Are SBIR “Mills” Truly a Problem?

Last year’s Small Business Innovation Research (SBIR) reauthorization debate was among the most contentious in the program’s four-decade history. Authorities lapsed for six months before April’s Small Business Innovation and Economic Security Act funded the program through 2031. As CSIS analysis noted during the debate, because the federal government does not produce standardized measures of SBIR performance, “it is easy for critics to cast doubt on the program’s efficacy.”

Nowhere was that clearer than in the debate over SBIR “mills.” Senator Joni Ernst (R-IA), whose INNOVATE Act typified the reform push, argued that firms with Beltway connections and grant-writing expertise “have been able to collect an outsized portion of the funding, with fewer results to show for it,” citing findings that the most frequent winners “were less productive in terms of commercialization, investments, and patents.” While Ernst’s proposed $75 million lifetime cap for grant recipients did not make it into law, similar reforms did, including agency-set annual proposal caps and a new premium award tier tied to matching private funds.

This prioritization of commercialization sits somewhat uneasily with the program’s origins. In 1945, Vannevar Bush, America’s wartime R&D cheif, published Science, the Endless Frontier, which argued that the government should fund the risky, long-term research that private industry would not. SBIR was an attempt to support such innovation outside of government labs and major corporations. The SBIR mill critique inverts this philosophy, treating the ability to attract private investment as the principal test of whether a firm deserves support from a program built to fund work the private sector would not.

Two business models coexist within the program today. Firms on the dual-use route develop commercial technology for both defense and civilian customers and finance that work partly with venture capital (VC). Firms on the defense route build for a single customer—the Pentagon—finance development through contracts, and are judged by the follow-on business they win. The two models are complementary: One imports commercial innovation into defense; the other sustains specialized capability that often has little commercial market. Any measure of efficacy must account for both.

Funding Efficiency Metrics Can Be Reconstructed from Public Records

Whether firms that win large numbers of SBIR contracts create demonstrably valuable innovation, as measured by conversion into follow-on defense contracts, patents, and private investment, sits at the heart of the debate. Conversion here means moving from subsidized awards to winning regular defense business, and it is the program’s own yardstick: SBIR’s statutory purpose is commercialization, and for defense technologies the paying customer is the Department of Defense (DOD). Fortunately, the question is answerable with records the government already publishes. Award histories sit on SBIR.gov. Federal contracts, including a flag identifying SBIR awards, appear on USAspending.gov. Patents and published measures of their significance are open data, and venture-style equity raises appear in Securities and Exchange Commission (SEC) Form D filings. New analysis links these records for all 10,280 firms that have won the more than 100,000 DOD SBIR awards since 1983. The combined dataset allows for the comparison of the 69 firms with 50 or more Phase II advanced-development awards (so-called SBIR mills) against awardees with three to nine awards (principally VC-backed firms).

Three Misconceptions, Measured

Critics have argued that SBIR mills exploit permissive size rules and bureaucratic expertise to capture disproportionate funding, crowd out small innovators, and weaken technology transitions. Others have claimed the program has become a subsidy for grant-dependent research houses rather than a launchpad for innovation. Such arguments have spawned proposals to impose award caps, first-time-recipient requirements, and commercialization-focused funding. This drive to restructure the program, however, is informed by three structural misperceptions regarding the SBIR program:

1. Mills win awards but rarely “demonstrate commercial traction or follow-on contracts with non-SBIR dollars.” By that standard, however, the heaviest winners actually lead the program. Within five years of its first SBIR award, 71 percent of the 50-plus award group won at least one DOD contract of $50,000 or more outside the SBIR program, compared with 44 percent of the 3–9 award group (Figure 1). Over its time in the program, the median “mill” has won roughly 50 cents in non-SBIR DOD contract dollars for every SBIR dollar received, compared with 16 cents for the median firm in the 3–9 group. Nine in 10 of the heaviest winners have won production contracts at some point, compared with fewer than half of the 3–9 group. Reaching procurement is the norm for these firms, not the exception. Essentially, the firms that critics call mills are more likely than other awardees to become substantial defense contractors themselves, but they cross the “valley of death” through follow-on procurement rather than private investments. The public records show such transitions as contract activity; no dataset currently captures development into programs of record.