Thank you Mr Chairman for the chance to present a statement about how the Congress can and should decide whether the Small Business Innovation Research program is meeting a sufficient performance standard to continue as it has for 15 years.
My qualifications for raising these questions are that I ran the SBIR/STTR programs in the Strategic Defense Initiative and its successor, the Ballistic Missile Defense Organization, from 1987 to 1996, and I have been a private consultant on SBIR to small high-tech companies since April 1996. I have published several papers on the subject in professional society publications, have participated in a National Research Council workshop on the subject, and have debated the subject endlessly in the SBIR community. On the non-SBIR side of R&D, before 1987, I served in the Department of Defense R&D in several positions from 1963 to 1987.
The first question for SBIR is whether the economic gain to the nation meets any rational investment standard? Did the program do anything that would have not happened anyway if the federal agencies had been left to manage their R&D programs without a mandated set-aside? Did the economic results from the companies funded with SBIR exceed by a decent margin the economic results of similar companies that had never heard of SBIR? In a nation that leads the world in innovation by a wide margin, the government should have a fear of doing more harm than good by pushing federal monies to companies whose technology does not seem to compete well in the open development market.
One of the many objectives of the SBIR is an economic contribution to national competitiveness. If such a contribution, large enough to justify government intervention in the market, has been made, it should be obvious after 15 years. A study of SBIR's economic return would ask what the return on investment has been to the national economy. If it does not meet some rational investment standard, it has failed its economic objective. Although there will be much debate over which standard(s) would work best, there should be no debate that some standard should be used that economists and investors can agree measures economic return. As a minimum such a study requires comparison with an external control group of companies with no SBIR as recommended by the work of Josh Lerner at Harvard.
An argument is often made that SBIR has enabled small companies to do innovative R&D. Which is true! And that R&D so performed has a societal benefit. Which is also true! But doing good is merely a necessary, not a sufficient, condition. A special program that micro-manages federal R&D must also prove to be much better than what federal R&D would have done anyway. The conditions that prompted Congress to start SBIR in 1982 have mostly disappeared and American small business is more competitive in the global marketplace than the small business of any other nation and better than many large American businesses. Venturesome capital is flowing to small company innovators in record amounts.
SBIR's advocates argue that SBIR enables more small business R&D. A warm but misguided argument. Federal R&D was sending 2.7% of its extra-mural R&D to small business before SBIR was ever invented and after 15 years that percent is essentially the same. Which means that no new money has gone to small business and that every SBIR dollar going to small business has been taken from other small R&D businesses. To what end? That is called a zero-sum game which means that more small business R&D is a mirage.
But that zero-sum game is not necessarily bad. If the transfer caused investment to shift from poor economic performers to the best economic performers, the program has succeeded. If not, the program has failed. If there is a net economic benefit, the small business community and the nation are winners. Whether you believe that such winning justifies government intervention in the world's best innovation markets is a purely political matter.
Measurement of economic results is not without its problems though. Innovation, especially disruptive innovation, often takes a long time to appear in volume markets. But since SBIR has been going for 15 years, even the long delay benefits should be showing up by now in economic data of some kind. Many economic benefits will appear in other companies than the one funded and accountability will be hard to trace. Return on investment will be hard to determine because there will be disputes over what constitutes a return. Clearly, to me, a measure of dollars of commercial sales by the performing company per dollar of SBIR carries little value as a measure of investment profit. If however, no better measure can be acceptably devised, some standard of a minimum average sales ratio should be established that would be a proxy for a competitive economic return to R&D investment. If, for example, an investor wanted a typical 30% return on investment and the products would yield a competitive profit margin of hypothetically 30%, commercial sales would have to be at least many multiples of SBIR investment, not multiples of about 1:1 as GAO is reporting for SBIR. Normal capital investors use either an income stream (not revenues) or an increased equity value.
Collecting such data systematically will be difficult. But not impossible. Government's collecting private business data from firms intrudes on commercial confidentiality in ways that may threaten the company's competitive position or assail a company's concept of protection from government. The National Research Council has already wrestled with the question and could be a valuable contributor to finding a solution.
If the program has failed its economic objective, and Congress insists that it meet an economic objective, Congress will have to decide whether to mend it or end it. Mending it will not be pleasant for the federal agencies. They would have to radically change their approach to picking SBIR winners and to apply economic standards to that portion of their R&D investment. Until now, most agencies have merely turned SBIR into their small business version of what they would have done anyway. They have shunted small business from one R&D funding window to another.
A deeper problem is the agencies have no incentive to devote 2.5% of their R&D program to general national economic benefit. Their Congressional oversight committees would not approve of such diversion were it a direct appropriation and no agency head has any economic interest in economic success. No stock options and early retirement for those bureaucrats! The constant turnover of agency heads within and across administrations further compounds the difficulty of giving them a strong interest in SBIR's success.
Anyway, we don't want government departments to act as businesses; we want them to pursue government objectives with good government methods. They could, however, take the view that Ballistic Missile Defense Organization has taken for many years that technological innovation with economic potential will efficiently produce new technology for the agency through the private sector. Then they would want to pursue the best economic innovations and SBIR's economic results would follow easily. If Congress wants economic gain, it will not only have to say so clearly in the SBIR legislation, it will have to convince agency heads that it is in their interest. A good economic evaluation scheme would prod them find the best ways to make competitive investments but would not in itself micro-manage their programs
The agencies could change their approach to SBIR by using productivity measures for its SBIR managers. Although productivity measures do not work for most government functions, SBIR would yield to such measurement. There are some easily recognized methods to do so; the mutual fund industry, for example, has a well-tested approach to portfolio management. Whatever they do, the results should produce economic returns that can be measured and compared with not having any such program at all.
If SBIR is worthy of a set-aside of federal R&D, an economic analysis will reveal its value. If not, then the Congress should consider returning management of the federal R&D to the agencies.
Thank you for the chance to present this statement. I would be happy to answer questions and participate with the Committee staff in finding a good evaluation scheme for SBIR.