A note on the statistic that built technology transfer

Technology transfer has a founding number. Fewer than 5% of 28,000 federal patents had been licensed. It was used to argue for the Bayh-Dole Act, and forty-five years later it still turns up in conference talks, policy papers and board decks. I have used it myself.

This year I went looking for where it came from.

What the trail actually shows

The most cited version of the figure sits in a report by the United States Government Accountability Office, GAO/RCED-98-126, dated 7 May 1998. That is eighteen years after the law it explains. The report states the number and gives no source for it.

Going backwards from there, the earliest use I could verify in print is a New York Times piece from 15 September 1980. Between that and today the figure drifts. The denominator appears as 26,000, as 28,000 and as 30,000. The share appears as 4% and as 5%. A number that moves like that is not being checked by the people repeating it, and I include myself in that.

Underneath the citations sits a study by Harbridge House for the Federal Council for Science and Technology, which looked at government-owned inventions patented in two years, 1957 and 1962.

That sample matters. Rebecca Eisenberg showed it was made up largely of Department of Defense patents that firms had already declined to take title to. What stayed in government hands was, in her words, presumably what had the least commercial potential. The statistic measured the licensing rate of inventions industry had already looked at and passed on, and it was then presented as the licensing rate of federally funded research in general.

The case that should not have worked

While that number was circulating, the most commercially successful university patent of its era was already running. Cohen-Boyer was filed on 4 November 1974, six years before the Act. It was licensed non-exclusively to 468 companies. It returned about US$255 million on more than US$35 billion of product sales.

Non-exclusive, pre-Act, and enormously successful. By the logic of the statistic used to justify the Act, it should not have happened.

The wider trend points the same way. Elizabeth Popp Berman counts roughly 100 university patents in 1968 rising to roughly 350 by 1980, with 37 universities holding Institutional Patent Agreements by 1971. University patenting was already climbing before the law arrived.

Why I am writing this down

I am not arguing that Bayh-Dole was a mistake. The Act did things the statistic was never needed to justify, and I run an office that exists inside the system it created.

My argument is closer to home. We inherited a number nobody audited, and with it we inherited its habit. Every technology transfer office I know reports disclosures, patents, licences and spin-outs. Not one of those denominators was chosen on purpose. They are what happened to be countable in 1980.

We tell researchers that evidence has to hold up. We would not accept a licensee's market estimate sourced the way this figure is sourced. The discipline we ask for should apply to the numbers we report about ourselves, and the first step is knowing which of ours would survive the same walk backwards.

What I could not establish

I could not find who first stated the statistic, or in which hearing. Several accounts attribute it to testimony in the late 1970s. I could not verify that at primary level, so I am leaving it open rather than guessing. If you can source it, I would like to see it.

---

Sources

• United States Government Accountability Office, Technology Transfer: Administration of the Bayh-Dole Act by Research Universities, GAO/RCED-98-126, 7 May 1998. Checked directly.
• Rebecca S. Eisenberg, Public Research and Private Development: Patents and Technology Transfer in Government-Sponsored Research, Virginia Law Review 82(8), 1996. Quoted and extended in So, Sampat, Rai, Cook-Deegan, Reichman, Weissman and Kapczynski, Is Bayh-Dole Good for Developing Countries?, PLoS Biology 6(10):e262, 2008.
• Elizabeth Popp Berman, Why Did Universities Start Patenting?, Social Studies of Science 38(6), 2008.
• Rajendra K. Bera, The story of the Cohen-Boyer patents, Current Science 96(6), 25 March 2009.
• Harbridge House, Government Patent Policy Study, for the Federal Council for Science and Technology.