2.3 Policy Issues and Current Debates
2.3.1 Current Policy Debates: Patents
Schumpeter (1934) defined our modern interpretation of innovation as an applied invention. Since then, researchers have sought to determine how innovation occurs, how it can be encouraged, how it affects economies and how firms interact in innovation. There are four main areas of research within patent policy, all of which have in mind the ultimate goal of maximizing social surplus.
These areas are: cumulative innovations, knowledge diffusion, patent scope and breadth, and harmonization. Cumulative innovation and knowledge diffusion are aspects of encouraging innovation, which need to be accounted for in IP policy.
Patent scope and breadth, along with international patent system harmonization, examine the design and implementation of policies.
Patents and Knowledge Diffusion
An important aspect of patent policy design is the disclosure incentive it
incorporates. In return for being granted a legal monopoly, the innovator must disclose information related to the innovation (as in Scotchmer and Green, 1990.) The goal of this disclosure policy is to increase the spread of knowledge and encourage cumulative innovation, which subsequently increases social surplus. As Scotchmer and Green, (1990) argue, the “social goal” of disclosure allows for “shared technical information [that] would help other innovators in their own research, reduce redundancy, and hasten the time to subsequent
innovation.”24 From the firm’s perspective, disclosure may not be desirable as it can benefit competitors, as noted in Scotchmer (2005.) The disclosure
surrounding patents reveals technological advances and can indicate the strategic plans of a company (Baker and Mezzetti, 2005.) With cumulative innovation, disclosure by one firm may allow a competitor to leapfrog (as in Fudenberg 1983.) This is one reason why trade secrets may be employed, as will be discussed later.
Breschi and Lissoni (2007) use patent data to study knowledge diffusion between researchers. Using patent data, they map knowledge networks and highlight the concept of technological gatekeepers, which are figures that sit at the centre of and link important research areas. Their findings, while
inconclusive, point to the concept that social and physical distance may have greater influence on knowledge diffusion than the disclosure of knowledge through patenting. This suggests limitations to the effectiveness of patents as a mechanism for disclosure.
Patent Scope and Breadth
In addition to examining patent policies as a whole, examining patent duration and breadth was also a popular research topic in the 1990s. Patent duration has obvious research appeal (Gold and Gruben, 1996) as it is merely the length of the patent (typically 20 years). Additionally, patent duration is dependent on the owner of the patent paying renewal fees, which can indicate the value of the patent (Shankerman and Pakes, 1984; Shankerman, 1998; and Lévêque and Ménière, 2006.) However, this represents a simplistic view as patent duration may not be the same as the market duration of an innovation. Thus, the focus of research shifted to effective patent duration, as in Denicolo (1996), which is more difficult to measure. The concept of effective patent duration highlights the fact that a multitude of factors, including market conditions, type of innovation, industry and cumulative innovation, can affect the optimal policy.
24 Scotchmer and Green (1990), p. 132.
A second focus of patent policy research is that of patent scope or breadth. This line of research examines how much a patent covered in terms of determining what similar innovations could be patented, what is considered infringement and how this is interpreted in the courts. From an economic perspective, breadth is the closeness of non-‐infringing substitutes; or, as Gallini (1992) identifies, the cost of inventing around a patent. This, however, presents an empirically challenging model as patent scope is a less tangible measurement and is often determined in court cases and practice rather than written policies (Chang, 1995.)
Patent length and breadth are two policy tools that counterbalance each other.
Longer patents are typically balanced by having narrower patent breadth. This argument is developed by Klemperer (1990), who examines both infinitely lived, narrow patents and short-‐lived, broad patents looking at product differentiation.
He argues that the market conditions of the demand side (consumer
preferences) determine the ideal policy. However, in the same issue of the same journal, Gilbert and Shapiro (1990) develop a similar argument, assuming
homogeneous products, which calls for infinitely lived patents offset by adjusting patent breadth. Yet both of these papers acknowledge the fact that they do not consider cumulative innovations, which could dramatically alter the results.
Gallini (1992) furthers the Klemperer and Gilbert and Shapiro models by allowing for the possibility of costly imitation. Contrary to their papers, she argues that the ideal patent length is shorter because longer patent lengths encourage others to invent around patents and/or infringe. Furthermore, the consequence of a longer patent length may actually be decreased R&D
expenditures as innovators face the increased likelihood of losing their monopoly as their competitors invent around the patent or infringe. Gallini concludes that the optimal policy is broad patents that limit imitation and adjustable patent lengths, which are adjusted to achieve the desired reward to the innovator.
Reid and Roberts (1996) focus on the scope of patenting as opposed to its
duration by examining the patent width, defined as the technological coverage of
the patent, and breadth, which they define as the scope for exploiting the market power which the patent rights confer on the innovator. They use this to analyse the patenting activity of the scientific instruments industry in the UK and
conclude that patent-‐active firms are typically larger than non-‐patent active firms and suggest that the patent activity has a non-‐linear relationship with firm size characterized by diminishing returns.
Scotchmer (2005) refines the breadth argument further by making an important distinction when analysing patent policy. She separates patent breadth from the inventive step, whereas other authors treat them singly. Scotchmer notes that the breadth determines how different another innovation must be in order to avoid infringement. The inventive step determines what is patentable. The product space of these concepts is represented in Figure 2-‐1, where policy determines the infringement and patentability of products. For product A, the inner circle is infringing and unpatentable substitutes due to the narrow patent.
The outer circle are substitutes that are not considered infringing, but do not meet the inventive step required. For product B, the inner circle is infringing due to the small inventive step, while the outer circle is considered infringing due to the broad patent.25
25 Scotchmer develops an example of this concept using the laser. The maser, which had a broad patent, was a blocking patent to the laser. The maser would be product B in Figure 1, while the laser would fall in the outer circle. See Chapter 3 of Scotchmer (2004) for further information.
Figure 2-1: Scotchmer’s Patent Breadth and Inventive Step
The concepts of patent length, breadth, scope and the inventive step are
important policy tools available to IP regimes. However, as this section notes, no consensus has been reached on the ideal design of IP policy. Therefore, IP
regimes often employ a mix of these tools and vary significantly across regimes.
Patent Law Harmonization
The movement towards harmonization of patent laws has also proven to be a fertile research ground. Harmonization seeks to standardize IP protection internationally, which is hoped to encourage investment in developing countries, promote innovation and reduce the costs of achieving IP protection (Pitkethly, 1999.) The signing of the Agreement on Trade Related Aspects of Intellectual Property Rights (TRIPS) in 1994 created a natural experiment, which allows for researchers to observe the effects of harmonization. Authors such as Beath (1990) model the effects of TRIPS on developing countries, international trade, etc. While the effects may not be seen for some time, they call attention to the conflicts surrounding harmonization. The main points of contention are the enforcement of IP laws and the level of protection provided. On a more subtle
Patent A: Narrow, large required inventive step
Patent B: Broad, small required inventive step
A B
Infringing, unpatentable Non-‐infringing, unpatentable Infringing, patentable
Patent Policy product space
note, the first-‐to-‐invent-‐rule26 of the US versus the international standard of first-‐to-‐file and the patentability of business methods and organisms, provide examples of underlying ideological conflicts, as examined in Kotabe (1992) and Kawaura (2005.)
An example of the efforts to harmonize national systems and develop an international standard is the European Patent Convention. This agreement created an international filing system that reduces foreign filing fees and is developing a unified opposition system (Caldirini and Scellato, 2003.) Unlike trademarks and geographical indicators, which are internationally unified within the EU and fall under the jurisdiction of the Office for the Harmonization of the Internal Market (OHIM), patents enjoy no such consistency. However, the British Patent Office’s Gowers Review (2006) of IP notes the high cost of translation fees associated with filing in multiple countries. Despite these costs, the fact remains that patent offices generate significant revenue for governments at the national level. The vested financial interests, national pride and ideological differences have created challenges to a unified system (Pitkethly, 1999.)
Patents and Cumulative Innovations
If I have seen further, it is only because I am standing on the shoulders of giants.
Newton (1676)
In his now clichéd phrase, Newton typifies the concept of cumulative innovation by stating that his advances were only possible due to the advances of other scientists before him. Innovations can thus be divided into two categories:
independent (i.e. one-‐off or isolated) and cumulative (i.e. sequential), as in Scotchmer (2005.) Cumulative innovations are a series of innovations that build on each other, whereas independent innovations do not spawn subsequent innovations.
26 The first-‐to-‐invent rule in the U.S. dictates that the first entity to invent the innovation should have the right to the patent. This differs from the first-‐to-‐file rule, in which the first entity to file the patent has the right to the patent.
There are numerous variations of cumulative innovations, as Scotchmer (1991) defines in her seminal paper on cumulative innovation. The first innovation may be a research tool without a commercial application. It may be a small or large innovation. The second innovation may be the commercially successful
application of the tool, a great leap forward (a radical innovation), or merely a minor improvement (an incremental innovation.) For example, Percy Spencer invented the microwave oven when he realized that the magnetron he was working with melted the chocolate bar in his pocket; Spencer developed a hugely successful commercial application for the magnetron that had been used
primarily for ship navigation.27 However, the first innovation is a necessary predecessor to the second. Therefore, the first innovator is only appropriately rewarded once the profit and social surplus of the second is taken into account (Scotchmer 1991; Encaoua and Lefoulli, 2006.)
Scotchmer also notes the difficulty in rewarding each generation appropriately.
In order to give the second innovator an incentive to invent, they must receive some of the surplus. However, the first innovator must also be rewarded with some of this surplus. This inherent conflict, referred to as a double
marginalization,28 makes the one-‐size-‐fits-‐all approach to patenting inefficient, as the surplus generated by the second innovation cannot be given to both
innovators. One of her suggested solutions, also addressed by others (Cugno and Ottoz, 1991; Choi, 2004; and Anand and Kahnna, 2000), is to allow more
collusion between firms through licensing and other means. This, however, poses anti-‐competition problems, as Chang (1995) also notes, due to the anti-‐
competitive nature of collusion agreements. Chang examines collusion in the context of cumulative innovation and licensing and concludes that allowing collusion may attract inefficient entry by imitators who invent around the original. Even Thomas Edison recognized the process of cumulative innovation in his quote, “I start where the last men left off.”
27 See Lemelson-‐MIT Program, 1996.
28 Scotchmer (1991), p. 34.