This is a great summary and interesting research. As a small firm that has been a recipient of a few SBIR grants in the past (both Phase I and Phase II), I'd like to outline a few thoughts on the relevance of patents as a metric for innovation.
The federal agencies persistently ask about patent filings as a way to measure long term innovation impact. But they also ask about revenue generated from the product, additional investment made by the small business in the product, 3rd party investments in the firm, employee count, asset sales (product was developed by the small business and then sold to another firm to develop further), and other metrics. Many of the agencies also ask for stories about the impact of the funding on developing new products and services. I think this attempt to gather a broad set of data about impacts is important, and, at least for us, patents wouldn't tell the story.
There is a lot of innovation for which a patent application is a poor match, and there are some firms that may prefer not to file a patent for a variety of reasons. As a software company, we have generally avoided filing for patents for a few reasons: 1) proving there is no prior art for a software invention has become more difficult over time (I think this is generally a positive change in the software ecosystem); 2) we have been advised by our IP attorneys that if we are going to file for a patent, we need to have the financial resources to defend the patent. Defending a patent in the United States is an extraordinarily expensive endeavor, and as a small privately held firm, we don't have the resources. Rather, we have put our resources into engineering, design, sales, and marketing.
Second, I think there is another nuance in the SBIR program that bears consideration. The federal agencies appear to genuinely and sincerely think of the SBIR program as an investment. However, they consider a broader definition of return on investment - similar, perhaps, to a social impact investor that may consider a continuum of returns that include financial as well community and societal impact. The private sector tends to be concentrated in a few technology domains that have a chance at disproportionate returns on capital within a limited period of time (5-10 years max). But the economy is broader than that, and there are lots of other potential innovations that would make our economy and society stronger and healthier that would not be seen as attractive to most private sector investors. These may include innovations in services, ecosystem management, reduction in pollution (often a market externality), and inventions with long timelines that exceed the patience or risk that exceeds the appetite of a private sector investor. In short even if it didn't return as many patents as other investment approaches, I think the SBIR program remains a powerful mechanism for the federal government to encourage an innovative and dynamic economy.
Finally, I think it's helpful to note that the SBIR program is a tiny program carved out of very large federal extramural research budgets that go predominantly to universities and other large institutions. It might be helpful to compare the return on investment from small firms with the 2.5% or so in SBIR budgets versus the return on investment from the other 97.5% that goes to large institutions. I suspect the relatively small investment in small firms would turn out to have outsized impacts when compared to the total extramural federal research budget.
Thanks for sharing. I'm somewhat reassured that SBIR isn't 100% focused on seeing patents - both for the reasons you say, as well as for the validity of my assumptions in the post. There's a report here that discusses some of the kinds of metrics you mention above (https://www.nap.edu/read/25674/chapter/1). The challenge is when this data is collected for grantees but not all applicants it gets harder to use for evaluation of the program, since you don't have the same level of data for people who didn't get the grant.
This is also a challenge in comparing SBIR to other R&D programs. I hope to cover some attempts to measure the returns to different kinds of R&D programs, but to be honest much of the evidence on other programs isn't up to the level of the evidence in this post. The SBIR is an unusually well studied program both because you can do these very nice quasi-experimental designs and because you have a decent idea what kind of thing to look for. Much harder to know where to look for the evidence of the efficacy of funding for, say, economics research!
In general, patents are a natural metric to apply to research - it is a poor one in my experience. Getting a patent is primarily a function of ones budget and knowing all the avenues of getting around a rejection. How much are you willing to spend. Many successful companies had a rather small patent portfolio until they attracted patent law suits and then generated or acquired patents. Know-how and trade secrets are probably more valuable (drugs IP is an exception). SBIRs have been around for awhile would be more interesting to look at add-on funding, revenue as metrics for how effective SBIRs are at picking/enabling winners. Similar to a VCs portfolio 10-year return.
Thank you for this informative piece! What I'm struggling to get is why governments are not offering an exit option for startups. Governments give capital, research labs and other investments in the critical early phases but don't get the financial return that a VC would when they invest a 10th of that at later stages.
Regarding SBIR in the US specifically - each agency tailors their solicitations to their priorities (eg, NSF runs SBIR very differently than how DoD runs it), but most of the $ across the government goes to meeting technical/scientific/engineering issues. There are SBIR cases where firms winning awards have gone on to massive success (Qualcomm is probably the biggest), but most SBIR awards serve 2 primary purposes: addressing a specific & solicited need, & cultivating the science base. The agencies, for the most part, aren't as focused on exits/returns as a VC would be. & that's part of how/why the program has been so successful.
Interesting. I haven't thought about this as much, but my first reaction is that because the government by default has upside exposure to any and all companies via taxation it doesn't have as much need for a vehicle to get equity in companies. In general, better to do whatever it takes to promote overall growth, including giving out grants that don't require firms to give up equity.
But as your post says, there could be rationales for companies where the government has an interest in promoting various goals that aren't incentivized by the market. I think it would be pretty context dependent - only a good idea when you have both a good rationale and a highly competent government org to run it.
Agree with you. There are currently two areas where acquisition would make sense and which are not mentioned in the piece. First the "hard" technologies that are deemed critical to the system e.g. many governments talk about grid control (and are already under state control) as well as cyber security.
The second - and more interesting - is the "soft" tech which affects society differently. E.g. recommendation tech, AI, cloud storage and others (quantum, EV...etc) that are "privately owned - public utility companies" (ie. the same rite de passage as energy/gas companies). The whole line of thought was kicked off/inspired by observing the TikTok/US acquisition debate.
In any case, thank you for your feedback above! Still formulating the argument/thoughts about this and am grateful if you have more feedback for an upcoming piece I'm writing.
Great piece, learned a lot!
This is a great summary and interesting research. As a small firm that has been a recipient of a few SBIR grants in the past (both Phase I and Phase II), I'd like to outline a few thoughts on the relevance of patents as a metric for innovation.
The federal agencies persistently ask about patent filings as a way to measure long term innovation impact. But they also ask about revenue generated from the product, additional investment made by the small business in the product, 3rd party investments in the firm, employee count, asset sales (product was developed by the small business and then sold to another firm to develop further), and other metrics. Many of the agencies also ask for stories about the impact of the funding on developing new products and services. I think this attempt to gather a broad set of data about impacts is important, and, at least for us, patents wouldn't tell the story.
There is a lot of innovation for which a patent application is a poor match, and there are some firms that may prefer not to file a patent for a variety of reasons. As a software company, we have generally avoided filing for patents for a few reasons: 1) proving there is no prior art for a software invention has become more difficult over time (I think this is generally a positive change in the software ecosystem); 2) we have been advised by our IP attorneys that if we are going to file for a patent, we need to have the financial resources to defend the patent. Defending a patent in the United States is an extraordinarily expensive endeavor, and as a small privately held firm, we don't have the resources. Rather, we have put our resources into engineering, design, sales, and marketing.
Second, I think there is another nuance in the SBIR program that bears consideration. The federal agencies appear to genuinely and sincerely think of the SBIR program as an investment. However, they consider a broader definition of return on investment - similar, perhaps, to a social impact investor that may consider a continuum of returns that include financial as well community and societal impact. The private sector tends to be concentrated in a few technology domains that have a chance at disproportionate returns on capital within a limited period of time (5-10 years max). But the economy is broader than that, and there are lots of other potential innovations that would make our economy and society stronger and healthier that would not be seen as attractive to most private sector investors. These may include innovations in services, ecosystem management, reduction in pollution (often a market externality), and inventions with long timelines that exceed the patience or risk that exceeds the appetite of a private sector investor. In short even if it didn't return as many patents as other investment approaches, I think the SBIR program remains a powerful mechanism for the federal government to encourage an innovative and dynamic economy.
Finally, I think it's helpful to note that the SBIR program is a tiny program carved out of very large federal extramural research budgets that go predominantly to universities and other large institutions. It might be helpful to compare the return on investment from small firms with the 2.5% or so in SBIR budgets versus the return on investment from the other 97.5% that goes to large institutions. I suspect the relatively small investment in small firms would turn out to have outsized impacts when compared to the total extramural federal research budget.
Thanks for sharing. I'm somewhat reassured that SBIR isn't 100% focused on seeing patents - both for the reasons you say, as well as for the validity of my assumptions in the post. There's a report here that discusses some of the kinds of metrics you mention above (https://www.nap.edu/read/25674/chapter/1). The challenge is when this data is collected for grantees but not all applicants it gets harder to use for evaluation of the program, since you don't have the same level of data for people who didn't get the grant.
This is also a challenge in comparing SBIR to other R&D programs. I hope to cover some attempts to measure the returns to different kinds of R&D programs, but to be honest much of the evidence on other programs isn't up to the level of the evidence in this post. The SBIR is an unusually well studied program both because you can do these very nice quasi-experimental designs and because you have a decent idea what kind of thing to look for. Much harder to know where to look for the evidence of the efficacy of funding for, say, economics research!
In general, patents are a natural metric to apply to research - it is a poor one in my experience. Getting a patent is primarily a function of ones budget and knowing all the avenues of getting around a rejection. How much are you willing to spend. Many successful companies had a rather small patent portfolio until they attracted patent law suits and then generated or acquired patents. Know-how and trade secrets are probably more valuable (drugs IP is an exception). SBIRs have been around for awhile would be more interesting to look at add-on funding, revenue as metrics for how effective SBIRs are at picking/enabling winners. Similar to a VCs portfolio 10-year return.
"High return" - relative to what?
Thank you for this informative piece! What I'm struggling to get is why governments are not offering an exit option for startups. Governments give capital, research labs and other investments in the critical early phases but don't get the financial return that a VC would when they invest a 10th of that at later stages.
Here's how it looks in the EU for example: https://shiftprint.writeas.com/european-technology-acquisition
Regarding SBIR in the US specifically - each agency tailors their solicitations to their priorities (eg, NSF runs SBIR very differently than how DoD runs it), but most of the $ across the government goes to meeting technical/scientific/engineering issues. There are SBIR cases where firms winning awards have gone on to massive success (Qualcomm is probably the biggest), but most SBIR awards serve 2 primary purposes: addressing a specific & solicited need, & cultivating the science base. The agencies, for the most part, aren't as focused on exits/returns as a VC would be. & that's part of how/why the program has been so successful.
What about public-private partnerships?
What's your thought on this "government-exit strategy/method"?
Interesting. I haven't thought about this as much, but my first reaction is that because the government by default has upside exposure to any and all companies via taxation it doesn't have as much need for a vehicle to get equity in companies. In general, better to do whatever it takes to promote overall growth, including giving out grants that don't require firms to give up equity.
But as your post says, there could be rationales for companies where the government has an interest in promoting various goals that aren't incentivized by the market. I think it would be pretty context dependent - only a good idea when you have both a good rationale and a highly competent government org to run it.
Agree with you. There are currently two areas where acquisition would make sense and which are not mentioned in the piece. First the "hard" technologies that are deemed critical to the system e.g. many governments talk about grid control (and are already under state control) as well as cyber security.
The second - and more interesting - is the "soft" tech which affects society differently. E.g. recommendation tech, AI, cloud storage and others (quantum, EV...etc) that are "privately owned - public utility companies" (ie. the same rite de passage as energy/gas companies). The whole line of thought was kicked off/inspired by observing the TikTok/US acquisition debate.
In any case, thank you for your feedback above! Still formulating the argument/thoughts about this and am grateful if you have more feedback for an upcoming piece I'm writing.