Here’s a table from the Harbridge House report, c. 1968. I’ve marked on it to call attention to some figures. First, when a contractor has experience and owns an invention, the commercial use rate is over 20%. Universities, however, as owners, fare worse–they are in the category at 6.6%. Under the IPA program from 1968 to 1978, universities were at 4.2%–4 out of 96 subject inventions.
This table reflects the Kennedy patent policy: if a contractor has a commercial position and capability, let the contractor hold title. Everyone else has to to make a pitch that they can do something, because the figures don’t indicate they generally don’t do so well. The Kennedy approach makes some sense, on the evidence.
Now here’s the thing. A university will aim to license patent rights to a company (almost always exclusively–though that’s just a choice university folks make). There are two choices. The university can license to a contractor with experience, or one without (such as a university startup). Let’s look at the Harbridge House numbers. If to a contractor with experience, the rate doubles–13.3%. But it’s still about half of what companies do if they also own. If to a contractor without experience, the rate drops by more than half.
The Bayh-Dole approach–especially in its faux version of mandated university ownership and exclusive licensing–at best operates at less than 50% the rate of contractors with experience and ownership, and for university startups, less than 10% of the rate for contractors with experience and ownership. If the Harbridge House figures are any guide, the worse possible thing we could do with federally supported inventions is have universities license them to startups without prior experience.
Why would anyone think Bayh-Dole is a really good way to do things? You’d think that the federal government should require universities to assign inventions to companies rather than forbid it. Giving companies ownership of inventions doubles their use rate. Or better, why not create an “experienced company vesting statute” that out-Bayh-Doles Bayh-Dole? Why should universities have any ownership in inventions, just to re-license them with overhead and bitterness and delay to a company that might not have any capability? Letting universities mess around this way adds costs to the system. If folks are determined to strip university inventors of their personal property rights in inventions and leave title to inventions to wiggle in the air until a bureaucratic decision is made, why not just hang the titles to subject inventions out there until one or more capable companies requests ownership? Vest title in the capable company directly. Screw all this university “marketing” and “licensing” complexity. If no company shows up in one year, initiate a make-use commons with option to sell. If no company shows up in five years, don’t pay the maintenance fee and let the patent lapse.
Have university inventors assign to whomever shows up with a statement of capability and a plan of use. Once one company has shown up, if within thirty days no other company shows up, then the first company gets assignment if it has presented a case for capability. If another or more companies with capability also request ownership, then they all vote on whether they require sole ownership or can get by with joint ownership. Vote decides. A tie means joint ownership–but all companies must commit to use of the invention. If the companies demand sole ownership, then it’s a competitive bidding war. They voted for that, so put them up to it.
For joint ownership, the first company manages patenting in consultation with the other joint owners for matters of claims and everyone pays their share of the costs. If no other companies show up in five years, then the companies with ownership have a monopoly for the life of the patent. Otherwise, the monopoly is reduced to eight years from the date the next company in requests access. Or something like that. If the companies stop using the invention for a year, the patent lapses. If a company sells or licenses its ownership position, then it must share 5% of the gross income from the sale with the university inventors and must pass the obligation down to the new owner or licensee.
Once we are down the vesting road–whether outright vesting or what NIST imagines it can do by requiring contractors to make their inventors assign all inventions to the contractors–we may as well get universities out of the loop altogether and make university inventors assign to capable companies. Get the university patent brokers out of the loop–their presence cuts commercial use rates in half or more.
According to the Harbridge House figures,
A university exclusive license means: cut the use rate in half.
A university exclusive license to a startup means: cut the use rate by 10x.
Why would anyone do this?
In the alternative–even better–let university principal investigators decide, and if they don’t make any ground rules as a condition of collaboration, then let the inventors decide. PIs and inventors can’t possibly do any worse than the university patent brokers–and they don’t create nearly so much overhead expense.
What the Harbridge House report doesn’t consider, of course, is the use rates where there are clear publications but no patents, or if patents then in a patent commons. We don’t know those use rates–but it may be that those rates are even better than that for patented inventions.