SPECIAL ARTICLEEconomic & Political Weekly EPW june 7, 200839Economic Development and Patenting BehaviourBiswajit Dhar, C Niranjan RaoThe capacity of countries to take advantage of the patent system bears a relationship with their stage of development. This paper explores the relationship between economic development and domestic and foreign patenting behaviour. The study uses a unique data set covering 55 countries and 24 years. It determines the association of domestic patenting with gross domestic product per capita and openness to trade, and the association of foreign patenting with these variables and with foreign direct investment as a proportion of GDP.The use of patents data as an indicator of inventive activity has long evoked considerable interest among students of technology over several decades. Although several ques-tions have been raised about the veracity of using patent data to measure technological progress across countries, this indicator has remained “as a mirage of wonderful plentitude and objectiv-ity” in a desert of data for answering questions about the “rate of technical and scientific progress” over “time and across indus-tries and national boundaries” [Griliches 1990]. The questions that the patent data help us answer are of considerable impor-tance in a world where technology has become the prime motive force for economic development.But while the levels of patenting activity across countries are of considerable significance for explaining the relative levels of technological advancement, what is equally important is to understand the determinants of the observed levels of patenting. This is an issue that requires to be investigated at some depth for two reasons. The first and the more obvious of the reasons is that countries have historically displayed considerable differences in their levels of patenting, which need some cogent explanations. The second reason, one that has its genesis in the ongoing dis-cussion on the evolving regime of intellectual property protec-tion, is that the levels of patent protection that the countries should provide, need to be in keeping with their development needs. What this argument alludes to, in other words, is that the capacity of the countries to take advantage of the patent regimes bears a relationship with their stage of development.1 These are some of the issues that the present paper would try to address by relating patenting activity in 55 countries over a 24-year period to a comparable set of development indicators. The paper is organised as follows. Section 1 reviews the litera-ture on determinants of domestic and foreign patenting activities. Section 2 deals with the limitations of patent data as a measure of inventive activity, inter-country comparability of patent data and using patent applications data. Section 3 provides a description of the database used in the study. In section 4 we discuss the models used and report the results of regressions. Section 5 concludes.1 LiteratureReviewIn this section we will present a brief review of relevant literature by referring to studies that explain either domestic or foreign patenting behaviour or both in an inter-country comparative framework.1.1 DomesticPatentingBehaviourVayrynen (1977) estimates Spearman’s rank-order correlation coefficients for 10 African countries between patents granted (both domestic and foreign) per capita and GDP andGDP per The work on this article was done at the Research and Information System for the Non-aligned and Other Developing Countries, New Delhi. We are thankful to the World Intellectual Property Organisation, Geneva, for funding the research project from which this paper was generated.Biswajit Dhar (bisjit@gmail.com) is with the Centre for WTO Studies, Indian Institute of Foreign Trade, New Delhi. C Niranjan Rao (niranjan@cess.ac.in) is with the Centre for Economic and Social Studies, Hyderabad.

SPECIAL ARTICLEjune 7, 2008 EPW Economic & Political Weekly40capita, both for 1971. While GDP has been taken as a measure of the market size, GDP per capita has been taken to represent the level of development. The correlation coefficients obtained for both sets of relationships, viz, patents and GDP and patents and GDP per capita were quite significant -0.90 and 0.85 respectively.Schiffel and Kitti (1978) try to explain the behaviour of domes-tic patent applications in a set of five countries, viz, Canada, France, West Germany, Japan and the United States (US). The explanatory variable taken for this exercise was research and development (R&D) expenditures, lagged by two years. The data covered the period 1963-74 for all countries except Japan for which the data covered the period 1967-74. The results showed that while for Canada and Japan, the R&D coefficient was positive and significant, for West Germany, it was negative and signifi-cant. For France and the US, the R&D coefficient was not signifi-cant. These results, according to the authors, were obtained because of the difficulties in measuring R&D expenditures.Soete (1981) explains domestic patenting in 19 Organisation for Economic Cooperation and Development (OECD) countries. The dependent variable taken for this exercise is the log of domes-tic patent applications per capita, averaged for the years 1976-78 and the independent variable isR&D spending per capita in the business enterprise sector for the year 1975. The author finds that R&D spending in the business enterprise sector is positive and significant in explaining domestic patents.Watanabe (1985) estimates correlation coefficients between the log of resident patent applications per capita, log of gross domestic manufacturing product per capita (for 53 countries, 16 of which are industrialised) and log of GNP per capita (for 60 countries, 21 of which are industrialised) for the year 1979. This study shows that a positive and significant relationship exists between the variables.1.2 Foreign Patenting BehaviourSchiffel and Kitti (1978) try to explain foreign patenting activity in theUS taking place from Canada, France, West Germany, Japan, the Netherlands, Sweden, Switzerland and the United Kingdom (UK). In this study, the independent variables are exports to the US originating in the concerned country and domestic patents granted in the concerned country. The period covered by this study is 1965-74 for all countries, except the Netherlands for which it is 1965-73. The results obtained show that domestic patenting is related in a positive and significant manner for Switzerland, while it is negative and significant for the Netherlands. For all other countries, it is not significant. The exports variable is positive and significant for all countries except for the Netherlands.In the same paper, Schiffel and Kitti also try to explain foreign patenting in Japan by applicants from Canada, France, West Ger-many, the Netherlands, Sweden, Switzerland, the UK andUS. As in the earlier case, the independent variables are exports to Japan from the concerned country and domestic patents in the con-cerned country. The period covered for this exercise is identical to that considered for the earlier case. In the case of Japan, domestic patents as an explanatory variable are found to be positive and significant for Canada and Switzerland and negative and significant for West Germany and Sweden. For all other countries, it is not significant. The exports variable is positive and significant for Canada, Sweden and Switzerland and not signifi-cant for all the other countries.The paper by Bosworth (1980) explains foreign patenting by theUS in 50 countries in the year 1974. The explanatory variables taken are GDP andGDP per capita in the receiving coun-try and exports of the US and foreign investment (the number of foreign firms being used as a proxy) from the US to the concerned country. The log-linear results show that all the independent variables are positive and significant but only exports and foreigndirect investment variables are significant at 1 per cent level of significance. Soete (1981) tries to explain patenting behaviour of foreign applicants from 17OECD countries in five of the major patent granting countries, viz, France, West Germany, Japan, theUK andUS. While the dependent variable is the number of patents granted averaged for the years 1976-78, the independent variable is theR&D expenditure of the business enterprise sector for the year 1975. It is interesting to note that domestic R&D expenditure is being used to explain foreign patents being taken in that par-ticular country. The log-linear results of regression between for-eign patents per capita and business enterpriseR&D per capita show positive and significant relationship for all the countries.Bosworth(1984)triesto explain the patenting activity of the UK abroad. He uses a cross section of data for 50 countries for the year 1974. The explanatory variables are GDP, GDP per capita, exports from the UK to the concerned country and size of multina-tional operations (the number of the UK subsidiaries in the con-cerned country is used as a proxy). The log-linear regression shows that all the independent variables are positive and significant.In the same paper, Bosworth also tries to explain foreign pat-enting activity in theUK. He uses a cross section of data for 22 countries for the year 1974. The explanatory variables are imports into theUK from the concerned country, size of multinational operations from that country (the number of subsidiaries from the originating country being used as the proxy) and domestic patenting in the originating country. The log-linear regression shows that the size of multinational operations from that country and domestic patenting activity of the concerned country are positive and significant, while imports from that country to theUK is not significant in explaining foreign patenting activity in theUK.Eaton and Kortum (1996) try to explain the flow of patent applications among 19OECD countries for the year 1988. The explanatory variables are human capital (average years of school-ing), imports of one country from the concerned country relative to one’s GNP, the ratio of R&D scientists and engineers to total labour force and relative productivity of one country in compari-son with the concerned country. The log-linear equation explain-ing the ratio of patent applications and total labour force shows that human capital, intensity of R&D scientists and engineers’ employment and relative productivities are positive and signifi-cant, while import intensity is not significant. This equation also contains four variables that can be used as an indicator of the strength of the patent system. The authors find that while

SPECIAL ARTICLEEconomic & Political Weekly EPW june 7, 200841countries providing strong patent protection attract more foreign patents, patent strength and domestic patenting activity are not so related.It is interesting to note that except one, all the studies were conducted during 1977-85. Considering the importance of the issue and the intense controversy it generated, one would have expected more studies, which explain determinants of domestic and foreign patenting. The studies reviewed above, which cover mostly developed countries are disparate and not comparable. Hence we cannot draw any overall conclusions. Our attempt in this paper is to address these limitations in the available literature.2 Limitations of Patent DataAny exercise that tries to explain cross-country patenting behaviour should necessarily look at the quality of data that areavailable. Comparison of patent data across countries, in our view has several problems, which we would discuss in this sec-tion. These problems could be grouped under three broad headings relevant for our study: (1) patents as indicators of inven-tive activity, (2) international comparability of patent data; and (3) use of patent applications data.2.1 Patents as Indicators of Inventive ActivityOne could expect that all patentable inventions would be pat-ented. But, in fact, not all patentable inventions are patented. Mansfield (1986) estimates that while in the pharmaceutical, oil and machinery industries more than 80 per cent of patentable inventions are patented, it is only 60 per cent in the case of pri-mary metals and automobile industries. The inter-industry dif-ferences in propensity to patent arise from the fact that while some industries such as pharmaceuticals depend on patent pro-tection as an appropriation mechanism, others do not. Taylor and Silberston (1973) showed that the fine chemical industry in general and the pharmaceutical industry, in particular, depend on the patent system. Ease of imitation is one of the reasons for this industry’s reliance on the patent system. The aircraft industry, for example, depends much less on the patent system, while they spend heavily on R&D, their output in terms of patents is very small.There are considerable differences in quality among patents. While a few patents represent significant advancements in tech-nology, most others do not appear to be so. One of the major con-tributors to the vastly varying quality of patents is the differences in interpretation of patentability criteria by patent offices.2 The patentability criteria, viz, novelty, non-obviousness and indus-trial applicability are difficult concepts to define and implement precisely. This point can be corroborated by the fact that a signifi-cant proportion of patents whose grant is challenged is held to be invalid by courts. One reason given is that patent offices do not have the resources to do a thorough examination of each and every patent,3 whereas a court is obliged to conduct a through examination of a patent before determining its validity [Engel 1985; Federal Trade Commission 2003].There have been attempts to quantify differences in the qual-ity of patents through the use of number of claims in a patent specification, renewal information and citations. The most successful of these attempts have been the patent renewal models. These models have tried to capture the requirement mandated in several national patent laws that a patent has to be renewed periodically for it to be effective through its complete term. The main reason behind this requirement seems to be to avoid giving a longer lease of life to the economically unviable patents. We can therefore assume that a patent, which has been renewed so as to complete the entire term of protection, is more valuable than the one, which was allowed to lapse. This information has been used to model patent renewal by Pakes (1986) and Schankerman and Pakes (1986). Using the patent renewal information and patent fee schedules, Schankerman and Pakes (1986) have shown that the distribution of private value of patents is highly skewed, while a few patents are very valuable a large majority of patents have lowvalue.Firm level differences in propensity to patent arise from the strategic perspectives of these firms towards R&D and patenting. Finally, these inter-industry and inter-firm propensities could change over time.2.2 International Comparability of Patent DataPatent data may not be comparable among countries because of differences in patent laws.4 There are vast differences in patent laws of different countries. That a country’s patent law would influence the patent output of that country has been accepted as a truism. Hence, it could be argued that patent data across coun-tries are not strictly comparable. In other words, if we have to compare patent data across countries, the patent laws in these countries should be the same. But in reality no two patent laws are the same. Further, these differences in patent laws are often accentuated by the differences in practices of the patent offices in interpreting these laws and ultimately the enforcement of these provisions.The differences in patent laws could arise because of various provisions. The more important among these are patentability criteria, coverage, duration, compulsory licensing and definitions of residents. Moreover, these laws could change over time. But the changes in patent laws are not very frequent.The generally accepted patentability criteria are novelty (new), non-obviousness (inventive step) and industrial applicability (usefulness).5 In addition, there are differences in the manner in which patent law defines novelty, either as novel in the world or only in the country concerned. Disclosure requirements, which are supplementary criteria of patentability, may also vary across countries. Some countries have only a patent registration system where there are only requirements of form to be fulfilled and no substantive examination of patent application takes place. Some of the former colonies had or have a dependent patent system linked to the former colonial power, for example, Hong Kong and Singapore. Australia, Germany, Japan and the Netherlands fol-lowed a “deferred examination system” where substantive exam-ination takes place only if specific request has been made within a specified period, which can be up to seven years from the date of application [UNCTAD 1975].Different countries have excluded specific sectors from the ambit of patenting at different times. Among the sectors that

SPECIAL ARTICLEjune 7, 2008 EPW Economic & Political Weekly42have been excluded were: pharmaceutical product patents, phar-maceutical process patents, food products, food processes, chemical products, methods of treatment of human or animal body, cosmetics, fertilisers, mixture of metals and alloys, agricul-tural machines and anti-contaminants. The other explicit exclu-sions found in some countries are animal varieties, plant varie-ties, biological processes for producing animal or plant varieties, micro-organisms and substances obtained by microbiological processes. Computer programmes and nuclear inventions have also been excluded by certain countries [WIPO 1988].The duration of patent protection is another important source of difference in the patent laws of countries. The patent term could not only vary among countries, it could also be different for different sectors within a country. Moreover, the date on which the term of the patent protection starts could vary across coun-tries, the most important of which are the date of application and the date of grant.There could be important differences in the way the compul-sory licensing provisions are incorporated in a patent law. While many countries in the world have compulsory licensing provi-sions in their patent law to check possible abuse of patent mono-poly, the US is the only country, which uses anti-trust provisions to the same effect. Many developed6 countries, which had strong compulsory licensing provisions in the past, diluted them in later years as they attained technological capability. Developing coun-tries, which inherited colonial patent laws, reformed their patent systems in the late 1960s and early 1970s, an important compo-nent of which was to have compulsory licensing provisions. As of the late 1980s, while developed countries had less stringent com-pulsory licensing provisions, many developing countries had strong compulsory licensing provisions. Non-working of the pat-ent is the main ground on which compulsory licences can be granted in a large number of countries (ibid). The other grounds on which a compulsory licence can be given are public interest, satisfaction of the reasonable requirement of the public, needs of export market and public health. The presence of compulsory licensing provisions may affect patent output, for instance Scherer (1977) in a survey of the US firms finds that anti-trust decrees involving compulsory licences discourage firms’ propen-sity to patent. Generally, the scope of patent rights covers making, selling and using the patented product in the case of product patent and using and selling a product directly obtained by a patented pro-cess. There are significant differences among countries in incor-porating importation as a right of the patent holder. While some countries explicitly include importation as a right of the patent holder, other countries do not explicitly mention it.Finally, there are differences in the definition of residents and non-residents among countries. While in the US “the concept of residence is determined by the place of residence of the first-named inventor rather than that of the applicant”, in Japan “the concept of residence is determined by the nationality of the applicant rather than the country of residence of the applicant”. All the other countries consider residence of the applicant rather than the nationality for making the distinction between foreign and domestic patentees.Similarly there may be differences in the quality of patents granted. This may be because of differences in the practices of the patent offices in interpreting the various provisions of the patent law. One reason for this could be the subjective element, which normally surrounds these provisions. Another reason could be the differences in capabilities of patent offices. Many a patent office in developing countries may not have the human or material resources, which are required to implement a patent law satisfactorily.There are differences in the enforcement of patent rights across countries. While many countries have only civil remedies in the case of infringement, a few countries have criminal provisions. The courts play an important part in the enforcement of patent rights and their practices differ among countries.Another source of differences in patent counts, which makes inter-country comparisons difficult, is the number of claims allowed for each patent. Japanese patent counts seem to be exag-gerated because of practice of granting a patent for each claim, while in case of other countries, patents contain multiple claims.The two most important factors affecting the patent output of a country could be the “strength” of the patent law and techno-logical capability. We can assume that a “strong” patent system would generate more patents than a “weak” one but two coun-tries at different levels of technological capabilities having the same “strong” patent system would not have the same patent out-put. Hence, technological capability has a strong influence on the patent output of a country. We assume that the technological capability of a country has more influence than the patent system in determining the patent output of that country.2.3 Use of Patent Applications DataWe propose to use data relating to patent applications rather than patents granted for our analysis. Although there are a few limita-tions of using the applications data, the most important of which being the fact that only some of the applications are granted pat-ents, there are a number of compelling reasons justifying their use in an exercise of the kind undertaken here. The first and the foremost reason for using data on patent applications is that while patent applicants decide the date of application, the practices of patent offices determine the date of grant.7 Furthermore, applica-tions for patents submitted in a year belong to that particular year but grants for a particular year contain applications made in different years. Data on patent applications are therefore much more comparable across countries than the data on grants are, [Eaton and Kortum 1999] and hence, our reliance on former data set. We recognise that the ideal data set would have been grant data arranged according to date of application but such a data set is not available. 3 Database for StudyThe source of data on patenting activity of different countries is the World Intellectual Property Organisation’s annual compilation Industrial Property Statistics. We have collected data on patent applications and grants for residents and non-residents8 for the period 1975-98.9 The choice of 1975 as the initial year for the study was driven by the consideration that patent laws were in the process

SPECIAL ARTICLEEconomic & Political Weekly EPW june 7, 200843of being changed in many countries, during the late 1960s and early 1970s and this process was completed by mid-1970s.Data on patents as is reported by theIndustrial Property Statis-tics include patents, inventors’ certificates, patents of importa-tion, patents of introduction, revalidation patents, precautionary patents and petty patents.The data set used for analysis in this study contains 55 coun-tries, out of which, 20 are developed countries and 35 are devel-oping countries.10 These countries are Argentina, Australia, Aus-tria, Belgium, Bolivia, Brazil, Canada, Chile, China, Colombia, Costa Rica, Denmark, Ecuador, Egypt, El Salvador, Finland, France, Germany11, Greece, Guatemala, Honduras, India, Iran, Ireland, Israel, Italy, Jamaica, Japan, South Korea, Malaysia, Mauritius, Mexico, Morocco, the Netherlands, New Zealand, Norway, Pakistan, Panama, Peru, the Philippines, Portugal, SouthAfrica, Spain, Sri Lanka, Sweden, Switzerland, Thailand, Trinidad and Tobago, Tunisia, Turkey, the UK, the US, Uruguay, Venezuela and Zimbabwe.Out of the 55 countries in our sample, 51 countries have com-plete data (for all the 24 years) and four countries have incom-plete data (for less than 24 years). The four countries with incom-plete data along with the period for which data are unavailable are: China (1975-84); Malaysia (1975-84); Thailand (1975-78) and Zimbabwe (1975-78). While China adopted a patent system only in 1984, Malaysia had a re-registration system until 1983. Thai-land enacted a patent law only in 1979 and Zimbabwe did not report during the period 1975-78. This makes our data set an unbalanced panel data.The patent data inIndustrial Property Statistics is based on vol-untary reporting by countries. In other words, if a country reports, data are included otherwise it is omitted. There is no mechanism to address the problem of non-reporting by countries. The extent to which countries provide patent data may be taken as an indicator of how important they felt the patent system to be. Some countries report only total applications without giving a break up of resident and non-resident applications. Only 20 coun-tries have provided data for each year, along with the break up between residents and non-residents during the entire period under study.Out of the 55 countries in our sample, 30 countries have not reported for at least one year and one country did not give a break-up between residents and non-residents for at least one year. Five countries did not report and did not give a break-up between residents and non-residents at least for one year. We have fitted a trend equation for available data and then estimated for the missing years.12 In cases where the break-up between resi-dents and non-residents are not available, we have used the aver-age proportions for the nearest available years. Out of a total 1,292 observations in our database, the non-reporting obser-vations were 158 (12.2 per cent) and no-break-up between residents and non-residents were 16 (1.2 per cent) making a total of 174 (13.4 per cent). The countries for which we had to estimate the missing data are: Argentina, Bolivia, Brazil, Chile, China, Colombia, Costa Rica, Ecuador, Egypt, El Salvador, Greece, Guatemala, Honduras, India, Iran, Iraq, Ireland, Italy, Jamaica, Japan, South Korea, Malaysia, Mexico, Morocco, Pakistan, Panama, Peru, the Philippines, South Africa, Sri Lanka, Thailand, Trinidad and Tobago, Tunisia, Turkey, Uruguay, Venezuela and Zimbabwe.Besides the problems with the data set mentioned above, there is yet another relating to Japanese patent data. The count of pat-ent applications in Japan is not comparable with patent counts in other countries for the following reason. While the practice of patent offices in all the other countries is to grant one patent with multiple claims, Japan has followed the unique practice of grant-ing patents covering each of the claims. Hence, the patent counts from Japan are inflated. Thus, “Okada (1992) finds that Japanese patents granted to foreigners contain on average 4.9 times as many inventive claims as those granted to Japanese inventors” [referred to in Eaton and Kortum 1996]. Following the above-mentioned studies, we have also made the necessary adjustments to make Japanese patent data comparable to that of the other countries. With these modified patent counts, Japan was placed third during 1975-79 and second during 1980-98 in the overall domestic patent ranking, it overtook Germany in 1980.4 Development and Patenting ActivityIn this section, we will analyse the relationship between selected macroeconomic variables and patents. The feature of the exer-cise that has been undertaken in this paper is that this is the first study, which tries to explain the patenting behaviour of a large cross section of countries through macroeconomic variables. One of the most contentious issues in the discussions on the international patent system that been addressed over several decades and has yet remained unresolved is whether countries at different stages of development can take advantage of the patent system by generating patentable inventions [Pretnar 1953]. The developing countries argued at various international fora that their stage of technological development, which is a manifesta-tion of their overall development, requires a “lesser” level of pat-ent protection than that provided for in the developed countries. These countries have also pointed to the fact that developed countries in their earlier stages of technological development adopted relatively “weak” patent regimes and it was only after they achieved a certain level of technological capability that they opted for a “strong” patent system. It is this argument that is the focus of our analysis here.The source of the data for the macroeconomic variables is the World Bank’s annual publication World Development Indicators. TheGDP per capita is in 1995 constant dollars. German unifica-tion in 1991 creates a break in its time series data on macro-economic variables. We have taken the patent figures only for West Germany. While the macroeconomic data pertains to West Germany from 1975 to 1991, the data pertains to unified Germany for the later years. We have 1292 observations. Our sample is unbalanced panel data. As the sample database is a panel, we use a fixed effects model. This will take into account the individual specific fixed effects.4.1 TheModelThe model, in two parts, explains the intensity of domestic and foreign patenting respectively.

SPECIAL ARTICLEEconomic & Political Weekly EPW june 7, 2008454.2 ResultsWe now interpret the results of the model.4.2.1 DomesticPatentsTable 1 gives the fixed effects estimate of the domestic patent equation. The results show that for the full sample 85 per cent of variation in domestic patent applications is explained by varia-tion inGDP per capita and foreign trade intensity variables. While both the explanatory variables are significant, GDP per capita is positive and foreign trade intensity is negative. Thus, GDP per capita taken as a measure of development has a significant and positive impact on domestic patent applications. This result is along expected lines – increase inGDP per capita raises the capacity of domestic entities to undertake R&D and hence, to seek patent protection. The negative sign of foreign trade intensity is interesting. The implication of this result is that the economies that are relatively more integrated with the global economy tend to produce a smaller number of domestic patent applications than closed economies. It can, however, be argued that this result is influ-enced by the following twin tendencies by countries appearing in our sample: (i) some of the industrialised countries, including theUS and Japan, which attract a large number of foreign patents display a lower trade to GDP ratio, and (ii) countries whose global integration have been at much higher levels, for instance those in south-east Asia or Latin America, have much lower levels of domestic patents.The equations estimated separately for developed and devel-oping countries, while displaying considerable similarity with that of the full sample, have some interesting differences as well. For developed countries, it was found that in the domestic patent applications equation, the variation in the dependent variable explained by the independent variables is 92 per cent but for the developing countries the corresponding figure is only 57 per cent. TheGDP per capita is positive and foreign trade intensity is nega-tive for both sets of countries. It is interesting to note that the coefficient of GDP per capita for developed countries is smaller than that for the developing countries. This result could be interpreted to mean that while an increase inGDP per capita would result in an increase in domestic patent applications, it does so less in developing countries as compared to developed countries. It may be pointed out that the above-mentioned conclusion is consistent with the reality. While countries with higherGDP per capita are in a position to allocate more resources in their R&D efforts, which results in more domestic patent applications, countries with low per capita income can be expected to have low output of domestic patent applications because of their inability to make significant investments inR&D activities.The negative relationship between domestic patent applica-tions and the openness index as measured by foreign trade inten-sity is a more interesting result. It implies that countries, which are more open tend to produce less domestic patent applications. One reason for this could be that countries, especially those in the developing world, which were found to be relatively more open, tend to depend more on foreign technology and hence, do not allocate sufficient resources for R&D effort. A comparison of coefficients of developed and developing country equations shows that a one per cent increase in foreign trade intensity will decrease domestic patent intensity more in developing countries than in developed countries.4.2.2 Foreign PatentsTable 2 gives the fixed effects estimate of the foreign patent equa-tion. The results show that for the full sample, 57 per cent of vari-ation in foreign patent applications is explained by variation in theGDP per capita, foreign trade intensity and FDI intensity vari-ables. All the three independent variables are positive and signifi-cant. Foreign patentees thus show the following behaviour while seeking patent protection: (i) they prefer countries with higher GDP per capita, which implies that they find countries that are relatively advanced countries more attractive; (ii) they prefer countries that are relatively more integrated with the global economy; and (iii) they seek patent protection in countries where the levels ofFDI are substantial. The last two tendencies are sig-nificant in that they corroborate the oft-repeated argument that inflow of foreign technologies would be more pronounced in countries that are relatively more open. An additional dimension that the results point to is that countries in which firms have already made investments are more likely to be the ones in which patents are taken. At the same time, however, the results indicate that countries at lower levels of development are not preferred by foreign patentees, a tendency that could have implications for these countries in terms of their access to technologies. This con-clusion is further strengthened by the estimates that we have obtained using separate equations for developing and developed countries, as indicated below.The equations estimating foreign patent applications in devel-oped and developing countries have interesting differences. While for developed countries the three independent variables explain 57 per cent of variation in foreign patent applications, for developing countries they explain only 42 per cent of the variation in foreign patent applications. For developed countries, all the three independent variables are positive and significant, while for developing countries only the GDP Table 1: Fixed Effects Estimate of Domestic Patent EquationVariable Full Sample Developed Countries Developing CountriesGDPpc 0.00001486 0.00001011 0.00007559 (14.692) (14.231) (12.871)EXIM/GDP -0.00043123 -0.00076608 -0.00109741 (-2.145) (-2.563) (-4.921)Adjusted R2 0.85 0.92 0.57Number of observations 1292 480 812Figures in the parenthesis are t-values. The dependent variable is DPit/Populationit.Table 2: Fixed Effects Estimate of Foreign Patent EquationVariable Full Sample Developed Countries Developing CountriesGDPpc 0.000120540.000055460.00015177 (13.914)(4.199)(12.552)EXIM/GDP 0.092941020.41423747-0.00429536 (7.410) (11.666) (-0.765)Net FDI/GDP 0.00352963 0.02033801 -0.00059613 (2.020)(3.840) (-0.765)Adjusted R2 0.57 0.570.42Number of observations 1292 480 812Figures in the parenthesis are t-values. The dependent variable is FPit/Populationit.

SPECIAL ARTICLEjune 7, 2008 EPW Economic & Political Weekly46per capita variable is positive and significant. For developing countries, the foreign trade intensity and the FDI intensity variables are not significant. This is an interesting result. It means that openness of the economy as measured by their foreign trade intensity is not related to foreign patenting. Foreign patentees do not show a preference for those developing countries, which are more open. In the same manner, FDI and foreign patenting are not related as far as developing countries are concerned. This means, in other words, that developing countries, which are the destination of foreign investment, do not seem to be destinations for foreign patenting. One reason for this could be that the FDI is not patent related but the fact that these three independent vari-ables are positive and significant for developed countries means that foreign patentees react differently to developed and devel-oping countries while making their patenting decisions.5 ConcludingRemarksThis paper makes an attempt to explain the observed patterns in patenting during the period 1975-98, using a set of development indicators in respect of a selected set of countries. We try to find the relationship between domestic and foreign patenting activity and the selected variables of the sample countries. For explaining domestic patenting activities we haveincluded two variables viz,GDP per capita and exports plus imports as per-centage ofGDP (as a measure of openness) were used for this exercise. Foreign patenting activity was sought to be explained using three variables, viz,GDP per capita, exports plus imports as percentage ofGDP (as a measure of openness) and foreign direct investment as a percentage ofGDP as explanatory variables. We found that while domestic patenting and GDP per capita are positively related, openness and domestic patenting are negatively related. On the other hand, while GDP per capita is positively related to patenting, both openness and FDI while positive and significant for developed countries is not significant in the case of developing countries. The present study represents the first step towards a better understanding of patenting behaviour of countries spread across the develop-ment spectrum.Notes1 This view was cogently reflected in the following statement made by the director general of the WIPO: “The international patent system must operate to the maximum benefit of the countries that participate in it, taking account of their widely varying stages of technological and economic development” [WTO 2001].2 The Federal Trade Commission, in its recent report, has highlighted the range of problems that United States Patent and Trademark Office (USPTO) faces in carrying out its activities, the consequences of which have been the grant of a large number of questionable patents. See Federal Trade Commission (2003).3 The United States Patent and Trademark Office receives 300,000 patent applications annually, which are examined by only 3,000 examiners.4 Watanabe (1985) says “Inter-temporal and inter-national variations, for example, in the product groups covered by the patent system, forbid use of readily available patent statistics for a very pre-cise comparative analysis. Still, one may be able to discuss the relative level of inventive activity among different nations, where observed gaps are large enough.”5 The rest of this sub-section is based on WIPO (1988). We refer to the pre-Trade Related Aspects of Intellectual Property Rights (TRIPs) situation (pre-1995) when such differences were possible.6 Throughout this paper we consider high-income economies identified as such by World Bank (1993) to be developed countries. High-income economies have GNP per capita of $ 7,620 or more in 1990 [World Bank 1993].7 See Griliches (1989) for a discussion on how the resources available to the patent office determine the grants. Some time these grants do not follow the trends in applications.8 We use resident and domestic; and non-resident and foreign as synonyms.9 The change in coverage of the Industrial Property Statistics in 1985 creates a discontinuity in the series. Since 1985 patent applications under Patent Co-operation Treaty (PCT) and European Patent Convention (EPC) were being reported separately.10We have excluded small countries (21), least developed countries (20), Organisation Africaine de la Propriete Intellectualle (15 member countries of which two have reported separately for some years), re-registration countries (2); former socialist countries and those that have emerged after their break-up (34), countries which have not reported for more than 11 years (11), countries with zero resident patent applications for more than 11 years (3) and countries with zero patent grants for more than 11 years (2). Iraq was excluded becauseWorld Development Indicators did not report data for most of the years coveredby our study, making for a total of 96 countries.11 For the period 1975-91 Germany refers to West Germany and for the period 1992-98 it refers to unified Germany.12 We substitute the estimated data for missing observations. Hence. we mix original data with estimated data.ReferencesBosworth, D L (1980): ‘The Transfer of US Technology Abroad’,Research Policy, Vol 9, pp 378-88. – (1984): ‘Foreign Patenting Flows to and from the United Kingdom’, Research Policy, Vol 13, pp 115-24.Eaton, J, and S Kortum (1996): ‘Trade in Ideas: Patent-ing and Productivity in the OECD’,Journal of International Economics, Vol 40, pp 251-78.–(1999): ‘International Technology Diffusion: Theory and Measurement’,International Economic Review, Vol 40, pp 537-70.Engel, R L (1985): ‘Patent Enforcement: The Uncer-tainties of Patent Litigation’ in G S Tolley, J H Hodge and J F Oehmke (eds),The Economics of R&D Policy, New York, Praeger.Federal Trade Commission (2003): To Promote Inno-vation: The Proper Balance of Competition and Patent Law and Policy, Federal Trade Commission, Washington DC.Griliches, Z (1989): ‘Patents: Recent Trends and Puzzles’, Brookings Papers on Economic Activity: Micro Economics, pp 291-319. – (1990): ‘Patent Statistics as Economic Indicators: A Survey’,Journal of Economic Literature, Vol 28, pp 1661-1707.Mansfield, E (1986): ‘Patents and Innovation: An Empirical Study’, Management Science, Vol 32, pp 173-81.Pakes, A (1986): ‘Patents as Options: Some Estimates of the Value of Holding European Patent Stocks’, Econometrica, Vol 54, pp 755-784.Pretnar, S (1953): ‘The International Protection of Industrial Property and the Different Stages of Economic Development of the States’,Industrial Property, Vol 69, pp 213-23 (translation supplied by WIPO of French original).Schankerman, M, and A Pakes (1986): ‘Estimates of the Value of Patent Rights in European Countries During the Post-1950 Period’,Economic Journal, Vol 96, pp 1052-076.Scherer, F M (1977): ‘The Economic Effects of Compul-sory Patent Licensing’,Monograph Series in Finance and Economics, No 1977-2, New York University, New York.Schiffel, D, and C Kitti (1978): ‘Rates of Invention: International Patent Comparisons’, Research Policy, Vol 7, pp 324-40.Soete, L L G (1981): ‘A General Test of Technological Gap Trade Theory’,Weltwirtschaftliches Archiv, Vol 117, pp 639-60.Taylor, C T, and Z A Silberston (1973): The Economic Impact of the Patent System: A Study of the British Experience, Cambridge University Press, Cambridge.United Nations Conference on Trade and Develop-ment (UNCTAD) (1975):The Role of the Patent System in the Transfer of Technology to Developing Countries, New York, UN (TD/B/AC.11/19).Vayrynen, R (1977): ‘The International Patent System and the Transfer of Technology to Africa’, Occa-sional Paper No 9, Tampere Peace Research Insti-tute, Tampere.Watanabe S (1985): ‘The Patent System and Indige-nous Technology Development in the Third World’ in J James and S Watanabe (eds),Technology, Institutions and Government Policies, Macmillan, London.World Bank (1993): World Development Report, Oxford University Press, Oxford.World Intellectual Property Organisation (WIPO) (1988): Existence, Scope and Form of Generally Internationally Accepted and Applied Standards/Norms for the Protection of Intellectual Property, Multilateral Trade Negotiations: The Uruguay Round, GATT, Geneva, (MTN/GNG/NG11/W/24).WTO (2001): WIPO and WTO Launch New Initiative to Help World’s Poorest Countries, World Trade Organisation, Geneva.