Chapter 10 A Solar City upon a Hill
My favorite conference happens every February in National Harbor, Maryland. The setting—a gorgeous convention center with an expansive, nineteen-story glass atrium that looks out onto the Potomac—evokes an ambience of sweeping possibilities. This is the ARPA-E Summit, the flagship event of a futuristic government agency tasked with finding and funding the most innovative energy technologies. Hundreds of scientists and entrepreneurs showcase their research at endless rows of booths brimming with posters, prototypes, and pitch decks. Like kids in a candy store, college students from all over the country furiously take notes as they chart careers in energy innovation.
I remember my first summit vividly. The year was 2012. I was still a graduate student, working on perovskite photovoltaic (PV) cells and eager to peer out of my technology silo to survey other technologies, from next-generation batteries to software that could manage the smart grid. But my top priority was to meet Dr. Arun Majumdar, the founding director of the Advanced Research Projects Agency-Energy (ARPA-E).
An acclaimed scientist from Berkeley in his previous life, Arun had deftly transitioned to the role of D.C. dealmaker in 2009, when President Barack Obama tapped him to run the new agency. He quickly drummed up strongly bipartisan support for the Advanced Research Projects Agency-Energy (ARPA-E), which was modeled after the Defense Department’s early-stage innovation arm, the Defense Advanced Research Projects Agency (DARPA), which, among many other successes, helped create the Internet. Energy and climate issues provoked deep disagreements between Democrats and Republicans at the time, and still do. Yet both sides of the aisle—from Senator Chris Coons (D-DE) to Senator Lamar Alexander (R-TN)—heaped praise on Arun and his new agency.1
I ended up cornering Arun at a bar at the end of a long day of sessions. I peppered him with questions about science, policy, and the future of energy, which he casually fielded while waving to old friends and cracking jokes. Looking back, I’m a little embarrassed for harassing the poor man, who probably just wanted to sip his beer and reflect on a successful summit. More broadly, ARPA-E’s model has enjoyed early success. Through 2016, less than a decade after its founding, the agency’s grants to researchers developing early-stage energy technologies had already stimulated even more investment—nearly $2 billion—in those projects from the private sector.
When I returned in February 2017 for the annual summit, the mood was different. The technology showcase was still impressive, and the speakers still presented optimistic visions of roads filled with autonomous electric vehicles (EVs) or a grid run by sophisticated artificial intelligence. But Donald Trump was now president, and the elephant in the room was the unanswered question of whether his administration would continue to prioritize energy innovation.
The answer emerged a month later: an emphatic “no.” The Trump administration’s budget proposal to Congress set out to slash funding for research and development (R&D) into renewable energy by 70 percent. Similarly, it aimed to eviscerate R&D related to the power grid and nuclear energy. And one proposal in particular was particularly concerning: ARPA-E would be completely defunded. The curt rationale in the president’s so-called skinny budget was that “the private sector is better positioned to finance disruptive energy R&D and to commercialize innovative technologies.”2
This is a statement of pure ideology. In fact, America’s history of sporadic support for energy innovation has demonstrated exactly the opposite. Whenever public support for energy innovation has waned, so too has private-sector investment, promptly stalling progress on developing and commercializing new technologies. Fortunately, the president’s proposal was sharply rebuked by lawmakers in his own party—a Senate subcommittee chaired by Sen. Alexander reported that it “definitely rejects this short-sighted proposal” to defund ARPA-E.3
American politicians are fond of calling the United States a shining “city upon a hill,” and when it comes to energy innovation, America has historically lived up to this billing. As I write this in 2017, the United States still outspends every other country on energy R&D, its researchers lead their global peers in securing patents for energy inventions, and its firms, such as General Electric, continue a proud tradition of investing in new technologies. But that primacy is not guaranteed. For nearly two decades, federal investment in R&D across all fields has flatlined, after half a century of brisk growth. Projections indicate that if public funding for energy innovation in particular stagnates in the coming years, China will surpass the United States (figure 10.1).
Historical and projected government funding for energy innovation in top countries. The black bars display current public funding for R&D and demonstration of energy technologies. The gray bars correspond to countries’ commitments under Mission Innovation to double their funding levels.
Source: Mission Innovation.
Stagnant U.S. funding would be disastrous on many levels. First, it could preclude the technological innovation that solar energy needs to realize its potential. Even if China and other countries step up their funding for energy R&D, the United States has by far the most well-developed innovation institutions and top-flight talent; therefore, without U.S. leadership, the global pace of energy innovation will slow. Then, such breakthrough technologies as perovskite PV, solar fuel generators, and high-temperature concentrated solar power (CSP) plants may languish on the drawing board, falling short of commercialization. Similarly, the constellation of enabling technologies for a flexible system that can accommodate volatile solar PV output—from flow batteries to advanced nuclear reactors—might also progress more slowly. Limited to existing silicon solar PV and a mediocre supporting cast, solar energy is likely to fall short of supplying a third of global power demand by midcentury or a majority of humanity’s energy needs by century’s end.
Canceling funding for energy innovation also would directly harm U.S. economic prosperity and competitiveness. That’s because even if solar energy does not achieve its full potential, solar PV is still on track to grow significantly, and the United States will fail to capture a significant piece of this growing pie if it does not invest in innovation. Bloomberg forecasts that even with no major technological shift away from silicon solar PV, fifteen times more solar power than the amount installed through 2016 will be added globally through 2040.4 More broadly, today’s global clean energy market is worth $300 billion, and it is set to grow rapidly. China has raced ahead to a daunting lead, not only dominating the production of solar panels, but also leading in the manufacture of batteries and wind turbines. Its dominance will only grow if unchallenged by innovative products from the United States. Not only would that constitute a missed economic opportunity, but the United States might even rely to an unhealthy level on imports of Chinese clean-energy products, reminiscent of U.S. dependence on foreign oil in decades past.
Skeptics of the need for U.S. leadership might counter that the United States has already lost the race to manufacture such clean energy products as solar panels, which China’s economy is structurally better adapted to produce. Moreover, solar panels are low-margin commodities, a business that the United States should want no part of. These arguments are certainly true, and the United States should not be competing to manufacture commodities. Rather, the U.S. economy shines when it invents new technologies, applies advanced manufacturing techniques, and constantly improves the performance of its products faster than the competition can keep up. Today’s solar industry—reliant, as it is, on silicon photovoltaic panels—may appear unsuited to U.S. strengths, but by investing in technological innovation, the United States has the ability to reshape the industry to better harness its core competencies.
To do so, U.S. policymakers’ first priority should be to boost R&D funding for next-generation solar technologies, as well as for the raft of other technologies that will be needed to support a solar-powered energy system. But federal dollars alone are insufficient—the private sector ultimately must provide the bulk of the investment to bring promising technologies to market. To adapt to this reality, the federal government will need to modernize the way it provides support for innovation, taking care to do so in a way that also attracts private sector investment. ARPA-E is a good start and should be expanded, not eliminated. In addition, the government will need to fill the gaping funding hole for demonstrating first-of-a-kind, commercial-scale energy projects, so the private sector knows when a technology is ready for prime time.
Technological innovation presents a tantalizing opportunity for the United States to develop new solar technologies to be used around the world. In addition, America can set an example for how to promote the other forms of innovation that solar needs to achieve its potential. For example, tweaking the U.S. tax code could accelerate financial innovation, opening the spigots of capital flow to deploy more solar. And to advance systemic innovation, federal and state governments could work together on improving the power grid’s ability to flexibly accommodate a high penetration of solar power—for example, by building long-distance transmission lines, modernizing markets, and encouraging a diverse, zero-carbon power mix. These proposals tend to get less airtime than others, such as putting a price on carbon emissions to create a level playing field between dirty and clean energy sources. But a carbon price is no panacea—just a sensible policy that should be accompanied by others that advance innovation.
Unfortunately, today’s solar policies mostly fail to advance any of the three types of innovation. Costly federal tax credits and state renewable energy mandates encourage more deployment of existing solar PV technology (and can even discourage the commercialization of emerging solar technologies) without preparing for a future in which the value of solar PV plunges below its cost. Today’s policies are at best an expensive effort to trim U.S. carbon emissions and create jobs involved in the installation of solar projects. Instead, by investing in the three types of innovation needed to continue solar’s rise, the United States can reap the economic rewards of an advanced solar industry and serve as a solar city upon a hill—an example for the rest of the world to follow to substantially displace fossil fuels with solar energy.
Still, it will take more than just domestic policies to enable solar to reach its global potential. In particular, commercializing a suite of new solar technologies could require public funding from countries around the world and a newly transformed solar industry willing to invest in innovation. U.S. leadership can make that happen. Through Mission Innovation, an international initiative to advance energy innovation, the United States should urge other countries to invest more in R&D and help them set up their own institutions like ARPA-E. (Before making recommendations abroad, however, the United States must invest in innovation at home, and President Trump should walk back his pledge to pull out of the Paris Agreement on Climate Change in 2020, a move that would destroy U.S. credibility with international partners.) And the United States should reduce trade barriers, opting instead to foster a globalized solar industry. These steps might appear to help the competition. But this is a competition that plays to U.S. strengths. Indeed, in other innovation-led industries, such as semiconductor manufacturing, U.S. firms command global supply chains and invest liberally in the next technology generation.
The United States is fully capable of transforming solar—from how it is financed, to what it looks like, to how flexible the system that it plugs into is—and enjoying the benefits of doing so. President Trump should rethink his administration’s priorities. Energy innovation remains a goal that is just as rational and bipartisan today as it was when both sides of the aisle resoundingly applauded the creation of ARPA-E.
Third Time’s the Charm?
The assertion in President Trump’s budget statement that the private sector is best positioned to finance applied energy R&D and commercialize technologies is not new. The roots of that sentiment stretch back more than half a century, to the final years of World War II. In late 1944, President Franklin Delano Roosevelt wrote a letter to Vannevar Bush, who had run the wartime Office of Scientific Research and Development, which honed radar, mass-produced antibiotics, and launched the Manhattan Project. In the letter, President Roosevelt wondered if the scientific innovation that had given the Allies the advantage—and would ultimately win them the war—might also “be used in the days of peace ahead for the improvement of the national health, the creation of new enterprises bringing new jobs, and the betterment of the national standard of living.”5
Absolutely, responded Bush. In a report entitled Science: The Endless Frontier, he argued that the government should fund “research in the purest realms of science,” which would generate “a stream of new scientific knowledge to turn the wheels of private and public enterprise.” He cautioned, though, that federal support must come with “complete independence and freedom for the nature, scope, and methodology of research.”6 A prolific scientist and inventor himself, Bush was no doubt aiming to secure a healthy funding stream for his fellow scientists that came with no strings attached, so that they could freely pursue interesting research. He would succeed wildly. From 1953 to 2012, federal funding for basic scientific research would increase from $265 million to $38 billion.7 (Although there is no hard-and-fast distinction, a rule of thumb is that basic research aims to fill gaps in the body of scientific knowledge, whereas applied research aims to solve practical problems in the real world.)
But the model of innovation that Bush proposed was seriously flawed. In the more than seventy years since the publication of his landmark report, it has become clear that just funding basic scientific research is not enough to guarantee new commercial technologies, especially in energy. Yet along the way, Bush’s philosophy has entered the conservative canon. President Ronald Reagan and his advisers fused Bush’s support for basic research with their free-market principles, concluding that federal support for applied R&D would encroach on the private sector’s territory. The Trump administration, focused on paring back what it perceives as a bloated government, is channeling Reagan’s aversion to applied R&D to hack off any government program that advances technology development with an application in mind.
The available evidence, however, does not bear out the thesis that government support for applied R&D cramps private investment. In fact, two short-lived waves of investment in applied energy R&D over the last half-century suggest exactly the opposite conclusion. The first wave, both initiated and abruptly ended by the federal government, demonstrated that unpredictable federal funding for applied R&D can chill the private investment climate in energy innovation. The second, led by private investors who ran out of money, demonstrated that emerging technologies face a gap in funding that the private sector is unwilling or unable to fill. Still, learning the lessons of these two failed waves of investment in innovation can inform a policy strategy to induce a third, sustained wave.
The first wave came in the 1970s in response to the oil crisis. Under President Jimmy Carter, the United States steeply ramped up funding for R&D into both solar PV and CSP (figure 10.2). These funds focused on applied activities to improve solar technology, spanning efforts to boost efficiencies in the lab and demonstrate solar technologies on a large scale. But as steeply as federal support rose, it plunged even more quickly. Spared an energy crisis by cratering oil prices in the 1980s, President Reagan had a free hand to slash applied R&D activities that did not comport with his vision of small government and free markets.
Historical U.S. DOE funding for applied R&D in solar technologies. Record of funding levels for DOE-sponsored applied R&D in solar PV and CSP. Note that DOE also funds basic science research that is relevant to solar technologies; and other U.S. federal agencies, like the National Science Foundation, also fund solar R&D. DOE funding, however, represents the bulk of funding dedicated to the development of new solar technologies.
Source: Schmalensee et al. (2015).
This move crippled solar innovation in the United States. It stranded federally funded research projects midstream, and it halted the $2 billion demonstration program before it could fully show the commercial potential of solar technologies.8 Rather than create space for the private sector to take over and invest in innovation, it led to a 50 percent decline in private investment in solar R&D from 1985 to 1995, when federal funding for applied R&D remained low.9
The salient lessons here are first, that unpredictable government funding for applied R&D can chill the private-sector investment climate; and second, that sustained low funding levels can keep the private sector away rather than creating space for it to invest in innovation. Notably, Reagan did largely spare basic scientific research, and his budgets would serve as a template for decades; basic science would account for 60 percent of all energy R&D into the late 1990s.
In the second wave of investment in solar innovation, driven by the private sector, Silicon Valley venture capitalists invested $25 billion in start-ups (many of them solar companies), from 2006 to 2011, and lost over half their money.10 These venture-backed start-ups failed in large part because of the terrible timing of China’s surge in silicon solar panel production, which produced a flood of cheap supply that washed away upstart technologies before they could reach commercial scale. The start-ups themselves weren’t blameless, and many made poor business decisions (for instance, to prematurely scale up production of products that weren’t yet ready for commercial prime time).
The start-up bust also stemmed in great measure from the inadequacy of private-sector support. My collaborator Ben Gaddy and I have looked at hundreds of clean energy technology start-ups that received funding from 2006 to 2011 and concluded that the venture capital model was a particularly bad fit for solar companies. Venture capital is great for getting software companies off the ground because they need limited capital and can return eye-popping sums to their investors if they succeed. The model also can work for more capital-intensive medical technology start-ups because large firms, like pharmaceutical companies, are often willing to coinvest or acquire start-ups.
But when it came to start-ups developing new solar materials—or battery chemistries, or biofuel syntheses—venture capitalists found themselves all alone. The major oil companies were exiting, rather than entering, the solar space. And government funding for applied R&D peaked in 2009, thanks to one-time stimulus funding, but then it dropped away (figure 10.2). Left to fund solar companies with capital needs in the hundreds of millions of dollars to set up production facilities and with commercialization timelines that might stretch a decade or more, venture investors often insisted that companies race to scale up and abandoned their bets once the companies ran out of money. The lion’s share of their losses on clean energy technologies came from start-ups developing new materials, chemicals, and processes. Their clean energy software bets, by contrast, paid off handsomely (figure 10.3).
Venture capital investments and returns by clean energy technology subsector. These figures were calculated from a database of all known clean energy technology start-ups receiving their first substantial (“A-round”) investments from venture capital firms between 2006 and 2011. For each subsector, the bars compare the total quantity of A-round investment in that subsector’s start-ups with the returns that those A-round investors realized when any of those companies were either acquired or taken public on the stock market. Each bar has an associated uncertainty range that arises from gaps in publicly available data.
Source: Gaddy, Sivaram, Jones, and Wayman (2017).
Some might say that this episode proved that the federal government has no business intervening in applied technology development. After all, the U.S. government lost a half-billion dollars on an infamous loan guarantee to Solyndra. But the overall federal loan program for energy projects actually ended up in the black, not the red (it also had notable successes, such as funding Tesla).11
On that count, the federal government might be faulted for actually not taking enough risk, rather than too much. Apart from a few risky bets, the loan program focused on funding large-scale demonstrations of well-understood technologies, such as solar PV farms. That money might have been better spent in smaller chunks than a half-billion dollars, and on demonstrating emerging technologies at a moderate scale to intrigue private investors. Because venture capitalists were unable or unwilling to foot the bill for such technology demonstrations, promising energy start-ups perished in the so-called Valley of Death, unable to reach the scale needed to attract more conservative, deep-pocketed investors.
The lesson from this discussion is not that all government funding for applied R&D and demonstration is effective. Rather, it is evident in hindsight that the Department of Energy (DOE) loan program should have targeted demonstration of earlier-stage technologies and spread out its funds among more projects. The surge in energy R&D funding in the 1970s produced boondoggles such as the $4.5 billion Synfuels program, which failed to produce any alternative to oil (though Laura Anadon and Greg Nemet argue in a recent retrospective that the program generated valuable technologies for other energy applications).12 Certainly, it is important to improve on past attempts to design more effective government schemes to support applied R&D and technology demonstration. But, without any government support at all, there is little chance that the private sector alone will carry new solar technologies through the Valley of Death.
Taking a step back from energy, the thesis that funding only basic science (with no intended application) is the best way to support innovation is out of touch with the reality across many fields. In fact, in the postwar period, the most prolific supporter of innovation has been the U.S. Department of Defense, which bets on technologies with clear national security aims in mind. The military’s goal-oriented approach to funding technology bets at the early stage (through DARPA) and continuing to support the development of those that show promise has resulted in many successes, including the deployment of precision missiles, stealth fighters, and drones.13 Some efforts have been plagued by cost and schedule overruns—notably, the ongoing development of a joint strike fighter—but the military can still boast that it played a principal role in enabling America’s technological primacy today.
Sometimes, as a welcome side benefit, the military’s targeted investments have ended up having broader application that was anticipated. For example, the military was the only buyer of integrated circuits back in the 1950s; it planned to use them in missile guidance systems but ended up spawning the entire semiconductor industry, whose chips are now in your iPhone.14 Similarly, the military’s objective in creating ARPANET—the precursor to the Internet—was to make it easier to share results from the few, spread-out research computers around the country. Obviously, that mission has now expanded to include sharing cat memes.
The Department of Defense’s successes are a powerful argument in favor of including applied R&D alongside basic science in the government’s portfolio of support for innovation. In fact, Venky Narayanamurti, former dean of engineering at Harvard, argues in his 2016 book that the distinction between basic and applied R&D is artificial.15 In 1945, Vannevar Bush’s model—that public support for basic science would lead linearly to private funding of technology development—might have been believable. Since then, too many examples of innovation’s twisted, nonlinear path have emerged to continue to separate basic and applied R&D. Venky’s book looks back at Bell Labs, a hotbed of innovation in the postwar period, and concludes that cross-fertilization of researchers thinking about theoretical, “pure science” problems with those turning these insights into real products, such as transistors, is the most effective way to develop technologies that improve society. Similarly, the best research labs today, such as the Howard Hughes Medical Institute, ignore the distinction between basic and applied R&D.
The right question to ask, then, is not if the federal government should support applied innovation. The case for doing so is open and shut. Much more important in the twenty-first century is figuring out how to do so most effectively.
How to Fund Energy Innovation
In 2016, I served on a review committee to evaluate postdoctoral researchers who had applied for grants from the DOE’s “SunShot” program. The goals were to pick dynamic, talented researchers who were studying innovative solar PV technologies, such as perovskite or quantum dot cells, and to prepare a new generation of leaders in this field.
The applicants rose to the challenge. Some projects had the potential to shatter existing efficiency records. Others focused on bread-and-butter studies of repeatable manufacturing processes. Some researchers had many years of experience, whereas others were fresh out of PhD programs, though with multiple top-tier journal articles to their names. I couldn’t help but think, “We should be funding every single one of these applicants.”
Reality set in when the committee convened to compare applicants. It soon became clear that only a handful of the researchers would actually receive funding, given the program’s strict budgetary limits. Most would be turned away. Although they might receive support from other U.S. funding sources, such as the National Science Foundation, those streams are also limited. Outside the United States, by contrast, they would be highly prized targets to be snapped up by other countries in Europe or Asia that would eagerly fund their cutting-edge research. Letting them go elsewhere would amount to the United States shooting itself in the foot—the economic returns that they were capable of generating would far outstrip the up-front investment in their research.
To accelerate solar innovation, the U.S. government’s first priority should be to dramatically boost funding for applied R&D into breakthrough solar technologies. Researchers at the Massachusetts Institute of Technology (MIT) point out that one way to do this is to reallocate funds that currently go to supporting less cutting-edge activities, such as research into reducing the soft costs of rooftop solar installations or making existing silicon solar panels live a little longer. That shift would be sensible. Out of the DOE’s more than $200 million budget for applied solar energy R&D, less than half goes toward new solar PV and CSP technologies.16 Although some R&D into incrementally improving existing technology might fill a gap that the private sector is failing to address, the government would be far more effective if it funded the next generation of technology development.
It would be a mistake, though, to fixate on shifting around nickels and dimes within the tiny budget that the federal government assigns to solar R&D. A generous accounting across all government agencies, lumping together everything from fundamental scientific research on semiconductor physics to economic analyses of solar manufacturing costs, yields an estimate for solar R&D of less than $400 million in 2015.17 By comparison, the federal government spends over ten times as much on tax breaks for the oil and gas industry.
U.S. policymakers should recognize that its paltry funding pales in comparison with the importance of solar innovation. This mismatch afflicts the broader state of support for energy innovation. As figure 10.4 illustrates, U.S. federal funding for energy innovation is several times lower than for health and space R&D. Left off the chart is funding for defense R&D, which is greater than that for all nondefense R&D combined.
U.S. federal investment in nondefense R&D, by function. Each bar breaks down the federal government’s budget for nondefense R&D, measured in billions of 2016 dollars. Note that “Nat. Res./Env” refers to “Natural Resources” or “Environment.”
Source: American Association for the Advancement of Science.
And neglect for energy R&D has been growing. From 1979 to 2017, energy R&D dropped from 13 percent to 2 percent of all federal R&D. Gravely concerned about the “economic, security, and environmental consequences” of underinvestment, several business leaders including Bill Gates called in April 2017 for the federal government to more than double funding for energy innovation to $16 billion annually.18
Still, just boosting funding for R&D is not enough. The federal government also needs to invest those dollars more intelligently. To do so, it should adopt many of the techniques used by ARPA-E, which takes a highly goal-oriented approach to funding “game-changing” technologies early in their development, with an eye toward their ultimate application.19 For example, recognizing that existing PV or CSP technologies harness only a fraction of incoming sunlight, ARPA-E has invested in a portfolio of “hybrid technologies” that generate both power and heat and use thermal energy storage to smooth out intermittency.20 Each ARPA-E portfolio of investments is linked by a pressing problem rather than siloed into a particular scientific discipline. And unlike most federal R&D funding streams, which are difficult to discontinue, ARPA-E is ruthless about cutting off funding to projects that are clearly missing their milestones.
Although ARPA-E is still less than a decade old, early analyses suggest its approach is effective. Anna Goldstein, a researcher at Harvard, was kind enough to share some of her preliminary findings with me.21 On average, researchers funded by ARPA-E are five times more likely to produce patents and publications, compared with those funded by other DOE arms that fund energy R&D. And when ARPA-E funds start-ups, they tend to receive more follow-on funding from the private sector than do energy start-ups funded by other parts of the government. Anna isn’t alone in lauding ARPA-E—in 2017, the National Academy of Sciences released a major report endorsing the agency’s early performance.22 Recognizing the early success of the model, President Obama’s administration urged Congress to increase ARPA-E’s budget from $300 million to $1 billion. 23
Even so, the federal government cannot bootstrap the commercialization of new energy technologies all by itself. Rather, the private sector ultimately must provide the lion’s share of investment—the point of government funding is to encourage private entrepreneurs, investors, and firms to make bets on promising technologies. In addition to ARPA-E, which has succeeded in attracting private follow-on investments for the technologies it selects, the Obama administration devised several other institutions to encourage collaboration between the public and private sectors. These deserve to be expanded.
The first new set of institutions, called “Energy Frontier Research Centers (EFRCs),” aims to connect fundamental scientific research with urgent technology needs. Several of the thirty-six existing EFRCs are directly relevant to solar energy conversion, including the Center for Advanced Solar Photo-Physics at Los Alamos National Laboratory. EFRCs are located in universities or national laboratories, where private firms can fund or collaborate with academic researchers.24
Another set of institutions that convene researchers from industry, academia, and federally funded laboratories is the network of Energy Innovation Hubs.25 These hubs are housed in U.S. National Laboratories, a network of seventeen crown jewels of American innovation whose origins date back to the Manhattan Project but which today are leaders in conducting energy R&D. Four hubs exist as of 2017, including the Fuels from Sunlight hub at Lawrence Berkeley National Laboratory and the Batteries and Energy Storage hub at Argonne National Laboratory.
Although these hubs can anchor a regional innovation ecosystem and convene the public and private sectors to intensely push toward solving major technology challenges, hubs received a total of less than $100 million in 2016—less than 2 percent of federal funding for energy innovation. It would be eminently sensible to expand the hubs and add new ones. These additional hubs might advance PV and CSP technology, as well as help develop flexible power grids to integrate high levels of intermittent renewable energy.
To make investment in energy innovation more palatable to the private sector, the federal government should provide shared facilities for private use in order to reduce the steep and redundant capital costs of each firm building its own facilities. The National Laboratories make their labs and equipment available for a fee, and they have recently rolled out innovative programs, such as Cyclotron Road, to fund and house entrepreneurs who can develop their technologies using the extensive resources on campus.26 This arrangement can fill the gap left by venture capital investors skittish about funding research into, for example, new solar materials.
Further down the line of technology development, the national network of federally funded manufacturing institutes (Manufacturing USA) offers facilities for firms to develop the processes to make new products.27 So, for example, a company might be emboldened to invest in developing a high-temperature CSP collector if it can experiment with the advanced equipment at a federal facility to figure out how to manufacture such a product at scale.
Finally, the biggest gap that the private sector alone has not filled is for demonstrating new technologies at commercial scale. Achieving successful demonstrations is absolutely crucial to bring down the risks perceived by private investors and firms weighing whether to undertake the time and expense to bring a technology to market. Unfortunately, the DOE’s Loan Programs Office, which made the infamous Solyndra bet, is politically embattled and unlikely to be the vehicle to fund demonstrations in the future.
There is no shortage of good ideas about what could replace it. The group of business leaders that recommended boosting funds for energy innovation to $16 billion annually also recommended the creation of a private corporation, funded by a single congressional appropriation of $20 billion over a decade, to invest in technology demonstrations.28 Because this corporation would be held at arm’s length from the government, it would be insulated from political meddling. And, it would take equity stakes in companies and projects rather than just insuring loans, so that the company could actually make money off its bets. To allay concerns that such a corporation would crowd out the private sector, it might be required to make each of its investments alongside private partners, which would attract rather than repel private investment.
Others have suggested recruiting state governments to guide the funding of demonstration projects that make the most sense from a regional context. And still others have proposed using prizes to attract private investment in demonstration projects. The National Aeronautics and Space Administration (NASA) has used such a strategy effectively (for example inducing Elon Musk’s SpaceX to meet a series of performance milestones).29 An energy prize might foster competition among firms and entrepreneurs to, for example, demonstrate a large-area installation of flexible, highly efficient PV coatings or generate substantial quantities of hydrogen from a solar fuel generator. Given the scale of the funding gap, all these different avenues for the government to support technology demonstrations are worth pursuing.
A likely objection, however, is that doing so would encroach on the free market and “pick winners and losers.” But this approach of providing government support—not just for fundamental science in the lab, but also for applied research, manufacturing scale-up, and commercial demonstrations—is exactly what the military has done, with striking results.
Its process starts with DARPA making numerous bets on high-risk, unproven technologies. Then the military can fund the most promising projects, and it has provided support for manufacturing and for demonstrating at scale for decades—from transistor factories in the 1950s to a planned refinery, announced in 2017, to produce biofuels for the U.S. navy.30 And the various service branches will often purchase the end product of a technology that the military has supported, giving firms the confidence that their products will ultimately find a buyer if they perform as desired.
In fact, the military could become the most effective supporter of clean energy innovation. New solar technologies in particular could provide direct security benefits for the military. For example, lightweight, flexible solar coatings could improve the operational versatility of U.S. troops and reduce the need for costly and vulnerable fuel convoys. Indeed, some scholars have argued:
The [Department of Defense (DOD)] is better placed for catalyzing rapid innovation in energy technologies than the DOE because the DOD is a major customer for energy-consuming systems and equipment for its roughly 500 permanent installations, as well as for operational equipment (spending $10 billion a year on liquid fuels alone). The scale of the resources that the DOD brings to technology development is impressive. It employs more than 30,000 engineers and scientists in R&D and procurement, and its annual R&D spending comes to about $80 billion, with procurement spending in excess of $100 billion. The DOD thus has the incentives and capacity to be a smart and demanding customer for new energy technologies.31
This two-pronged model of inducing innovation—funding technology development and creating a market for the products that emerge—has worked for the military and could accelerate innovation in solar technology, as well as for a range of other energy technologies needed for a solar-powered future. Clearly, the federal government currently underfunds R&D and demonstration of new technologies and should boost its support.
One might be forgiven for assuming that the government is doing well on the other prong, given that its subsidies to deploy solar power vastly outweigh its expenditures in research, development, and demonstration. Unfortunately, it turns out that U.S. policies to create the right market conditions for solar to thrive are also in need of reform.
No Taxation without Innovation
The problem with U.S. public policies that seek to encourage market adoption of solar power is not how much money is spent, but what that money is spent on. The goal of solar policy should be to encourage financial, technological, and systemic innovation that enables a sharply rising share of solar energy to decrease customer costs and carbon emissions at home, while providing an example of what works to the rest of the world. Today’s policies do not accomplish this. And even widely endorsed proposals to tax carbon emissions are not policy silver bullets; a carbon tax would need to be paired with support for new technologies to stimulate innovation most effectively.
At present, the largest federal subsidy for solar power is the investment tax credit, worth 30 percent of the up-front cost of a solar installation. In 2015, the federal government extended this credit through 2020, at which point the tax credit steps down to 10 percent over the next two years. As a result, the U.S. Treasury forecasts that the federal government will spend over $2 billion per year subsidizing solar power. Another federal tax incentive, which allows solar project owners to apply an accelerated depreciation schedule, could cost another $200 million annually.32
These tax incentives promote the wrong things, however. They almost exclusively benefit mature technologies—mostly silicon solar PV panels. Given the choice between an emerging, innovative technology and a commercially established one, both of which are equally eligible for the same tax credit, an investor will prudently choose the established one, even if the emerging technology might be cheaper and more efficient when produced at scale.33
A better approach would be to design incentives that fell in value as a technology got cheaper, weaning it off subsidies that it would not need and focusing support on the deployment of newer technologies that need an early boost. A rational policy would also cut support for technologies that failed to improve their cost or performance over time. Finally, none of these incentives should be disbursed through tax breaks, which can be hard to monetize in the first place; instead, direct cash incentives could be twice as effective at stimulating deployment without raising the fiscal burden on the federal government.34 These are all promising policy directions for Congress to pursue rather than choosing to extend the current investment tax credit beyond 2020, when the phase-out begins.
Critics of eliminating tax credits might contend that the credits have been crucial to establishing a thriving U.S. solar market. Indeed, through 2016, the United States had installed over 40 gigawatts (GW) of solar and was second only to China as the largest annual market in the world. Domestic subsidies played a large role in these achievements. Yet, I would counter that, from here on out, subsidizing existing solar technology will do little to reduce the U.S. carbon footprint, let alone the world’s. It will speed arrival of the day when existing solar technology penetrates as much of the market as it ever will, but solar will never go further if various forms of innovation do not materialize.
From an economic point of view, the solar industry is fond of trumpeting that it has created over a quarter of a million jobs in the United States.35 But most of those jobs are in the installation of projects using imported solar panels; government support for those jobs provides limited knock-on economic benefits. By contrast, investments in innovation could deliver a much higher return on public spending, generating nearly five additional jobs for the economy from stimulating a single new job in advanced manufacturing.36 Therefore, from both environmental and economic perspectives, any subsidies would be more effective if they were pointed toward innovation.
Critics might also argue that if fossil fuels receive billions in tax breaks, it is only fair for solar power to receive similar handouts. This observation raises an important point. The U.S. government should eliminate the $4 billion in tax breaks for oil and gas companies, which—as research I commissioned at the Council on Foreign Relations demonstrated—does not actually improve U.S. energy security or reduce prices at the pump.37 Still, even if those fossil fuel subsidies persist, it will still make sense to reform solar subsidies to encourage innovative new technologies that will be even more competitive with fossil fuels.
In addition to the federal solar investment tax credit, many states have supported solar deployment. One way they do that is through a mandate known as a “renewable portfolio standard,” which sets a quota for the share of electricity the state must procure from renewable sources such as wind and solar. Again, these policies can justifiably claim important achievements in deploying solar power, most notably in California. But, as with the investment tax credit, they do little to support technological innovation, mostly supporting the deployment of existing silicon solar panels instead.
Moreover, renewable portfolio standards obstruct systemic innovation by discriminating against nonrenewable, zero-carbon sources, such as nuclear power or fossil plants with carbon capture and storage. This discrimination runs counter to the goal of building a diverse mix of power resources to enable deep decarbonization. So, in California, although the share of wind and solar is rocketing upward, the last remaining nuclear plant is slated to close, in a state where nuclear contributed 15 percent of its power as recently as 2011.38
By contrast, in 2016, New York enacted policies to encourage both renewable energy and nuclear power, aiming to get most of the state’s electricity from clean sources by 2030.39 Policies like these, which do not discriminate against nuclear and other nonrenewable clean energy resources, are far more sensible than renewable portfolio standards. More states should enact them to ensure that a high penetration of solar power is made possible by the continued existence of flexible base resources, which, as chapter 9 noted, are plants that can be ramped up and down to compensate for fluctuating renewable energy output.
Mandates may not even be necessary to encourage renewable energy deployment. Texas provides an instructive case study. Although the state had mandated 10 GW of renewable capacity by 2025, the energy industry paid little attention to the mandate, and Texas blew through the target in 2012, over a decade early (although most of the installed capacity was wind rather than solar). The actual driver of renewable energy growth, rather than the renewable portfolio standard, was Texas’s buildout of transmission lines that linked faraway, windy regions like the Texas Panhandle—known as a “Competitive Renewable Energy Zone (CREZ)”—to demand centers in major cities. Emboldened by the CREZ transmission lines, developers eagerly added generation capacity, and Texas is now the biggest wind powerhouse in the country.40 Texas’s growth was motivated by the favorable economics of renewable energy once sufficient transmission capacity was in place, rather than by the government mandate.
Investments in expanding the electricity grid exemplify policies that advance systemic innovation to lay the groundwork for much higher penetrations of solar power in the future. Therefore, the federal government and states should work together to install a network of long-distance, high-voltage direct current (HVDC) transmission lines across the country to expand access to remote but rich renewable energy resources. The Trump administration has already expressed its eagerness to put new infrastructure projects in the ground. It could do so by exerting federal regulatory authority to speed up the siting and permitting process for new transmission lines.41
In tandem with this kind of initiative, state governments should work together to link their power markets, so renewable energy generated in one state can travel via long-distance transmission to demand centers in other states. For example, if California moves forward with its embattled plan to link its power market with those of its neighbors in the West, it may not have to throw away thousands of megawatts of surplus renewable power that exceed in-state demand. Transmission expansion and power market reforms are steps that, in the near term, may support only the expansion of existing technologies. But by promoting systemic innovation that enables a flexible grid capable of hosting more solar power, these policy steps would set the stage for emerging technologies to ultimately succeed existing ones.
Similarly, U.S. policymakers should undertake policies that spur financial innovation as well. Again, although this approach may benefit only existing technologies in the near term, financial innovation that unlocks vast pools of capital could make it possible to scale up advanced technologies rapidly in the future. A step that lawmakers could take in this direction would be to pass the MLP Parity Act. Recall from chapter 4 that Master Limited Partnerships (MLPs) are financial vehicles used extensively to finance oil and gas infrastructure such as pipelines and pay no corporate income tax. Renewable energy firms attempted to recreate the MLP with the growth-oriented YieldCo model (bundling together a portfolio of renewable energy projects to be traded on the stock market), which blew up in their faces. But the bipartisan MLP Parity Act, first introduced in 2012 by Senator Chris Coons (D-DE) and Senator Jerry Moran (R-KS), would enable institutional investors to invest in listed MLPs comprising solar power assets. Safe, tried-and-true vehicles, these MLPs could speed the flow of capital into solar.42 And down the road, if newer technologies can achieve commercial scale, they would benefit from the availability of this financing structure to tap into institutional investors’ deep pockets.
Finally, no discussion of clean energy policy would be complete without mention of pricing carbon. A carbon tax, a trading system for carbon pollution with an overall cap, or some other system to create economic pain for those entities that cause greenhouse gas emissions would indeed be a welcome correction to existing markets, which do not penalize polluters.43 In addition to a steady chorus of just about every economist singing its praises, prominent politicians on both sides of the aisle support a price on carbon. In the first year of the Obama administration, the U.S. House of Representatives passed the Waxman-Markey Bill to create a carbon cap-and-trade scheme, enticing eight Republicans to cross the aisle (the bill, however, was never taken up by the Senate). More recently, an all-star delegation of senior Republican statesmen paid a visit to the White House in 2017 to convince President Trump of the merit of a revenue-neutral carbon tax. Their plan would apply a price on carbon emissions across the U.S. economy, scrap the entire patchwork of clean energy deployment subsidies and mandates that drive conservatives crazy, and deliver a politically attractive tax break to working-class families.44 This solution would seem to be elegant and efficient, right?
Probably not on its own. As the MIT economist Daron Acemoglu and colleagues argue, a carbon tax solves one important market failure but leaves another unaddressed. It corrects the market’s failure to put a value on the damage from carbon emissions, which levels the playing field between clean and dirty sources of energy. That adjustment sets up the market to cost-effectively match energy supply and demand while limiting pollution.
Whereas this policy is ideal for efficiently dispatching existing technologies, however, it is not ideal for accelerating innovation.45 Absent government support for R&D—which creates scientific knowledge and intellectual property for the good of all, but which the market does not fully value—technological innovation proceeds more slowly than it would with government support. Acemoglu and colleagues find that even with a price on carbon, delaying investment in technological innovation can set back a clean energy transition and steeply raise its cost.46
Pricing carbon is an elegant, important policy, but it is by no means a panacea. Rather, an ideal policy portfolio for smartly promoting clean energy in the United States would direct a substantial fraction of the proceeds from a nationwide carbon tax to boost funding for technological innovation. This policy would be a grand bargain that would exchange ineffective mandates and clean-energy subsidies with a single carbon tax. (Importantly, environmental regulations, such as pollution controls on coal power plants—which are justified by health concerns—should remain.) This new approach could still be politically palatable, with substantial funds left over to compensate those who might be hardest hit by a carbon tax. It is crucial that support for energy innovation is prioritized in such a grand bargain. Otherwise, the United States might forego the game-changing technologies that it and the world need.
Executing Mission Innovation
I will never forget the Paris Climate Change Conference of 2015. For two weeks, the world’s attention fixated on the diplomats hammering away at a historic agreement. I caught glimpses of U.S. secretary of state John Kerry emerging from the negotiating room in the middle of the night, somehow energized, even as his staff trailed sleepily behind him. A brightly lit Eiffel Tower bore the ominous message: “No Plan B.” At last, on the final day, weary ministers joined hands and raised them in triumph, having signed a nonbinding agreement to limit global warming to 2°C. That agreement set up a framework for the countries of the world to curb their emissions, hold one another accountable, and ratchet up the ambition of their targets over time.
But for all the drama, the most important element of the Paris summit was decided on Day 1, on the sidelines of the formal negotiations. That day, the triumvirate of Bill Gates, President Obama, and Prime Minister Narendra Modi of India led the unveiling of an initiative to boost global energy innovation: Mission Innovation. All told, twenty world leaders signed up to double their countries’ public funding for energy R&D within five years. And twenty-eight billionaires, led by Gates, pledged to invest billions to bring breakthrough energy technologies to market.
This was the crucial breakthrough that happened in Paris, for the development of radically superior technologies is what will enable countries to make increasingly ambitious emissions pledges. And both the public and private sectors will need to come together to fulfill the promise of accelerating innovation.
Unfortunately, early actions by the Trump administration threaten to undo recent progress. In June 2017, the president announced his intention to withdraw the United States from the Paris Agreement in 2020. Following through on that pledge would wipe out the international goodwill President Obama and Secretary Kerry earned from negotiating the accord, damaging U.S. credibility on global energy and climate issues (and destroying diplomatic capital America needs to advance other foreign-policy priorities). In addition, President Trump’s determination to renege on the U.S. commitment to double its R&D spending threatens to further erode U.S. credibility as a leader on energy innovation. Other countries, including China, are eager to advance innovation in America’s stead, though, both increasing R&D funding at home and coordinating the direction of energy innovation abroad.47
Trump is misguided; if he continues to exclude energy innovation from his “America First” agenda, America could well end up last. The United States has compelling reasons to lead internationally on energy innovation. In particular, reshaping solar into a global, technology-driven industry would make use of and enhance U.S. strengths. Today, both the production and deployment of solar energy are concentrated in Asia. The combination of Chinese scale and tried-and-true silicon solar technology has proved unbeatable so far. The United States should not attempt to try and beat China at its own game.
Rather, the United States should spearhead a global push to commercialize new solar technologies. My colleague Dan Sanchez at Stanford and I have urged the United States to share lessons with other countries from its experience setting up ARPA-E to support innovative technologies.48 And researchers at Columbia University have proposed an initiative for some of the top funders of solar R&D—the United States, South Korea, Japan, and Germany—to freely share information and coordinate their funding priorities to advance new solar technologies rapidly. Importantly, such a collaboration would intimately involve private-sector firms, encouraging them to invest corporate dollars in R&D and seeking to bring new technologies to market.49
Precompetitive, international research collaborations play to U.S. strengths, because of the surfeit of talented scientists and engineers and boldly innovative firms in the United States. In the semiconductor industry, U.S. firms collaborate with international competitors all the time, often sharing the capital cost of R&D and factories so that they can afford to invest proportionally more of their revenue in advancing innovation. By spearheading a major solar research collaboration effort, the United States could inculcate norms currently lacking in the global solar PV industry. For example, the average Chinese solar company invests just 1 percent of its revenues in R&D, whereas the average American semiconductor firm invests more than ten times as much.50
Naysayers might contend that an energy industry will never resemble a high-tech sector. Energy is dominated by lumbering utilities and state-owned behemoths, it is politically bedeviled, and its core products—kilowatt-hours of electricity or gallons of gasoline—are low-margin commodities. Tech companies, by contrast, are nimble, write the rules as they go along, and use the high margins from computer chips, iPhones, and the like to fund ongoing R&D.
It would be better instead, the critics might continue, to employ aggressive trade policies to protect domestic solar manufacturers beleaguered by cheap Chinese imports. And indeed, the United States has rightfully imposed tariffs to retaliate against Chinese solar dumping that wiped out innovative Silicon Valley start-ups. At the time of writing, the Trump administration was preparing to erect additional, sweeping trade barriers to imported silicon solar cells and panels from around the world, relying on the rarely invoked Section 201 of the 1974 Trade Act. But trade barriers are blunt, destructive tools that should be only a small component of America’s innovation strategy. Deployed recklessly, they can provoke trade wars that might stymie U.S. chances of taking the helm of an innovative industry with a global supply chain along which goods, services, and capital can flow efficiently across borders.51
Rather than retreating within its borders, America should embrace the chance to reshape the global solar industry in its innovative image. There are important differences between solar panels and semiconductors, and the business model of the latter may not translate exactly to the former. But that is no excuse to accept the current industry’s culture of painfully slow technology progress, especially given the exciting advances emerging from top labs around the world.
Remember, the great solar game is not over; it is just beginning. To meet the goals set out by this book, solar power will have to grow over thirtyfold by midcentury, and much more beyond that horizon. Although the United States finds itself at a daunting disadvantage to China’s early lead, America can win back market share and deliver planet-saving technologies by leading a global push for innovation.
I’ve spent a decade studying solar. Today’s industry is barely recognizable from the one I started in. But for humanity to finally tame the sun, solar technology and the solar industry must become even more unrecognizable in the decades to come. Now, that is a future I look forward to.