Why Japan boasts of the most patent applications in the world. Minoru Masujima and Yoshitoshi Tanaka*

Transcription

1 Int. J. Liability and Scientific Enquiry, Vol. 4, No. 3, Why Japan boasts of the most patent applications in the world Minoru Masujima and Yoshitoshi Tanaka* Graduate School of Innovation Management, Tokyo Institute of Technology, Shibaura, Minato-ku, Tokyo , Japan *Corresponding author Abstract: Japan is the second largest recipient of patent filing in the world. This paper will investigate the reason why there are more residential patent applications in Japan than in the USA or Europe. We focused on the number of patent applications which did not request examination (non-request examination). We created multiple regression expression on the number of residential patent applications in Japan, and analysed the influence volume of some factors. The result of our study showed that the past number of patent applications with non-request for examination is the most important determinant in the number of residential patent applications in Japan. In addition, we also found that the differences in the industry field has almost no influence. The results of this will be helpful for Japanese enterprises to manage patent applications more efficiently. Keywords: intellectual property; patent applications; examination request system; multiple regression analysis; Japanese patent law system. Reference to this paper should be made as follows: Masujima, M. and Tanaka, Y. (2011) Why Japan boasts of the most patent applications in the world, Int. J. Liability and Scientific Enquiry, Vol. 4, No. 3, pp Biographical notes: Minoru Masujima graduated from Waseda University with a Bachelor s degree in Engineering in 2000, and then joined Siemens Japan KK as an Engineer in the same year. Since 2007, he has been working at the Japan Patent Office as a Patent Examiner. At the same time, he is currently a graduate student at the Graduate School of Innovation Management, Tokyo Institute of Technology. Yoshitoshi Tanaka is a Professor at the Graduate School of Innovation Management, Tokyo Institute of Technology. He graduated at the Tokyo Institute of Technology and joined Japan Patent Office as a Patent Examiner in He worked for the Japanese Government and Science and Technology Agency until He was a Research Fellow at UCLA in the USA. He is a registered Patent Attorney since He is a member of Japan Intellectual Property Academic Association, Japan MOT Society, etc. Copyright 2011 Inderscience Enterprises Ltd.

2 186 M. Masujima and Y. Tanaka 1 Introduction A patent is an exclusive right granted by law to applicants/assignees to make use of and exploit their inventions for a limited period of time (generally 20 years from filing). The patent holder has the legal right to exclude others from commercially exploiting his invention for the duration of this period. In return for exclusive rights, the applicant is obliged to disclose the invention to the public in a manner that enables others, skilled in the art, to replicate the invention. The patent system is designed to balance the interests of applicants/assignees (exclusive rights) and the interests of society (disclosure of invention). In recent years, there has been a significant increase in world patent activity. The number of patent applications in Japan was 391,002 in 2008, second only to the USA. However, although Japanese companies have been ahead in the number of patent applications submitted so far, Japan was truly a filing great power rather than real-life patent power (Arai, 1999). Japanese lag behind the USA and European community in terms of holding patents (Nakata and Zhang, 2009). Patent applications are applied by domestic applicants and foreign applicants. We define the patent applications filed by domestic applicants as residential patent applications, and the patent applications filed by foreign applicants as non-residential patent applications. Resident filing refers to an application filed at an office of or acting for the USA in which the first-named applicant in the application concerned has residence. Likewise, non-resident filing refers to an application filed at an office of or acting for the State in which the first-named applicant in the application concerned does not have residence. Table 1 shows the number of residential patent applications in major regions in The number of residential patent applications in the USA is 231,588, in European Patent Office (EPO) by the residents in European Patent Convention (EPC) member countries; 72,183, and, in Japan; 330,110, in In the number of residential patent applications in each region, Japan is the top ranking nation. Table 1 Resident EPO*: resident in EPC signatory state The number of residential and non-residential patent applications in major regions in 2008 Japan USA EPO KR CN 330, ,588 72, , ,579 Non-resident 60, ,733 74,378 43,518 95,259 Total 391, , , , ,838 Source: Japan, the USA, KR and CN: World Intellectual Property Organization (2009), European Patent Office (EPO, 2009) Why does Japan have more patent applications than the rest of the world? Some studies claim that the number of patent applications is correlated with the growth of R&D expenditures (Pakes and Griliches, 1984) or the size of enterprises (Brouwer and Kleinknecht, 1999) in the USA. However, these factors are not sufficient to explain the large number of patent applications in Japan. Table 2 shows the total R&D expenditures, the number of enterprises and the number of employment in major regions. The Japanese total R&D expenditures, the number of

3 Why Japan boasts of the most patent applications in the world 187 enterprises and the number of employment are not bigger than the USA and the European Union (EU); however, the number of residential patent applications in Japan is bigger than the USA and EU. This indicates that there are significant factors other than R&D expenditures and the size of enterprises to explain the large number of residential patent applications in Japan. Table 2 Total R&D expenditures, the number of enterprises and the number of employment in major regions Total R&D The number of enterprise expenditures 2008 (mil) US$ Year SMEs Large Total The number of employment Japan 149, ,197,719 12,351 4,210,070 40,126,980 USA 398, ,004,056 18,071 6,022, ,917,165 EU 276, ,409,000 43,000 20,452, ,805,000 Source: Total R&D expenditures: Organization for Economic Development (2010); the number of enterprise and the number of employment: European Commission (2009); The Small and Medium Enterprise Agency (2010): US Small Business Administration (n.d) This study aims to determine what factors influence the large number of patent applications in Japan. It is widely accepted that patent statistics are a reliable (although not perfect) indicator of innovative activity. For this study, we will focus on the number of request for examination system under Patent Law in Japan. According to the Japanese Patent Law, Japan Patent Office (hereafter JPO) examiner will not investigate patent application until the patent applicants request for examination and pay examination fee (Article bis). The examination can be requested by the applicants themselves, as well as by a third party within three years from the filing date (Article 48ter). The examination request was introduced in Japan in 1971 and the purpose of this system is to give the applicant a certain period to reconsider whether he should proceed with his request to have his application examined. Request for examination and payment of examination fee are needed for an application to be examined (Article 195, paragraph 2). 1 The duration for the examination request is three years and during this period, the applicant may or may not proceed with his request for examination of his invention. Under this system, if no request for examination is made within this request time (three years after then); the patent application is deemed to have been withdrawn (Article 48ter, paragraph 4). Under such a first-file system, entitlement in the case of competing applications by independent inventors is established on the basis of the filing or priority date of the application, regardless of the date of actual invention. Many countries have the same three year rule, except the USA which does not have a first-file system but a first-invention system, wherein a patent is granted to the person who first conceived and practiced the invention, rather than to the person who first filed the invention with authorities. Article 39 of Japan Patent Act states that a person who is the first-to-file an application for a patent for an invention may obtain that patent, rather than a different person who is the first-to-invent the same invention.

4 188 M. Masujima and Y. Tanaka Between 2004 and 2005, there was a sharp increase in the number of pending applications at the JPO. In 2006, there were around 836,801 patent applications awaiting examination at the JPO. However, the increase at the JPO was mostly due to the shortening of the time limit for request for examination, from seven years to three years, which has created an increased examination workload for a period of several years ( In this paper, we refer to patent applications which have not been requested for examination as patent applications with non-request for examination. Table 3 shows the ratio of patent applications with request for examination in major regions. The ratio in the United States Patent and Trademark Office (USPTO) is 100% because under the first-invention system, the filing implies a request for examination, in sharp contrast to the EPO, JPO and Korean Intellectual Property Office (KIPO), where applicants still have to formally request a patent examination. The latter is less time consuming and cheaper compared to the USPTO system when challenges to a patent arise. The first-to-file system is based on an objective criterion which can be easily determined on the face of the documents without recourse to extraneous evidence and without costs. This leads to procedural certainty as the filing date of an application can very rarely be challenged. In Japan, opposition procedure after an examiner s decision to grant a patent was abandoned in 2003; instead a trial for invalidation (Article 23) serves as the alternative. Table 3 The ratio of patent applications with request for examination Year JPO USPTO EPO KIPO Note: The number of patent applications with request for examination / the number of patent applications) Source: Japan s Patent Office (2009b) The EPO s deadline is six months after the date of issuing the search report. In JPO, these are three years after the date of filing patent application. It is five years in KIPO. However, even if we consider the difference in the deadline period, it is clear that the ratio of request for examination in Japan is very low compared with other regions. In Japan, patent applications with non-request for examination after the deadline is deemed as withdrawn without being examined. Since the number of patent applications in Japan is about 400,000 every year (Japan Patent Office, 2009a), this means that about 35%, 140,000 patent applications are withdrawn without request for examination every year. In KIPO, patent applications which are withdrawn without request for examination are about 30,000, and those in EPO are only about 10,000. The patent applications with non-request for examination, rather than R&D expenditure or the size of enterprise, might explain the high incidence of patent applications in Japan. Under the current system of request for examination, the applicant does not need to make the decision for request for examination at the time of filing, and they can evaluate the invention within the three-year deadline period for requesting examination. Figure 1 shows the trend of the number of patent applications in each industry in Japan. In the electric industry which includes 92 enterprises, the number of residential

5 Why Japan boasts of the most patent applications in the world 189 patent applications is over 100,000 every year. The number of residential patent applications of 92 enterprises is about 1/3 of the patent applications in Japan. In this Figure 1, it is understood that the number of patent applications is different in each industry. Figure 1 The trend of the number of patent applications in each industry in Japan (see online version for colours) Note: Top 299 enterprises of the number of patent applications in Japan Hasegawa et al. (2001) reported the importance of patent application management in Japanese enterprises and said that the objective of patent application is different in each industry. Therefore, this study took into account the growth of the number of residential patent applications factor may be different in each industry. We define the growth of the number of residential patent applications per the growth of a factor as the influence volume of a factor. This research also tries to clarify how different the influence volume of the influential factor is in each industry. From these backgrounds, we found out the reason why Japan has a large number of residential patent applications despite the Japanese total R&D expenditures, the number of enterprises and number of employment being lower than those of the USA and EU. We focused on the number of patent applications with non-request for examination because the number in Japan is larger than those in the USA and EU. In addition, we looked at how the past number of patent applications with non-request for examination influenced the total number of residential patent applications in Japan. In addition, we showed the difference in the influence volume in each industry. The aim of this study is to determine whether the number of patent applications with non-request for examination is the most influential factor in the high number of residential patent applications. If Japanese enterprises want to reduce their residential patent applications for efficient management, they should reduce the number of patent applications with non-request for examination. This sudy will proposed that the request for examination system in Japan should be changed similar to the European system, which has a search report before request for examination, or through an amendment of the patent law in connection with the request for examination system.

6 190 M. Masujima and Y. Tanaka 2 Literature review Pakes and Griliches (1984) made the functional form of the relationship between patents and past R&D expenditures. They analysed this form for the patenting behaviour of the 121 US enterprises during As a result, they revealed the long-term relationship between 1% of R&D expenditure and 0.61% of patent. However, though Japanese total R&D expenditures is lower than those of the USA and EU (Table 2), the number of residential patent applications in Japan is more than those in the USA and EU. Therefore, only by using their factors, we cannot reveal the reason why Japan has more number of residential patent applications than those of the USA and EU. Brouwer and Kleinknecht (1999) conducted multiple regression analysis for the number of EPO patent applications. As the objective variable, the number of EPO patent applications was defined. As the independent variables, he size of enterprise (log of number of employees), dummy of enterprise engaged in R&D collaboration, dummy of high technological opportunity sectors, R&D intensity level (R&D expenditures / net sales), etc were defined. The larger the size of enterprise is, the larger the number of patent applications is. However, it is unable to explain everything by this. Though the number of enterprises or employment in Japan is lower than those in the USA and EU (Table 2), the number of residential patent applications in Japan is more than those in the USA and EU. Therefore, only by using their factors, we can t explain the reason why Japan has more number of residential patent applications than those of the USA and EU. Yamauchi and Onishi (2008) analysed the factor which has the influence on the ratio of patent applications per R&D expenditures. They compared the characteristics of intellectual property management system for SMEs and large enterprises. They conducted multiple regression analysis based on ordinary least square. Independent variables set up market share, intellectual property intensity level (intellectual property action expenditures / net sales), R&D expenditures, the ratio of patent utilisation (the number of utilised patents / all patents), and industry dummy. As a result, they concluded that intellectual property intensity level was the significant factor of the ratio of patent applications per R&D expenditures in both SMEs and large enterprises. However, the past number of patent applications with non-request for examination is not included in their valuables. This study failed to explain how the number of patent applications with non-request for examination influences the number of residential patent applications in Japan. Kato et al. (2006) questioned the large number of patent applications in Japan, and they explained the importance of the evaluation items in deciding whether a patent application is filed by questionnaire for some Japanese enterprises. These items are novelty, inventive step, the effect of invention, the own execution possibility, the other companies execution possibility, violation discovery, strategic necessity. However, these evaluation items assume acquiring rights after patent applications. Therefore, they did not take into account the existence of patent applications with non-request for examination. Wang (2008) researched on the notification of reasons for refusal in the examination process for a granted patent for Japanese enterprises and EU enterprises in determining

7 Why Japan boasts of the most patent applications in the world 191 the reason of the large number of patent applications in Japan. The result of her study showed that patents having citations are more improvement-oriented patent, and patents having no patent citations are more basic-oriented patents. The result of statistical analysis reveals that Japanese companies have a higher percentage of more improvement-oriented patents whereas European companies retain a higher percentage of more basic-oriented patents. However, her study s approach, in contrast to this study, to the number of patent applications is different. Her approach is based on the acquired patent rights and the quality of patents while our approach is based on the issue of intent in requesting examination before acquiring patent rights and the number of patent applications. Tanaka (2010) used deductive thinking to determine factors behind the large number of patent applications in Japan. He found the following: licensing-in from western enterprises, practical behaviour reflected by utility model law, cultural issues like uncertainty avoidance, intense inter-enterprise competition, lack of collaborations between IP department and other functional departments. He focused on the lack of collaborations, conducted questionnaires, and analysed. He said: Regarding IP activities which make influences to the generation of un-utilized patents, such as patent strategy making, decision for patent applications, patent practical works, standardization, patent inventory, it was clarified that there is a lack of collaboration between IP departments with other functional departments, which will generate the situation having many unutilized patents. By using the number of patent applications with non-request for examination which are generated from the lack of collaboration between the IP departments with other functional departments, we reveal statistically how this situation influences the number of patent applications. According to previous researches, the number of patent applications is influenced by the R&D expenditure, the size of enterprise, R&D intensity level and market share etc. However, we cannot attribute the reason why Japan has the large number of residential patent applications though those factors R&D expenditure, the size of enterprise, R&D intensity level and market share etc. in Japan are lower than those in the USA and EU. Moreover, the previous researches mentioned also the contents of invention or the decision filing patent applications as important factors. In contrast, this study takes a different perspective by focusing on the number of patent applications with non-request for examination. We assumed that the past number of patent applications with non-request for examination is one of the factors on the number of residential patent applications in Japan. We analyse how they influence the number of residential patent applications. We hypothesise the following: Hypothesis 1 Hypothesis 2 The past number of patent applications with non-request for examination is the most influential factor that explains the large number of residential patent applications in Japan. This factor is more comprehensive than other factors such as R&D expenditure, the size of enterprise, R&D intensity level (R&D expenditures / net sales), and market share etc. The influence volume of the most influential factor behind the number of residential patent applications is different in each industry, in Japan.

8 192 M. Masujima and Y. Tanaka 3 Methodologies 3.1 Procedure The procedure of this study is as follows. We selected each variable (3.2). Then, we verified the Hypothesis 1 (4.1). We analysed the significant factor for the number of residential patent applications in Japan. We conducted correlation analyses between all variables, and select independent variables which is significant. We then created multiple regression expression about the number of residential patent applications in Japan. To create multiple regression expression, we used stepwise procedure. Objective variable sets up the number of residential patent applications in each enterprise while independent variables set up some indicators including the past number of patent applications with non-request for examination. Comparing each variable s coefficient values, we could understand what kind of variable was the most influential factor in the number of residential patent applications in Japan. Moreover, we checked the possibility of multicollinearity. Finally, we created a path diagram of the number of residential patent applications in Japan. As the next step, we verified Hypothesis 2 (4.2). We analysed the difference in the influence volume of the most influential factor in each industry. We conducted correlation analyses between variables selected at 4.1 in each industry. Thereafter, we calculated each industry s coefficients by using the most influential factor which clarified by multiple regression expression at 4.1. After comparing the value of same influential factor in each industry, we could understand the difference of the influence volume of the most influential factor in each industry. 3.2 The selection of each variable We selected each variable. At first, the objective variable sets up the number of residential patent applications in each Japanese enterprise. Japanese 100 enterprises which keep being published in the 200 ranks of the use of Japanese patent system in Japan Patent Office annual report were selected. The data source of the number of residential patent application is the Japan Patent Office Annual Report The application year used The total number of residential patent application 2007 in selected Japanese 100 enterprises is 164,663. This provides a hit of 49.4% of the total number of residential patent applications in Japan 2007 (333,498). Table 4 shows industry classification of selected Japanese 100 enterprises. Independent variables used indicators in 2006, because the uncertainties in the flow of time can be lost in the causal relation. Each independent variable is as follow: The number of patent applications with non-request for examination 2006: We assumed that the past number of patent applications with non-request for examination as the factor of the number of residential patent applications in Japan. Therefore, we used this variable. Japanese deadline of request for examination is three years after filing patent application. We selected the number for residential patent applications in 2003 because these periods of request for examination were ended in These numbers used the data from the Japan Patent Office (2008).

9 Why Japan boasts of the most patent applications in the world 193 Domestic net sales 2006 financial year: This variable was used by previous researches (Yamauchi and Onishi, 2008). As the indicator of existing market share in each enterprise, we used this variable. The data sources are the enterprise s annual report or summary reports on financial result from each enterprise s websites. R&D intensity level 2006 (R&D expenditures / net sales): Natural logarithm of the number of employees 2006: these variables were used by previous researches (Brouwer and Kleinknecht, 1999). As the indicator of the size of enterprise, we use natural logarithm of the number of employees The data sources are each enterprise s annual report or summary reports on financial result from each enterprise s websites. R&D expenditures, capital expenditures: We considered that the invention to acquire patent right is produced by behaviour within not only R&D but also capital expenditure. Therefore, we selected both variables. The data sources are each enterprise s annual report or summary reports on financial result from each enterprise s websites. The ratio of global patent application 2006 (the number of overseas applications / residential applications), the ratio of overseas net sales 2006 (overseas net sales / all net sales): We used these variables to investigate how each enterprise s global intention has the influence on the number of residential patent applications. The data source of the ratio of global patent application 2006 is Japan Patent Office (2008). The data source of the ratio of overseas net sales 2006 is the enterprise s annual report or summary reports on financial result from each enterprise s websites. The ratio of the US net sales 2006 (the US net sales / all net sales), the ratio of European net sales 2006 (European net sales / all net sales): We used these variables to investigate how the difference of main overseas market influences the number of residential patent applications in Japan. Table 4 Industry breakdown of the selected enterprises Industry The number of enterprise The number of patent applications Electric 28 90,661 Transport 13 22,917 Chemistry 15 14,868 Machinery 11 10,254 Manufacturing NEC 10 8,951 Primary metal 7 4,984 Precision machinery 4 4,950 Textile machinery, glass products 5 3,869 Non manufacturing 7 3,809 Total ,663

10 194 M. Masujima and Y. Tanaka 4 Analyses and results 4.1 The influential factor of the number of residential patent applications in Japan We conducted correlation analyses on selected variable. We measured the strength between two variables by Spearman rank correlation coefficient. Tables 5, 6 and 7 show the results of correlation coefficients with significance. The number of residential patent applications in 2007 had strong correlations with the number of patent applications with non-request for examination 2006 (0.766) and natural logarithm of the number of employees 2006 (0.723), and had middle correlations with domestic net sales 2006 financial year (0.426) and R&D intensity level 2006 (0.526) (Table 5). The other variables had middle correlations, except for the domestic net sales 2006 financial year and the R&D intensity level Table 5 The result of correlation analyses 1 (1) The number of residential patent applications 2007 (2) The number of patent applications with non-request examinations (3) Natural logarithm of the number of employees 2006 (1) (2) (3) (4) (5) 1.766** 1.723**.688** 1 (4) Domestic net sales 2006 financial year.426**.503**.631** 1 (5) R&D intensity level **.370**.409** (R&D expenditures/net sales) Notes: *significant at 5% level (two-sided tests) **significant at 1% level (two-sided tests) The natural logarithm of the number of employees 2006, R&D expenditures 2006, and capital expenditures 2006 had strong correlations (0.898, 0.775, and 0.833) (Table 6). We selected one variable of natural logarithm of the number of employees 2006, except for the other two variables from the following analyses, because it should consider the possibility of multicollinearity. Table 6 The result of correlation analyses 2 (1) Natural logarithm of the number of employees 2006 (1) (2) (3) 1 (2) R&D expenditures ** 1 (3) Capital Expenditures **.833** 1 Notes: *significant at 5% level (two-sided tests) **significant at 1% level (two-sided tests) The ratio of overseas net sales in 2006 had middle correlations with the ratio of global applications 2006 (0.407), and strong correlation with the ratio of the US net sales 2006 (0.799) and European net sales 2006 (0.754) (Table 7). We selected one variable of the ratio of overseas net sales 2006 and exclude the ratio of the US net sales 2006 and

11 Why Japan boasts of the most patent applications in the world 195 European net sales from the following analyses, because of the possibility of multicollinearity. In addition, the ratios of the US net sales 2006 and European net sales 2006 had middle correlation (0.692). Therefore, we considered it difficult to classify Japanese enterprises by their main overseas market and we found that enterprises of the high ratio of the US net sales have high ratio of European net sales too. Table 7 The result of correlation analyses (1) (2) (3) (4) (1) The ratio of overseas net sales (2) The ratio of global applications ** 1 (3) The ratio of the US net sales ** (4) The ratio of European net sales **.298*.692** 1 Notes: *significant at 5% level (two-sided tests) **significant at 1% level (two-sided tests) We created multiple regression expression on the number of residential patent applications 2007 by using selected independent variables. Each independent variable does not have strong correlations with others. Table 8 shows the result of multiple regression analysis. In this multiple regression expression, multiple correlation coefficient was 0.91, multiple determination coefficient was 0.83, and corrected multiple determination coefficient was As p-value is 0.000, this multiple regression expression had high significance. Table 8 The result of multiple regression analysis (p-value is 0.000) Coefficient Standard error Standardised partial regression coefficient E p-value (Intercept) 3.18E The number of patent 1.23E E applications with non-request for examination 2006 Natural logarithm of the 3.26E E number of employees 2006 Domestic net sales E E financial year R&D intensity level 2006 (R&D expenditures/net sales) 1.30E E As the result of this multiple regression analysis, standardised partial regression coefficient of the number of patent applications with non-request for examination 2006 was Natural logarithm of the number of employees 2006 and domestic net sales in 2006 financial year were the same standardised partial regression coefficient (0.18). And, standardised partial regression coefficient of R&D intensity level was The remainder variables did not become significant variables (the ratio of overseas net sales 2006 and the ratio of global applications 2006). To these standardised partial regression coefficients, we found that the number of patent applications with non-request for examination 2006 has the highest influence volume to the number of residential patent applications in selected independent variables.

12 196 M. Masujima and Y. Tanaka Moreover, we checked the possibility of multicollinearity between selected independent variables. Objective variable set up one selected independent variable, and we conducted multiple regression analyses by using other independent variables. We checked each multiple correlation coefficients (Table 9). As these coefficients are lower than 0.7, we found that the possibility of multicollinearity between selected independent variables is low. Table 9 Multiple correlation coefficient of each independent variable Multiple correlation coefficient of each independent variable The number of patent applications with non-request for 0.58 examination 2006 Natural logarithm of the number of employees 0.69 Domestic sales 2006 financial year 0.60 R&D intensity level 2006 (R&D expenditures / net sales) 0.37 Figure 2 shows the path diagram of the result of 4.1. The factors which have the influence on the number of residential patent application 2007 are the number of patent applications with non-request for examination 2006 (0.67), natural logarithm of the number of employees 2006 (0.18), domestic net sales 2006 financial year (0.18), and R&D intensity level (0.14). The number of patent applications with non-request for examination 2006 was the most influential determinant. These factors had middle correlations except between domestic net sales 2006 financial year and R&D intensity level We found that the possibility of multicollinearity between these factors is low. Figure 2 Path diagram of the number of residential patent applications in Japan The influence volume of the most influential factor in each industry We analysed the difference in the influence volume of the influential factors in each industry. The industries selected are the electric industry, transport industry, and chemistry industry, because these industries are more patent applications oriented than the other industries. We conducted correlation analyses in each industry. These variables selected the number of residential patent applications 2007, the number of patent applications with non-request for examination 2006, natural logarithm of the number of employees 2006, domestic net sales 2006 financial year, and R&D intensity level We measured of the strength between two variables by Spearman rank correlation coefficient.

13 Why Japan boasts of the most patent applications in the world 197 Table 10 The result of correlation analyses in each industry Electric (N = 28) Transport (N = 13) Chemistry (N = 15) (1) (2) (3) (4) (5) (1) (2) (3) (4) (5) (1) (2) (3) (4) (5) (1) (2).854** 1.560* 1.675** 1 (3).813**.753** 1.692**.786** 1.654**.657** 1 (4).741**.614*.842** 1.621*.819**.892** 1.539*.839**.764** 1 (5) ** ** Notes: (1) The number of residential patent applications 2007 (2) The number of residential patent applications with non-request for examination 2006 (3) Natural logarithm of the no. of employees 2006 (4) Domestic sales 2006 financial year (5) R&D intensity level 2006 (R&D expenditures / net sales) *significant at 5% level (two-sided tests) **significant at 1% level (two-sided tests)

14 198 M. Masujima and Y. Tanaka Table 10 shows the result of correlation analyses in each industry. We compared each result with the result of 100 enterprises (Table 5). The number of residential patent applications 2007 in each industry had the strong or middle correlations with the other variables except R&D intensity level 2006 in electric industry. The other variables in the 100 enterprises had only the middle correlations. However some of these in each industry had strong correlations. In electric and transport industries, the number of patent applications with non-request for examination 2006 and natural logarithm of the number of employees 2006 had the strong correlations. In the transport and chemistry industries, the number of patent applications with non-request for examination 2006 and domestic net sales of the 2006 financial year had strong correlations. In all industries, we could confirm the strong correlations between natural logarithm of the number of employees 2006 and domestic net sales As the result of these correlation analyses, we could not conduct multiple regression analysis in the number of residential patent applications in each industry because the possibility of multicollinearity is high. We focused instead on the most influential independent variable to the number of residential patent applications. Then, we conducted simple regression analyses on the number of residential patent applications in each industry. Object variable sets up the number of residential patent applications 2007 in each industry. Independent variable sets up the number of patent applications with non-request for examination 2006 because this variable had the highest influence volume to the number of residential patent applications in 100 enterprises (Table 8). Table 11 shows the standardised partial regression coefficient of single regression analyses in each industry. In all industries, the number of patent applications with nonrequest for examination 2006 had a high significance standardised partial regression coefficient to the number of residential patent applications 2007 (0.893, 0.778, and 0.974). We found that the number of patent applications with non-request for examination 2006 is the most influential factor of the number of residential patent applications 2007 in all industries. However, because the influence volume is high in each industry and the difference of the influence volume was small, we could not determine how the influence volume of the number of patent applications with nonrequest for examination 2006 is different in each industry. We found that the influence volume of this factor is not different in each industry. Table 11 Standardised partial regression coefficient Standardised partial regression coefficient in each industry Note: **significant at 1% level Electric industry (N = 28) Transport industry (N = 13) Chemistry industry (N = 15) 0.893** 0.778** 0.974** 5 Discussion We conducted two analyses to verify the two hypotheses. At first, we analysed the influential factors to explain the number of residential patent applications in Japan (4.1). As the result, the factors which influence the number of residential patent applications 2007 are the number of patent applications with non-request for examination 2006 (0.67),

15 Why Japan boasts of the most patent applications in the world 199 natural logarithm of the number of employees 2006 (0.18), domestic net sales 2006 financial year (0.18), and R&D intensity level 2006 (0.14). We found that the past number of patent applications with non-request for examination is the most important determinant, in comparison to other factors such as the size of enterprise, R&D intensity level, and market share that influences the number of residential patent applications. Therefore, Hypothesis 1 is accepted. As a next step, we analysed whether the influence volume of the past number of patent applications with non-request for examination is different than to the number of residential patent applications in Japan in electric, transport, and chemistry industry (4.2). As the result, the influence volume of the factor in electric industry was 0.893, the influence volume in transport industry was 0.778, and the influence volume in chemistry industry was We could not judge how the influence volume of this factor is different in each industry because the influence volume in each industry was small. Therefore, Hypothesis 2 was not clearly verified. We considered why Hypothesis 1 was accepted, and Hypothesis 2 was rejected. As Hypothesis 1 is accepted, we found that the past number of patent applications with non-request for examination is the most important determining factor in the number of residential patent applications in Japan. Hypothesis 2 was rejected and we found that the influence volume of the past number of patent applications with non-request for examination is not different in any industry. Therefore, we found that the past number of patent applications with non-request for examination is the most influential factor in any industry. We found that in Japanese enterprises, most of the applicants do not have the intention of requesting for examination. We can mention that Japanese enterprises have a major tendency of filing patent applications even before they decide whether the invention should acquire patent rights. We considered two reasons from the legal viewpoint in Japanese patent law. In Japan, the deadline for requesting examination is three years after filing the patent applications. When Japanese enterprises file patent applications, they do not think it is necessary to decide whether they could acquire patent rights. After filing the application, it is only then that they investigate the existence of prior arts and evaluate whether the invention should be commercialised or not. Only after this process do they then decide whether they should acquire patent rights. As long as the number of patent applications with non-request for examination in Japan is large compared with those of other regions, we can surmise that there is a tendency for Japanese enterprises to file patent applications without sufficient evaluation on its commercialisation at the timing of filing. Second, we considered the fee structures of patent application and the request for examination. In Japan, the fee for filing is quite low, but the fee for requesting an examination is quite expensive, and therefore, there is a tendency for the Japanese enterprises to file patent applications without having enough evaluation before filing. When enterprises file patent applications in Japan, they only pay a filing fee to JPO at the time of the filing, and in the case of request for examination, during the filing for request. However, in the USA, enterprises must pay filing fee, search fee, and examination fee together at the USPTO. In Europe, they must pay filing fee, search fee, and designation fee to EPO at the time of the filing. In addition, from the viewpoint of management, Japanese enterprises might have the tendency of filing patent application without the intention of acquiring patent rights at the time of filing. We can consider two explanations: First reason is to make other enterprises

16 200 M. Masujima and Y. Tanaka be unable to acquire patent rights, even though the invention or the patent right might not be important to him or his business. The second reason is to encourage and motivate employees to innovate. When an enterprise doesn t allow filing patent application of an in-house invention, the inventor s morale might be lowered. Therefore, the enterprise tends to file without the intention to acquire rights itself. These reasons explain the incident of high number of patent applications in Japan. From a legal and management perspective, we can consider that Japanese enterprises have the strong tendency of filing patent application without the intention or the purpose of acquiring patent rights. Japan is a filing great power rather than real-life patent power. In other parts of the world, if Japanese companies behave similarly, it would cost a substantial amount of money. To reduce the number of patent applications with non-request for examination, the Japanese enterprises have to avoid filing patent applications without the intention or the purpose to acquire patent rights. Therefore, before filing patent applications, the enterprise must decide whether the invention should acquire patent rights at the timing of filing. However, the law must also be amended to discourage this wasteful exercise of resources. Under the current law, the applicant has to make an evaluation on the invention within three years, without receiving any information from JPO, such as prior arts search report. In contrasy, applicants in the EPO receive the search report from EPO and then decide whether it should be requested for examination within six months from the receipt of the search report. Fee structures for the fee for filing and the fee for request for examination must be paid together. This will deter nuisance filing. Other important aspects remain. Patent applications with non-request for examination is opened to the public after 18 months from the date of filing. This means that so many technical information will be disclosed without being protected, resulting in the outflow of technical information through 18 months publications. It is necessary for enterprises to be able to file patent application efficiently and at the same time, have their invention protected. 6 Conclusions The result of our study shows that there is a huge number of patent applications in Japan, but fewer right holders. The discrepancy is due to the ease and low cost of filing patent applications. We concluded, using analytical data of high regression coefficient with the number of residential patent applications, that the number of patent applications with non-request for examination is the most critical determinant in explaining the high number of patent applications. The Japanese enterprises tend to file patent applications without intention of actually acquiring patent rights. At the same time, this conduct results in unexpected outflow of technology through 18 months of unprotected publications. In order to make the Japanese patent protection more efficient, it is necessary to amend existing provisions in the Patent Law so that filing fee and request for examination fee are paid at the time of the filing of the application.

17 References Why Japan boasts of the most patent applications in the world 201 Arai, H. (1999) Intellectual property policies for the twenty-first century: the Japanese experience in wealth creation, Policy Advisory Commission of The World Intellectual Property Organization (WIPO). Brouwer, E. and Kleinknecht, A. (1999) Innovative output, and a firm s propensity to patent: an exploration of CIS micro data, Research Policy, Vol. 28, No. 6, pp European Commission (2009) First section of the annual report on EU small and medium-sized enterprises. European Patent Office (2009) Annual Report Hasegawa, K., Nagata, A., Hirata, T., Sasaki, T. and Toyama, R. (2001) Current status of intellectual property management of Japanese enterprises, Proceedings of The Japan Society for Science Policy and Research Management 2001, Vol. 16, pp Japan Patent Office (2008) Japan Patent Office Annual Report Japan Patent Office (2009a) Japan Patent Office Annual Report Japan Patent Office (Ed.) (2009b) Four Office Statistics Report 2008 Edition. Japan Patent Office (2010) Japan Patent Office Annual Report Kato, K., Ishi, K. and Sugawa, S. (2006) Research of invention evaluation for patent application decision making support, Journal of Information Processing and Management, Vol. 49, pp Nakata, Y. and Zhang, X. (2009) A Survival Analysis of Patent Examination Request in Japanese Electronic and Electrical Manufacturers, available at epip04/files/zhang_xingyuan.pdf (accessed on 16 September 2010). Organization for Economic Co-operation and Development (2010) Main Science and Technology Indicators: 2010/1 Edition. Pakes, A. and Griliches, Z. (1984) Patents and R&D at the firm level: a first look, in Griliches, Z. (Ed.): R&D Patents and Productivity, pp.55 72, University of Chicago Press. Tanaka, Y. (2010) Research on the factors how to avoid un-utilized patents to support strengthening technology management, Proceedings of PICMET 2010, pp The Small and Medium Enterprise Agency (2010) 2010 white paper on small and medium enterprises in Japan. US Small Business Administration (n.d.) Firm Size Data, available at advo/research/data.html, (accessed on 16 August 2010). Wang, J. (2008) Comparative study of Japanese and European companies behavior in patent application strategy of selection and concentration in patent acquisition, Project Report of Graduate School of Innovation Management, Tokyo Institute of Technology, 16 August 2010, available at World Intellectual Property Organization (2009) World Intellectual Property Indicators Yamauchi, I. and Oonishi, K. (2008) Research on intellectual property management system for SMEs, Research Report Concerning Application Trend etc. of Industrial Property Right etc. in Our Country in 2007 Fiscal Year, pp Notes 1 The Japan Patent Office s interpretation of the patent law related to examination procedure is provided for in the Examination Guidelines for Patent and Utility Model in Japan.