Comparative patent quality in the United States and Europe

I know well the difficulty of drawing a line between the things which are worth to the public the embarrassment of an exclusive patent, and those which are not.

Thomas Jefferson, 1813

The job of a patent examiner is to award patents to inventions that are novel, inventive, adequately disclosed and useful. Easier said than done. For nearly as long as there have been patents, there have been challenges in carrying out this seemingly straightforward task. The Patent Quality Review Office at the US Patent and Trademark Office (USPTO) was created in 1974 to address quality concerns, while quality has long been a priority for the European Patent Office (EPO). Patent examination as we know it today was introduced in 1836 in the United States to remedy the previous system’s defect of registering patents without applying any quality filters at all. The patent registration system that preceded it, in turn, was devised as an alternative to the examination that the nation’s first patent examiners – including then-Secretary of State Thomas Jefferson – had trouble carrying out given their demanding schedules.

A few hundred years later, patent quality concerns are once again in the spotlight. Current, past and likely future leadership at the USPTO and EPO have made patent quality a priority. However, quality concerns reverberate far beyond the patent office. Recent research has found that for a young company, getting a patent issued spurs jobs creation, sales growth, innovation and financial performance. Not getting a patent, or getting it after a delay, can do the opposite, in some industries sealing the fate of a young company. Mistakes in the other direction can be equally devastating. The assertion of dubious patents – some later invalidated for a lack of novelty decades after they were issued – has raised serious policy concerns in the United States and attracted the attention of members of Congress, government regulators and judges.

Since at least 2010, IAM has asked its readers about the comparative quality of patents produced at various offices. While admittedly unscientific, these efforts provide one of the few ways to measure, if not absolute patent quality, at least public perception of comparative patent quality. Respondents have consistently ranked the EPO most highly in terms of quality and service.

However, what has not been explored in depth are the drivers of quality and how the particulars of each patent system, each patent office and each patent practice contribute to quality. Nor have previous surveys considered what patent holders view as their top priorities when it comes to patent quality and where they believe the greatest opportunities for improvement lie.

This article discusses both the perceptions and, to some degree, the realities of patent quality across the EPO and USPTO. It reports on a survey conducted earlier this year by IAM of about 650 practitioners with experience of the USPTO and EPO, which probed various aspects of patent quality. It also draws upon my larger forthcoming study of comparative patent quality (“Comparative Patent Quality”, Ariz St Law J (forthcoming), available at papers.ssrn.com/sol3/papers.cfm?abstract_id=2833980), as well as several other empirical studies, to explore what is actually happening with respect to patent quality and how perception and reality compare.

Before doing so, it is important to acknowledge the ways in which the USPTO and the EPO are not comparable, at least currently, including size and scale, evolution of the law and the competitiveness of examiner positions in the labour market. In 2015 the USPTO received about 600,000 newly filed utility applications, to be processed by roughly 8,000 examiners. The EPO received only about one-quarter of the number of applications, although it is staffed with more than half the number of examiners. This means that the USPTO receives about double the number of applications per examiner. The law pertaining to patentable subject matter has evolved substantially over the last few years in the United States. Examiner salaries are determined according to governmental paygrades, but domestic market forces determine what other, potentially more lucrative opportunities may lure existing or prospective patent examiners outside of the organisation.

These realities shape, but are largely beyond the control of, each patent office. Perfect harmonisation between the offices is probably neither possible nor desirable, although greater coordination between them makes this a good time to consider what each can learn from the other.

With these caveats in mind, then, what did respondents say should be priorities for the USPTO and for the EPO?

US priorities: consistency and predictability

According to survey responses, the USPTO should focus on consistency and predictability in order to improve patent quality. Among the initiatives which respondents thought the USPTO should prioritise, greater consistency was the top-ranked answer (76%) and predictability came in second (56%) (Figure 1). Consistency was also the area of greatest difference with respect to the performance of the offices (Figure 2); predictability again came in second.

Figure 1. What should the USPTO prioritise as it moves to improve quality?

Figure 2. EPO v USPTO, comparative rating (USPTO ranks best in customer service, timeliness and cost)

Consistency and predictability in patent quality have multiple dimensions. Three of the most important are:

• consistency across people (ie, examiners within an art unit);
• consistency across the process (ie, limiting rework); and
• consistency in the application of the evolving patent law.

Although some have cautioned against overinvesting in patent quality when so few patents are enforced, the uncertainty which inconsistent processes inject into applicant experiences affects every single patent application and arguably those applicants which have fewer applications much more than others.

Consistency across people

Expressing the sentiment of many, one respondent said: “There is too much variability in quality of examination, and skill level in examiners.” Another pointed out that: “Examiners are not consistence (sic). So, application future depends on examiner not on subject matter.” Differences in how the USPTO and EPO evaluate obviousness ranked third highest in terms of the offices’ relative performance. The comments also drilled down to the lack of consistency in obviousness rejection: “The rating of adequate for evaluation of specification and obviousness is due to inconsistency between examiners. Some are excellent; some seem to have difficulty understanding the material.”

The academic literature has documented the variation in US examiner grant rates based on the analysis of millions of data points. In a 2011 study, Shine Tu identified two problematic examiner populations – one guilty of rubber stamping applications and the other potentially applying too harsh a filter. Recent work by Melissa Wasserman and Michael Frakes published in the Duke Law Journal has documented that much of the difference depends on when an examiner was hired, not just the merits. Indeed, the variability between examiners is perceived as so great that economists model the patent application process as a lottery or, to use the words of commenters, “a crap shoot” or “truly a mixed bag”. Services such as Patent Advisor (paid) and Examiner Ninja (free) allow applicants to engage in patent prosecution moneyball.

However, if performance is more consistent across the EPO, what contributes to this consistency? One important check is the practice of assigning three examiners to an application – two examiners and a chair – at the examination stage, to make decisions as a group. This limits the risk of outlier decisions, as the group must reach consensus before taking action. In addition, although the search process at the EPO is carried out by a single examiner, he or she must consult with other examiners if his or her search report is positive, to avoid improperly raising the expectations of the applicant. In the United States, an application is also assigned to an art unit or group of examiners. However, the examination is largely the responsibility of a single examiner, under the guidance of a supervisory patent examiner.

The comments reflect the sense that EPO patent examiners are more rigid in their rejections, to the detriment as well as benefit of applicants. As one comment relayed: “The EPO has high-quality examiners, but they are too much sticklers on literal wording of applications and minor word differences between priority documents.” Several others complained about the support requirements, with the EPO taking “too hard a line on requiring near-verbatim spec support for claims”, leading to what the respondent perceived as “massive inefficiency in terms of completely worthless redundant disclosure to meet technical rules that serve no real purpose”. However, another suggested that while EPO examiners “stick to their guns”, this makes a patent “harder to obtain but also more reliable. Also the continuation system is not as subject to abuse”. The use by US examiners of “ready-make texts from the MPEP” and “boiler-plate content which is not specific” was also cited as unhelpful for patent quality.

The EPO is also proud of its ISO 9001 quality certification, which requires it to have procedures for detecting and dealing with ‘non-conforming products’. One practice: when the patent is challenged, the original examination team which issued the patent participates in the opposition and can learn from the prior art which surfaces.

The USPTO’s Enhanced Patent Quality Initiative (EPQI) is poised to emulate some of these best practices. The USPTO’s Master Review Form – a single, comprehensive tool to evaluate examiners according to a uniform set of metrics – should enable examiners and others to become more aware of the differences within art units and for corrective action to be taken. The agency has also rolled out a Post-Prosecution Pilot (P3) which, among other things, provides examiners with prior art and other materials submitted during trial proceedings. Based on statistics presented by Deputy Commissioner Valencia Martin-Wallace at the September patent quality conference at Santa Clara University, in almost half of child cases, the examiner referred to prior art provided in the trial petition of the parent case. These initial results are promising and can be built upon.

Consistency across processes and across patent law

Much of the frustration with respect to inconsistency at the USPTO was directed at the process of examination, rather than the examiners themselves, with prior art often being revealed in stages rather than comprehensively at the beginning. In the words of one commenter: “Examiners often engage in piecemeal examination. Often multiple RCEs are required which is costly… Once amendments are made to overcome [a rejection,] often new art, which is not necessarily more relevant, is cited. Sometimes the process goes smoothly, other times (for no apparent reason) it is very drawn out.”

The piecemeal nature of US examination, due to the practice of continuing applications, was echoed in a number of comments. Another added: “It appears that issues could have been terminated in the first round, but the examiner chooses to dig up more prior art necessitating a second or third round – almost as if the examiner avoided making a decision.”

As I have argued elsewhere, the drawn-out nature of the patent process creates a greater tolerance for mistakes on the front end, as applicants know that they can circumvent a poor examiner or rejection by filing for a request for continued examination or continuation. As one respondent put it: “The unnecessarily high costs are the result of the extremely poor evaluation of obviousness and the expense connected with the continuation practice (which is per se a pain and should be abolished altogether).”

The contrast with European practice is clear here – Europe follows a ‘once and done’ model, in which the initial decision to grant or deny is more or less final; whereas in the United States, a final rejection is often followed by additional prosecution. However, the EPO’s curtailed process is arguably supported by consistency on the front end – applicants are less incentivised to continue prosecution because they know that the treatment that they will get will not vary; whereas in the United States, continuation practice provides a safeguard against poor-quality or inconsistent examination.

Another major source of frustration among respondents was Section 101. While some comments were directed at the examiners’ inconsistent application of eligibility case law, others placed blame on the Supreme Court: “It is impossible for the examiners to do a good job when the Supreme Court has issued so many irrational patent decisions.” Among respondents to an earlier IAM survey from this year (see “Hard times in the global patent market; but hope too, in places”, IAM Issue 78, May 23 2016) (Figure 3) who found patent quality problematic (59%), 47% indicated that uncertainty around eligibility was a leading factor, second only to patent office staffing (50%). While the USPTO has taken proactive steps and conducted training geared at ensuring consistency in the application of Section 101 across art units and technology centres, the EPO is perceived as “advantaged by [a] relatively much slower pace of change in laws and less vague underlying laws”.

Figure 3. The main factors considered to be negatively affecting current patent quality (private practice)

Within this environment, it remains to be seen how far the USPTO’s current interventions with respect to consistency and predictability can go. For example, the USPTO’s new post-prosecution pilot initiative enables resolution of cases after final rejections are issued, but before a new cycle begins and has received accolades. However, for the reasons discussed earlier, the USPTO’s efforts to encourage compact prosecution would seem to fall short unless accompanied by an increase in quality and consistency on the front end. If the desire is to effect a shift in the default towards a compact, once and done model, the USPTO will likely need to substantially change the fees, if not the law.

Likewise, the level of consistency among examiners is a product of their incoming skills, tenure and turnover. At the EPO, patent quality is viewed largely as a by-product of strong human capital and training as incoming examiners must have a master’s degree in science or engineering. Technical competence was mentioned in several comments, with some describing USPTO examiners as being “excellent”, while others remarked that a “lack of confidence and understanding of the technology on the examiner’s part result in some weird applications/combinations of prior art”, and that “Examiners often do not understand the technology presented to them”. In addition to having graduate-level science or engineering training, EPO examiners must be able to work in three languages. It takes four years to fully train an EPO examiner, who receives 51 days of classroom training in the first two years. The EPO has had lower turnover, which increases consistency. The USPTO also boasts a strong training programme, and recent efforts to increase the technical training that examiners receive from practising scientists and engineers in the field should improve the baseline of knowledge. However, in contrast to the EPO, the USPTO has, over time, had to grow its ranks relatively rapidly in order to match growing applications, and operates in a job market in which it must compete for technical talent with start-ups, technology companies, law firms and other prospective employers.

EPO priorities: cost, timeliness

Turning the tables, the EPO is perceived to lag behind the USPTO in several respects – primarily cost and timeliness. In fact, the greatest difference in quality ratings between the two was based on cost differences (Figure 4). What is the difference? In the EPO, the search fee is €1,300 and the examination and designation fee is €1,635 – a total of €2,935. In the United States, examination costs are lower, although not significantly – the application fee is $1,600 and the issue fee is $960. However, in the EPO, annuities start at year three from filing and are levied every year, rising from €470. The USPTO, in contrast, requires maintenance fees only every three and a half years, and these begin only at three and a half years post-issuance, starting at $1,600 per patent. Many of the comments were directed at what respondents perceived as too high annuity and divisional fees.

Figure 4. Reasons for withdrawal of EPO applications

In fact, relatively low fees have been a deliberate feature of US prosecution from the beginning of the US patent system; and indeed, per gross domestic product, US fees are the lowest in the world (Figure 5). In 1790, fees for the US patent system were set below UK fees, consistent with the idea of a democratic patent system, which contrasted with the system of privileges established by the British monarchy. For the first half of the 19th century, fees in the United States remained lower than those in most European countries.

Figure 5. Patent costs relative to market size (total fees per $billion GDP)

Source: Colleen V Chien, “Comparative Patent Quality”, Ariz St Law J (forthcoming), available at papers.ssrn.com/sol3/papers.cfm?abstract_id=2833980.

It is also the case that in Europe, fees perform a filtering function. In related research, I and others have found that patent applications filed at both the EPO and the USPTO are more likely to become US patents than EPO patents. Upon investigation, it appears that the reasons for doing so are counterintuitive – rather than EPO examiners rejecting patent applications at a higher rate, applicants are withdrawing their applications at a higher rate. To probe applicants’ motivations for withdrawing, the survey included a question asking those who had withdrawn their EPO applications why they did so. The majority (59%) chose cost considerations, while substantive considerations – in particular, the projected patentability of the application, based on the search or examination report – were viewed as relatively less important (27% and 33% respectively).

The finding that the top reason for withdrawing an EPO application is that it is too expensive is striking. How to interpret it is less clear. On the one hand, the sensitivity of applicants to cost factors means that the fees are working as intended – to weed out less valuable applications. On the other hand, if cost sensitivity is more acute among small and medium-sized enterprises, EPO fees – which are not currently calibrated to the size of the enterprise, as they are in the United States – may be acting as a regressive filter, discouraging participation in the patent system by smaller and less well-funded entities.

The other area of comparatively worse performance at the EPO pertains to timeliness. Comments such as the following were typical: “Timeliness is good when requesting accelerated examination (PACE); otherwise terrible” and “[m]ost examiners are excellent. But too late to get a patent”. In one sense, the relatively slower examination in the EPO is unsurprising, as prosecution proceeds – as it does in most of the rest of the world – in distinct stages. In contrast, search and examination are combined in the USPTO.

In fact, based on recent research carried out by Jay Kesan and colleagues, the gaps between the USPTO and EPO are narrowing with respect to pendency. The researchers considered patents that had been filed in both jurisdictions, with priority dates ranging from 2002 to 2008, and corrected for the delays that are introduced into the EPO process by the bifurcation of search and examination. While EPO applications based on 2002 filings (based on the filing date of the US application) tended to pend for between two and four years more than their US counterparts, for 2008 EPO applications (based on the filing date of the US application), the difference in pendency narrowed dramatically (regardless of whether priority was claimed in the USPTO and EPO), to within a month or so (Figure 6).

FIGURE 6.Trend of average pendency

However, another problem mentioned by commenters was the lack of predictability with respect to when actions were expected and the requirement to pay annuities in the interim. As one commenter relayed: “EPO’s examination process is very slow, and applicant cannot predict when the office action is mailed. After applicants reply to the office action, sometimes the next office action could be mailed within a few months. But the next office action could not be mailed after several years. But the applicant must still pay the annuity for application annually without receiving the next office action. And there is no system for saving lost and delayed time during examination.”

In fact, applicants are informed when examiners have started to work on an application a month or two in advance of the issuance of the search report or the first communication from the examination division, with the search report deadline set at six months after an application is received. Examiners are also supposed to issue an office action four months after receipt of the examination request or the applicant’s submission – although compliance with this requirement is less than complete.

Investing early in quality

The USPTO’s quality initiative is targeted along the entire lifecycle of a patent application – from the initial application search to the post-grant phase. The survey asked respondents which, among existing USPTO initiatives, deserved the most attention. Initiatives at the front end of the patent process – namely search and training initiatives such as automated search and awareness of the Science and Technical Information Centre, and training to improve the clarity and reasoning of office actions – received the most votes (47%). Prosecution-focused interventions – including clarity of the record initiatives and interview specialists, on which the USPTO has arguably focused most of its energy – received the second-highest number of votes. Of lower priority to respondents were evaluation and post-grant enhancements, each rated as important by 7% of respondents (Figure 7).

Figure 7. Which existing initiatives should the USPTO prioritise?

Search, and in particular the consideration of non-patent literature (NPL), is viewed as an area of comparative strength for the EPO (Figure 2). In the EPO, search is described as the “cornerstone” of the examination process, with the agency devoting as much as 80% of its resources to it. The EPO claims to have the largest collection of documents available for prior art searching and hires employees who can read and apply documents in at least three languages. Because search is carried out once, examiners search the specification rather than searching the claim terms. The search report is prepared with the goal of enabling the applicant to rely on the report and opinions contained therein to build a strategy for future prosecution at the EPO and other offices. Indeed, Batistelli has claimed that a large proportion of applicant withdrawals follow the search report, although the findings described earlier show that cost considerations are also strongly driving withdrawals.

In related research, I have explored the differences in NPL citation rates of the two – as measured by the difference between whether any NPL is included in the EPO examination report and whether a US examiner relies on NPL in his or her examination (Figure 8). The differences are dramatic – across all industry areas tested and among applications filed in both jurisdictions, EPO search reports were much more likely to include NPL in their examinations. Although NPL is often provided by applicants to the office, previous work by Mark Lemley and Bhaven Sampat confirms what respondents intuitively know: that examiners rely on their own searches, primarily of patent references. However, given the tools and the time, US examiners would likely cite a more diverse set of references. Starting last year, Google’s Prior Art Search tool combines both patent and non-patent references to enable searching across both types of document. Companies including Microsoft, IBM, SAS and Cisco have followed the lead of the financial services industry and moved to make their non-patent prior art available for searching by the USPTO in response to White House and USPTO executive action. Cisco has led the way in developing a database which is available both to the public and to USPTO examiners.

Figure 8. The PTAB is more likely than the EPO or USPTO to cite non-patent literature in its decisions

What next?

This article does not presume to provide a silver bullet to a problem that is as old the US patent system – how to best ensure patent quality. However, it is clear that what the public expects and desires is a quality product, which is produced through a reliable, predictable and affordable process and can withstand legal challenges after it is issued. As the EPO and USPTO continue to work on various quality initiatives, it is important that they design and assess their performance against these metrics and communicate their progress along the way.