Growth in patenting corresponds to a new organisation of research that is less centred on the individual firm and more based on knowledge networks and markets: innovation processes throughout the OECD area have become increasingly competitive, co-operative, globalised, and more reliant on new entrants and technology-based firms. Market mechanisms play a more central role in technology diffusion. Businesses have been demanding more and more patents to accommodate these new conditions.
At the same time, patent regimes themselves have experienced major changes that have encouraged an increase in patenting. Not only have new types of inventions (software, genetic, and business methods) been deemed patentable by some patent offices, but the ability of patent holders to protect and enforce their rights has also increased, leading many to call the past two decades a pro-patent policy era. There is little doubt that many of these policy changes have helped the patent system to cope with changes in innovation systems by attracting more private-sector funding for R&D and supporting the development of markets for technology to help diffuse patented knowledge. In that sense, the patent system has been instrumental in the recent waves of innovation which have occurred in the fields of biotechnology and ICT.
This strengthening of patent systems in the European Union, Japan and the United States has, however, raised new concerns and exacerbated old ones. There have been numerous claims that patents of little novelty or excessive breadth have been granted, allowing their holders to extract undue rents from other inventors and from customers. This has been of particular concern in software, biotechnology and business methods, where patent offices and courts have had most difficulties in responding to rapid change, building up institutional expertise, evaluating prior art and determining correct standards for the breadth of granted patents. More basically, it has also been asked whether patentability might hamper the diffusion of knowledge, and therefore innovation, notably in these new areas. Other concerns have been raised about access to basic technologies, and research tools, which seems to have been hindered sometimes by patent holders exercising their right to exclude. As universities are becoming more likely to patent and commercialise their own inventions, exemptions for research use of existing inventions are under threat, with the danger of public research being faced with rising costs and difficulties of access.
Addressing these concerns and ensuring that patent systems continue to fulfil their mission of both stimulating invention and promoting diffusion of knowledge requires careful examination of broader issues. This report summarises OECD work to date on the relationships between patents, innovation and economic performance. It aims to place major changes in patenting patterns and patent regimes in the economic context, and to review the evidence regarding the links between patenting, innovation and diffusion in areas of particular interest (PROs, biotechnology, software and services). It provides policy-relevant conclusions based on existing analysis, and identifies policy issues and options for further consideration.
Empirical evidence tends to support the effectiveness of patents in encouraging innovation, subject to some cross-industry variation. In a series of surveys conducted in the United States, Europe and Japan in the mid-1980s and 1990s, respondent companies reported patents as being extremely important in protecting their competitive advantage in a few industries, notably biotechnology, drugs, chemicals and, to a certain extent, machinery and computers. Companies in other industries reported that patents play a secondary, if not negligible, role as a means of protection for their inventions, as they tend to rely more on alternative means such as secrecy, market lead, advance on the learning curve, technological complexity and control of complementary assets (Levin, Klevorick, Nelson and Winter, 1987; Cohen, Nelson and Walsh, 2000).
However, patent protection may also hamper further innovation, especially when it limits access to essential knowledge, as may be the case in emerging technological areas when innovation has a marked cumulative character and patents protect foundational inventions. In this context, too broad a protection on basic inventions can discourage follow-on inventors if the holder of a patent for an essential technology refuses access to others under reasonable conditions. This concern has often been raised for new technologies, most recently for genetic inventions (Bar-Shalom and Cook-Deegan, 2002; Nuffield Council on Bioethics, 2002; OECD, 2003a) and software (Bessen and Maskin, 2000; Bessen and Hunt, 2003).
In addition, as has long been recognised, the main drawback of patents is their negative effect on diffusion and competition. As patents are an exclusive right that creates a temporary monopoly, the patent holder can set a market price higher than the competitive price and limit the total volume of sales. This negative impact on competition could be magnified as patent holders try to strengthen their position in negotiations with other firms, in an attempt to block access by competitors to a key technology, or inversely, to avoid being blocked by them (Shapiro, 2002). Such strategic patenting seems to have developed over the past 15 years, notably in the electronics industry (Hall and Ziedonis, 2001).
Nevertheless, patents can also have a positive impact on competition when they enhance market entry and firm creation. Not only is there evidence of small companies being able to assert their right in front of larger ones thanks to their patent portfolio, but patents may also be a decisive condition for entrepreneurs to obtain funds from venture capitalists (Gans, Hsu and Stern, 2002). Moreover, patents may enhance technology diffusion. Patenting means disclosing inventions which might otherwise be kept secret. Industrial surveys show that the reluctance of firms to patent their inventions is primarily due to the fear of providing information to competitors. This has been confirmed in the OECD/BIAC survey on the use and perception of patents in the business community, sent to firms in OECD countries in 2003 and in which respondents indicated their intensive use of patents as a source of information (Box 2; Sheehan, Guellec and Martinez, 2003). Patents also facilitate transactions in markets for technology: they can be bought and sold as property titles or, more frequently, be subject to licensing agreements which allow the licensee to use the patented invention in return for payment of a fee or royalty (Arora, Fosfuri and Gambardella, 2001; Vonortas, 2003). Finally, enhancing technology diffusion has been the goal put forward by governments to encourage universities to patent their inventions, with the objective of licensing them to businesses that will further develop and commercialise them (OECD, 2003b).
In summary, the traditional view of patents as a compromise between incentives to innovate and barriers to technology diffusion, if not incorrect, presents a rather partial picture, as patents can either encourage or deter innovation and diffusion, depending on certain conditions. In fact, the effect of patents on innovation and diffusion depends on particular features of the patent regime. Patent subject matter, patenting requirements and patent breadth are three basic tools for policy makers involved in the design of patent regimes that could be used to enhance both innovation and diffusion (Encaoua, Guellec and Martinez, 2003):
Since 1998, software-related inventions (and mathematical algorithms in general) are patentable in the United States as long as they produce a "useful, concrete and tangible" result, in addition to the usual criteria (novelty, non-obviousness and industrial application). However, in Europe and to some extent in Japan, they are only patentable if "sufficiently technical in nature" (which excludes business methods in particular), a position which has been recently confirmed in Europe, although the legislative process is still ongoing (Hall, 2003; Motohashi, 2003).
Following permissive patentability trends, patents for software and business method inventions have increased rapidly in recent years in the United States. Various estimates indicate that the number of software patents granted by the USPTO grew from fewer than 5 000 per year in 1990 to approximately 20 000 in 2000, or approximately 15% of all US patents granted in that year (Hunt and Bessen, 2003). In contrast, business methods patents represent a very low share of the total number of grants, with around 1 000 grants per year in the US since 1998. Interestingly, software publishers account for only a fraction of software patents (only 6% of software patents according to one recent study) with the majority of software patents owned by large firms in the ICT manufacturing and electrical machinery sectors. Large software consultancies and other service-sector firms also account for a small, but growing, number of patents to date. This pattern reflects the increasing role of software and services business units within large ICT firms, as well as the growing pervasiveness of embedded software in a range of electrical and electronic devices.
Growth in software and business methods reflects both increased innovative activity and changes in patenting behaviours. R&D spending by software and ICT firms has grown rapidly over the past decade. Microsoft�s R&D expenditures alone grew from USD 270 million in 1991 to USD 4.4 billion in 2002. More than three-quarters of ICT firms responding to the OECD/BIAC survey reported that they were generating more inventions now than ten years ago (Sheehan, Guellec and Martinez, 2003). Nevertheless, the patenting strategies of these firms have also changed. More than three-quarters of ICT firms in the survey reported that they now patent technologies they would not have patented a decade ago even if the technology had been patentable then. Software and ICT firms see patents as an important bargaining chip in negotiating alliances with other firms and as a means of generating additional revenue via licensing. Indeed, more firms in the ICT sector than in other sectors reported increases in outward licensing and cross-licensing over the past decade. Other research has also demonstrated the key role of strategic patenting in the semi-conductor industry (Hall and Ziedonis, 2001).
Software programmes tend to be complex, modular products that combine multiple functions, each of which may be the subject of a different patent. Increased patenting may therefore inhibit follow-on innovation or the assembly of complex programmes as it increases transaction costs. Interoperability also needs to be high, meaning that open standards and interfaces are critical to ensuring innovation and market entry. On the other hand, if patents give more protection, they also could require more disclosure, which can be helpful for reducing the exclusion effect generated by patents. Network effects are also strong in the software sector, and switching costs can be high, locking-in customers to dominant products, especially if interoperability cannot be assured. In this context, patents might contribute to enhancing competition and innovation by allowing new market entrants to defend their technological position against incumbents.
In summary, when addressing the issue of software protection, the following points should be considered:
The paucity of economic evaluation of the patent system is striking. Most of the changes to patent regimes implemented over the past two decades were not based on hard evidence or economic analysis. It is necessary to develop economic analysis in this domain that would inform the policy debate, giving governments a clearer view beyond the arguments put forward by pressure groups. Such analysis should rely notably on quantitative evidence: an effort to build and make available to analysts the corresponding databases has been initiated notably by the OECD, but this work needs to be broadened. In addition, more information is needed on the ways in which patents are used by their holders, for instance as regards in-house implementation, licensing contracts and business strategies...