ELF Working Paper: # 11
www.elf.unige.ch
Language policy and innovation.
An evaluation of the language regime of the PCT system
Michele Gazzola
Table of Contents
Abstract .......................................................................................................................................................... 2
1. Introduction ........................................................................................................................................... 3
2. The Languages of the PCT...................................................................................................................... 4
3. Language Policy and Evaluation ............................................................................................................ 8
4. Evaluating the Fairness of the PCT Language Policy ........................................................................... 10
5. Evaluating the Cost-effectiveness of the Language Policy of the PCT ................................................. 14
6. Conclusions and Policy Implications .................................................................................................... 17
7. References ........................................................................................................................................... 19
8. Appendix .............................................................................................................................................. 22
PROVISIONAL DRAFT
PLEASE DO NOT QUOTE WITHOUT PERMISSION OF THE AUTHOR
© Michele Gazzola December 2014, Humboldt-Universität zu Berlin, Germany
Comments are welcome
Contacts: www.michelegazzola.com
1
Abstract
This article provides an evaluation of the language policy of the World Intellectual Property
Organisation (WIPO). It focuses on the reform occurred in 2008, when the Korean language
was given the status of a language of publication of the Patent Cooperation Treaty (PCT).
Results show that the 2008 reform entailed an average reduction of at least 54 percent in the
costs of access to the PCT procedures for Korean-speaking applicants, and it generated aggregate savings for them of about €24 million between 2009 and 2011. The additional costs
borne by the International Bureau of WIPO due to the reform were €2.3 million. The new
language policy has led to a more balanced distribution of admission and interaction costs
among applicant countries. It is likely that the 2008 reform brought about a transfer of information costs from Korean-speaking countries to English-speaking countries and inventors
fluent in English as a second language, but this effect was offset by exogenous factors. Finally, adding the Korean language is likely to have had a positive impact on the costeffectiveness of the language policy of the International Bureau of WIPO. It is estimated that
between 2009 and 2011 the reform increased the number of International patent applications
from Korean-speaking applicants by about 21.6 percent, and that the extra fee income collected by the International Bureau after the reform was enough to cover its extra translation costs.
Our results suggest that more multilingualism in international organisations can be at the same
time more cost-effective and fairer than monolingualism if comprehensive and adequate notions of cost and benefit are used.
Keywords
World Intellectual Property Organisation; Patent Cooperation Treaty; Translation costs;
Language policy; Korean language
JEL classification:
H40, H89, K11, O30, O31, O34, O38, O39, O50, P48, Z18, Z19
2
1. Introduction1
The role of language policy in the area of Intellectual Property (IP) rights has remained
relatively under-explored in literature. The few existing contributions in this area focus most
often on the estimation of translation costs to validate European patents granted by the European Patent Office (EPO),2 or on the translation costs borne by applicants who enter into the
national phase of the Patent Cooperation Treaty or PCT (WIPO 2008). However, the language
policies (or “language regimes”) of IP organisations, such as patent offices, do not affect only
the translation costs just mentioned. Provisions regarding the set of official languages of IP
organisations and the related translation requirements affect a whole range of language-related
costs arising at different stages of the patenting procedures, from the moment in which a patent application is filed to opposition and appeal procedures. For example, applicants seeking
IP protection may have to bear preliminary translation costs when they file an application
with an IP authority if their working language is not one of the official languages of the authority concerned. In addition, patent offices’ language policy influences the process of
knowledge diffusion because they determine in what language patent applications and patent
granted must be published and possibly translated. This affects the diversity of languages in
which the stock of new patent information is available to other inventors, that is, the features
of one of the most important inputs on which inventors build innovation (Ceccagnoli et al.
2005, Guellec and Van Pottelsberghe 2007).
The potential economic impact of language regimes should not be underestimated.
Recent empirical evidence shows that language policies can bring about cost biases among
competing inventors. For example, the language regime of the EPO, based on three official
languages, namely, English, French and German, makes the average cost of accessing patenting procedures for European applicants whose first language is not English, French or German at least 30% higher than the average cost borne by English-, French- or German-speaking
applicants. Besides it generates the paradox that it is de facto cheaper for a US or Canadian
company to patent an invention in Europe than it is for a Spanish or Polish inventor (Gazzola
and Volpe 2014). This difference can be explained by the fact that English-, French- or German-speaking applicants do not have to pay for any translation costs in order to have access to
the EPO’s patenting procedures.
Generally speaking, language policies, just as is the case for any type of public policy,
are a form of regulation whose effects must be evaluated (Grin 2003). This article aims at
contributing to interdisciplinary research in the economics of innovation and IP rights, by
addressing the question of the evaluation of the cost-effectiveness and fairness of the language
policy of the World Intellectual Property Organisation (WIPO), and more specifically that of
the PCT system. In line with the literature in economics and policy analysis, there is no moral
or ethical content in the notion of fairness (Just et al. 2004). Assessing fairness simply implies
identifying who loses, who gains and (if possible) how much, and how the costs of alternative
policies are shared among individuals or groups.
1
List of non-standard abbreviations: EPO (European Patent Office), EPC (European Patent Convention), International Bureau (IB), International Preliminary Report on Patentability (IPRP), International Search Authority (ISA), International Search Report (ISR), Korean Institute of Intellectual Property
(KIPO), PCT (Patent Cooperation Treaty), Return on Fee Income Index (RFI), Representation Index
(RI), Receiving Office (RO), WIPO (World Intellectual Property Organisation), Written Opinion
(WO).
2
See among others Van Pottelsberghe and Mejer (2010), Harhoff, et al. (2009), Guellec and Van
Pottelsberghe (2007).
3
The official languages (or languages of publication) of the PCT system are Arabic, Chinese, English, French, German, Japanese, Korean, Portuguese, Russian and Spanish. The language regime of the PCT was reformed in 2008, when Korean and Portuguese were included
amid the languages of publication. It is therefore necessary to provide an evaluation of this
change, by discussing its effect on the effectiveness, cost-effectiveness (or efficiency) and
fairness of the PCT system, which is administered by International Bureau of WIPO (IB).
This article focuses on the addition of the Korean language, and it is organised as follows:
Section 2 briefly describes the language regime of the PCT and presents some unpublished
data on the use of languages in the PCT system, showing that technology is more multilingual
than is commonly believed. This section is a necessary stepping stone to the following sections, as it clarifies the current linguistic trends in the PCT system and the context in which
the 2008 reform took place. Section 3 explains the methodology used in this article, which is
based on standard policy analysis tools adapted to the evaluation of language policies in international patent organisations. In section 4 the distributive effects of the 2008 reform are evaluated, whereas section 5 focuses on the evaluation of efficiency. Results reveal that the 2008
reform is likely to have improved the cost-effectiveness of the PCT language policy, and that
it brought about a more balanced distribution of costs among applicants. Hence, increasing the
level of linguistic diversity in international organisations is not necessarily less efficient and
fair than monolingualism. Section 6 summaries and concludes the article.
2. The Languages of the PCT
2.1 The language regime of the PCT system
The International Bureau of WIPO, or Secretariat, is responsible for managing the PCT.
The IB is not a fully fledged patent office because it does not carry out the substantive examination of patent applications and it does not grant patents. Nevertheless, it performs several of
the functions of traditional patent offices such as receiving international patent applications
and publishing them. The IB is the central node of a complex network of IP authorities in the
world that act at different stages of the PCT procedures. The language provisions of the PCT
are strictly linked to patenting procedures, and they must be presented together. Without going into details,3 suffice to say that applicants file an “International patent application” (or
PCT application) with a Receiving Office (RO), typically a national IP authority (such as the
US Patent and Trademark Office) or regional patent office (such as the EPO). The IB itself
acts as a RO. The international patent application is subsequently sent by the RO to an eligible
International Search Authority (ISA). ISAs carry out an “International search”, that is, technical research of prior art. The result of the international search is the “International Search
Report” (ISR), a report containing citations of the relevant documents and a list of the fields
searched. ISAs also issue a non-binding “Written Opinion” (WO) as to the patentability of the
invention. The WO is the basis for the issuance by the International Bureau, on behalf of the
ISA, of the “International Preliminary Report on Patentability” (IPRP).
If the international patent application is not withdrawn, the IB publishes the international patent application together with the ISR no later than 18 months from the priority date (that
is, the date of the filing of an international patent application or the date of the filing of a national or regional patent application if the applicant claims priority from such application).
The end of the PCT procedures coincides with the beginning of the “national phase” in one or
3
See, among others, Tritton and Davis (2014) for a more extensive discussion.
4
more states or regions designated by the applicants. The applicant files a national or regional
patent application in the designated country or group of countries, and the competent patent
office will carry out the substantive examination of the patent applications and, in case, grant
a patent. The IPRP is made available for public inspection after the expiration of 30 months
from the priority date, typically at the beginning of the national phase the IB communicates.
The IPRP is useful to national or regional offices during the substantial examination of the
application carried out during the national phase.
Let us now turn to language requirements. The international patent application must be
filed in a language permitted by the competent RO. If the application is not drafted in one of
the languages of publication of the PCT, the applicant must translate it into one of these languages according to the RO rules. For example, let us assume that an applicant files an international patent application with the Italian Patent and Trademark Office acting as a RO. This
Office accepts international patent applications in English, French, German or Italian, and it
designates the EPO (and only the EPO) as the competent ISA. As the official languages of the
EPO are English, French or German, applications filed in Italian must be translated into one
of these three languages. No further translation is required because English, French and German are also the languages of publication of the PCT. The ISR, the WO and the IPRP must be
written in the same language and this language must be the language in which the international application will be published. The IB publishes the international patent application, the
WO, the IPRP, the ISR and other relevant documents in the language of publication chosen by
the applicant (and only in this language). If the ISR has not been drafted in English, the IB
provides a translation into this language. The IB also provides a translation of the title and the
abstract of all published international patent applications into English and French.
2.2 The linguistic profile of applicants
The PCT came into force in 1978. In that year the IB published 19 PCT applications. In
2011, 163,669 PCT applications were published. During the last two decades, the number of
international patent applications published per year has increased by 13% on average. Contrary to what is commonly believed (e.g. “technology speaks English only”), the degree of linguistic diversity in the system has increased. The addition of Korean and Portuguese in 2008
must be viewed and interpreted in the light of recent trends in the global patent system. It is
therefore necessary to sketch out the evolution of the use of languages in the history of the
PCT before turning to the evaluation of the 2008 reform.
If not specified differently, all tables and figures in this article have been elaborated directly by the author on the basis of a comprehensive dataset kindly provided by the statistical
unit of WIPO at the end of 2012. Counts are based on the patent publication date (that is, usually six months after the date of filing the international patent application or 18 months from
the priority date). As there is a certain time lag between the moment of filing and the date of
publication, the year of filing may not coincide with the year of publication. Applications
filed but not published (that is, withdrawn applications) are disregarded, as they do not become part of existing knowledge.
In 1978, there were only five languages of publications (English, French, German, Japanese and Russian). Spanish was added in 1985, Chinese in 1994, Arabic in 2006 and Korean
and Portuguese in 2008. Since 1998 applicants are allowed to file an international patent application in a language which is not a language of publication, but within one month they
must provide a translation into one of the languages of publication designated by the compe5
tent RO. Table 1 shows the percentage of PCT applications published, by the language in
which the PCT application was filed from the origin of the PCT system. Only the first ten
most important languages are considered. In order to track recent trends, Table 1 presents
percentages not only for the whole period but also from 2009 to 2011.
Table 1: Percentage of PCT patent applications published, by the language in which the PCT patent application was filed. Top 10 languages
Language
English
Japanese
German
French
Chinese
Korean
Swedish
Spanish
Russian
Italian
Source: WIPO statistical database
1978-2011
67.1
15.6
13.3
4.3
3.0
2.3
0.8
0.8
0.6
0.6
2009-2011
55.6
18.0
10.6
3.5
5.1
3.7
0.2
0.9
0.5
0.9
English, Japanese and German are the most frequently used languages in filing, reflecting the importance of the United States, Japan and Germany as countries of innovation. During the last decade, however, China and the Republic of Korea have also become intensive
users of the PCT system. The rapid increase in the importance of Asian innovators is gradually changing the linguistic profile of PCT applicants. The percentage of published PCT applications that were originally filed in English, for example, has been steadily decreasing during
the last 15 years (see Figure 1). Only part of this trend can be explained as a result of the 1998
reform, which made it possible to file PCT application in any language. After a peak in 1996
(69.4%), the percentage of PCT applications filed in English decreased to 52.4% in 2011.
WIPO’s most recent estimates show that the percentage of PCT applications filed in English
is gradually approaching 50% (WIPO 2013a: 57).
6
Figure 1: Percentage of published PCT applications by language used in filing (1979-2011).
75,00%
65,00%
55,00%
45,00%
35,00%
25,00%
15,00%
5,00%
-5,00% 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011
Japanese
German
French
Chinese
Korean
English
Source: WIPO statistical database. The list of the six most important languages used in filing is based on
2011 data.
By contrast, the role of Asian languages ― and, although to a much lesser degree, other
languages such as Spanish or Italian (which is not a language of publication) ― is rising. The
number of PCT applications published in 2011 that had been filed in Chinese or Korean is
higher than those field French. Recent estimates published by WIPO reveal that the share of
international applications filed in Korean and Chinese is approaching 5% and 9% respectively, a percentage which is very close to that of German (WIPO 2013a: 57). While applicants
from the Republic of Korea still use English quite often in filing (23% in 2011), Chinese and
Japanese inventors indisputably prefer to file PCT application in the official language of the
country in which they reside (in 2011, only 6.9% and 7.9% of the PCT patent applications
filed, respectively, by Chinese and Japanese applicants were in English). Recall that PCT applications are classified according to the country of origin of the first author of the application.
German, and to a lesser extent French, are also important carriers of technical
knowledge, even if during the last decades the share of published PCT applications that were
filed in these two languages, in proportional terms, has declined more than the share of published PCT applications that were filed by applicants resident in, or national of, countries
where French or German are official languages.4 German, French, and in particular Swiss,
inventors today tend to use English in filing more often than in the past. For example, in 1996
only 3.5% and 5.4% of the PCT applications filed by residents in, or national of, Germany
and France respectively were in English. 15 years later this percentage was 18% for Germany
and 27.9% for France. This trend is linked to several factors, including the internationalisation
4
Section 4 clarifies how the number of PCT application is shared when dealing with multilingual
countries.
7
of R&D activities, new firms’ patenting strategies and the effects of a reduction in the requirements for translations into German and French that have been introduced by the London
Agreement for applicants following the Euro-PCT route (Gazzola 2014).
Not surprisingly, international patent applications filed in a language of publication are
published in the same language, whereas if they are filed in a language which is not a language of publication of the PCT (e.g. Swedish or Dutch), they are usually published in English. The percentage of international applications published in English is always higher than
the percentage of published PCT applications that were originally filed in this language (in
2011, 54.3% and 52.4%, respectively). The change in the language policy of the PCT that
occurred in 2008 must be placed within the perspective of these trends.
Before concluding this section, let us note that figures presented in Table 1 are average
values that do not fully highlight the complexity of multilingualism in the PCT system. Table
5 in the Appendix presents the percentage of PCT applications by technological sector and by
language of publication (as opposed to the language used in filing, as in Table 1). Table 5
shows that the importance of a language as a carrier of technical knowledge, measured by the
percentage of PCT applications published in it, depends on the technological sectors considered. For example, from 2009 to 2011, 28.9% of all international patent applications published in the formally defined technological sector “digital communication” were in Chinese,
whereas only 5.1% of all PCT applications on average were published in this language in the
same period (see Table 1). By the same token, Korean is particularly important in “microstructural and nano-technology” (9.8% of all PCT applications published between 2009 and
2011), Japanese in “optics” (38.9%), German in “mechanical elements” (29%) and English in
“pharmaceuticals” (80.3%). This has an important consequence: the linguistic needs of inventors seeking access to existing patent information and agents involved in patent intelligence
depend on the technological area in which they work.
3. Language Policy and Evaluation
It is well known that patents are a policy tool designed to stimulate innovation and
hence economic growth. The state grants a monopolistic power to the inventors, but it requires in exchange a diffusion of the content of the invention through the publication of patent
applications. The effectiveness of alternative language policies adopted by a patent office
must be evaluated with respect to these general goals.
Public policies are usually evaluated in the light of standard criteria such as effectiveness, efficiency (often interpreted as cost-effectiveness in applied research) and fairness (Just
et al. 2004, Boardman et al. 2006). Language regimes can also be evaluated on the basis of
such criteria (Grin and Gazzola 2013). To the best of our knowledge, there are no detailed
data or well-established theoretical models that can be used to evaluate language policies in
international patent organisations. Nevertheless, policy evaluation provides an analytical
framework that is at the same time general and flexible enough to be adapted and applied to
situations in which few data are available. As a general rule, some data are better than no data
at all, and this article provides a first attempt to use such data to provide a first evaluation the
language policy of the PCT. Results, therefore, ought to be interpreted as an order of magnitude rather than precise estimates.
8
Generally speaking, all other things being equal, a language regime A of a patent office
is more effective than B if A contributes more than B to the main goal of such an office, that
is, fostering innovation and guaranteeing an orderly sharing of knowledge. Effectiveness is
evaluated and compared through outcome indicators (see section 5). As regards the costs of a
language regime, a distinction must be made between primary and implicit costs. The primary
costs of a language regime consist of the sum of its direct costs (e.g. translation and interpreting services), plus indirect costs, such as a given share of common administrative structures,
or overheads directly associated with translation and interpreting services. If the outsourced
language services are present in the budget of the organisation considered, they can be included in primary costs. The primary costs of a language regime should be put into relation with
effectiveness indicators in order to compare the cost-effectiveness (or efficiency) of alternative language regimes.
Implicit costs or external costs are defined as language-related costs borne by individual
inventors, companies, universities, and research institutions that arise from interacting with a
patent organisation. Following Gazzola and Volpe (2014), three types of implicit translation
costs are identified. First, let us mention “admission costs”, that is, costs to translate patent
applications into one official language of the patent organisation at the time of filing. The
second type of implicit costs are “interaction costs”, i.e. translation and interpreting costs
connected to the flows of communication between patent applicants and third parties (let us
call them “decentralised agents”) on the one hand, and the IP organisation considered on the
other hand (metaphorically defined as the “central authority”). Interaction costs arise during
intermediate communication during the procurement procedure, amendments to claims, and
opposition and appeal procedures. Third, “information costs” are costs related to accessing the
existing patent literature which is available in a language that users either do not understand
or do not speak fluently. Information costs arise for monitoring the state of the art, freedomto-operate analysis, patent intelligence, patent monitoring with the purpose of opposition.
The strategy followed in this article to compare the fairness of alternative language regimes consists of evaluating how the implicit costs of a language policy are distributed among
applicants.5 In other words, the fairness of a language regime is evaluated in terms of its distributive effects on different groups of applicants, identified according to their first language
in the case of persons and to the main working language if we are dealing with legal persons
(e.g. corporate actors). In practice, the regulations of patent offices define the language of
applicants on the basis of the official language(s) of the state where they reside or of which
they are national (for example, see article 14 of the European Patent Convention ― EPC).
Although this criterion does not fully account for multilingual individual inventors and does
consider that corporate actors do not have a “mother tongue”, it has an undisputed practical
applicability. Applied linguistics, in fact, showed that corporate actors are not monolingual
entities ― even when they claim to have one official language ― but complex collaborative
spaces where multilingual repertoires are used (see, among others Berthoud et al. 2013). In
this article, the possibility that corporations and research centers do not work in the official
language of the country where they are based is not ruled out. However, the multilingual nature of corporate actors does not change substantially the usefulness of the territoriality criterion, as working in a foreign language entails several types of costs that can be viewed as implicit costs. First, multilingual staff is more expensive than monolingual staff (Grin et al.
2010: 55-71). In addition, the output of staff working in a foreign language is often subject to
5
Clearly, external costs can be viewed as a component of the total cost of a language policy and therefore as a key element in the evaluation of efficiency. This question is explored further in section 5.
9
editing, which is also a form of cost. Finally, drafting documents in a foreign language has an
opportunity cost, as it usually requires more effort and more time.
There have been four changes in the number of the languages of publication in the PCT
(i.e. 1985, 1994, 2006 and 2008). The 2008 reform is the most suitable for a comparative
analysis. Arabic, Spanish and Chinese were added as they are official languages of the United
Nations (UN), and WIPO is part of the UN system. Spanish, for example, became a language
of publication in 1985, but only 20 PCT applications from Spain — the most important Spanish-speaking applicant country — were published between 1985 and 1989. Chinese was not
used before 1994 because China was not a contracting state of the PCT. Arabic was not used
before April 2013 because no ISA accepted it for the purpose of an international search.
Hence, the most interesting case is the 2008 change because both the Republic of Korea and
the largest Portuguese-speaking countries (Portugal and Brazil) were already contracting
states of the PCT before 2008. This article focuses on the Korean language because more data
are available. The number of PCT applications filed in Portuguese is, in fact, still relatively
low.
4. Evaluating the Fairness of the PCT Language Policy
Let us start with the characterisation of the distributive effects of the reform of the PCT
language regime implemented in 2008. Table 2 presents the average global cost of filing a
PCT application for a Korean-speaking applicant.
Table 2: Average global cost of access to the PCT system for a Korean-speaking applicant
Fees and admission costs
Euro
Transmittal fee
Average international filing fee
41
1,093
Search fee
Admission translation costs
293
1,700
TOTAL
3,127
Admission translation costs on total
54%
Source. Table compiled by the author using the official fees published by WIPO in 2011 and European Commission (2011). The international filing fee charged by WIPO did not changed between 2008 and 2011
Transmittal fees are paid to the competent RO. In Table 2, the RO is the Korean Institute
of Intellectual Property (KIPO). The international filing fee is paid directly to the IB. The bulk
of the income of the PCT system comes from the international filing fees paid by applicants
(WIPO 2013b). The international filing fee charged by the IB is €1,093.6 The search fee is
paid to the ISA and it does not become part of the budget of the IB. Table 2 reports the search
fee charged by the KIPO as the ISA for an international search carried out in Korean.
Admission translation costs are the average costs of translating an average international patent application into a language of publication. Following the estimations provided by the
6
The filing fee is set at 1,330 Swiss francs (the official currency of WIPO). This article adopts an exchange rate 1 Euro = 1.216 Swiss francs, following the exchange rate used by the IB for its fees on the
1 December 2012.
10
European Commission for the project of a European unitary patent (European Commission
2011, adapting Roland Berger 2004), the conservative assumptions are made that the average
patent application is made up of 20 pages, including all notations for drawings and that the
average translation cost per page is €85 (the average admission cost, therefore, is €1,700).7
The 2008 reform had a first distributive effect at the level of admission costs. Table 2
shows that admission translation costs for Korean-speaking applicants represent 54% of the
general costs to have access to the PCT procedures. After 2008, any inventor can file an international application in Korean without bearing these costs before publication, provided that
the competent RO designates an ISA that accepts this language for the purpose of international search (e.g. the KIPO). Admission costs for other applicants do not change. From 2009 to
2011, 14,024 international patent applications were published in Korean. Hence, the total reduction of admission costs (and thus savings) for Korean-speaking applicants was about €24
million.
One could argue that these translation costs must be borne anyway if an applicant wishes
to enter the national phase in at least one country whose official language is not Korean, such
as the US or Japan. This argument is not fully convincing, however. First, one of the main
advantages of the PCT route is that applicants “buy time”, because they postpone the moment
at which various fees for the substantial examination of the patent application, the fees for
granting and renewing a patent have to be paid (WIPO 2011). If a Korean-speaking applicant
seeks protection only in the Republic of Korea and chooses the PCT route to buy time, translating a patent application into another language is a waste of resources. Furthermore, international patent applications may be withdrawn before publication for several reason (e.g. if the
ISR reports bibliographical references that clearly point out that the invention is not novel). In
this case, translation costs incurred before publication would be a deadweight loss. Finally,
applicants prefer to postpone translation costs as much as they can, as they have no incentive
to provide translations before the beginning of the national phase.
The second distributive effect of the reform is a clear reduction of interaction costs for Korean-speaking applicants. After having received the ISR and the WO from the competent ISA
that accepts Korean for the purposes of an international search, the applicant can propose
amendments to the claims of his application in Korean before the publication by the IB.
The third distributive effect of the 2008 reform regards information costs. Quantifying
such costs is more complicated than estimating admission costs. International patent applications filed in Korean will no longer be translated into any of the previous eight languages of
publication. As a result, the information costs for all Korean-speaking applicants who wish to
have direct access to applications published in this language drops to zero. By contrast, it is
likely that the reform of 2008 increased the information cost for English-speaking inventors
who do not speak Korean. Between 1998 and 2008, the vast majority of PCT applications
filed in Korean were eventually published in English (and to a lesser extent in Japanese).
Thus, English-speaking countries are obviously those most affected by the 2008 reform.
Clearly, the impact on inventors (or potential inventors) who are not native speakers of Eng7
Other authors estimate a higher average number of pages of patent applications, for example, Van
Zeebroeck et al. (2009: 1008) Van Pottelsberghe and Mejer (2010: 230) and Archontopoulos et al.
(2007: 126). According to a report published by WIPO (2008: 51), the average translation cost of an
average patent are higher than €85 per page. However, the choice has been made to adopt the conservative assumption suggested by Roland Berger (2004). The cost estimate provided in this article,
therefore, must be viewed as a lower bound.
11
lish depends on the distribution of skills in English among the population of the non-Englishspeaking country considered (see European Commission 2012 for continental Europe).
To our knowledge, no reliable data on the language skills of a representative sample of international applicants and inventors exist. Hence, it is not possible to estimate whether the
reduction in information costs for Korean-speaking inventors is enough to offset the increase
in the information costs borne by inventors who are proficient in English. As the knowledge
of English in Korean-speaking countries such as the Republic of Korea is more widespread
than the knowledge of Korean in the English-speaking world, and it is likely that on balance
there has been an increase in the aggregate information costs.8 Yet, it is risky to rely on anecdotal evidence. In order to provide a first tentative quantitative assessment of the distributive
effects among countries of the 2008 reform on information costs, two indicators are used. The
first indicator is the Representation Index (RI) defined as RI=Ax/Cx, that is, the ratio between
the percentage of all international patent applications published in language X in a given period (Ax) and the percentage of international patent applications published in the same period
that were filed by residents or national of countries where X is an official language (Cx). An
RI larger than one means that the X-speaking countries benefit from an informative advantage, because the percentage of PCT applications published in the official/national language of these countries is higher than the percentage of published international patent applications that were filed by resident in, or national of, such countries.
The second indicator is the Return on Fee Income Index (RFI), defined as: RFI=Ax/Fx. This
is the ratio between the percentage of applications published in language X in a given period
(AX) and the percentage of the fee income of the IB that can be ascribed to X-speaking countries (FX). If the RFI is higher than one, it means that, all other things being equal, the share of
patent information easily available to applicants who are resident in X-speaking countries is
relatively higher than their contribution to the budget of the PCT. Recall that WIPO is a selffunding organisation; the international filing fee paid by applicants to have access to the PCT
procedures (€1,093, see Table 2) constitutes the bulk of the WIPO’s budget.
Table 3 below shows the values of the RI and RFI before and after the 2008 reform using a
three-year time span. To keep the analysis tractable, only the top 25 PCT applicant countries
are considered.9 These countries make up 98% of all international patent applications published. Countries are grouped according to their official/national language. The first row includes six English-speaking countries (Australia, Ireland, Singapore,10 the UK and the US and
part of Canada),11 the second row four French-speaking countries (France and part of Belgium, Canada and Switzerland), and the third row includes four German-speaking countries
8
For a discussion of the position of English as a foreign language in education in Korea, see Jung and
Norton (2002). On the position of Korean as a foreign language in the US higher education institutions, see Furman et al. (2010). Note, however, that data on foreign language learning do not include
information about the number of Korean-speaking engineers and scientists who moved to Englishspeaking countries, especially in the US.
9
Countries are ranked according to the number of applications published in 2011. The list includes the
United States, Japan, Germany, China, the Republic of Korea, France, the United Kingdom, the Netherlands, Sweden, Switzerland, Italy, Canada, Finland, Australia, Israel, Denmark, Belgium, India,
Austria, Russia, Norway, Singapore, Ireland and Brazil.
10
The government of Singapore recognises four official languages, that is, Malay, Mandarin, Tamil
and English. The Singaporean government has adopted an English-medium education policy since the
1960s (Chua 2010). As a result, English is the language in which children acquire literacy.
11
Results do not change substantially if New Zealand is added.
12
(Austria, Germany and part of Switzerland and a small share of the Belgian population). For
multilingual countries, using census data or other secondary sources,12 the number of applications has been divided on the basis of the percentage of the population speaking official languages as their mother tongue. The last row (“others”) includes all countries among the top 25
whose official language is not one of the PCT languages of publication. Arabic is excluded, as
it is not used in practice.
Table 3: Representation Index and Return on Fee Income Index (2006/2008 and 2009/2011).
Group of countries or
country/ language
RI
(2006/8)
RI
(2009/11)
RFI
(2006/8)
RFI
(2009/11)
English-speaking
French-speaking
1.54
0.63
1.60
0.60
1.52
0.63
1.57
0.59
German-speaking
Japan/Japanese
0.82
0.91
0.79
0.92
0.81
0.89
0.78
0.91
Spain/Spanish
0.78
0.78
0.77
0.77
China/Chinese
Russia/Russian
Korean/R. of Korea
Brazil/Portuguese
Other countries
0.88
0.93
-
0.92
0.90
0.57
0.34
-
1.16
2.65
-
1.06
2.78
0.56
0.52
-
Source: Table compiled by the author
Results reveal that the RI is much higher than one just for English-speaking countries,
meaning that English is over-represented. This entails a de facto form of implicit redistribution of the burden of patent information costs from native speakers of English towards nonnative speakers of English, confirming the problem of “free-riding” in international communication (Grin 2005 and Van Parijs 2007). This is due to the fact that several applicants who
are resident in, or national of, non-English-speaking countries choose English as the language
of publication for their applications because the official language of their country is not a language of publication. However, some applicants lodge PCT applications in English even if
they could use the official language of their country (e.g. German). One could argue that these
applicants are likely to have a good command on English, and therefore that a RI higher than
one does not prove that the redistributive effect pointed out before is significant. Nevertheless,
it is worth noting that published patent information is a global public good (Stiglitz 2007) that
can be used by all actors potentially interested in such information and not only by former
applicants, and the potential audience includes individual researchers and staff working in
small and medium enterprises who are not necessarily proficient in English.
The RFI is below one for all countries or groups of countries except for the Englishspeaking countries, China and Russia. Recall that according to the PCT individual applicants
resident in some countries such as Russia and China are entitled to a 90% reduction in their
international filing fee. For this reason individual applicants from these countries contribute
less then proportionally to the budget of WIPO. This explains why the RFI for Russia and
China is higher than one. By contrast, the RFI is below one for all other countries. This indicates an implicit transfer of resources in favour of the group of English-speaking countries,
and to a lesser extent in favour of China and Russia.
12
For Canada, see Cardinal (2005).
13
The RI and the RFI for the Republic of Korea and Brazil, the only Korean- or Portuguesespeaking countries included in our list, have obviously increased after 2008. The interesting
result is, however, that the value of the RI and the RFI for the English-speaking countries has
not decreased. Clearly, if Korean and Portuguese had not been given official status, the RI and
the RFI for English-speaking countries would have been higher. Yet, after the reform the value of these two indicators remained essentially stable. This can be ascribed to two factors. The
percentage of PCT applications published in English has declined in percentage terms less
than the percentage of published PCT applications filed by English-speaking countries. The
second factor, related to the first one, is the increasing use of English as a language of filing
by some non-English-speaking countries, especially the Nordic countries. Table 4 summarises
the results of this section. It focuses on the position of English-speaking countries because, as
shown before, English was the language of publication most often chosen by applicants who
filed PCT applications in Korean before 2008.
Table 4: Distributive effects resulting from the 2008 reform
Indicator
Effects on Korean-speaking
applicants
Effects on Englishspeaking applicants
Cost for admission to patenting procedures
Decrease by €24 million between
2009 and 2011
No effect
Interaction costs with WIPO and
ISAs
Costs of access to existing knowledge
Decrease
No effect
Decrease
Increase
Representation index
Increase from 0 to 0.57
Not substantially affected
thanks to exogenous factors
Return of income fee index
Increase from 0 to 0.56
Not substantially affected
thanks to exogenous factors
Source: Table compiled by the author
To conclude, it is plausible that the 2008 reform has brought about a transfer of information costs from Korean-speaking countries to English-speaking countries and inventors
fluent in English as a second language, even if some factors could help in alleviating this effect (e.g. the translation into English of the ISR, IPRP and the translations into English and
French of the abstract and the title, machine translation and translations provided by applicants to national or regional patent offices at the beginning of the national phase). Nevertheless, preliminary evidence based on the RI and the RFI indexes points out that the expected
negative effect of the 2008 reform on the information costs borne by English-speaking inventors (if any) has been offset by exogenous factors.
5. Evaluating the Cost-effectiveness of the Language Policy of the PCT
Let us now turn to the impact of the 2008 reform on the costs and the effectiveness of the
language regime of the IB. Generally speaking, the change of language regime in 2008 may
have three effects on effectiveness. All other things being equal, the reduction of admission
costs for Korean-speaking applicants can encourage them to file international patent applications, instead of protecting their inventions through secrecy, or to choose the PCT route instead of alternative routes (e.g. the Paris route). Both effects would increase the number of
new PCT applications, thus contributing to the goals of the PCT, that is, promoting innovation
and contribute to the diffusion of technical knowledge. Second, the availability of information
in Korean makes cheaper and faster the access to technical knowledge for inventors who un14
derstand this language but do not know or are not fluent in the other languages of publication.
This knowledge can be used as an input for new inventions. Finally, the introduction of Korean may increase the information cost for inventors who are proficient in English but not in
Korean (recall that Korean-speaking applicants inventors used to translate into English applications lodged in Korean before 2008). From the point of view of those who do not understand Korean, adding this language to the languages of publication of the PCT is a partial
strategy of secrecy that can delay or make more costly the availability of information to be
used as an input for further innovation.
If the additional number of PCT applications filed because of the reform, ceteris paribus, is
higher than the number of PCT applications that have not been filed because of additional
information costs, then the 2008 reform increased the effectiveness of the language regime.
Yet, estimating the net effect of the change in the language regime on the number of new international applications filed after 2008 can turn out to be a daunting task. In order to assess
the net effect of the reform, one would need ad hoc survey data on a representative sample of
companies that make it possible to isolate the effect of different factors on the applicants’ filing decisions. Such data, nevertheless, are not available and this question cannot be explored
further. In this section, therefore, the change in the effectiveness of the IB language regime is
assessed on the basis of a simple (albeit admittedly imprecise) indicator, that is, the number of
international patent applications published per year that have been filed in Korean. This indicator is used as a proxy to assess the effectiveness of a language regime after the 2008 reform.
Hence, results must be viewed as an order of magnitude rather than precise estimates.13
In order to estimate the change in the cost-effectiveness of the language regime of the IB,
we have first estimate the variation in the primary costs. The title and the abstract of applications published in Korean have to be translated into English and French. The ISR, the WO
and the IPRP must be translated into English at the expense of the IB.14 Before 2008 WIPO
only had to provide translations into French of the application abstract and title. Between
2009 and 2011, the average unit cost for processing a PCT application was €655 (WIPO
2013a: 63, p. 63). On average 29.6% of this cost can be attributed to translations (WIPO 2006,
Appendix 3). The actual additional unit translation cost of the 2008 reform is not €194, however. If applications in Korean had been published in English, the abstract would have been
translated into French anyway. On average the price for an outsourced translation of an abstract into French from English is €34 (WIPO, 2007b: 3, p. 3). Hence, the additional average
unit translation cost per PCT applications filed in Korean is €160. From 2009 to 2011, 14,024
international patent applications were published in Korean. Thus, the additional primary costs
of the PCT language regime can be estimated at €2.24 million, which will be rounded to €2.3
million out of caution. Recall that the corresponding decrease in admission costs for Koreanspeaking applicants between 2009 and 2011 was estimated at €24 million. These results tend
to support the claim made by Pool and McFann, according to which “it is wrong to claim that
13
There are several indicators that can be used to assess the effectiveness of a language regime of a
multilingual patent office, provided that data are available. For example, one may assess whether the
2008 reform has had an impact on the average number of citations of PCT applications published in
Korean (per 100 PCT applications published). A net increase in citations made by residents in the Republic of Korea of applications published in Korean could be interpreted as the effect of an easier access to these documents.
14
A further consequence of the reform is that ISAs have to deal with more linguistic diversity when
they carry out an international search on prior art. However, this effect, strictly speaking, is not related
to the primary translation costs of the IB. However, ISAs can rely on translations (if any) provided by
applicants to national or regional patent offices at the beginning of the national phase.
15
having many official languages is necessarily inefficient. As more native languages are made
official, translation costs rise but adoption costs [i.e. implicit costs in the terminology used in
this article] fall. The tendency to regard multiple official languages as inefficient may reflect a
state-centred neglect of costs incurred by individuals in adapting to language policies.” (Pool
and McFann 1991, p. 9).
Let us now turn to the evaluation of effectiveness. An increase in the number of the languages of publication does not necessarily entail a net increase in primary translation costs, if
additional primary costs are outweighed by additional fee income. WIPO is a self-funding
organisation; adding new languages to the set of existing languages of publication can entail
an increase in the fee income of the organisation if the change of language regime brings
about an increase in the use of the PCT system by applicants. This was indeed one of the
strongest arguments put forward by the Republic of Korea in 2007 to convince the other PCT
contracting parties to accept the inclusion of Korean among the languages of publication
(WIPO 2007a). There are some economic reasons behind this claim. Empirical evidence
available for national or regional patent offices shows that, all other things being equal, a reduction of fees has a positive and significant impact on the number of new applications filed
(see De Rassenfosse and Van Pottelsberghe de la Potterie 2013: 705 for a review).15 De
Rassenfosse and Van Pottelsberghe de la Potterie (2012), for example, analyse the demand for
patents at the European, US and Japanese patent offices from 1980 to 2007 and carry out an
econometric analysis through dynamic panel data models. They conclude that the average fee
elasticity of the demand for patents is −0.3. Other studies, using different cross-section datasets converge towards a value around −0.5 (see De Rassenfosse and Van Pottelsberghe de la
Potterie 2013: 705 ).16 Hence, patent are an inelastic good, but nevertheless fees have an effect on the number of applications filed. Using patent fee elasticity as an estimate of the expected effects of a change of patenting costs on the number of patent applications filed and on
innovation is a common practice in the literature (see, for example, Van Pottelsberghe and
Mejer 2010 on the expected effects of an implementation of the London Agreement in the
European patent system). By analogy, it is reasonable to assume that a reduction in admission
translation costs has a similar effect on the number of PCT applications, as the cost for translating a patent application is implicitly a form of fee (Gazzola 2014).
With the data available, it is not possible to estimate precisely which percentage of all applications filed by Korean applicants and published between 2009 and 2011 would not have
been filed via the PCT system (instead of alternative routes, such as the Paris route) or would
not have been filed at all if Korean were not a language of publication. However, on the basis
of the data available, we can carry out a simulation of the impact of a reduction of admission
costs on the number of applications.17 Table 2 shows that adding Korean to the languages of
publication entailed in 2008 a decrease of 54% in the general costs of accessing the PCT procedures for Korean-speaking applicants up to the beginning of the national phase. Assuming
an average value of the fee elasticity equal to −0.4, that is, the mean between the two values
15
See also Van Pottelsberghe and Mejer (2010), Danguy and Van Pottelsberghe de la Potterie (2011),
De Rassenfosse and Van Pottelsberghe de la Potterie (2012).
16
A distinction is often made between short-term fee elasticity and long-term fee elasticity (see De
Rassenfosse and Van Pottelsberghe de la Potterie 2013: 705-706). The former refer to fee elasticity of
patents within a 12-month period after the change in fees, whereas the latter refers to the value of the
elasticity in a 3 or 4-year period. Hence, for the purposes of this article, it is clearly more appropriate
to use the long-term fee elasticity.
17
The estimation of the effect (if any) of a change in information costs on the number of international
patent applications filed is left to future research.
16
pointed out above. All other things being equal, the number for PCT applications filed in Korean and published between 2009 and 2011 is expected to increase by 21.6% with respect to
the three years before the entrance into force of the reform (2006-2008). As from 2006 to
2008, the total number of published PCT applications that had been filed in Korean was
10,625, the predicted increase amounts to 2,295 PCT applications.
The question is how many additional PCT applications were necessary to cover extra primary costs borne by the translation services of the IB linked to the inclusion of Korean as a
language of publication from 2009 to 2011 (i.e. €2.3 million, see above). In order to generate
such an additional fee revenue during the period considered 2,086 applications at the average
filing fee of €1,093 were necessary. 2,086 applications represent 12% of the total number of
published PCT applications filed in Korean between 2009 and 2011 (17,416).18 In other
words, if 12% of all applications filed in Korean and published between 2009 and 2011 had
actually been lodged as a result of the reform, then the extra fee revenue collected by the IB
would have been enough to cover the additional primary translation costs of the PCT language regime. As the number of additional PCT applications filed in Korean estimated on the
basis of the patent fee elasticity (2,295) is higher than the number of extra PCT applications
needed to cover the extra-primary costs of the language regime (2,086), it is possible that the
cost-effectiveness of the language regime of the IB improved. The extra primary costs of its
language regime could have been offset by additional income fee, whereas the number of PCT
applications in Korean (that is, the indicator of effectiveness used) increased. Obviously, results must be interpreted with caution, as it was not possible to isolate the effect of exogenous
factors on the number of PCT applications filed.
6. Conclusions and Policy Implications
The original contribution of this article consists of an evaluation of the PCT language policy, that is, a form of regulation in the area of language status and use. Such evaluation is
based on standard policy analysis criteria such as effectiveness, cost-effectiveness and fairness. The methodology and the concepts presented in this article, however, can be adapted for
further research to the evaluation of the language policy of other multilingual organisations,
including patent offices or multilingual organisations dealing with trademarks and industrial
design.
Two language regimes adopted by WIPO at different moments in time are compared, that
is, the language regime before and after 2008, when Korean (and Portuguese) were added as
languages of publication in the PCT. As regards fairness, results show that the 2008 reform
entailed a clear reduction in admission costs for Korean-speaking applicants of at least €24
million between 2009 and 2011 without increasing admission costs for other applicants. In
addition, it has reduced interaction costs for Korean-speaking applicants. From this perspective, the reform has increased fairness, in the sense that existing inequalities have been reduced without negatively affecting no one. The reform has also reduced information costs for
Korean-speaking applicants, but it has entailed a transfer of information costs from them to
English-speaking applicants and inventors. The overall impact on the aggregate information
costs is ambiguous, but it has probably increased. However, the results of our analysis show
that (i) the transfer of information costs from Korean-speaking applicants towards English18
Some of the international applications filed in Korean were published in English between 2009 and
2010, but this is due to the fact that these two years were a transitory period. After 2011, all PCT applications filed in Korean were subsequently published in Korean.
17
speaking inventors due to the 2008 reform was simply a (partial) counterbalance for the implicit information cost savings from which English-speaking inventors had already benefited
before 2008, and (ii) on aggregate this transfer was offset by exogenous factors.
As regards cost-effectiveness, simulations based on patent fee elasticity suggest that the
possible effect of the 2008 reform was an increase in the number of PCT applications filed by
Korean-speaking applicants from 2009 to 2011 equal to 2,295. However, this article shows
that an increase of 2,086 during the same period would already be enough to generate an extra
fee income for the IB that offsets the increase in the primary costs of the PCT language regime. As the costs do not change, but the number of PCT applications filed is likely to have
increased, then it is also likely that the cost-effectiveness of the PCT language regime after
2008 has improved.
Of course, further research and better data are needed to corroborate these results, but this
article provides circumstantial evidence that the effects of linguistic regulations adopted by IP
organisations are not negligible. The claim that, generally speaking, multilingualism as such
has a negative effect on efficiency and that it hampers development should be put under closer
scrutiny. Obviously, this does not mean that the level of linguistic diversity in international
organisations should tend towards infinity. Nevertheless, an intermediate solution is probably
wiser than aiming at monolingualism. The general conclusion of this article, which can be
extended beyond the particular case analysed, is that adding new languages in an international
organisation can entail extra costs for the central budget of the organisation itself, but it can
contribute at the same time to reducing the implicit costs borne by different decentralised
agents. This, in turn, may have not only positive distributive effects but improve the costeffectiveness of the multilingual organisation analysed.
Finally, in contexts where limiting the number of official languages can be justified by
practical reasons, decision-makers could consider the possibility of designing a system of financial transfer to compensate those who must bear the implicit translation costs. The EPO,
for example, applies a discount of 30% to several fees for applicants19 who are resident in a
Contracting State of the EPC that does not share an official language with the EPO (e.g. Poland or Spain). WIPO could introduce some form of financial compensation to reduce the
implicit costs borne by applicants who do not have a language in common with the PCT system. This could contribute not only to a more balanced distribution of costs among the contacting states of the PCT but would also probably have a positive effect on the costeffectiveness of the language policy of the PCT and on innovation in general.
Acknowledgments The financial support from the DYLAN Project (“Language dynamics and
management of diversity), 6th Framework Programme, and the Research Executive Agency of
the European Commission (Project number PIEF-GA-2012-327225) is gratefully acknowledged. The author wishes to thank François Vaillancourt and Alessia Volpe for their useful
remarks, and to express his gratitude to the statistics service of WIPO. Any errors and all interpretations are all my own.
19
These applicants are small and medium-sized enterprises, natural persons, or non-profit organisations, universities or public research organisations. Large companies are excluded.
18
7. References
Archontopoulos, Eugenio, Dominique Guellec, Niels Stevnsborg, Bruno Van
Pottelsberghe de la Potterie, and Nicolas Van Zeebroeck (2007) "When small is
beautiful: Measuring the evolution and consequences of the voluminosity of patent
applications at the EPO", Information Economics and Policy, 19, pp. 103-132.
Berthoud, Anne-Claude, François Grin, and Georges Lüdi (eds.) (2013) Exploring the
dynamics of multilingualism: The DYLAN project. Amsterdam: John Benjamins.
Boardman, Anthony E., David H. Greenberg, Aidan R. Vining, and David L. Weimer
(2006) Cost-benefit analysis. Concepts and practice (3rd). Upper Saddle River (NJ):
Prentice Hall.
Cardinal, Linda (2005) "The ideological limits of linguistic diversity in Canada", Journal
of Multilingual and Multicultural Development, 26 (6), pp. 481-495.
Ceccagnoli, Marco, Alfonso Gambardella, Paola Giuri, Georg Licht, and Myriam Marian
(2005) Study on evaluating the knowledge economy what are patents actually worth?
The value of patents for today’s economy and society. Final report for lot 1. Brussels:
European Commission.
Chua, Siew Kheng Catherine (2010) "Singapore's language policy and its globalised
concept of Bi(tri)lingualism", Current Issues in Language Planning, 11 (4), pp. 413429.
Danguy, Jérôme and Bruno Van Pottelsberghe de la Potterie (2011) "Cost-benefit analysis
of the Community patent", Journal of Benefit-Cost Analysis, 2 (2), pp. 1-41.
De Rassenfosse, Gaetan and Bruno Van Pottelsberghe de la Potterie (2012) "On the price
elasticity of demand for patents", Oxford Bulletin of Economics and Statistics, 74 (1),
pp. 58-77.
De Rassenfosse, Gaetan and Bruno Van Pottelsberghe de la Potterie (2013) "The role of
fees in patent systems: Theory and evidence", Journal of Economic Surveys, 27 (4),
pp. 696-716.
European Commission (2011) Impact assessment. Accompanying document to the
proposal for a Council regulation implementing enhanced cooperation in the area of
the creation of unitary patent protection with regard to the applicable translation
arrangements, SEC(2011) 482/2. Brussels: European Commission.
European Commission (2012) Europeans and their languages, Special Eurobarometer
386. Brussels: European Commission.
Furman, Nelly, David Goldberg, and Natalia Lusin (2010) Enrollments in languages other
than English in United States institutions of higher education. New York: Modern
Language Association of America.
Gazzola, Michele (2014) The evaluation of language regimes. Theory and application to
multilingual patent organisations. Amsterdam: John Benjamins.
Gazzola, Michele and Alessia Volpe (2014) "Linguistic justice in IP policies: Evaluating
the language regime of the European Patent Office", European Journal of Law and
Economics, 38, pp. 47-70.
Grin, François (2003) "Language planning and economics", Current Issues in Language
Planning, 4 (1), pp. 1-66.
Grin, François (2005) L'enseignement des langues étrangères comme politique publique,
19. Paris: Rapport au Haut Conseil de l'évaluation de l'école.
Grin, François and Michele Gazzola (2013) "Assessing efficiency and fairness in
multilingual communication: theory and application through indicators", in Berthoud,
Anne-Claude, François Grin, and Georges Lüdi (eds.) Exploring the dynamics of
multilingualism, pp. 365–386. Amsterdam: John Benjamins.
19
Grin, François, Claudio Sfreddo, and François Vaillancourt (2010) The economics of the
multilingual workplace. London: Routledge.
Guellec, Dominique and Bruno Van Pottelsberghe (2007) The economics of the European
patent system: IP policy for innovation and competition. Oxford: Oxford University
Press.
Harhoff, Dietmar, Karin Hoisl, Bettina Reichl, and Bruno Van Pottelsberghe (2009)
"Patent validation at the country level - the role of fees and translation costs",
Research Policy, 38, pp. 1423-1437.
Jung, Sook Kyung and Bonny Norton (2002) "Language planning in Korea: The new
elementary English program", in Tollefson, James W. (ed.) Language policies in
education. Critical issues, pp. 245-266. Mahwah (NJ): Lawrence Erlbaum Associates.
Just, Richard E., Darrell L. Hueth, and Andrew Schmitz (2004) The welfare economics of
public policy: A practical approach to project and policy evaluation. Cheltenham:
Edward Elgar.
Pool, Jonathan and Brian McFann (1991) "The language auction: a nondiscriminatory
method of choosing official languages", Paper presented at the International
Symposium on Linguistic Human Rights, Tallin, Estonia, 12-16 October 1991. [Utilika
Foundation, Seattle, Washington (USA): http://panlex.org/pubs/etc/auction.pdf]
Roland Berger (2004) Study on the cost of patenting. Munich: Roland Berger Market
Resarch.
Stiglitz, Joseph (2007) "Knowledge as a global public good", in Abbott, Frederick M.,
Thomas Cottier, and Francis Gurry (eds.) International intellectual property in an
integrated world economy. The Hague: Wolters Kluwer.
Tritton, Guy and Richard Davis (eds.) (2014) Intellectual property in Europe. London:
Sweet & Maxwell.
Van Parijs, Philippe (2007) "Takling the anglophones' free ride: Fair linguistic cooperation
with a global lingua franca", AILA [Association Internationale de Linguistique
Appliquée] Review, 20, pp. 72-86.
Van Pottelsberghe, Bruno and Malwina Mejer (2010) "The London Agreement and the
cost of patenting in Europe", European Journal of Law and Economics, 29, pp. 211237.
Van Zeebroeck, Nicolas, Bruno Van Pottelsberghe de la Potterie, and Dominique Guellec
(2009) "Claiming more: The increased voluminosity of patent applications and its
determinants", Research Policy, 38, pp. 1006-1020.
WIPO (2006) Progress report on the follow up of the Joint Inspection Unit’s
recommendations as contained in its report “review of management and
administration in WIPO: budget, oversight and related issues”. Geneva: World
Intellectual Property Organisation (WIPO).
WIPO (2007a) Addition of Korean as a language of publication, PCT Union. Assembly,
Thirty-Sixth (16th Ordinary) Session Geneva, September 24 to October 3, 2007.
PCT/A/36/8. Geneva: World Intellectual Property Organisation (WIPO).
WIPO (2007b) Flexibility formula for administration of the PCT, PCT/A/36/5. Geneva:
World Intellectual Property Organisation.
WIPO (2008) World patent report. A statistical review. Geneva: World Intellectual
Property Organisation (WIPO).
WIPO (2011) PCT Applicant’s Guide – International Phase. Geneva: World Intellectual
Property Organisation (WIPO).
WIPO (2013a) PCT Yearly Review. Geneva: World Intellectual Property Organisation
(WIPO).
20
WIPO (2013b) WIPO Program and Budget for the 2012/2013 Biennium. Geneva: World
Intellectual Property Organisation (WIPO).
21
8. Appendix
Table 5: Languages in which international patent applications have been published between 2009 and
2011, by technological sector (according to the International Patent Classification).
Language
Field of technology
I - Electrical engineering
Electrical machinery,
apparatus, energy
Audio-visual technology
Telecommunications
Digital communication
Basic communication
processes
Computer technology
IT methods for management
Semiconductors
English
Japanese
German
French
Chinese
Korean
Other
Total
43.6%
28.8%
15.7%
2.8%
4.7%
3.3%
1.0%
29,489
47.7%
35.3%
4.2%
2.1%
5.6%
4.6%
0.6%
17,284
56.1%
55.0%
59.1%
18.9%
9.5%
25.5%
3.0%
1.5%
6.0%
2.1%
2.1%
3.1%
12.7%
28.9%
4.2%
6.7%
2.7%
1.5%
0.6%
0.4%
0.6%
15,778
31,259
3,876
70.6%
78.7%
15.5%
9.2%
3.5%
2.0%
2.4%
1.4%
4.9%
1.7%
2.5%
6.2%
0.6%
0.8%
30,294
6,610
48.6%
35.2%
7.5%
1.8%
2.3%
4.2%
0.4%
17,981
45.8%
55.1%
78.3%
38.9%
18.9%
10.4%
6.3%
14.8%
4.3%
2.4%
5.5%
2.4%
2.6%
2.5%
1.0%
3.2%
1.8%
2.2%
0.7%
1.5%
1.3%
13,057
19,748
5,439
55.5%
74.9%
16.5%
10.4%
16.0%
6.9%
3.9%
2.3%
3.2%
1.6%
2.4%
2.2%
2.5%
1.6%
6,679
31,705
67.9%
79.3%
80.3%
52.4%
13.0%
9.7%
8.0%
27.2%
9.0%
2.9%
2.7%
12.6%
4.2%
1.9%
2.2%
3.7%
3.0%
1.8%
2.8%
1.5%
1.8%
2.4%
2.1%
1.9%
1.1%
1.9%
2.1%
0.7%
16,493
15,751
24,021
9,001
67.1%
64.7%
16.3%
15.4%
4.2%
11.8%
2.9%
2.7%
2.3%
2.2%
3.5%
1.9%
3.7%
1.3%
4,626
14,211
42.2%
48.2%
29.0%
30.9%
13.5%
11.8%
6.1%
3.7%
3.6%
2.2%
3.1%
2.1%
2.6%
1.1%
8,848
7,555
62.9%
11.4%
10.1%
3.0%
1.0%
9.3%
2.3%
1,055
59.0%
52.9%
13.8%
20.0%
14.9%
12.8%
4.9%
5.6%
3.0%
2.8%
2.4%
3.3%
2.0%
2.5%
11,064
6,855
IV - Mechanical engineering
Handling
58.0%
Machine tools
43.6%
Engines, pumps, tur43.4%
bines
14.1%
21.6%
18.5%
16.3%
23.8%
23.2%
4.1%
3.7%
6.6%
2.4%
3.4%
3.3%
2.5%
2.3%
2.2%
2.5%
1.6%
2.7%
11,442
8,710
13,741
II - Instruments
Optics
Measurement
Analysis of biological
materials
Control
Medical technology
III - Chemistry
Organic fine chemistry
Biotechnology
Pharmaceuticals
Macromolecular chemistry, polymers
Food chemistry
Basic materials chemistry
Materials, metallurgy
Surface technology,
coating
Micro-structural and
nano-technology
Chemical engineering
Environmental technology
22
Textile and paper machines
Other special machines
Thermal processes and
apparatus
Mechanical elements
Transport
55.0%
19.5%
16.5%
2.5%
3.3%
2.1%
1.1%
6,100
54.7%
47.8%
17.3%
19.9%
14.8%
15.7%
5.3%
4.5%
2.2%
5.8%
2.8%
3.7%
2.8%
2.6%
11,974
7,412
42.0%
40.2%
18.6%
18.5%
29.0%
24.3%
4.6%
9.3%
2.7%
2.8%
1.6%
2.7%
1.6%
2.3%
12,637
17,586
V - Other fields
Furniture, games
Other consumer goods
Civil engineering
61.2%
57.2%
66.6%
9.3%
11.2%
5.8%
12.0%
14.1%
12.5%
3.7%
4.2%
4.2%
5.6%
4.0%
3.6%
5.3%
6.3%
3.7%
3.1%
3.0%
3.6%
9,578
9,174
13,595
Average
58.2%
18.0%
10.6%
3.5%
5.2%
3.0%
1.5%
23