Key takeaway: In a research project, I am relying on i3 data to understand the impact of Department of Energy grants to cleantech startups. This research, and accompanying report, should be completed towards the end of 2014.
In the clean tech world, everyone has a different opinion about what role, if any, government should play in energy innovation. Regardless, the fact on the ground is that the government has invested significant sums in university research, new energy technology procurement, and research grants to private sector firms. Do these programs work? Do they generate inventions and commercial success that would not have happened if the private sector were solely responsible for energy R&D? The answer for most of the basic research programs is unambiguously yes. It’s very hard for private actors to capture the benefits of basic research, so they don’t do very much of it on their own.
The answer is muddier when it comes to government grants to private firms to do later-stage research and commercialization. The argument for the programs is that the energy sector and small startup firms both suffer from market failures. The carbon costs of fossil fuels are not included in their prices. It’s often very hard for outside investors to asses the quality and risk level of a new high-tech energy startup, much less accept the long lead time and capital intensity of the typical new energy technology. Together these and other market failures cause a perfect storm of insufficient investment in cleantech innovation. Thus, the argument goes, the government should step in and subsidize those willing to take on the intimidating task of innovating in a sector with a lot of incumbent infrastructure.
But what if the government picks firms that would have gotten plenty of private capital anyway? What if the grantees would have filed just as many patents if they hadn’t gotten the grant? Across a sector of interest, like solar, do the grants speed overall innovation?
I’m trying to shed some light on these questions. I’m working with the Department of Energy to evaluate the impact of a certain class of grants that tends to go to research-intensive startups. The Small Business Innovation Research program (SBIR) began in 1983, and requires every agency to commit 2.7% of its extramural research funding to small, privately-owned, U.S.-based companies. For my research, I’m looking at a database of SBIR grants and grant applications within DOE’s Energy Efficiency and Renewable Energy (EERE) office, the primary channel for commercializing new, clean energy technology.
I’ve also gathered some useful private transaction and patent databases in order to evaluate the outcomes of companies that win the grants. A critical resource is the i3 database, which includes information about companies in the cleantech world that is very hard to find anywhere else.
Venture capital (VC) is an important enabler for the establishment of innovative, high growth firms, but it is volatile (see the graph below) and has trended away from funding firms in the seed stage. In theory, the DOE SBIR program might help fill the seed and early VC “Valleys of Death.” I’ve interviewed nearly 20 VC, angel, and corporate VC investors, as well as about 10 startup entrepreneurs. They tell me very different things about the signaling value of a DOE SBIR grant, suggesting that my research project could result in interesting, new information about how well DOE grant recipients perform in aggregate.
Global VC Clean tech Investment by Sector:
SBIR grants consist of two phases. The “Angel” Phase 1 grants, which are very competitive (only 2-12% of applicants win), fund proof of concept exercises, and today are typically $150,000. Phase 2 is “VC style” prototype development of up to $1.5 million, and requires that the company be a Phase 1 recipient. . About half of Phase I recipients win in Phase II.
The graphs below showing annual EERE SBIR awards highlight the abnormality of the Stimulus. Before 2009, EERE awarded on average $2.5 million (adjusted for inflation to 2012 $) in Phase 1 SBIR awards each year, an amount that increased by a factor of 10 in 2010. Similarly, Phase 2 awards averaged $11.5 million before 2009, but were $128 million in 2010. Some investors in the space have hypothesized to me that this was far too much money chasing too few good ideas.
The goal of my project is to learn about the impacts of SBIR grants along both innovation and financial dimensions. The basic research problem is that it’s really hard to know what would have happened to a firm or an industry in the absence of the grant program. It’s also likely that in general, the firms selected for a grant are different (hopefully better) than the losers, and also different from firms who don’t apply at all.
My solution is to assume that the runners-up (firms that lost, but were most highly ranked) and the lowest-ranked winners in a competition had the same potential for success ex-ante. This generates the equivalent of a random experiment, where we took a handful of high-quality firms and randomly gave some a grant. I’ll compare in aggregate the winners with the runners-up, while controlling for things like the specific topic of the grant and the year. Economists call this strategy a regression discontinuity design.
The i3 database is critical to my effort, because it helps tell me when awardees and applicants get subsequent private capital and when they enter sales or licensing relationships with other entities. I matched companies in the i3 “Deals” and “Companies” data sets to my EERE SBIR data. First, it is interesting that for this group of companies, data coverage essentially starts around 2000 (there are no matched deals between 1983-1992). This is in part because there weren’t very many clean tech deals before the late 1990s. The graph below shows a timeline of the matched deals with SBIR applicant companies, by deal type.
I observe 9,705 applications, submitted by nearly 4,661 unique firms, of which only 527 ever won. Below, a table provides some statistics about my data and how it interacts with the i3 database. I can match 554 firms in my SBIR applicant data to the i3 database, for which i3 has non-grant financing events (investment, acquisition, IPO, or a relationship) for 260. Amazingly, 19.4% of SBIR winners have private financing in i3, compared to only 3.8% for firms who never won an EERE SBIR grant.
This suggests that in general winning an SBIR is correlated with private capital. In the future, my research will incorporate other financial transaction databases, and will look only at the runners-up in comparison with winners, rather than all applicants. Stay tuned for the results!
Table 1: Summary Statistics using only i3 Data for Private Financing, looking only at Phase 1 DOE EERE SBIR Applicants
Let’s look at some specific companies in my data.
- Ocean waves contain tremendous energy – in theory, EPRI estimates that recoverable wave energy along the U.S. outer continental shelf is about one-third the total power used in the U.S. in terms of terawatt-hours per year. We are just beginning to try to harness this energy, and companies of all sizes are testing diverse mechanisms. One of the startups in this race is Oscilla Power, founded in Utah in 2009. According to i3, “Oscilla Power is developing renewable energy and energy harvesting technologies. The company’s iMEC technology platform allows magnetostrictive alloys to convert mechanical energy into electrical energy. The company’s technology can be applied to utility-scale wave energy generation and downhole power for the oil & gas industry.” The company raised a $1.6 million series A round in late 2011, and won its first EERE SBIR Phase 1 grant that same year. In 2012, Oscilla won a Phase II grant of $1 million. Subsequently, Oscilla raised a $0.76 million series B round from Angels with Attitude in early 2013. In the middle of 2013, Oscilla borrowed $0.94 million in structured debt. Also in 2013, Oscilla won a second EERE SBIR Phase 1 grant. Oscilla is still in the “project development” stage according to i3, and has 5 employees. Their website describes ongoing tests of its relatively inexpensive wave power generation system, and proudly advertises its DOE SBIR grants.
- Renewable energy is often associated with generation devices, like wind turbines. Less sexy but equally necessary is the electrical equipment that manages the power and transforms it into something the grid can handle. Satcon Technology, according to i3’s overview, “designs and delivers power conversion solutions that enable producers of renewable energy to convert clean energy into grid-connected electrical power. It also offers system design services and solutions for management, monitoring, and performance measurement to improve capital investment, and quality and performance over the lifespan of an installation.” Satcon is the largest provider of inverters for PV systems. Founded in 1985, this Massachusetts-based company followed the standard successful VC funding and successful exit path. The table below shows Satcon’s financing evolution. The company received four SBIR grants from EERE before it first got VC funding in 1999. Over time, and coincident with economic downturns, Satcon successfully applied for four more SBIR grants. i3 documents 28 relationships (sales or project development) that Satcon participated in with companies as diverse as Wipro EcoEnergy in India, Q Cells in Germany, and SoCal Edison in California. The company had 246 employees, and was selling globally in early 2012. Unfortunately, Satcon was never really able to achieve profitability. It reported losses for 22 straight quarters as of October 2012, when it filed for bankruptcy. Although the fall in European demand for solar certainly played a role, inverter prices never had the same oversupply problem that PV panel makers faced.
Table 2: Satcon Technology i3 Financial Transactions & DOE EERE SBIR grants
- Cerahelix, based in Maine, is trying to develop a novel filtration membrane to purify water in industrial applications. Industry globally uses three times as much water as is used residentially. Conventional water purification is done by reverse osmosis, which is quite energy-intensive. Cerahelix’s membrane, with pores of less than 1 nanometer, is a patented technology they call helix NFM. They use a DNA template to make a nano-ceramic coating capable of filtering solutions at low pressure and with greater efficiency, in theory reducing the cost of filtering. Cerahelix was founded in 2011 and received a single SBIR grant in 2012. In 2011, the fledging company got $10,000 in seed funding, followed by a further $255,000 in seed funding from the Maine Technology Institute in 2012. In mid-2013, Cerahelix was awarded a $1 million EERE SBIR Phase 2 award. In future posts, I’ll let you know the conclusions of my research on the outcome of government grants to firms like Oscilla, Satcon and Cerahelix.
 Sabrina Howell is a PhD candidate at Harvard University in the PEG Economics program.
 Primarily through the Departments of Defense and Energy, as well as the National Science Foundation.
 Based on the economic literature. For an excellent explanation, see Jaffe, Adam B. “The Importance of “Spillovers” in the Policy Mission of the Advanced Technology Program.” Journal of Technology Transfer Vol. 23 (2): 11–19. For the original argument, see Arrow (1962). “Economic Welfare and the Allocation of Resources for Invention.”
 The individual grant competitions are limited to a very specific topic, like thin film solar, and tend to have only one to a few winners.
 As applicant data can only be published in aggregate, I discuss winners and their successful SBIR applications.