The America Invents Act, signed into law

patents Funding and performance at the US Patent and Trademark Office Deepak Hegde Despite recent patent law reforms, the US Patent and Trademark Office s ability to deal with inefficiencies in patent examination will continue to rely on the annual Congressional appropriations process. npg 2012 Nature America, Inc. All rights reserved. The America Invents Act, signed into law on September 16, 2011, enacts sweeping changes to the US patent system 1. Besides switching US patent priority from the present first-to-invent system to a first-to-file system, expanding post-grant review options and prioritizing examination for technologies important to US competitiveness, the Act provides for the creation of a US Patent and Trademark Office Public Enterprise Fund in which to deposit the fees paid by users (patent and trademark applicants) of the US Patent and Trademark Office (USPTO). The fund would be set aside for the USPTO s exclusive use so that Congress could not appropriate the deposits for other purposes. Another important change gives the USPTO authority to set its user fees in order to recover the cost of its operations. These last two changes seek to give the USPTO more control over its revenues so that the agency can deal with mounting concerns about the speed and quality of patent examination 2. A timely and accurate patent examination is critical for resolving uncertainty about inventors property rights and promoting innovation. However, as noted by a recent Nature editorial, USPTO s authority to set fees and draw from the fund remains circumscribed by Congressional oversight, and according to some critics, the changes are insufficient to deal with concerns about patent examination 3,4. What challenges does the USPTO face in a rapidly changing economic, innovation and intellectual property (IP) environment? How does the Congressional appropriations process Deepak Hegde is at New York University, Stern School of Business, New York, New York, USA. e-mail: dhegde@stern.nyu.edu Amount ($ billions) 2.1 1.9 1.7 1.5 1.3 1.1 0.9 0.7 0.5 1991 1993 1995 1997 Appropriations Revenue from user fees Budget request Figure 1 Appropriations, revenues (from user fees) and budget requests at the USPTO during FY1991 2010 (in 2010 billion $). Source: USPTO Performance and Accountability Reports. influence the agency s ability to deal with these challenges? And what are the implications of the recent enactment to the relationship between the Congress and the USPTO, and the agency s responsibilities as the steward of an increasingly global innovation and IP landscape? This article addresses the above questions by providing historical facts related to the USPTO s funding and performance. I focus on the speed and quality of patent examination as measures of the agency s performance 5. I analyze data from 1991 onwards because the USPTO was obligated under the Omnibus Reconciliation Act of 1990 to fully cover the costs of its operations including prior art search, patent and trademark examination, and administration through the fees it charges users with no funding from US taxpayers. 1999 2001 2003 2005 2007 2009 Funding the USPTO Each February the USPTO, in consultation with the President s administration, presents its budget request for the upcoming fiscal year, which runs from October through September, to Congress. These requests are based on the USPTO s forecasts of revenues (from user fees) and operational costs (expenditures on wages, rents and equipment) for the upcoming fiscal year. Congress, with input from the Congressional Budget Office about USPTO s requests, appropriates funds that become available for spending at the start of the fiscal year. Figure 1 shows that USPTO s budget requests (based on forecast revenues) were below actual revenues for the fiscal years between 1993 and 2002, and above revenues between 2003 and 2009. However, for 15 of the 20 years between 1991 and 2010, appropriations 148 volume 30 number 2 february 2012 nature biotechnology

patents npg 2012 Nature America, Inc. All rights reserved. to the USPTO have been lower than its revenues from user fees. The USPTO does not have the authority to spend fee collections that exceed its appropriations for the fiscal year, even when those fees are for additional work that the agency must perform. These additional revenues from its users have been unavailable to the USPTO Congress diverted $1.04 billion, or 4% of the agency s revenues during 1991 2010, to the Treasury s general fund for other purposes. Challenges at the USPTO Patents are designed to encourage innovation by awarding inventors the right to exclude others from infringing on their inventions. Delays in patent examination can hinder innovation by delaying inventors investments in product development until their property rights are clarified, and by distorting competitors investments in R&D for fear that their inventions may infringe on pending patents. One measure of delay in patent examination is patent pendency the time between the filing of a patent application and its grant or abandonment. Alternatively, first-action pendency measures the time between the filing of an application and the first action, or the examiner s formal communication to the applicant of the preliminary decision regarding the patentability of an application. Unlike patent pendency, which includes the time applicants take to respond to examiners questions, first-action pendency discounts delays caused by applicants and, thus, more accurately reflects delays caused by the USPTO s patentexamination process. Average patent pendency at the USPTO nearly doubled from 18.2 months in 1991 to 35.3 months in 2010. During the same period, average first-action pendency nearly quadrupled, rising from 7.6 months to 25.7 months Number of patents (thousands) 1,300 1,100 900 700 500 300 100 1991 1993 1995 1997 Issued Pending Filed Awaiting action by examiner (Supplementary Table 1 displays pendency statistics for each year between 1991 and 2010). In comparison, the European Patent Office (EPO) took six months for first action and 43 months for a final decision in 2009; at the Japanese Patent Office (JPO), total pendency ranges from five to six years. Delays are longest for technologies characterized by short lead times and rapid innovation cycles. Average patent pendency is 40 months at the USPTO s Computer Architecture, Software, and Information Security technology center and 43 months at its Communications technology center (Supplementary Table 2 displays pendencies at the eight different technology centers at the USPTO). Why does patent examination take longer now than it did two decades ago? Figure 2 shows that patent applications at the USPTO have surged from 178,000 in 1991 to 509,000 in 2010. However, applications have not grown at a steady rate. For instance, the number of incoming applications jumped 17.4% in 1995, 15% in 1997, 8.2% in 1998, and 12% in 2000, well above the average annual growth rate of 5.8% during 1991 2010. Trends in the technology mix of applications have also been volatile: for example, applications at the USPTO s Biotechnology and Organic Chemistry technology center jumped 13.8% from 2001 to 2002 (compared to an average annual growth rate of 4% during 2001 2010) (Supplementary Table 3 displays the number of patent applications at the different technology centers at the USPTO for each year between 2001 2010). Scholars have offered many explanations for the recent growth in patenting, including: the 1990s boom in information and communication technologies, a change in the term of US patents in 1995 (from 17 years after issue date to 20 years after filing date), a 1998 court decision that upheld the patentability of business 1999 2001 2003 2005 2007 2009 Figure 2 Patents awaiting examiner action, pending, filed and issued at the USPTO during 1991 2010 (in thousands). Source: USPTO Performance and Accountability Reports. method inventions, and procedural changes to continuation applications in 1995 and 2000 that made it easier for applicants to request continued examination of their rejected applications 6,7. Although the USPTO uses sophisticated forecasting models to calculate budget requests, increases in applications due to judicial, legislative and technological shocks are impossible to accurately forecast. Also, because the USPTO submits its requests to Congress in February, and Congress enacts appropriations in October, appropriations are based on forecasts that are ten months old a lag that amplifies the effect of unavoidable errors in forecasting. And as explained above, Congress has diverted a large portion of the agency s revenues from fees for other purposes. Therefore, in years when the application rate was higher than forecast, the USPTO could not use the revenues from additional fee collections to deal with the unanticipated flow of applications, resulting in a mounting backlog of unexamined applications. The USPTO s response to the unanticipated growth in applications was further hampered by: (i) a growing demand in the 1990s for private-sector scientists and engineers that increased examiner attrition and made it more difficult for the agency to hire replacements, especially in emerging technological areas (such as business methods and biotech); (ii) new examiners requiring at least three years of training before they can process applications at full capacity and, thus, replace examiners that leave (Supplementary Table 4 shows that examiner attrition peaked during the years when pending patent applications were experiencing above-average growth rates and that the annual growth rate of examiners has been significantly lower than the growth rate of pending patent applications during 1991 2010); and (iii) the lack of borrowing authority to finance multi-year capital needs (such as IT infrastructure) required to deal with the near doubling of pending applications between 1997 and 2001. Hence, unanticipated increases in patent applications and inadequate resources at the USPTO to deal with the increases, precipitated by Congressional diversions of the agency s revenues during the last 20 years have contributed, at least in part, to the explosion of pending applications and delays in patent examination. Several scholars argue that the increasing burden on examiners has led to a decrease in the quality of patent examination at the USPTO 8. An incomplete patent examination process can harm innovation by creating uncertainty about property rights, distorting investments in R&D and forcing inventors to nature biotechnology volume 30 number 2 february 2012 149

patents npg 2012 Nature America, Inc. All rights reserved. incur substantial litigation expenses to either defend or challenge patents 9. Although no single metric captures the quality of patent examination, quality can be thought of as the probability that an examiner s decision either to grant a patent or to reject an application survives closer scrutiny. To gauge the outcomes of scrutiny, I assembled data on the probability that the US courts and the Board of Patent Appeals and Interferences (BPAI) ruled against examiners decisions regarding patentability. Only a minority of lawsuits are eventually settled by the courts (a majority of the cases reach out-of-court settlements), and among those that were, the courts ruled against the USPTO s decision to allow patents in 17.4% of the cases, on average, between 1993 and 2009 (Supplementary Table 5 presents data on the number of patent-related lawsuits, court decisions and outcomes). However, litigation data do not reveal a systematic trend suggesting either an increase or a decrease in the quality of patent examination during the period of this study. The BPAI, in response to ex parte appeals, reversed, on average, 38.6% of examiners decisions to reject applications, but these data also do not provide evidence for either an increase or a decrease in the quality of examination during the last two decades (Supplementary Table 6 reports trends in disposals of ex parte appeals by the BPAI). I also found no evidence to suggest that owing to increasing workloads, patent examiners are taking less time to examine patents or that they have become overly generous in allowing patents. On the contrary, the average time spent by an examiner on an application increased from 18 h in 1991 to 22.4 h in 2010, and patent-allowance rates dropped from 65.5% in 1991 to 45.6% in 2010 (Supplementary Table 7 displays statistics for the number of allowances and examiner s time spent in prosecution for each of the years during 1991 2010). Hence, data on various aspects related to quality do not support the view that the quality of examination has declined during the last two decades 10. This finding, however, does not undercut the importance of an accurate examination system for promoting innovation. Policy implications How can the USPTO deal with concerns about inefficiencies in its patent examination system and be responsive to an ever-changing innovation and IP landscape? Motivated by the recent legislative change, I have focused on how the annual appropriations process may have constrained the USPTO s timely response to judicial, legislative and technological shocks that affected its workload in the last two decades. The America Invents Act appears to give the USPTO greater control over its revenues from user fees, and hence some leeway to deal with future shocks, but the Act retains Congressional oversight of the agency s purse. The user fees collected by the USPTO reflect a complex balance between the costs to the public of granting monopoly rights and the benefits of encouraging innovation, and cannot be solely tied to the quality of service offered by the agency to its users. Yet, a slow and questionable patent examination process, at arguably the world s most influential patent office, not only harms users, but imposes a significant drag on innovation in the United States and beyond. Hence, Congress has a responsibility to ensure that the USPTO is not stifled in its ability to deal with concerns about inefficiencies in its patent examination process due to vagaries in its workload 11,12. This is particularly important now, as the USPTO has made a commitment to overhaul its procedures, and these changes, when implemented, are likely to affect the agency s workload in unexpected ways. More control of its finances alone, however, is unlikely to ameliorate the quality and timeliness of the USPTO s patent examination process in the long run. A rough calculation suggests that even if workloads were constant at current levels, the agency would need an additional sum (over and above the funding required to process the new applications) of $450 million (or nearly 23% of its annual appropriations) over the next three years to bring down first-action pendency to 1991 levels (7.6 months). The USPTO needs to re-examine its internal procedures and practices, such as the application-continuation process and the wages and performance incentives of examiners, which may be contributing to inefficiencies in patent examination. And with foreign inventors share of utility patent applications at the USPTO rising from 46% in 1991 to 51% in 2009, the sharing of patent examination work (such as prior-art search) with other international patent offices might also shield the USPTO from vagaries in its workload. Note: Supplementary information is available on the Nature Biotechnology website. ACKNOWLEDGMENTS D.H. is grateful to Stuart Graham for facilitating this study. The author also thanks Dianne Buie, Dana Colarulli, Todd Dickinson, Talis Dzenitis, David Fitzpatrick, Bronwyn Hall, Bruce Lehman, Alan Marco, Gerald Mossinghoff, Christopher Pearson, Michelle Picard, Tom Stoll and Gerard Torres for helpful discussions and suggestions. Part of this research was conducted when D.H. was a visiting scholar at the Chief Economist s Office in the USPTO. The analysis and recommendations in this manuscript, however, are solely D.H. s and do not necessarily reflect the views of the USPTO. All data reported in this paper are archived in the annual USPTO Performance and Accountability Reports (http://www.uspto.gov/web/ offices/ac/ido/oeip/taf/ann_rpt_intermed.htm). COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Leahy-Smith America Invents Act, H.R. 1249, August 23, 2010. <http://www.opencongress.org/bill/112- h1249/text> 2. Leahy, P. The America Invents Act: promoting American innovation, creating American jobs, growing America s economy. <http://leahy.senate.gov/imo/media/doc/ PRESS-Summary-OnePager-FINAL.pdf> (23 August 2010). 3. Anonymous. Nature 477, 249 250 (2011). 4. Quinn, G. Lack of commitment on PTO funding is killing patent reform. IP Watchdog <http://ipwatchdog.com/2011/06/22/lack-of-commitment-onpto-funding-is-killing-patent-reform/id=17813/> (22 June 2011). 5. Bessen, J. & Meurer, M.J. Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk (Princeton University Press, 2008). 6. Hall, B.H. J. Technol. Transf. 30, 35 48 (2005). 7. Hegde, D., Mowery, D.C. & Graham, S.J. Manage. Sci. 55, 1214 1226 (2009). 8. Jaffe, A.B. & Lerner, J. Innovation and Its Discontents: How Our Broken Patent System is Endangering Innovation and Progress, and What to Do About It (Princeton University Press, 2004). 9. Shapiro, C. Berkeley Technol. Law J. 19, 1017 1018 (2004). 10. Lemley, M.A. & Sampat, B.N. Emory Law J. 58, 415 427 (2008). 11. Committee on Intellectual Property Rights in the Knowledge-Based Economy, National Research Council. A Patent System for the 21st Century (National Academies Press, Washington, DC, 2004). 12. Anonymous. US Patent and Trademark Office: Transforming to Meet the Challenges of the 21st Century (National Academy of Public Administration, Washington, DC, 2005). 150 volume 30 number 2 february 2012 nature biotechnology

Supplementary Online Material Funding and Performance at the U.S. Patent and Trademark Office Deepak Hegde 1 This file includes Supplementary Tables 1, 2, 3, 4, 5, 6 and 7 Supplementary Table 1: Average total pendency and first action pendency at the USPTO in months, 1991-2010 Year Average First Action Pendency Average Total Pendency 1991 7.6 18.2 1992 7.8 19.1 1993 7.6 19.5 1994 7.7 19.0 1995 8.4 19.2 1996 10.5 20.8 1997 11.0 22.2 1998 12.6 23.8 1999 13.8 25.0 2000 13.6 25.0 2001 14.4 24.7 2002 16.7 24.0 2003 18.3 26.7 2004 20.2 27.6 2005 21.1 29.1 2006 22.6 31.1 2007 25.3 31.9 2008 25.6 32.2 2009 25.8 34.6 2010 25.7 35.3 Notes: The Table displays average First Action Pendency (time between application filing and examiner s first action) and average Total Pendency (time between application filing and patent issue or abandonment) during the years 1991-2010 at the USPTO. Source: Workload Tables under the USPTO s annual Performance and Accountability Reports available at: 1 New York University; Email: dhegde@stern.nyu.edu 1

Supplementary Table 2: Pendency statistics for different technology centers at the USPTO in months, 2010 Average First Average Total Technology Center Action Pendency Pendency Biotechnology & Organic Chemistry 22.8 36.0 Chemical and Materials Engineering 25.7 37.4 Computer Architecture, Software & Information Security 29.3 42.5 Network, Multiplexing, Cable & Security 27.5 42.7 Communications 32.0 42.9 Semiconductor, Electrical, Optical Systems & Components 20.7 30.6 Transportation, Construction, Agriculture & E-Commerce 25.2 35.7 Mechanical Engineering, Manufacturing & Products 27.7 38.4 Notes: The Table displays average First Action Pendency (time between application filing and examiner s first action) and average Total Pendency (months between application filing and patent issue or abandonment) during the year 2010, for the nine technology centers at the USPTO. Source: Workload Tables under the USPTO s annual Performance and Accountability Reports available at: 2

Supplementary Table 3: Number of Utility patent applications at the USPTO s Technology Centers, 2001-2010 Technology Center 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Biotechnology & Organic Chemistry 36,590 41,641 38,613 38,164 38,994 41,756 44,971 47,805 45,387 47,384 Chemical and Materials Engineering 49,122 49,636 49,585 49,334 50,329 57,368 56,477 57,033 57,439 62,100 Computer Architecture, Software & Information Security 39,352 29,475 29,388 34,653 38,648 44,425 49,691 60,693 40,355 38,236 Network, Multiplexing, Cable & Security - - - - - - - - 30,207 43,895 Communications 42,920 42,144 41,528 48,210 58,445 65,974 70,637 73,301 63,981 57,138 Semiconductor, Electrical, Optical Systems & Components 72,203 74,012 75,437 81,144 88,221 94,851 95,009 99,150 92,430 92,458 Transportation, Construction, Agriculture & E-Commerce 34,944 44,641 45,306 47,489 49,776 51,661 54,772 58,834 58,630 61,386 Mechanical Engineering, Manufacturing & Products 50,950 52,139 53,595 56,533 59,815 63,725 70,080 71,853 72,495 78,886 Notes: The Table displays the number of patent applications at the nine technology centers at the USPTO for the fiscal years 2001 through 2010. TC s occasionally undergo reorganizations due to which the examination of patents belonging to one technology center may be assigned to another. For example, the examination of patents belonging to Business Methods moved from Computer Architecture, Software & Information Security to Transportation, Construction, Agriculture & E-Commerce in FY2002. The Technology Center "Network, Multiplexing, Cable & Security" was formed in FY2009 from art units in the Technology Centers Computer Architecture, Software & Information Security and Communications. Source: Workload Tables under the USPTO s annual Performance and Accountability Reports available at: 3

Supplementary Table 4: Number of Patent Examiners and attritions at the USPTO, 1991-2010 Patent Attrition Year Attritions Examiners Rate 1991 1835 210 11.44% 1992 1898 166 8.75% 1993 1979 131 6.62% 1994 2036 161 7.91% 1995 2162 162 7.49% 1996 2366 190 8.03% 1997 2329 239 10.26% 1998 2785 259 9.30% 1999 3205 375 11.70% 2000 3143 437 13.90% 2001 3296 263 7.98% 2002 3803 250 6.57% 2003 3850 241 6.26% 2004 3959 336 8.49% 2005 4471 425 9.51% 2006 5150 510 9.90% 2007 5800 543 9.36% 2008 6417 583 9.09% 2009 6585 415 6.30% 2010 6579 281 4.27% Notes: The Table displays the number of patent examiners at the USPTO and the number of attritions for the years between 1991-2010. The counts include both patent examiners and supervisory patent examiners (SPEs). Source: USPTO s annual reports available at: 4

Supplementary Table 5: Lawsuits related to U.S. patents, rulings, and reversals, 1993-2009 Total Total Reversals Reversal Year Law Suits Rulings or Remand Rate 1993 1,047 111 14 12.61% 1994 1,066 92 15 16.30% 1995 1,081 87 14 16.09% 1996 1,233 81 25 30.86% 1997 1,530 82 10 12.20% 1998 1,635 44 9 20.45% 1999 1,602 83 25 30.12% 2000 2,036 42 7 16.67% 2001 7,290 83 7 8.43% 2002 4,334 67 19 28.36% 2003 2,897 55 5 9.09% 2004 2,613 63 7 11.11% 2005 2,115 65 11 16.92% 2006 2,205 66 13 19.70% 2007 4,927 62 11 17.74% 2008 4,512 60 7 11.67% 2009 6,111 82 15 18.29% Notes: The second column of the Table displays the number of notices for lawsuits filed in selected US courts (US District Courts and US Courts of Appeals) for the corresponding fiscal year. Most of the lawsuits are settled out-of-court and the Courts issue final judgments on only a small proportion of the lawsuits. Column 3 displays the number of decisions issued by the Courts during the corresponding fiscal year. Column 4 displays the number of decisions that either reversed the USPTO s decision on patentability or remanded the Agency (for granting the patent). Column 5 displays reversal rates calculated as the percentage of reversals or remands out of the total. Source: Workload Tables under the USPTO s annual Performance and Accountability Reports available at: 5

Supplementary Table 6: Ex parte appeals and outcomes at the Board of Patent Appeals and Interferences, 1991-2009 Year Disposals by the BPAI Affirmations Affirmations in Part Reversals & Remands Dismissals 1991 4405 2514 541 1322 28 1992 3870 2264 404 1188 14 1993 3719 2082 416 1198 23 1994 2863 1468 365 1013 17 1995 2322 1123 300 868 31 1996 2138 1050 272 781 35 1997 2213 936 299 942 36 1998 3350 1530 418 1296 106 1999 3656 1335 505 1668 148 2000 4145 1476 533 1930 206 2001 3938 1504 456 1876 102 2002 4114 1530 466 2065 53 2003 3517 1359 406 1704 48 2004 3234 1245 408 1480 101 2005 2851 1145 348 1259 99 2006 2932 1304 369 1180 79 2007 3719 2126 490 1052 51 2008 5090 2945 731 1315 99 2009 6794 3733 1008 1905 148 Notes: The second column of the Table displays the number of disposals of ex parte appeals at the Board of Patent Appeals and Interferences (BPAI) for the corresponding fiscal year. The Board of Patent Appeals and Interferences (BPAI) is an administrative law body of the United States Patent and Trademark Office (USPTO) that hears appeals against examiners decisions to reject applications (either rejections of entire applications or two or more claims in the applications). Column 3 displays the number of decisions by the BPAI that affirmed the examiner s decision. Column 4 displays the number of decisions by the BPAI that partially affirmed the examiner s decision. Column 5 displays the number of decisions by the BPAI that either reversed the examiner s decision, or remanded the examiner. The last column displays the number of appeals dismissed by the BPAI. Source: These data were collected from the USPTO s Patent Application Location and Monitoring (PALM) system. 6

Supplementary Table 7: Allowance rate of patents and hours per balanced disposal at the USPTO, 1991-2010 Allowance Hours per Year Rate balanced disposal 1991 65.50% 18 1992 63.40% 18 1993 63.10% 18.6 1994 62.20% 18.8 1995 61.60% 19 1996 67.50% 19.2 1997 68.80% 19.3 1998 70.40% 19.6 1999 70.80% 20 2000 70.90% 19.8 2001 69.70% 19.8 2002 66.00% 20.4 2003 66.20% 20.5 2004 62.50% 20.8 2005 58.70% 21.9 2006 53.60% 21.7 2007 50.90% 22.1 2008 44.20% 21.7 2009 41.30% 21.6 2010 45.60% 22.4 Notes: The second column of the Table displays patent allowance rates at the USPTO for the years 1991-2010. Allowance rates are the percentage of applications reviewed by examiners that are allowed. The third column displays the number of hours consumed by an average examiner at the USPTO to achieve a balanced disposal. A balanced disposal is calculated by summing up the total number of first actions on the merits and the total number of disposals (allowances, abandonments and Examiner's Answers on Appeal) and dividing by 2. Source: Workload Tables under the USPTO s annual Performance and Accountability Reports available at: 7