of trade. It took extensive and aggressive trade policies to prevent the loss of the U.S. semiconductor industry to Japan and Korea, as well. However, the U.S. loss of most of the video- and audio-manufacturing industry did not prevent Apple from leapfrogging into entirely new product categories. And of course, the actual products are manufactured mainly in Asia.
Solar does provide domestic jobs, but relatively few are in production. According to the Solar Foundation’s 2015 survey, there are 209,000 solar workers in the U.S., an increase of 20 percent since 2014 and 123 percent since 2010. Only about 15 percent, however, are in production; with automation, that is not likely to increase much, even if the U.S. stays in the game. Of the remaining jobs, about 57 percent are in installation, with the rest in sales, distribution, and project development. So the rationale for more investment in innovation and production is not simply about jobs.
Solar—both installation and manufacturing—epitomizes the technology of the future economy. The debate will continue about whether America truly needs to produce solar cells in order to ramp up generation of solar power. But one thing is clear. If we want U.S. companies to stay in the production business, we need far more coherent policies—on trade, on subsidies, on technology, and on the connection between production and installation—than the ones we have now.