Bayh-Dole and worldwide IP metastasization

Tags: #<Tag:0x00007f7cc21381e8> #<Tag:0x00007f7cc213ff60> #<Tag:0x00007f7cc213fd80> #<Tag:0x00007f7cc213fa38>

Reconsidering the Bayh-Dole Act and the Current University Invention Ownership Model (working paper pdf; 2-column author self-archive published version; summary) by Martin Kenney and Donald Patton is from 2009, but recently highlighted on the P2P Foundation blog.

Bayh-Dole is a 1980 U.S. law that gave recipients of federal funds (mostly universities) ownership of funded research. The bulk of the paper is an analysis of the “[University Technology Licensing Office]-inventor relationship describing the contradictory goals, information asymmetries, and perverse incentives resulting from the university ownership model”, which is interesting enough on its own. But what I want to note here is the story the authors tell, scattered throughout the paper, of the interests that lobbied for Bayh-Dole and its worldwide metastasization, and their two proposals, one of which is commons-favoring.

Brief dis-contiguous excerpts on interests. I’ve stripped the references for readability here, but please go read the whole paper to see them and the rest of its context:

Based on minimal evidence, it was believed that government-owned patents were insufficiently utilized.

The popular press and others hailed the model of university ownership of these inventions as a boon to society.

a process of global organization mimesis is underway, as other nations adopt B-D-like models patterned in the belief that it is the best way to ensure the commercialization of university inventions.

B-D was the result of lobbying efforts by interested parties – in this case, corporations and university licensing officials hoping to monetize these inventions. For the universities, the desire to appropriate the fruits of their employees‟ federally-funded research was undoubtedly fueled by the emergence of the biotechnology industry, whose promise of riches to invention owners culminated with the spectacular 1980 Genentech initial public offering.

In many respects B-D was a formalization of an extant movement, but it 7 served to alert faculty and administrators still operating under the previous more Mertonian social ethos that conditions were changing, and it was socially desirable for university researchers to patent inventions. With visions of a new income source, universities that did not have a TLO soon established one.

There is no evidence that technology transfer is the primary goal of university TLOs. In many cases, TLOs are unnecessary for the commercialization of university inventions. Further, at least, some TLOs have become so focused on the pursuit of licensing revenues that they are willing to tax the research enterprise itself. The evidence, as we will show, suggests that the only transfers most TLOs are interested in are ones from which they can extract licensing fees. To do this they are willing to impede technology “transfer” to unlicensed users.

For administrators TLO income is attractive because the funds, though meant to be used for further research, are, in fact, largely unencumbered, giving administrators wide discretion on how to spend them. Often the support monies for TLO personnel can originate from public funds, either federal or state. This asymmetry offers a powerful incentive – restricted funds can be spent to operate the TLO, while earnings are far less restricted. The strength of this incentive is difficult to measure, but it may be considerable as more flexible funds are invariably in short supply.

As a bureaucratic entity, the TLO may also be the victim of university political decisions as decisions to patent may not be made purely on merit. To illustrate, if an invention is not patented and marketed, inventors may threaten to leave, taking their 21 laboratory and grants with them. Resignation by professors with large federal grants results in the loss of significant overhead income. In an effort to retain faculty members who attract large grants, the TLO‟s superiors may demand favorable but inefficient decisions.

TLOs and their personnel are, for good reason, often measured in terms of revenue. For these TLOs, the emphasis naturally shifts to extraction of the greatest amount of income.

There are many perverse incentives. Larger, more successful TLOs can have a longer-term perspective and to maximize return may use patent troll-like strategies such as pursuing “submarine” patents. An excellent example is Columbia University‟s secret efforts to extend the Axel transformation patents through a clever strategy of asking for Patent Office continuations, and then getting the patent issued immediately prior to the initial patent expiration . This strategy suggests that the primary goal of Columbia’s TLO is not to transfer technology but to maximize revenue. Efforts to extend the patents, by whatever means necessary, are entirely logical when one considers that Columbia faced losing annual revenues exceeding $50 million. With such revenues, contributions to knowledge transfer and social welfare are deemed not as important, though they may be by-products.

many university TLOs are risk-averse bureaucracies focused on short-term revenue maximization. Oddly enough, knowing this, recently there have been recommendations that TLO managers be provided with incentive pay, which can only to lead to increased short-termism with high up-front licensing fees discouraging entrepreneurship and encouraging greater aggressiveness on the part of the TLO personnel in encouraging/demanding disclosure.

Frustrating entrepreneurs through difficult financial and contractual demands is likely to be so costly in terms of future donations that it far outweighs the gains from licensing. As with any organizational unit, TLOs are boundedly rational. They may pursue their office’s interest to the detriment of the university’s interests.

From an economic perspective, the inventor’s position is curious. First, university employees are legally obligated to disclose inventions. And yet, enforcement is difficult. Link et al. recognize this difficulty, and they conclude that “it also seems prudent for universities that place a high priority on formal technology transfer to place a higher value on patenting, licensing, and start-up formation in promotion and tenure decisions". These authors are suggesting a wholesale transformation of the university incentive structure to suit the needs of the flawed current TLO model. In other words, research, teaching, and contributions to the general societal knowledge pool would vie with patent-generation and firm-formation as university goals. They would raise a minor appendage of the university and an insignificant source of funds to a central goal of the university – and this without any evidence that such a radical change in policy would contribute either socially desirable effects or generate greater economic activity. This structure is a recipe for transforming first-tier research universities into a hybrid of corporate contractor and small firm incubator.

Organizational arrangements are outcomes of social and political choices, but most arrangements soon develop a facticity and aura of normalness that discourages critical evaluation. Today, the university ownership model is framed as the natural method for organizing the interface between university inventions and inventors and the economic realm.

It is so “natural” that, with a few exceptions, academic research on university spin-offs uses the data provided by TLOs, which almost always omits entrepreneurial firms not using the TLO.

One area of interest is how key technology licensing personnel and their professional organization, the Association of University Technology Managers (AUTM), shaped the current system through political action and operational decisions that framed what has now become an institutionalized model. AUTM and TLO employees from various universities, though stating that they do not represent their universities, appear to be speaking for the university when they testify and lobby for an ever more restrictive patent model, which Jaffe and Lerner criticize as inhibiting innovation and commerce. This desire for an ever more restrictive patent regime is the direct result of the university‟s status as a non-profit organization, which no research university would like to relinquish, and the university‟s structural inability to commercialize inventions directly. For this reason, university TLOs find their economic interests aligned with and, in some cases, operate as “patent trolls,” even while the university is dedicated to and predicated upon free flows of information.42 Such archival research could help answer the question of how such a small unit within universities developed such salience in defining the way in which commercial knowledge is transmitted to the greater economy.

This article examined the foundations of the university ownership model and built the case that the model itself is fundamentally flawed. We proposed two quite different new foundations for handling university inventions. In part, we reacted to the triumphalism with which the university ownership B-D model has spread globally. Apparently, some policy-makers now subscribe to a cargo cult-like belief that passing new regulations mimicking U.S. models will deliver entrepreneurship and new “Silicon Valleys,” even to the point that the most successful university technology-based entrepreneurial region in the world outside the United States― Cambridge, England―abandons its successful model. Such beliefs are epitomized in an OECD paper finding that “one of the most urgent tasks is still to raise awareness of and support for university patenting and related activities.”

Concentrated paragraphs on the open alternative:

Another model that has received interest is a mandation that all inventions generated through federal support are de jure placed in the public domain or, less radically, only licensed on a non-exclusive basis (Eisenberg 1996; Nelson 2003; Rai 2005). Since non-exclusive licensing is a “tax,” and shifts the invention rents from one actor to another, an open source model would go further by socializing the value of the inventions (Rhoten and Powell 2007). For basic process innovations, even in biology, an “open” strategy is as effective as, or even more effective than, either exclusive or nonexclusive licensing in encouraging technological transfer and progress. In many engineering-based technologies, patents are not normally considered to be of great significance except to ensure cross licensing (Cohen et al. 2000; Mansfield 1986). The greatest concern in a non-patenting model would be for proprietary pharmaceutical compounds that might not be developed absent exclusive patent protection (Levin et al. 1987; Mansfield et al. 1981). Even here there have been alternatives. For example, there were no patents on the anti-cancer drug Taxol, and yet it was successfully commercialized (U.S. General Accounting Office 2003). The case of penicillin suggests that the Taxol case might not be as exceptional as many believe (Kingston 2001; Neushul 1993).

It is possible that, absent patent rights, small biotechnology firms might not be able to compete with the large established pharmaceutical firms having many complementary assets, thereby limiting entrepreneurial startups based upon university biological science. Possibly, a small firm could be established to commercialize university findings, and, as it operates, create commercially valuable proprietary knowledge. Alternatively, once the knowledge is in the commons, entrepreneurs may capitalize on the knowledge and form yet other firms. This dynamic operated in the open source software field as cost of entry is very low. Whether the IT model would work equally well in pharmaceuticals is uncertain.

For the university, open source would ameliorate current concerns about commercialization influencing its mission or unduly influencing faculty. In many cases, it would lower the cost and uncertainty of using new university-developed technologies and thereby accelerate their adoption. Though a radical response to the difficulties of the current model, considering an open source model provides an alternative reference point for considering other ownership models. Thus, using an open source model may lead to better solutions compared with using the “perfect world” assumptions underlying most economic models.

In TLOs Bayh-Dole has created another hard-core concentrated interest for intellectual property. Putting them out of business (so to speak) surely would be a good thing, even by means of the authors’ non-commons alternative: researcher/inventor ownership, for these constitute a more diffuse interest with more diverse objectives, not only revenue maximization. But the bad public understanding, that more knowledge ownership would be a boon to the economy, preceded the new hard-core IP interest.

The public understanding needs to change: treating knowledge as a commons is necessary to obtain a good future. Also, concentrated interest groups for knowledge commons need to be created. Perhaps the ideal research funding reform would proceed from the former and institutionalize the latter.

The paper critiques TLOs as a barrier to technology transfer, but could have gone further by explicitly bringing in the concept of equality, both international, e.g., “a substantial, although not very visible, share of the redistribution from rich to poor countries has operated through free (or low-price) IP transfers” (Tirole) and national, e.g., “the main force in favor of greater equality has been the diffusion of knowledge and skills” (Piketty). I suspect doing so is necessary in order to place TLOs in total opposition to the university mission and become the target of political activism. Security ought also be brought into the picture, though that would be multiple steps from anything the authors stated.

It’s commonly assumed that patents work better or are more necessary for drugs than other fields, and this paper does too, though perhaps admits for more possibility for “open source” drug development than usual. I think this assumption is a pillar of contemporary support for IP and needs to be directly attacked. But another common belief is that IP should be “tailored” for different fields. I think a fairly natural next step from what the authors already stated would be experimenting with such tailoring through research funding policy, presumably much easier than doing so through changing patent law directly. Each of the ownership schemes (university, inventor, and commons) could be defaulted to depending on the field of research, and some research governance body would periodically assess the results and reassess the default for each field.

The English Wikipedia article on Bayh-Dole could really use sections on criticism and copycat policies in other jurisdictions.