This study presents the results of an empirical test of two hypotheses concerning the quality of business method patents. The hypotheses are motivated by two frequently voiced criticisms of those patents: that their scope is overly broad and that they cite too little prior art. Using a sample of over 3,500 data processing, software, and internet patents granted between 1975 and 1999, I find little support for these criticisms. Rather, I find that business method patents are not broader and do not cite less prior art than comparable patents. While these findings don't completely exonerate business method patents of the charges of inferior quality, they do suggest that, at a minimum, they are no worse than comparable patents, at least along these two dimensions of quality.
"There are persistent reports that patents in the software area, perhaps especially, patents for 'business methods' implemented in software, are of extremely poor quality.
-Robert Merges, UC Berkeley Law Professor
"The burden of proof is not for the people who defend property rights, but for those who want to take them away."
-Jay Walker, founder of Walker Digital, an Internet R&D laboratory
Although patents for business methods implemented in software have been granted for a several decades (United States Patent & Trademark Office, 2001), they gained considerable notoriety and acceptance after the 1998 "State Street" decision laid to rest longstanding, and "ill-conceived" objections to them (State Street Bank & Trust Co. v. Signature Financial Group, Inc. , Fed. Cir. 1998). The court's affirmation of the patentability of mathematical algorithms performed by computers which provided "useful, concrete, and tangible" results were many. New applications for business method patents more than sextupled, climbing from 1320 in 1998 to nearly 8000 by the year 2001. There was also an sharp increase in the quantity, amplitude, and range of the concerns raised in the press (Krigel 1998; Sandburg 1999; Gleick 2000; Dorny 2001) and by legal scholars (e.g. Merges 1999; Thomas 1999; Dreyfuss 2000, 2001 ; Bagley 2001 ; Meurer 2002) about patents on methods of doing business, especially those involving the conduct of e-commerce, e.g. Amazon.com's "1-click" patent. In the spring of 2000, under mounting pressure, the United States Patent & Trademark Office (USPTO) announced a patent quality improvement initiative which incorporated many of the changes proposed by its harshest critics and its staunchest defenders (Dickinson 2000).
Impatient and distrustful of the USPTO's willingness and ability to reform the examination of business method patents, new legislation was passed which limited how patents on methods of doing business could be used against alleged infringers (e.g. American Inventors Protection Act of 1999). The Business Method Patent Improvement Act of 2000, a bill which never emerged from committee, proposed that business method patents, and only business method patents, meet new and higher statutory requirements. Also in 2000, Amazon.com founder Jeff Bezos, relenting to harsh criticism about his firm's decision to enforce it's "1-click" patent against Barnes & Noble.com, sponsored a web-site known as "Bounty Quest" which offered money to on-line sleuths to uncover examples of prior art which could be used to invalidate several well-known and many less known business method patents (Felton 2001). For many, however, these changes and recommendations were too little done too late to prevent what, for most, had become a foregone conclusion: that patents on business methods were possessed of substandard quality and would, as a result of that low quality, eventuate more harm than good for the software industry and the broader economy, introduce more rather than less subjectivity into these patents' examination, and increase the amount of litigation in this …