The Technology Adoption Problem

Established corporations (hereafter firms) make decisions about when and which technologies to adopt to increase revenues and stay competitive. The types of technologies they adopt are diverse, from information technology infrastructures that facilitate communication and customer service to manufacturing equipment that increase output and precision. The process a firm undergoes in deciding which technologies to adopt can be complex, but has been modeled by McCardle.[i]

McCardle posits that when a firm is faced with a technology adoption decision it will first estimate that technology’s potential profitability. Let this estimate be denoted p, to be contrasted with the technologies’ actual profitability p*. From this starting point the firm will undergo a process of (not costless) information gathering in order to increase the precision of its initial profitability estimate.

Information is gathered until p passes the firm’s preference-based upper or lower bound of profitability signaling adoption or rejection, respectively. As gathered information increases the precision of the estimate, the thresholds narrow. Put differently, an imprecise but high p value and a precise, but more modest p value may both trigger adoption. If information is exhausted or too costly, a decision will be made based on the estimate and its precision.

The Real Options Solution

Firms may gather information about external R&D products through two means. First, the firms may continually seek information from, in this case, the University and faculty inventor. Second, the firm may engage in a real options contract.[ii] Purchasing a real option creates the right but not the obligation to make a larger purchase (the technology license) at some future time. When a firm has internal R&D, the option is created through internal R&D investment. When R&D is external, e.g. already completed in the University, the option is purchased through contract. A small options payment gives the firm an options period, characterized by the resolution of exogenous and endogenous uncertainty, over which the license decision may be made.

The options contract creates a different framework than the straight forward pay-for-information scenario. Under both the options framework and pay-for-information scenario exogenous uncertainty is resolved through information exchange and waiting. However, only under the real options framework are endogenous uncertainties resolved through in-house (firm) evaluation. This allows the firm to discover and evaluate the costs of developing the technology to interface with its infrastructure or sales goals. Accordingly, the options framework allows the firm a high degree of precision in estimating the profitability of a nascent technology without substantial sunk costs.

Implications for the University

This process is often further complicated, as Lippman and McCardle note, because firms may evaluate more than one technology at a time. [iii] Accordingly, from the perspective of the University, it is important to establish a high initial profitability estimate, or demonstrate a high likelihood of an increase in p through technology-infrastructure interfacing during the options period.

Further, the University should recognize that pursuing the options contract may attract licensees where there would not have been under the sunk-cost licensing framework. Not only could this increase University revenues, but may diversify the pool of firms that idiosyncratically develop the nascent technology and explore its potential.

There may be, however, drawbacks to the options contract. First, when the options period expires the firm must either license the technology or abandon the associated research. The firm may decide to abandon the license but, with intimate knowledge of the innovation, invent around the University’s patent. Second, where there are not high sunk costs, licensing firms may be more incentivized to prematurely abandon the development of a technology.

Bilski v. Kappos is a 2010 Supreme Court case concerning process patents, as allowed under section 101 and defined under section 100 as a “process, art or method, include[ing] a new use of a known process, machine, manufacture, composition of matter, or material.” At issue in Bilski is whether patent protection should be granted over a series of steps that, when followed, are said to hedge risk in the energy commodities market.

The typical test of patent eligibility is the so-called machine-or-transformation test. When an innovation does not include a particular machine, it must perform “an act, or a series of acts, upon the subject matter to be transformed and reduced to a different state or thing.” However the Supreme Court dismisses the notion that this is the exclusive test of patentability.

Rather section 101 allows the term “process” to be broadly construed to include certain types of business methods, methods being found in the section 100 definition of process. As such, business methods for the Supreme Court are not categorically barred. However, the Court in Bilski decided that the process patent application, which seeks protection over (i) the concept of hedging risk and (ii) and application of that concept to a particular market, is an abstract idea that falls outside the section 100 and 101 definitions.

The Court found the proposed patent similar to two previous rulings. First, in Benson, the Court explained that a principle in the abstract is a fundamental truth; it is something that cannot be claimed as an exclusive right. It used this to reject protection over an algorithm that converted numerals to binary code. Second, in Flook, the court decided that even if an algorithm were limited to application in a particular technological environment it still could not be granted patent protection.

Only in Diamond v. Diehr did the court find that “an application of a law of nature … to a known structure or process may [deserve] patent protection,” and only when it limits its reach to a particular, inventive application of the law. Contrast with a 2012 case, Mayo v Prometheus, makes clear the difference between an application of natural law to a technological environment, and an application of natural law to a particular structure.

The patent at issue in Mayo is a three step process patent; step 1 instructed the doctor to administer a certain autoimmune disease fighting medication; step 2 instructed the doctor to measure the resulting and correlated metabolite levels in the patient’s blood; step 3 gave a scale that indicated metabolite concentrates above which there may be harmful side-effects and below which the drug is ineffective. The patent over this process was voided on the ground that “the three steps add nothing specific to the laws of nature other than what is well-understood, routine, conventional activity.” Put differently, using metabolite concentrations to check drug efficacy in a patient is not an “innovative concept.”

Contrast this ruling with Diehr, where at issue was also a process that employed a mathematical equation or rule. In Deihr, however, the patent did “not seek to pre-empt the use of that equation, except in conjunction with all of the other steps in their claimed process.” The process, for the first time, monitored the temperature of rubber during heating in order to achieve optimal post curation functionality. The temperature reading was fed into a computer that calculated the optimal time with a well-known mathematical formula relating variables such as pressure, thickness and temperature. The patent for this process was granted, standing for the proposition that the implementation is not a bar to patentability of an otherwise innovative process.

This distinction makes clear why the patent at issue in Bilski was dismissed. The concept of hedging was the centerpiece of the desired patent, just as it was in Mayo (though Mayo came later). The concept may have been applied in a new context, but the patent sought protection over the idea of hedging itself. Perhaps if the concept of hedging were used in an otherwise innovative process, as the mathematical formula was in Diehr, the patent would have been granted.

          This article attempts to describe the contemporary landscape of University Technology Transfer. Part 1 presents a brief history of the field. Part 2 summarizes the Litan-Mitchell model of university tech transfer and licensing optimization. Part 3 augments this optimization model. Part 4 supplements the licensing gap with entrepreneurial considerations.

I.                    University Technology Transfer: An Historical Primer

          Universities did not always hold the property rights to the innovations of their faculty. Not until the Bayh-Dole Act in 1980 did the federal government – which, Catherine Ives notes, funds the “overwhelming bulk of university research -” grant these rights to the University itself (Ives, 2012).[i] The grant was, however, conditioned upon a government, non-exclusive license to the invention, a restriction of sale of the invention abroad, and a right of entry for the public interest.

          Dr. Ives continues that this current OTT model was established only after several historical paradigms. Relations between industry and academia in the early 20^th century generated continuous innovation, especially in the pharmaceutical field. But in the 1960’s and 70’s successive presidential administrations claimed title to over 28,000 patents, only licensing 4%. This commercialization deficit sparked the era of IPAs, which allowed select Universities to claim title to innovations conditioned on their creation of Technology Transfer Offices. Finally, through personal favor over political merit, the Bayh-Dole bolstered the ratio of inventions disclosed to licenses executed to 4 to 1.

II.                  Suggested Improvement: The Litan-Mitchell Model

          Having run on Bayh-Dole for nearly 30 years, Robert Litan and Lesa Mitchell suggest that it may be time to update the university technology transfer model. Litan and Mitchell have envisioned a licensing system they believe would optimize the social benefit flowing from novel academic innovations. Their theory goes that under the current commercialization paradigm, an academic innovator is constrained by his duty to seek product development through his or her home University’s technology transfer office. Litan and Mitchell analogize this paradigm to economic monopolization, and suggest, true to economic theory, that a more competitive framework would increase commercialization efficiency and consumer welfare. They suggest that innovator free agency – i.e. allowing novel technologies to be licensed, further developed and marketed by any institution regardless of the home university – will usher in an era of such competition.

III.                Augmenting The Litan-Mitchell Model

          University Technology Transfer offices must content with highly complex knowledge. Often times the research of their University’s professors is so intricate and sophisticated that the communication costs – the time and energy necessary to understand and convey their discovery – may prohibit the rapid comprehension by potential licensees. The cost of the formal search process has lead Bessen to argued that “a division of labor applied to the search process ensures that many new discoveries will be tested even without specific prior knowledge” (Bessen, 1998).[ii] In other words, he contends that industry firms would be best suited to absorbing new technologies by forming specialized departments to search out innovations. Until such improvements are systematically in place, search costs are a substantial consideration and may favor entrepreneurship over licensing.

          In light of this insight, a free agency free-for-all may be inefficient. Rather than Professors seeking out Universities on a case by case basis, Tech Offices themselves should establish relationships with one another. Rather than cut inventors loose from their home universities, universities should communicate their intellectual assets with each other while retaining the right to innovations they have supported. Universities with an innovation in one industry should communicate with universities with strong industry ties in that field. Having universities with specialized knowledge of a given industry, through past transactions, would reduce the search costs of innovators looking for licensing or spin-out funding. This would allow universities to employ their resources not just to the limited number of innovations produced by their own faculty, but add value to innovation projects from universities that are not so endowed in a given sector. These assisting universities should also be given cut of the licensing from the home university. Such a division of labor will more efficiently employ resources and better connect innovators with those who commercialize. And this would take no legislation to allow.

IV.                University Entrepreneurship: Relevant Considerations

          Technologies originally developed by academic researchers and intended for commercialization may take two paths. Either the technology is licensed to an established firm or is the centerpiece of an entrepreneurial venture. The licensing path of an academic innovation may be optimized according to the Litan-Mitchell model. However, the dearth of licensing that the Litan-Mitchell model seeks to remedy may be supplemented with entrepreneurial activity.

          A.      When Does, and Should, Entrepreneurship Take Place?

          Di Gregorio and Shane found that two university technology transfer policies had a statistically significant influence on the rate of university entrepreneurship (Di Gregorio, Shane, 2002).[iii] First, the inventor’s share of royalties was inversely related to entrepreneurship, causing the authors to posit that an increased opportunity cost, in the form of high royalty rates, made innovators tend towards licensing. Second, entrepreneurial activity was positively correlated with a willingness of universities to take an equity share in the start-up, causing the authors to theorize that alleviating patent and licensing expenses decreased the start-ups capital constraints.  Among statistically insignificant factors include the geographic proximity of venture capital firms and technology incubators, the proportion of industry funding a university received and the availability of university venture capital. Their study shows that an innovator’s opportunity costs and the availability of equity financed capital are relevant to determining whether licensing or entrepreneurship is the appropriate strategy.

          B.      When Should Entrepreneurship Be Avoided?

          The structure of the relevant market has an impact on the likelihood of a start-ups survival, based on two characteristics of the innovation. The first characteristic is the innovations radicalness, i.e. whether it is an incremental step towards efficiency gains, readily adaptable to existing infrastructure, or where the adoption of the technology will requires a knowledge base overhaul. The second characteristic of the innovation is the scope of the intellectual property. Nerkar and Shane found that the more radical a technology, the more fragmented a market must be in order for a university start-up to survive (Nerkar, Shane, 2003)[iv] Put differently, the more concentrated a market is, the less likely it is that a radical technological start-up will survive. Similarly, broader scope patents have the same effect in those markets. A concentrated market, therefore, should signal a strategy of licensing over entrepreneurship and vice versa.