ideas-bayeDoleCompulsoryLicensing

The Baye-Dole act says that universities can own the IP from grant-funded research. Perhaps the intent was to encourage universities to form closer partnerships with industry, doing more "tech transfer", but the result seems to be the opposite. I've heard from various different sources that people who want to form startups based on their own research have trouble because the universities want such a big cut that it scares away investors (see for example http://groups.google.com/group/lean-startup-circle/browse_frm/thread/6132b73cdb2e755e?hl=en&pli=1 , but i've heard the same thing from other sources). I don't know whether this is because the university IP office people are incompetent negotiators (who end up letting deals fall through rather than accept a small cut of a barely profitable business), or, what seems more likely to me, if the university is rationally maximizing its profit by accepting a few big deals rather than many small ones. For whatever reason, the Baye-Dole act is actually reducing innovation by killing startups that would otherwise have existed.

Also, I've heard that many university IP offices actually lose money! That is, that the yearly cost of the running a university IP office and paying the people who work there is often greater than the amount of money that it takes in per year from deals.

So, it seems like the obvious thing to do is repeal Baye-Dole and make the results of all government-grant-funded research public domain. But in case vested interests prevent that from happening, I propose amending Baye-Dole to add the following requirements:

• at the sole option of the people who actually worked on the particular grant (professors, students, postdocs, etc), the IP may be placed in the public domain at any time
• unless an exclusive license has been granted, anyone may obtain a non-exclusive license to grant IP for no more than .2*(the value of the grant that generated the IP).
• the maximum price of an exclusive license to grant IP is .5*(the value of the grant that generated the IP). Companies in which the people who worked on the project have a "large" stake get priority.
• the maximum price is additionally discounted by 50% if the licensee is 100% owned by researchers who participated in the project. If the licencee is only partly owned by these researchers, then the license fee is discounted by a proportional amount (that is, the discount is .5*(ownership fraction of researchers)

Of course, the university may offer the IP on other terms as well (in exchange for a stake or a cut), as long as it is possible for licensees to choose to pay cash according to the above terms instead.

This will ensure that the results of grant-funded research is not prevented from being commercially applied by overly large university licensing fees, and will give researchers a particularly strong incentive to get involved in the commercialization of their own work. So, the commercialization of research will be encouraged rather than penalized.

I suggest that universities implement these restrictions themselves even if the law is not changed. It's better for society, and a "startup-friendly" atmosphere will attract talent (and eventually lead to rich alums). I suggest that individuals decline to donate to universities that don't have some policy along these lines.