Energy Innovation Doesn't Just Happen: How Government Policies Destroyed and Regenerated the U.S. Wind Turbine Industry, Twice
- December 8, 2014
- 405 views
After a decade of annual near-death experiences, the production tax credit (PTC)—a tax benefit for generating electricity from certain renewable sources like wind—was allowed to expire at the end of last year. Like other policies such as investment tax credits and renewable portfolio standards (RPSs), the PTC was designed primarily to help shift the country’s energy supply towards more renewable sources. The response to these policies has largely come from wind generation, which now contributes over 4% of U.S. electricity supply. A different set of policies in various European countries has enabled wind to meet over 7% of E.U. demand.
There are generally two ways to increase electricity generation from wind: we can construct more wind plants, or we can make the wind plants we build more productive. Arguments for the PTC tend to focus on the former: counting new project starts and added capacity, and noting ancillary benefits like jobs and economic stimulus. This makes sense, since it’s easy to count wind turbines cropping up over the landscape, and the relationship between the presence of incentives and the pace of construction can be readily seen.
The second means of increasing production is through technology innovation, and, in contrast to construction projects, it can be difficult to measure—innovations are not as easily countable as turbine towers, and the response time between implementation of an innovation policy and the appearance of any resulting technological advances in the commercial market is not immediate. Furthermore, when considering incentives for innovation, policymakers face difficulties of determining the counterfactual, i.e., would the advance occur even without the incentive? Nonetheless, innovation has been an immensely important part of the wind electricity success story: turbines today are larger and have higher capacity factors than their predecessors, reducing the cost of wind electricity to levels more or less competitive with conventional generation. If advances in technology had not happened, then wind projects would likely always need subsidies to be economically viable. Thus, it’s worth looking beyond building wind plants to think about how government policies incentivize technology innovation as well.
Innovation policy design
There are two general theories of innovation policy. Under technology-push policies, governments aim to reduce the costs of investing in innovation by providing direct subsidies—most often, R&D funds. Under demand-pull mechanisms, policymakers hope to increase the payoffs of investing in innovation by enhancing the market for the technology. The PTC and RPSs fall under this category—they create a demand for wind capacity, thereby (perhaps) inducing wind turbine suppliers to invest in creating better technology. Within each of these categories, policies can either be market interventions, like the PTC, that give one technology preferred treatment in the market, or command-and-control regulations, like the RPSs, in which a certain standard must be met. We often think of these respective options colloquially as carrots and sticks.
Command-and-control policies can be particularly effective at promoting innovation when they are stringent enough to be technology-forcing, that is, when meeting them is difficult using current technology. Notable examples of technology-forcing regulations in the U.S. occurred with emissions control technologies for power plants and automobiles.
Incentivizing the wind
So which policies, if any, have been successful in the U.S. at not only incentivizing construction, but also inducing innovation? The history of wind technology development in the United States provides a clear example of how important the policy environment can be for innovation.
Many of us think of serious wind development in the U.S. beginning in the 1970s, but the first wind buildup actually occurred in the early half of the twentieth century. Enterprising rural farmers connected parts from water-pumping windmills, generators, and batteries to provide household electricity, and by the 1930s mass-produced “windchargers” could be purchased (via the Sears catalog, among other outlets). Hundreds of thousands of these small turbines soon dotted the American Midwest, but the establishment of the Rural Electrification Administration in 1935 and its subsequent aggressive push to connect these isolated farms to the electric grid essentially killed this industry by the 1950s.
The promise of atomic energy and cheap fossil fuels in the post-WWII era meant little interest in wind. By the latter half of the 1970s, however, stubbornly high nuclear plant costs and the oil embargo sowed the seeds for the first wind renaissance. The U.S. government took two significant policy actions to encourage wind technology. First, the Large Wind Turbine (LWT) R&D program administered by DOE and NASA attempted to develop commercial, utility-scale (multi-megawatt) wind turbines (Figure 1). Second, the government issued what amounted to a 25% investment tax credit on wind turbine construction. Combined with state-level policies, the tax credit in California totaled 50%, and the wind rush was on. Instead of investing in large, complex machines, the order of the day was to build small and build fast. By 1985, 13,000 150-250 kW plants were constructed (Figure 2).
Figure 1: Boeing MOD 2 turbines at Goodnoe Hills, Washington, developed as part of the Large Wind Turbine research program. Erected 1980, dismantled 1986. Photo: US Government/Public Domain.
Were these policies successful? The LWT program failed to produce a turbine that made it to commercial production. The large plants proved technically complex and generally unreliable, and while some prototypes met with limited success, many met the fate of the Boeing MOD 2 turbine at Medicine Bow, Wyoming, which was dynamited and sold as scrap for $13,000 just five years after being built at a cost of $6 million.
In California, the tax credit incentivized construction of the turbines, but there was no incentive for those to then produce electricity. As a result, many plants were unreliable: the worst wind farms had capacity factors of less than 10%, and a not insignificant portion were later removed. The most successful turbines were of an older, simple, Danish design.
Figure 2: Old turbines installed at Altamont Pass, CA. Photo: David J. Laporte / CC-BY-2.0
By some accounts, these two policies were high-profile failures that set the industry back a decade as people became disillusioned with wind. To be fair, though, these programs at least established a foothold for wind generation in the U.S. A DOE report found that the R&D program laid the technological foundation for later growth, while in California, the state retained enough working turbines in 1990 to produce over 2.5 TWh of electricity annually.
A second wind renaissance began just before the turn of the millennium. Coincident with the PTC (enacted in 1992) and state-level RPS policies—which mandate a certain proportion of electricity generation come from renewable sources—established beginning in the late 90s, the U.S. has seen dramatic growth in wind production over the past fifteen years.
We have mentioned four main policies involved in the history of U.S. wind generation: the level of federal R&D spending, the investment tax credits of the 1980s, the PTC, and the state-level RPSs—one technology-push policy and three demand-pull policies. It seems reasonable to conclude that the investment tax credits were not “technology-forcing,” and thus unlikely to have spurred innovation. But how do we assess policy effectiveness? First, we need to a way to measure innovation.
Economists use patent counts as one possible proxy for innovation, for reasons mainly related to data availability. Notwithstanding the well-documented limitations of this metric, it is not an unreasonable way to get a rough idea of the relative level of innovation activity within a particular technology area over time.
You can see a chart of patent counts juxtaposed with these policy variables in Figure 3 and can probably make a few reasonable hypotheses about the relationship between each policy and patenting activity. If we can also represent these policy variables (and additional “control” factors) quantitatively, we can set up a regression equation to determine which policies are most correlated with innovation. This is exactly what we do in a recent paper, to which we refer you for all the modeling details.
Figure 3: Wind patenting rate over time (blue) juxtaposed with various wind policies: level of federal wind turbine R&D funding (top, green); counts of state-level renewable portfolio standards (bottom; red); and periods during which various tax credits were in place (bottom, shaded). Image from our paper (Horner et al 2013 Environ. Res. Lett. 8 044032).
Our results support the conclusion that the investment tax credits had no effect on innovation. Federal R&D had a significant, but small correlation, supporting the idea that these investments have been marginally effective. Perhaps the most interesting finding is that the state-level RPS policies, and not the PTC, are most correlated with patenting activity.
These results make sense for several reasons. First, of these policies, the RPS is the only command-and-control mechanism and is thus most likely to be technology-forcing. Wind farm operators desire to meet renewable generation mandates as efficiently as possible. Some of the costs of building and operating a wind farm, such as land acquisition and tower construction, would be expected to scale on a per-turbine basis. Larger turbines also utilize better quality wind at a higher altitude. Therefore, larger turbines likely achieve lower per-MWh costs. The experience of the federal R&D program indicated the need for technological advances to achieve reliability in these larger sizes, and thus there was an incentive to invest in innovation.
Second, the lack of an effect from the PTC in spurring innovation may have had a lot to do with how the credit was implemented. Beginning in 2000, the credit was renewed for periods of only one or two years and was allowed to expire briefly three times. Because there is a lag between innovation investment and delivery to the market, this short renewal period did not inspire confidence in suppliers that a payoff for investment would still exist in the future. RPSs, in contrast, provided a stable, long-term signal that there would be a market for better turbine technology.
The history of wind turbine technology is quite fascinating, and there’s much more that can be said about how policies have affected its development. However, we’d like to end with two takeaways. First, transformative innovations often need government support to transition from the high-risk early period to a state where industry can take over. The fact that wind turbine technology has needed policy intervention to achieve its current level of success should not be seen as an indictment of it. After all, the IT and aerospace industries are full of technologies now providing huge societal benefit but that needed government support early on.
Second, the manner of this support—policy design—is critically important. Successful policies balance between rolling out existing technology and incentivizing investment in the next generation of technologies. The investment tax credits of the 1980s were mis-targeted, and a careful survey of the technology landscape could perhaps have provided a more fruitful direction for the LWT program. One important policy attribute is time horizon: when dealing with innovative technology areas, often a longer-term view is warranted. The lack of predictability in the PTC renewal schedule likely hindered its effectiveness in inducing innovation (though it clearly led to turbine construction), and an expiration date further in the future would perhaps have made it a more effective driver of technological progress.
Policy decisions arguably killed wind in the U.S. in the 1930s and 1980s, and policy decisions brought it back both times. While it doesn’t look like wind is headed for a third death in the near future, the talk of rolling back RPS policies in some states does threaten to bring the same sort of instability that plagued the PTC. In any case, the lessons learned from the history of wind policy should be useful as we look to incentivize development of other energy technologies.
 For an excellent history of U.S. wind energy through the early 1990s, see Robert Righter’s Wind Energy in America: A History (University of Oklahoma Press, 1996).
 Only one multi-megawatt turbine had ever been built—forty years previous! The 1.25 MW Smith-Putnam turbine successfully fed electricity into Vermont’s electricity grid in 1941, but suffered two equipment failures, and interest faded in the postwar environment.
 Admittedly, this is a simplistic characterization of how RPSs are deployed in practice. RPSs vary drastically from state-to-state; many allow for at least some of the mandate to be fulfilled via purchase of renewable energy credits and have cost-effectiveness requirements. These attributes can make RPSs less stringent than typical command-and-control policies, and thus less effective technology-forcers. For a detailed look at how each state’s RPS is set up, see dsireusa.org.
Nathaniel Horner is a doctoral student in the Department of Engineering and Public Policy at Carnegie Mellon University, where his research includes innovation policy, energy systems, and energy use in the information technology sector.
Inês Azevedo as an Associate Professor in the Department of Engineering and Public Policy at Carnegie Mellon University, where she also serves as Co-Director for the Climate and Energy Decision Making (CEDM) Center.