Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 217
Patents in the Knowledge-Based Economy Patents in Software and Biotechnology
OCR for page 218
Patents in the Knowledge-Based Economy This page in the original is blank.
OCR for page 219
Patents in the Knowledge-Based Economy Intellectual Property Protection in the U.S. Software Industry1 Stuart J. H. Graham and David C. Mowery Haas School of Business University of California, Berkeley INTRODUCTION The software industry is a knowledge-intensive industry whose output is information, the coded instructions that guide the operations of a computer or a network of computers. Both the inputs and much of the output of this industry consist of intangibles, the prices of which contain considerable Schumpeterian rents. The rewards to innovators in the software industry of the 1980s and 1990s were extraordinary, as illustrated by the meteoric rise of William Gates III to control of the largest personal fortune in the world. The modern computer software industry thus is an extreme example of an industry in which the returns to innovators’ investments, and in many cases market structure, are influenced by the ownership of intellectual property. As such, it is hardly surprising that the legal framework establishing and regulating ownership of such property has attracted considerable attention and debate. The “modern” computer software industry of the twenty-first century differs from the software industry of the 1950s or 1960s, most notably in the growth of mass markets for so-called packaged software. These differences are reflected in the central importance of formal protection of intellectual property. The increased importance of formal intellectual property rights protection, as well as the changing economic and legal importance of different instruments for such protection, create significant challenges for U.S. intellectual property rights policy. 1 We are grateful to participants in the STEP Board conference on “The Operation of the Patent System,” participants in the U.C. Berkeley Innovation Seminar, and to Professors Rosemarie Ziedonis, Wesley Cohen, and Brian Silverman for comments on the paper. We also appreciate assistance with our analysis of patenting data from Arvids Ziedonis. This chapter draws on research supported by the Andrew Mellon Foundation and the National Research Council.
OCR for page 220
Patents in the Knowledge-Based Economy Although the computer software industry is a global industry, significant differences remain among the software industries and the associated intellectual property regimes of the industrial economies. Domestic lobbying for the creation or modification of legal regimes covering this relatively new form of intellectual property has contributed to differences in the level and characteristics of intellectual property rights for computer software among major industrial economies. The recent controversies over business methods patents and the response by both Congress and the U.S. Patent and Trademark Office (USPTO) to these controversies (see below) are only the latest examples of this endogenous character of national intellectual property rights regimes. This chapter surveys intellectual property rights policies and controversies in the U.S. computer software industry. Immediately below, we discuss the historical development of the U.S. software industry, highlighting the ways in which the role, structure, and importance of formal intellectual property rights have changed over the course of the industry’s development. We then present data on the (limited) portion of the software industry for which reliable indicators of the intensity of patenting activity during the 1980s and 1990s can be computed, focusing on patenting by specialized packaged software firms. These indicators cover the “propensity to patent” (patents per R&D dollar) and provide some evidence on change over time in the “importance” of these firms’ patents. We also discuss patenting by large electronics systems firms in the same patent classes and compare the patenting behavior (and the “importance” of their patents) of the electronics systems firm that for many years was also the leading vendor of software, IBM, and the largest specialized packaged software firm, Microsoft. After a brief discussion of the changing prominence of U.S. universities as patenters in software, we examine the changing importance of copyright and patent protection of software-related intellectual property during the 1980s and 1990s. Our conclusion considers some of the policy implications of this analysis. THE HISTORICAL DEVELOPMENT OF THE COMPUTER SOFTWARE INDUSTRY The growth of the global computer software industry has been marked by at least four distinct eras spanning the 1945-2001 period. The first era (1945-1965) covers the development and commercialization of the computer. The gradual adoption of “standard” computer architectures in the 1950s supported the emergence of software that could operate on more than one type of computer or in more than one computer installation. In the United States, the introduction of the IBM 650 in the 1950s, followed by the even more dominant IBM 360 in the 1960s, provided a large market for standard operating systems and application programs. The emergence of a large installed base of a single mainframe architecture occurred first and to the greatest extent in the United States. Nonetheless,
OCR for page 221
Patents in the Knowledge-Based Economy most of the software for mainframe computers during this period was produced by their manufacturers and users. During the second era (1965-1978), independent software vendors (ISVs) began to appear. During the late 1960s, producers of mainframe computers “un-bundled” their software product offerings from their hardware products, separating the pricing and distribution of hardware and software. This development provided opportunities for entry by independent producers of standard and custom operating systems, as well as independent suppliers of applications software for mainframes. Unbundling occurred first in the United States and has progressed further in the United States and Western Europe than in the Japanese software industry. Although independent suppliers of software began to enter in significant numbers in the early 1970s, computer manufacturers and users remained important sources of both custom and standard software in Japan, Western Europe, and the United States during this period. Some computer “service bureaus” that had provided users with operating services and programming solutions began to un-bundle their services from their software, providing yet another cohort of entrants into the independent development and sale of traded software. Sophisticated users of computer systems, especially users of mainframe computers, also created solutions for their applications and operating system needs. A number of leading suppliers of traded software in Japan, Western Europe, and the United States were founded by computer specialists formerly employed by major mainframe users. During the third era (1978-1993), the development and diffusion of the desktop computer produced explosive growth in the traded software industry. Once again, the United States was the “first mover” in this transformation, and the U.S. domestic market became the largest single market for packaged software. Rapid adoption of the desktop computer in the United States supported the early emergence of a few “dominant designs” in desktop computer architecture, creating the first mass market for packaged software. The independent vendors that entered the desktop software industry in the United States were largely new to the industry. Few of the major suppliers of desktop software came from the ranks of the leading independent producers of mainframe and minicomputer software, and mainframe and minicomputer ISVs are still minor factors in desktop software. Rapid diffusion of low-cost desktop computer hardware, combined with the emergence of a few “dominant designs” for this architecture, eroded vertical integration between hardware and software producers and opened up opportunities for ISVs. Declines in the costs of computing technology have continually expanded the array of potential applications for computers; many of these applications rely on software solutions for their realization. A growing installed base of ever-cheaper computers has been an important source of dynamism and entry into the traded software industry, because the expansion of market niches in ap-
OCR for page 222
Patents in the Knowledge-Based Economy plications has outrun the ability of established computer manufacturers and major producers of packaged software to supply them.2 Estimates of the relative size of the “packaged” and “custom” software markets are extraordinarily scarce, reflecting the failure of public statistical agencies to collect reliable data on this rapidly growing component of the “information economy.” Nonetheless, the few existing estimates suggest that the market for “packaged” software exceeded that for “custom” software by the mid-1980s. Data reported in Mowery (1996), which summarize surveys compiled by the OECD and the International Data Corporation (IDC), indicate that global consumption of “packaged” software amounted to roughly $18 billion in 1985 (current dollars) versus $11.6 billion for “custom” software. U.S. consumption of “packaged” and “custom” software, both of which were overwhelmingly domestic in origin, amounted to $12.6 billion and $4.2 billion, respectively, in 1985. Global consumption of packaged software in 1996 reached $109 billion, according to IDC estimates published in the Department of Commerce’s 1998 U.S. Industry and Trade Outlook, and the Department estimated that global consumption would amount to more than $221 billion by 2002 (U.S. Department of Commerce, 1998, p. 28-3 et seq.). More recent estimates of the size of U.S. or global consumption of “custom software” unfortunately are unavailable; but most studies of the computer software industry (e.g., OECD, 1998) suggest that consumption and shipments of packaged software have grown much more rapidly than those for custom software during the 1985-2002 period. The packaged computer software industry now has a cost structure that resembles that of the publishing and entertainment industries much more than that of custom software—the returns to a “hit” product are enormous, and production costs are low. And like these other industries, the growth of a mass market for software has elevated the importance of formal intellectual property rights. An important contrast between the software industry and the publishing and entertainment industries, however, is the importance of product standards and consumption externalities in the software market. Users in the mass software market often resist switching among operating systems or even well-established applications because of the high costs of learning new skills as well as their demand for an abundant library of applications software to complement an operating system. These switching costs typically are higher for the less-skilled users who dominate mass markets for software and support the development of “bandwagons” that create de facto product standards. As the widespread adoption of desktop computers created a mass market for software during the 1980s, these de facto product standards in hardware and software became even more important to the commercial fortunes of software producers than was true during the 1960s and 1970s. 2 Bresnahan and Greenstein (1996) point out that a similar erosion of multiproduct economies of scope appears to have occurred among computer hardware manufacturers with the introduction of the microcomputer.
OCR for page 223
Patents in the Knowledge-Based Economy The fourth era in the development of the software industry (1994 to the present) has been dominated by the growth of networking among desktop computers within enterprises through local area networks linked to a server and/or the Internet, which links millions of users. Networking has opened opportunities for the emergence of new software market segments,3 the emergence of new “dominant designs,” and, potentially, the erosion of currently dominant software firms’ positions. Like previous eras in the industry’s development, the growth of network users and applications has been more rapid in the United States than in other industrial economies, and U.S. firms have maintained dominant positions in these markets (see Mowery and Simcoe, 2001). How has the growth of the Internet changed the economics of intellectual property protection in the software industry? At least three different effects are apparent thus far in the Internet’s development. First, the widespread diffusion of the Internet has created new channels for low-cost distribution and marketing of packaged software, reducing the barriers to entry into the packaged software industry that are based on the dominance of established distribution channels by large packaged software firms. In this respect, the Internet expands the possibilities for rapid penetration of markets by a “hit” packaged software product—in the jargon of the software industry, a “killer app[lication]”—which enhances the economic importance of protection for these types of intellectual property. The Internet also is an important factor in the growth of patents on software-embodied “business methods,” many of which concern tools or routines employed by online marketers of goods and services. But the Internet has also provided new impetus to the diffusion and rapid growth of a very different type of software, “open source” software. Although so-called shareware has been important throughout the development of the software industry, the Internet’s ability to support rapid, low-cost distribution of new software and, crucially, the centralized collection and incorporation into that software of improvements from users has made possible such widely used operating systems as Linux and Apache (see Kuan, 1999 and Lerner and Tirole, 2000). The Internet thus has increased the importance of formal protection of some types of software-related intellectual property while simultaneously supporting the growth of open source software, which does not rely on such formal instruments of intellectual property protection. THE EVOLUTION OF INTELLECTUAL PROPERTY RIGHTS POLICY AND PRACTICE IN THE U.S. SOFTWARE INDUSTRY This study is primarily concerned with intellectual property rights in software that combine some grant of limited monopoly in exchange for an element of 3 For example, the operating system software that is currently installed in desktop computers may reside on the network or the server.
OCR for page 224
Patents in the Knowledge-Based Economy disclosure or public use. As such, it is appropriate to examine copyright and patent protection, because software has been brought underneath the umbrella of each of these regimes during the last several decades. In the near future, however, the use by software innovators of legal protections in the areas of trade secret,4 misappropriation,5 trademark,6 and even the Semiconductor Chip Protection Act7 will remain important. Copyright Copyright protection for software innovation was singled out by policymakers during the 1970s as the preferred means for protecting software-related intellectual property (Menell, 1989). In its 1979 report, the National Commission on New Technological Uses of Copyrighted Works (CONTU), charged with making recommendations to Congress regarding software protection, chose copyright as the most appropriate form of protection for computer software (CONTU, 1979). Because copyright protection adheres to an author-innovator with relative ease and has a long life—now upwards of 120 years for works created for hire—the Commission determined that copyright was the preferred type of intellectual property protection for software. Congress adopted the Commission’s position when it wrote “computer program” into the Copyright Act in 1980.8 The federal judiciary’s application of copyright to software in the aftermath of the CONTU initially promised strong protection for inventors. Apple Computer, Inc. v. Franklin Computer Corp.9 is an early and important case of copyright litigation in packaged software. Although the federal judiciary had long held that copyright protected only “expression” in works,10 the court in Apple Computer held that Apple’s precise code was protected by its copyright. The 4 A trade secret is formally some information used in a business that, when secret, gives one an advantage over competitors. The secret must be both novel and valuable. Metallurgical Industries, Inc. v. Fourtek, Inc., 790 F.2d 1195 (1986). 5 Collectors of valuable information can prevent competitors from using the information. International News Service v. Associated Press, 248 U.S. 215 (1911). 6 Protects names, words, and symbols used to identify or distinguish goods and to identify the producer. Zatrains, Inc. v. Oak Grove Smokehouse, Inc., 698 F.2d 786 (5th Cir., 1983). 7 Protection is available for software embodied in semiconductor chips—so-called mask works. E.F. Johnson v. Uniden Corp. of America, 653 F. Supp. 1485 (D. Minn. 1985). 8 17 U.S.C. sec. 101, sec. 117 (as amended 1980). For a more complete discussion, see Menell (1989). 9 714 F.2d 1240 (3rd Cir. 1983). Consistent with its position as a leading firm in the packaged software industry Microsoft, which supported stronger formal protection for software-related intellectual property, filed an amicus curiae brief on behalf of Apple in this case. 10 Historically, a major distinction in the copyright law has been that ideas are not protected, only expressions are. Baker v. Selden, 101 U.S. 99 (1879).
OCR for page 225
Patents in the Knowledge-Based Economy court concluded that efforts by a “follower” firm to use the copyright holder’s code for purposes of achieving compatibility with the original software were in-consequential to the determination of whether infringement had occurred. This decision strengthened copyright protection considerably, making it possible for one firm’s copyrighted software to block the innovative efforts of others. Subsequent decisions—the so-called “look and feel” cases—extended traditional copyright protection of “expression” to such “nonliteral” elements of software as structure, sequence, and organization.11 Subsequent court decisions, however, narrowed the protection provided by copyright for software-related intellectual property. The sweeping interpretation of copyright protection in Apple Computer was narrowed and weakened considerably in a series of copyright infringement cases brought by Lotus Development. Lotus successfully sued Paperback Software International over the latter’s alleged imitation of the “look and feel” of Lotus’s spreadsheet software in a case that Lotus won in 1990. Lotus then sued Borland International over the alleged infringement by Borland’s “Quattro” software of the “look and feel” of Lotus’s 1-2-3 spreadsheet software in a case that lasted for six years, producing four opinions in a federal District Court and appeals to both the Court of Appeals and the U.S. Supreme Court. The District Court found that Borland had infringed Lotus’s 1-2-3 spreadsheet software. Borland rewrote its software to achieve partial compatibility with elements of Lotus’s 1-2-3 software, but this modification also was met with infringement findings by the District Court and a permanent injunction banning its sale.12 The Court of Appeals ultimately reversed some of the District Court’s conclusions, arguing that “second movers” in the software industry must be allowed to emulate and build on parts of the innovator’s code and methods.13 The decision of the Court of Appeals was affirmed in 1996 by the Supreme Court in a 4-4 decision.14 The Borland decision weakened the strong protection for software inventions provided by Apple Computer, Inc. v. Franklin Computer Corp, and along with other decisions affirming the strength of software patents may have 11 Computer Associates Int’l v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992); Whelan Associates v. Jaslow Dental Laboratory, 797 F.2d 1222 (3rd Cir. 1986). 12 Lotus Development Corp. v. Borland Int’l, Inc., 788 F. Supp. 78 (D. Mass. 1992)(finding Quattro a virtual copy of Lotus’s menu structure); Lotus Development Corp. v. Borland Int’l, Inc., 799 F. Supp 203 (D. Mass. 1992); Lotus Development Corp. v. Borland Int’l, Inc., 831 F. Supp. 202 (D. Mass. 1993); Lotus Development Corp. v. Borland Int’l, Inc., 831 F. Supp. 223 (D. Mass. 1993). 13 Lotus Development Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995). 14 116 S. Ct. 804 (1996).
OCR for page 226
Patents in the Knowledge-Based Economy contributed to increased reliance by some U.S. software firms on patents in the 1990s.15 Patents In contrast to copyright, federal court decisions since 1980 have broadened and strengthened the economic value of software patents. Although some early cases during the 1970s supported the initial stance of the U.S. Patent and Trademark Office (USPTO) in stating that software algorithms were not patentable,16 judicial opinions have shifted since then to support the use of patents in software (Samuelson, 1990).17 In the cases of Diamond v. Diehr18 and Diamond v. Bradley,19 both decided in 1981, the Supreme Court announced a more liberal rule that permitted the patenting of software algorithms, strengthening patent protection for software (Merges, 1996). The economic value of these patents was highlighted in several high-profile cases during the 1990s. For example, a 1994 court decision found Microsoft liable for patent infringement and awarded $120 million in damages to Stac Electronics. The damages award was hardly a crippling blow to Microsoft, but the firm’s infringing product had to be withdrawn from the market temporarily, compounding the financial and commercial consequences of the decision (Merges, 1996). As the USPTO adopted a more favorable posture toward applications for software patents, the ability of patent examiners to identify “novelty” in an area of technology in which patents historically had not been used to cover major innovations was criticized well before the surge of “business methods” software patent applications in 1998 and 1999. The celebrated “multimedia” patent issued by the USPTO to Compton Encyclopedias in 1993 is one example of the difficulties associated with a lack of patent-based prior art. On November 15, 1993, Compton’s Newmedia announced that it had won a “fundamental” patent for its 15 Ironically, in light of subsequent controversies over the role of software patents, Menell’s influential 1989 analysis of intellectual property protection of software, written in the wake of the strong judicial interpretation of copyright embodied in Apple Computer, Inc. v. Franklin Computer Corp., argued that patents had significant advantages over copyright as a means for protecting computer applications software: “The patent system’s threshold requirements for protection—novelty, utility, and nonobviousness—are better tailored than the copyright standard to rewarding only those innovations that would not be forthcoming without protection” (Menell, 1989, p. 47). As we note below (see also Merges, 1999), the debate over software patents centers on precisely these issues—Is the USPTO able to apply these requirements with sufficient rigor to prevent the issue of low-quality patents? 16 Gottschalk v. Benson, 409 U.S. 63 (1972). 17 Samuelson (1990) argues that the USPTO was at odds with the Court of Customs and Patent Appeals (CCPA) throughout the 1970s over the patentability of software and concludes that the CCPA’s views in favor of patentability ultimately triumphed. 18 450 U.S. 175 (1981). 19 450 U.S. 381 (1981).
OCR for page 227
Patents in the Knowledge-Based Economy multimedia software that rapidly fetched images and sound.20 The patent was quite broad, covering a database search system that retrieves multimedia information in a flexible, user-friendly system. The search system uses a multimedia database consisting of text, picture, audio and animated data. That database is searched through multiple graphical and textual entry paths.21 Compton’s president, Stanley Frank, suggested that the firm did not want to slow growth in the multimedia industry, but he did “want the public to recognize Compton’s Newmedia as the pioneer in this industry, promote a standard that can be used by every developer, and be compensated for the investments we have made.” Armed with this patent, Compton’s traveled to Comdex, the computer industry trade show, to detail its licensing terms to competitors, which involved payment of a 1 percent royalty for a nonexclusive license.22 Compton’s appearance at Comdex launched a political controversy that culminated in an unusual event—the USPTO reconsidered and invalidated Compton’s patent. On December 17, 1993, the USPTO ordered an internal re-examination of Compton’s patent because, in the words of Commissioner Lehman, “this patent caused a great deal of angst in the industry.”23 On March 28, 1994, the USPTO released a preliminary statement declaring that “[a]ll claims in Compton’s multimedia patent issued in August 1993 have been rejected on the grounds that they lack ‘novelty’ or are obvious in view of prior art.”24 This declaration was confirmed by the USPTO in November of 1994.25 Patents in “Business Methods” Recent federal judicial decisions have continued to support the rights of patentholders and have expanded the definition of “software” subject to protection by patent. On August 23, 1998, the Court of Appeals for the Federal Circuit (CAFC) upheld the validity of a “business methods” software patent in State 20 Peltz, J. “Compton’s wins patent covering multimedia,” Los Angeles Times, November 16, 1993, D:2. The Compton’s patent was entitled “Multimedia Search Systems Using a Plurality of Entry Path Means Which Indicate Interrelatedness of Information.” Markoff, J. “Patent Office to Review A Controversial Award,” The New York Times, December 17, 1993, D:2. 21 Abstract, United States Patent Number 5,241,671, August 31, 1993. 22 Abate, T. “Smaller, faster, better; Tech firms show off their latest wonders at trade show and foretell a user-friendly future,” San Francisco Examiner, November 21, 1993, E:1. 23 Markoff, J. “Patent Office to Review A Controversial Award,” The New York Times, December 17, 1993, D:2. 24 Riordan, T. “Action Was Preliminary On a Disputed Patent,” The New York Times, March 30, 1994, D:7. 25 Orenstein, S. “U.S. Rejects Multimedia Patent,” The Recorder, November 1, 1994, 4.
OCR for page 248
Patents in the Knowledge-Based Economy substitutes, in the protection of software-related intellectual property. Copyright protection of the software code (the expression) could complement patent protection of the underlying technical advance. Although neither Menell (1989) nor Lemley and O’Brien (1997) give serious consideration to a complementary relationship between patent and copyright protection in software, it is possible that commercial software developers are indeed using both, rather than substituting patents for copyright. In this section, we examine new data on software copyright registrations in a preliminary analysis of the changing relationship between copyright and patent protection in software. Just as we did in the examination of patent data for large packaged software firms above, we seek to develop measures of the “copyright propensity” of large packaged software firms during the 1987-1997 period. A finding that this propensity remained constant or increased would constitute evidence of complementarity between the use of copyright and the use of patents to protect software-embodied intellectual property, because these firms have increased their patent propensities during this period. A finding that the copyright propensity has declined, however, would provide preliminary support for the hypothesis that copyright and patent protection are substitutes, consistent with the Lemley-O’Brien argument cited earlier, and that commercial software firms now are relying more heavily on patents than copyrights to protect their intellectual property. Copyright Data Our data on copyrighting of computer programs by packaged software firms are drawn from the U.S. Library of Congress (LOC) collection of registered U.S. copyrights. The LOC has data on all materials45 that have been registered for copyright with the U.S. Copyright Office since 1978. Each record includes the identity of the entity requesting registration of copyright, a unique registration number, and the media type. Three dates are recorded for each registration: the date of creation of the work; its date of publication; and its date of copyright registration. As of January 2001, the LOC copyright database included over 13 million records. Using the list of the largest packaged software firms in 1997 provided by Softletter, we searched these LOC records for uniquely numbered copyrights registered on “computer programs.” Computer software can be designated as such by the author on the copyright registration form, and the Copyright Office assigns an internal “computer program” code to the relevant pieces of intellectual prop- 45 Including books, maps, sound recordings, computer files, dramatic works, toys, games, jewelry, technical drawings, photographs, multimedia kits, sculptural works, textiles, motion pictures, and choreography, among others.
OCR for page 249
Patents in the Knowledge-Based Economy erty. We rely upon this latter internal code when defining a registered computer program copyright. Although copyright provides some protection for a piece of written software regardless of whether it is registered with the Copyright Office,46 there are additional incentives for pursuing registration of a copyright. The registration procedure is quick and inexpensive, and the legal strength of the resulting protection is greater for a registered copyright.47 Registration within 5 years of original publication gives the copyright a presumption of validity under law.48 Infringement actions cannot be brought in the courts until a copyright is registered.49 The holder of a registered copyright is entitled to the recovery of attorney fees and statutorily defined damages, including those for willful infringement, only for the period after registration. Ordinarily, the owner cannot collect these damages for the period between the time of publication and the time of registration of the copyright, but the law offers an incentive for registering early: Damages are available from the date of publication only if the owner registers the copyright within 3 months of publication of the work.50 Faced with these incentives, it is plausible that rational actors in a crowded commercial space that rely on copyright to protect their software-related intellectual property will register the copyright on their software soon after creation. We therefore use data on registered copyrights to analyze trends in the use of copyright to protect software-related intellectual property. Our use of registered copyrights means that we are examining trends in the use by firms of copyrights for which some positive action and (modest) expenditure on the part of the “inventor” are required, rather than simply counting the copyrights that are created more or less automatically with the development of a new piece of software. Although all software is copyrighted at the moment of its creation, all software does not receive registered copyrights, and only registered copyrights provide a basis for the filing of a suit against an alleged infringer. Copyright Propensities Among the Leading U.S. Packaged Software Firms, 1987-1997 As in our analysis of software patenting among the largest U.S. packaged software firms, we restrict the sample of firms to include only firms for which R&D spending data are available, enabling us to compute “copyright propensities” for these firms. Our working definition of “software” in this analysis is 46 The 1976 Copyright Act, in accord with the international Berne Convention, gives copyright protection to authors regardless of registration status. 47 As of March 2001, registration required a two-page filing and fees totaling $30 US. 48 17 U.S.C.A. §410 (2000). 49 17 U.S.C.A. §411 (2000). 50 17 U.S.C.A. §§412, 504, 505 (2000).
OCR for page 250
Patents in the Knowledge-Based Economy much broader than that employed in the examination of patent propensities, because the LOC does not provide any disaggregated copyright class information for its registered software copyrights. As before, however, we limit our sample of firms to those for which we can obtain R&D spending data, to compute a “copyright propensity.” The data in Figure 17, a weighted 3-year moving average of the “copyright propensity” for the same 15 large packaged software firms for which patent propensity data were plotted in Figures 7 and 8, tends to support the Lemley-O’Brien assertion that copyright protection has been supplanted by the use of patents in software, at least among these leading producers of packaged software. As Figure 18 shows, excluding Microsoft from this sample does not substantially alter the conclusion that the copyright propensity of these firms has declined. In data not displayed in these figures because of space limitations, the copyright propensity data for Novell, Microsoft, and Adobe all display declines in the number of copyrights registered per $100 million of (constant-dollar) R&D spending during 1987-1997. Novell and Microsoft in particular exhibit sharply contrasting trends in patents/R&D$ and copyrights/R&D$; both firms display increases during this time period in patenting propensity and a downward trend in the propensity to copyright their intellectual property. Adobe, which exhibited little or no consistent time trend in its patent propensity, also displays a downward trend in its copyright propensity. A comparison of the copyright behavior of “incumbent” and “entrant” firms among the Softletter 100 (defined as above) also yields little indication of contrasting behavior among these two groups in their copyright propensities. Both incumbents and entrants decreased their use of copyright, relative to R&D spending, to protect their intellectual property during the 1980s and 1990s. FIGURE 17 Copyright propensity, 15 largest packaged software firms (1997), 3-year moving average, 1987-1997.
OCR for page 251
Patents in the Knowledge-Based Economy FIGURE 18 Copyright propensity, 15 largest packaged software firms (1997), excluding Microsoft, 3-year moving average, 1987-1997. Because we lack software-related R&D investment data for our sample of electronic systems firms, we are not able to examine changes during 1987-1997 in these firms’ copyright propensities. However, we do have software-related R&D investment data for IBM for the 1992-1997 period, and Figure 19 compares trends during 1992-1997 in the copyright propensities of Microsoft and IBM. As in the case of these firms’ patent propensities, IBM obtains substantially more registered copyrights per $100 million in R&D than does Microsoft throughout this period, and in contrast to their patent propensities, the gap has widened by 1997 (IBM’s copyright propensity is roughly twice as large as that of Microsoft in 1992 and more than four times as great in 1997). But both firms are reducing their copyright propensity through this time period, consistent with the “substitute” relationship posited above. Because our coverage of software copyrights differs somewhat from that of our software patents data, direct comparison of these trends in patent and copyright propensity must be interpreted cautiously. Nevertheless, our preliminary analysis of packaged software firms’ use of copyright to protect software-related intellectual property suggests that patents have increased in importance relative to copyright as a means for the protection of software-related intellectual property during 1987-1997. Moreover, a decline in copyright propensity during the 1990s is apparent as well in our limited comparison of a leading packaged software firm and a leading electronic systems firm. As we noted above, a shift from copyright to patent protection was once seen as an important step to raise the threshold for protection of software-related intellectual property, and it is ironic that increased patenting by software firms has been accompanied by a chorus of concern over “junk patents.” Junk patents may indeed be a problem (although our limited evidence on citations does not support this claim for the patents of large
OCR for page 252
Patents in the Knowledge-Based Economy FIGURE 19 Copyright propensity, Microsoft and International Business Machine, 1992-1997. packaged software firms), but any such problem might have been more severe had firms continued to rely heavily on copyright in preference to patents. Why might firms have shifted from copyright to patent protection? As we noted above in our discussion of the evolution of the software intellectual property “regime,” the treatment of copyright by the U.S. federal judiciary has changed to limit the sweeping rights originally claimed by copyright holders.51 This shift in judicial opinion may reflect the lack of a specialized appeals court that would support copyright holder rights as vigorously as the CAFC has done for patentholders. Certainly, software patents have enjoyed a more supportive judicial climate during the past decade than copyright. In addition, patents may better support the types of “defensive” intellectual property strategies that Hall and Ziedonis (2001) describe in the semiconductor industry—cross-licensing of portfolios of 51 Lemley and O’Brien note in their discussion that “…the courts have cut back the scope of protection rather dramatically in the past five years” (1997, p. 280).
OCR for page 253
Patents in the Knowledge-Based Economy patents may be less difficult than similar transactions in copyrighted material.52 The use of software-related patents to support markets in intellectual property (suggested by Lemley and O’Brien, 1997) and/or as a complement to defensive intellectual property strategies remains an important issue for future research. Nonetheless, to the extent that transactions in intellectual property are facilitated by reliance on patent rather than on copyright, and to the extent that the (admittedly limited) quality controls imposed by the USPTO on the issue of patents enforce a higher average “quality level” among software patents than is true of copyrighted material, a shift from copyright to patent protection may well be a desirable development. CONCLUSION The U.S. and global computer software industries have been transformed during the past 20 years as a result of the explosive diffusion of the microcomputer and the development of the Internet. No longer are the business activities and revenues of leading firms dominated by sales of products that incorporate high levels of user-specific customization. Instead, the dominant firms in the U.S. software industry, enterprises that account for a leading global market share as well, rely on sales of packaged software to mass markets. Accordingly, formal instruments for intellectual property protection have assumed much greater importance, despite the hazy and evolving legal status of these instruments. In the United States, which can be broadly categorized as an economy characterized in recent years by relatively strong protection for intellectual property rights, copyright protection for software-related intellectual property has been supplemented, and appears to have been supplanted, by patent protection. The U.S. judicial and legislative arenas have strengthened the rights of owners of intellectual property in a number of industries since 1980, including computer software. The strong protection for intellectual property provided in the United States is followed by that in Western Europe, where the European Commission has provided somewhat more lenient treatment for “reverse engineering” of software for purposes of complementary invention, and Japan, where protection for software-related intellectual property historically has been relatively weak (see Merges, 1996). These contrasting regional or national systems of intellectual property policy have evolved in parallel with the software industries in each area. Indeed, the furor over the Compton’s multimedia patent, as well as the more 52 It is possible that software firms are choosing not to register copyrights because such early registration no longer is necessary to support litigation against alleged infringers, a possibility that would indicate greater judicial deference to copyright. This possibility seems unlikely, however, in view of the more circumscribed role accorded to copyright by the federal bench since the late 1980s that we described in earlier sections of this paper.
OCR for page 254
Patents in the Knowledge-Based Economy recent controversy over business methods patents, provides additional evidence of the influence of industry-led political action on U.S. patent policy.53 Although U.S. intellectual property rights policy has influenced the development of its software industry, the reverse also is true. In other words, the relationship between the development of the domestic software industries and the intellectual property rights regimes of the United States, Western Europe, and Japan is best characterized as one of “coevolution,” involving mutual causation and influence (Nelson, 1994; Merges, 1996; Khan and Sokoloff, 2001).54 Relatively large firms in the U.S. packaged software industry are shifting toward a more “patent-intensive” approach to the protection of their intellectual property, as the largest firms increase their patent propensities. Moreover, the evidence of increased patenting is strongest for older (and, in most cases, larger) firms within the U.S. industry. We observe no tendency for entrants to seek patent protection more intensively than incumbent firms. No evidence suggests that entry by specialized packaged software firms has been curtailed by these policies, however, and much more information is needed on entry, profitability, and the long-term evolution of industry structure before such a conclusion is warranted. The analysis also highlights the fact that despite increased use of patent protection by packaged software firms, large electronic systems firms are more important in overall software patenting. A comparison of the patent propensities of IBM and Microsoft suggests that the “patent propensity gap” between these two firms narrowed during the 1990s, but IBM continues to patent more intensively, relative to its R&D spending, than does the world’s largest packaged software specialist. The limited evidence on the “importance” of the patents obtained by the largest U.S. packaged software and electronic systems firms does not support a characterization of these patents as “junk patents,” by comparison with software patents generally. Moreover, large packaged software firms appear to be 53 The filing by Microsoft of an amicus brief in Apple Computer v. Franklin Computer was noted above. The Business Software Alliance, a group enlisting Microsoft and other large packaged software firms (its members are Adobe, Apple, Autodesk, Bentley Systems, Borland, CNC Software/ Mastercam, Macromedia, Microsoft, Symantec, and Unigraphic Solutions; additional members of its Policy Council include Compaq, Dell, Entrust, IBM, Intel, Intuit, Network Associates, Novell, and Sybase), also was active during the 1990s in appearing before congressional committees and filing amicus briefs, all in favor of stronger formal protection for software-related intellectual property. 54 The endogenous nature of software-specific intellectual property rights policy, as well as intellectual property rights policies more generally, has been widely noted. In his 1996 discussion of software-related intellectual property policy in the United States, Western Europe, and Japan, Merges notes: “…[H]ow has the intellectual property regime affected the development of the software industry in these major markets? As we shall see, the answer must be incomplete unless one considers the converse question—how has the industry affected the legal regime?” (p. 275). In their discussion of the evolution of nineteenth century patent policies in the United States, Britain, France, and other nations, Khan and Sokoloff (2001) conclude that “…scholars who try to relate patterns of invention to patent system characteristics should be cognizant of, or take care in dealing with, the likelihood that those patent system characteristics are not exogenous with respect to the invention” (p. 28).
OCR for page 255
Patents in the Knowledge-Based Economy substituting patent for copyright, based on a comparison of trends in patent and copyright propensities. Computer software as a product of inventive effort is nearly 50 years old, but the application of intellectual property rights to these products is relatively recent. Although patents were originally viewed by some experts as preferable to the extensive reliance on copyright for protection of software-related intellectual property (Menell, 1989) because of the higher threshold for patent protection, the expanded use of patents to protect software-related intellectual property has also sparked controversy. The 1998 State Street Bank decision extended patent protection into the previously unexplored area of “business methods,” and growth since State Street in this class of patenting may trigger additional litigation over validity and infringement. Software patents, especially business methods patents, raise unusual challenges to the U.S. patent system, which relies on inventors and patent examiners for searches of “prior art” rather than allowing for interested parties to challenge patents before their issue in a formal pre-grant opposition process. Because of the historical lack of software patents, a primary source of software-related “prior art” scarcely exists, and this contributes to the issue of patents (such as the “multimedia” patent discussed above) of potentially sweeping breadth and limited validity. As the multimedia patent example suggests, there are few cases thus far of such broad patents being issued and upheld by either the USPTO or the courts. But the general problem is a serious one—how can searches of prior inventions be undertaken in a technology where patents have only recently become common? Innovation in software generally is a cumulative activity, and individual software products frequently build on components from other products. As a result, some industry experts argue that software developers may become aware of a related patent only after they have completed development of a new product.55 But this type of problem (which is not unique to software) is associated with the transition to a new, patent-based regime of intellectual property protection in software and may decline in severity as expanded software patenting expands the body of prior art that can be searched by patent examiners. Increased publication of patent applications after 18 months also should reduce the severity of this problem somewhat, and the liberalized “prior use” defense embodied in the AIPA also could reduce the incidence of litigation over infringement. The costs of the transition to patent-based protection of software-related intellectual property nevertheless could be high, because of the reliance on litigation to establish the validity of this growing body of prior art. The leading alternative mechanism in the 55 Dan Bricklin, a pioneer in the packaged software industry and developer of the first spreadsheet program, argues that a typical software product may involve literally thousands of patentable processes, which creates enormous hazards for independent or small firm inventors who may belatedly discover that important components of their newly developed product are in fact patented by others (Merges, 1997, pp. 119-120).
OCR for page 256
Patents in the Knowledge-Based Economy U.S. system for challenging the validity of patents, re-examination, is utilized less extensively than the opposition process in the EPO. Nonetheless, current evidence on the EPO opposition system does not suggest that this process operates quickly or cheaply to resolve questions of patent validity (see Graham et al., 2003). The computer software industry provides a fascinating “laboratory” for observing the transition from a relatively open intellectual property regime to one in which formal protection, especially patents, figures prominently. The cross-national differences in domestic patent systems, combined with cross-national differences in the structure of domestic software industries and domestic software markets, provide additional rich material for comparative studies of the interaction of intellectual property systems, innovation, and industrial development. Current research, including this chapter, has scarcely scratched the surface of this rich subject. REFERENCES Abate, T. (1993). “Smaller, Faster, Better; Tech Firms Show Off Their Latest Wonders at Trade Show and Foretell a User-Friendly Future.” San Francisco Examiner, November 21, E:1. Aharonian, G. (1993). “Setting the Record Straight on Patents.” Communications of the ACM 36: 17-18. Association of University Technology Managers. (2000). The AUTM Licensing Survey: Executive Summary and Selected Data, Fiscal Year 1999. Norwalk, CT: Association of University Technology Managers: http://www.autm.net/surveys/99/survey99A.pdf. Bresnahan, T., and S. Greenstein. (1996). “The Competitive Crash in Large Scale Commercial Computing.” In R. Landau, T. Taylor, and G. Wright, eds., The Mosaic of Economic Growth. Stanford, CA: Stanford University Press. CONTU (National Commission on New Technological Uses of Copyrighted Works). (1979). Final Report. Washington: USGPO. Graham, S. J. H., B. H. Hall, D. Harhoff, and D. C. Mowery. (2003). “Patent Quality Control: A Comparison of U.S. Patent Re-examinations and European Patent Oppositions.” In W. Cohen and S. Merrill, eds., Patents in the Knowledge-Based Economy. Washington, D.C.: The National Academies Press. Hall, B. H., and R. M. Ziedonis. (2001). “The Patent Paradox Revisited: An Empirical Study of Patenting in the U.S. Semiconductor Industry, 1979-1995.” Rand Journal of Economics, 32(1): 101-128. Hart, R., P. Holmes, and J. Reid. (1999). “The Economic Impact of Patentability of Computer Programs: A Report to the European Commission.” London: Intellectual Property Institute. Khan, B. Z., and K. L. Sokoloff. (2001). “The Innovation of Patent Systems in the Nineteenth Century: A Comparative Perspective.” Presented at the Franco-American conference on the Economics, Law, and History of Intellectual Property Rights, UC Berkeley, October. Kortum, S., and J. Lerner. (1999). “What Is Behind the Recent Surge in Patenting?” Research Policy 28: 1-28. Kuan, J. (1999). “Understanding Open Source Software: A Nonprofit Competitive Threat,” unpublished manuscript, Haas School of Business, UC Berkeley. Lemley, M., and D. O’Brien. (1997). “Encouraging Software Reuse.”Stanford Law Review 49(2): 255-305.
OCR for page 257
Patents in the Knowledge-Based Economy Lerner, J., and J. Tirole. (2000). “The Simple Economics of Open Source,” unpublished manuscript, Harvard Business School. Magrab, E. B. (1993). “Patent Validity Determinations of the ITC: Should U.S. District Courts Grant Them Preclusive Effect?” Journal of the Patent and Trademark Office Society 75(125): 127-135. Markoff, J. (1993). “Patent Office to Review A Controversial Award.” The New York Times, December 17, D:2. Menell, P. (1989). “An Analysis of the Scope of Copyright Protection for Application Programs.” Stanford Law Review 41: 1045-1096. Merges, R. P. (1996). “A Comparative Look at Intellectual Property Rights and the Software Industry.” In D. C. Mowery, ed., The International Computer Software Industry: A Comparative Study of Industry Evolution and Structure. New York: Oxford University Press. Merges, R. P. (1997). Patent Law and Policy. Charlottesville, VA: Michie. Merges, R. P. (1999). “As Many as Six Impossible Patents Before Breakfast: Property Rights for Business Concepts and Patent System Reform.” Berkeley Technology Law Journal 14: 578-615. Mowery, D. C. (1996). “Introduction.” In D.C. Mowery, ed., The International Computer Software Industry: A Comparative Study of Industry Evolution and Structure. New York: Oxford University Press. Mowery, D. C. (1999). “The Computer Software Industry.” In D. C. Mowery and R.R. Nelson, eds., The Sources of Industrial Leadership. New York: Cambridge University Press. Mowery, D. C., R. R. Nelson, B. Sampat, and A. A. Ziedonis. (2001). “The Growth of Patenting and Licensing by U.S. Universities: An Assessment of the Effects of the Bayh-Dole Act of 1980.” Research Policy 30: 99-119. Mowery, D. C., and B. N. Sampat. (2001). “University Patents and Patent Policy Debates: 1925-1980.” Industrial and Corporate Change 10: 781-814. Mowery, D. C., and T. Simcoe. (2001). “The Origins and Evolution of the Internet.” In B. Steil, R. Nelson and D. Victor, eds., Technological Innovation and National Economic Performance. Princeton, NJ: Princeton University Press. National Public Radio. (1998). “Patenting Math Formulas.” All Things Considered August 11, tran-script 98081115-212. Nelson, R. R. (1994). “The Co-Evolution of Industries and Institutions.” Industrial and Corporate Change 3: 47-64. OECD. (1998). “Measuring Electronic Commerce: International Trade in Software.” Directorate for Science, Technology, and Industry. Orenstein, S. (1994). “U.S. Rejects Multimedia Patent.” The Recorder, November 1, 4. Peltz, J. (1993). “Compton’s Wins Patent Covering Multimedia.” Los Angeles Times, November 16, D:2. Riordan, T. (1994). “Action Was Preliminary On a Disputed Patent.” The New York Times, March 30, D:7. Samuelson, P. (1990). “Benson Revisited: The Case Against Patent Protection for Algorithms and Other Computer Program-Related Inventions.” Emory Law Journal 39: 1025-1122. Softletter. (1986). “The 1986 Softletter 100.” Softletter Trends & Strategies in Software Publishing 3: 1-8. Softletter. (1991). “The 1991 Softletter 100.” Softletter Trends & Strategies in Software Publishing 8: 1-24. Softletter. (1996). “The 1996 Softletter 100.” Softletter Trends & Strategies in Software Publishing 12: 1-24. Softletter. (1998). “The 1998 Softletter 100.” Softletter Trends & Strategies in Software Publishing 14: 1-24.
OCR for page 258
Patents in the Knowledge-Based Economy Steinmueller, W. E. (1996). “The U.S. Software Industry: An Analysis and Interpretive History.” In D.C. Mowery, ed., The International Computer Software Industry: A Comparative Study of Industry Evolution and Structure. New York: Oxford University Press. Trajtenberg, M. (1990). “A Penny for Your Quotes: Patent Citations and the Value of Innovations.” The Rand Journal of Economics 21: 172-187. U.S. Bureau of the Census. (1995). Service Annual Survey: 1995. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce. (1998). U.S. Industry and Trade Outlook 1998. Washington, D.C.: U.S. Government Printing Office. U.S. International Trade Commission. (1995). A Competitive Assessment of the U.S. Computer Software Industry. Washington, D.C.: U.S. International Trade Commission. Waldmeir, P., and L. Kehoe. (1999). “E-commerce Companies Sue to Protect Patents: Intellectual Rights Given Legal Test.” Financial Times, October 25, 16.
Representative terms from entire chapter: