MASTER THESIS

MASTER THESIS

The Impacts of the Trade-Related Intellectual Property Rights

Agreement on Innovative Activities in Developing Countries

Phillip Aulmann

Matriculation Number: s3207196

p.aulmann@student.rug.nl

University of Groningen

Faculty of Economics and Business

Supervised by Dr. A.A. Erumban

Abstract

This master thesis examines the impacts of the Trade-Related Intellectual Property Rights Agreement (TRIPS) on innovative activities in developing countries. The TRIPS sets minimum standards of intellectual property rights to ensure an innovator the protection from imitation to encourage innovative activities. A panel analysis with two different time periods, before and after the implementation of the TRIPS, is developed to observe the impacts of higher protection in developing countries. Our results suggest, that protection itself does not explain domestic innovative activities in developing countries. However, it explains innovative activities by non-residents in these countries.

Table of Content

1 INTRODUCTION

... 1

2 HISTORICAL PERSPECTIVE

... 3

3 THEORETICAL BACKGROUND OF PATENT PROTECTION AND

THE CREATION OF KNOWLEDGE

... 6

4 LITERATURE REVIEW

... 8

4.1.

I

MPACTS OF

F

REE

T

RADE ON

D

EVELOPING

C

OUNTRIES

8

4.2

I

MPACT OF

P

ATENT

P

ROTECTION ON

I

NNOVATIONS AND

E

CONOMIC

D

EVELOPMENT

10

4.3

P

ROTECTION AND

D

IFFUSION OF

K

NOWLEDGE

12

4.3.1

T

RADE

12

4.3.2

F

OREIGN

D

IRECT

I

NVESTMENT AND

L

ICENSING

13

4.4

G

AP IN THE

L

ITERATURE

14

5 METHODOLOGY AND DATA DESCRIPTION

... 16

Table of Figures

T

ABLE

1

-

S

UMMARY

S

TATISTICS

... 22

T

ABLE

2

(

A

)

–

L

N

P

ATENT

A

PPLICATIONS BY

R

ESIDENTS

... 28

T

ABLE

2

(

B

)

-

L

N

P

ATENT

A

PPLICATIONS BY

N

ON

-R

ESIDENTS

... 29

T

ABLE

3

(

A

)

–

L

N

P

ATENTS

G

RANTED TO

R

ESIDENTS

... 30

T

ABLE

3

(

B

)

–

L

N

P

ATENTS

G

RANTED TO

N

ON

-R

ESIDENTS

... 31

APPENDIX ... 36

T

ABLE

A1

(

A

)

–

L

N

P

ATENT

A

PPLICATIONS WITH

WTO

D

UMMY

... 36

T

ABLE

A1

(

B

)

–

L

N

P

ATENTS

G

RANTED WITH

WTO

D

UMMY

... 37

T

ABLE

A2

(

A

)

–

L

N

P

ATENT

A

PPLICATION BY

R

ESIDENTS

(R

ANDOM

E

FFECTS

) ... 38

T

ABLE

A2

(

B

)

–

L

N

P

ATENT

A

PPLICATION BY

N

ON

-R

ESIDENTS

(R

ANDOM

E

FFECTS

) ... 39

T

ABLE

A3

(

A

)

–

L

N

P

ATENTS

G

RANTED TO

R

ESIDENTS

(R

ANDOM

E

FFECTS

) ... 40

T

ABLE

A3

(

B

)

–

L

N

P

ATENTS

G

RANTED TO

N

ON

-R

ESIDENTS

(R

ANDOM

E

FFECTS

)

... 41

T

ABLE

A4

(

A

)

–

L

N

P

ATENT

A

PPLICATION BY

R

ESIDENTS

(R

OBUST

) ... 42

T

ABLE

A4

(

B

)

–

L

N

P

ATENT

A

PPLICATION BY

N

ON

-R

ESIDENTS

(R

OBUST

) ... 43

T

ABLE

A5

(

A

)

–

L

N

P

ATENTS

G

RANTED TO

R

ESIDENTS

(R

OBUST

) ... 44

T

ABLE

A5

(

B

)

–

L

N

P

ATENTS

G

RANTED TO

N

ON

-R

ESIDENTS

(R

OBUST

) ... 45

C

OUNTRY

L

IST

... 46

1 Introduction

Since the last decades, there exist a discussion about the impacts of patent protection on developing countries. This topic got more debated since the commence of the Agreement of Trade-Related Intellectual Property Rights (TRIPS) which is part of the World Trade Organization (WTO) since 1995. The TRIPS agreement provides a minimum level of patent protection and consequently raises the overall level of protection. The debate about harmonizing patent protection among countries emerged since developing countries started increasing their importance in world trade by getting more integrated into the world economy (International Monetary Fund, 2001). Since the 1980’s, the global production structure has become more fragmented and thus more globally dispersed (Los et al., 2015). Developing countries primarily gained from technology diffusion through imitation which was affected by trade with developed economies (Connolly, 2002). Thus, they got the chance to catch up with industrialized countries by having access to foreign (high) technologies. At this time, imitation was easily feasible since developing countries did not provide protection for intellectual property, which presented an issue for developed nations. The TRIPS agreement is ought to remedy this issue because it is a commitment for all member countries to provide protection after joining the WTO. The agreement envisages, by granting an innovator the right to exclude others from the commercial use and imitation of his invention, to encourage economic agents to invest in research and development. This is assumed to increase the production of new technologies and hence stimulate economic development.

The critique about protection in developing countries has already been in place since Chin and Grossman (1988) developed the ‘North-South’ trade model concerning intellectual property. That model defines the ‘South’ (developing countries) as the global imitator that is not in the position of producing innovations but has the ability to copy innovations produced by the ‘North’ (industrialized countries). Since protection hampers developing countries from imitating it is doubtful how they can benefit from it. Therefore, the agreement comprises the obligation by developed nations to engage in technology transfer towards developing countries to foster whose economic development.

The purpose of this thesis is to find out how aligning the national laws to the regulation of the TRIPS has impacted developing countries on innovative activities. Therefore, it will only be tested for countries which fully implemented the TRIPS requirements at the same time.

2 Historical Perspective

Providing exclusive rights for the owner of intellectual property (IP) is an idea that has been in existence for a long time. For instance, agreements like the Paris Convention1 _{(1883) and} the Berne Convention (1886) were aimed to regulate the protection of intellectual property in the past. However, these agreements were not compulsory and did not require minimum standards globally. The World Intellectual Property Organization (WIPO) was founded in 1967 and became an integrated component of the United Nation in 1974. Today the WIPO oversees all existing agreements of intellectual property protection. Moreover, the WIPO was ought to support international cooperation with respect to intellectual property rights in addition to the constitution of IP protection systems in developing countries. According to Liebig (2001) since developing countries have a majority in the WIPO several attempts to strengthen the rules concerning intellectual property protection were rejected2. The urgency for industrialized countries to increase the protection of IP arose mostly due to the more integrated world economy as well as the rapid advances in technology. Over the last decades, developing countries gained more economic power and their ability to imitate high technology products poses a threat to the competitive positions of the industries of developed countries. For instance, since China introduced its economic reforms in 1978 the country had on average real annual GDP growth rates of approximately 10 percent until its slowdown in 2012 (Morrison, 2013). According to the World Bank (2017) China could raise its GDP from 191.15 Billion in 1980 to 360.86 Billion in 1990, which already represents a doubling within ten years. In 2000 Chinas GDP reached levels of 1.2 Trillion and in 2015 they further raised it until 11.07 Trillion3. Referring to Nelson and Pack (1999) a similar development was experienced by the so-called ‘Four Asian Tigers’4. These countries were able to increase their GDP per capita fourfold since the 1960’s (1965-1999), by devoting large investments to structural change and industrial development. Developing high technology goods typically demand high investments in R&D, but once these technologies are developed, they are often imitated easily by others. Consequently, infringements of IP rights or no grants of IP protection used to be a conflict pool in international trade policy.

1 _{One of the first official agreements of the protection of intellectual property rights like} patents, utility models, trade marks etc. The Berne Convention is an agreement of the protection of artistic works like copy rights nowadays.

2 _{Especially in industries in which industrialized countries have a comparative advantage} (high technology industries like the pharmaceutical industry) called for a stronger intellectual property protection.

3 _{Numbers are in current US Dollars.}

A well-known case is the implementation of the so-called ‘Section 3015_{’ which is part of US} trade laws. Countries that particularly threatened the US trade policy were Brazil, India, China, South Korea and Thailand, which did not provide sufficient protection for goods that were imported from the US and thus jeopardized the competitive position of US firms (Müller-Langer, 2009). These accrued conflicts finally led to the negotiations of Uruguay and resulted in the foundation of the WTO in 1995.

The so-called Uruguay-Round (1986-1994) included the General Agreement on Tariffs and Trade (GATT), the General Agreement on Trade in Services (GATS) 6, as well as the Trade- Related Aspects of Intellectual Property Rights (TRIPS) into the newly founded World Trade Organization (WTO). The purpose of the WTO is to monitor trade rules to encourage the global trade. The institutional obligations of the WTO aim to increase the living standards of its members as well as their economic development (Kommerskollegium, 2004). Whereas the GATT governs all responsibilities with respect to spur trade and to reduce tariffs, TRIPS sets the minimum rules of intellectual property rights.

After the final negotiation of Uruguay, every member of the GATT (after 1995: WTO) and following members had to provide minimum standards of intellectual property protection. The most important intellectual property rights (IPR), which are covered by the agreement, are patents, industrial designs and copyrights. We will mainly focus on patent protection. Before TRIPS commenced, all GATT members were free to choose their particular level of intellectual property protection. Hence, there was no minimum standard and no harmonization between countries on the level with respect to IP protection. The exact level of protection gets determined by each country but they must be within the scope of the requirements of the WTO. Therefore, there still exist differences across countries. Since most developing countries did not apply IP protection7 until signing the obligations of Uruguay, the contract incorporates transition periods for developing and least developed countries. Referring to TRIPS Article 65, each country was not required to apply the TRIPS regulations during these transition periods but to transform the national law according to the TRIPS requirements. The transition period for developing countries was from 1st January 1995 until 1st January 2000. Least developed countries’ deadline was first scheduled on 1st January 2005 but extended to the 1st July 20138. Referring to the Declaration of Doha (2001), patents for pharmaceutical products were intermitted until 2016 for least developed countries.

The TRIPS agreement can be viewed as a milestone in the history of trade policies since it is the first time that an agreement settles minimum standards of intellectual property protection compulsory for every member country. It is important to mention that the WIPO as the

5 _{Part of the ‘Omnibus Foreign Trade and Competitiveness Act’ of the US Trade Law (1988):} It allows the US to impose import restrictions on countries which don´t provide sufficient protection for US products.

6 _{Since the topic of this thesis is IPR we are not going to examine the effects of GATT too} deeply.

7 _{At least not that comprehensive and strict as under the TRIPS obligations.}

managing organization was replaced by the WTO after 1995. Consequently, a shift of power towards the industrialized countries compared to the former situation occurred. Referring to the WTO (2016) the organization was founded by 128 countries and grew over time to 164 members with representatives of the developed- developing and least developed world. As previously presented, developed countries demanded higher international protection due to rising competitive pressure whereas developing countries primarily aimed differing goals. According to World Bank (2001), the adherence of developing countries to the international harmonization of intellectual property protection was considered as a chance to gain access to markets in developed economies9_{. Moreover, developing countries recognized the opportunity} of an access to a dispute settlement system instead of bilateral conflict regulation (Kommerskollegium, 2004).

3 Theoretical Background of Patent Protection and the Creation of Knowledge

Schumpeter (1912) already discussed the importance of innovations as a (main) component of economic development. He describes the production of innovations as results of the entrepreneurial activities (activities of the so-called ‘dynamic entrepreneur’). A feature of his model, which got picked up later by Kline (1985), is the idea of imitation and diffusion of innovations. Kline further evolves Schumpeter´s insights in his chain-linked model which takes feedback processes into account. Though, Kline´s focus is the production of knowledge in general and not necessarily an innovative outcome. These models provide a more comprehensive explanation of economic growth as the predominant models of the neo-classic growth theories as for instance the Solow-Swan model (1956). The neo-classic approach explains economic growth in the long-run by capital accumulation, population growth, and an increase in productivity. In the short run, an economy records investments in capital until a steady state is reached. New investments only make sense after a technological change occurs. These models are also called ‘exogenous growth models’ since technological change is not endogenous to the system. Representatives of the neo-classic growth theory do not provide a proper explanation how this technological change comes off.

represent a market failure since an underinvestment in R&D, and consequently, a lack of creation of new knowledge occurs.

The dominant theoretical view regarding patents is that, without protection, no member of the society has an incentive to create new knowledge. This causality is problematic because, according to Arrow (1962), the production of information (R&D) is a risky process in which the producer has no certainty of his success or applicability of the invention. Hence, various obstacles to produce new knowledge without having an incentive exist, despite demand for innovation in society.

In line with the findings of Arrow is the statement by the World Bank (2001) which declares IP protection as being necessary to find an appropriate solution between the needs of the society, creation of new knowledge, and the encouragement of economic agents to innovate. Patents give innovators an incentive to invest in R&D by granting them the rights to exclude others from the commercial use of the innovation and thereby reduce the risk of investing in R&D. In exchange, the innovator must disclose the newly generated knowledge that the society can use, i.e. for further research and inventions. This trade-off is regarded as necessary since both possible solutions, that is under perfect competition as well as under monopolistic competition, entail market failures. By granting an innovator a temporary monopoly, the government provides a so-called second best solution. The temporary and static inefficiencies10 (market failure) that occur in a monopoly are accepted for the dynamic benefits11. The innovator, now the monopolist, can charge the monopoly price, which is substantially higher than under perfect competition, providing a smaller amount of the good to the market. The interdependency between smaller amount of provided goods and higher demanded price create the loss of welfare. In exchange, he ensures the dynamic benefits which arise by the provision of new knowledge to society over time. Referring to the TRIPS agreement, innovations need to be new, non-obvious, and commercially applicable to be accepted as a protectable good. The objective of the TRIPS agreement is to increase both, the number of domestic innovations as well as the transfer of knowledge from developed to developing countries. In other words, enhancing innovative activities and the diffusion of newly generated knowledge. The reason why this topic is highly debated in the economic literature is that countries are different, especially with respect to their technologic capabilities12. Hence, it is doubtful whether all countries benefit from a harmonized patent system in similar ways.

10 _{Occur by the misallocation of resources at a certain time. In a monopoly, a higher price is} charged for a smaller amount of the product. Thus, the monopolist obtains a higher rent on the expense of the consumer surplus. Moreover, by charging the monopoly price and the loss of consumer surplus, a deadweight loss arises.

11 _{These accrue over time by the provision of new products, knowledge, processes etc.} Contrarily to static inefficiencies, they cannot be observed to a certain time. Hence, they are difficult to measure.

4 Literature Review

After the theoretical considerations about protection of intellectual property were shown, the next section examines recent findings from the relevant literature. These findings comprise several aspects which belong to the topic of intellectual property rights with respect to the TRIPS agreement. Firstly, the impacts of free trade on developing countries, in general, will be discussed. Secondly, the topic of free trade will be linked to technological change and economic development, which can be affected by both, domestic innovative activities and the adoption of foreign knowledge. In the end, recent relevant findings from the literature about impacts of intellectual property protection on domestic innovation, as well as diffusion of knowledge, will be presented. We want the reader to understand the complex matter of intellectual property protection and economic development after the review of literature. Based on findings from the review, we will conduct an analysis of the effects of intellectual property protection, according to the TRIPS, on innovative activities in developing countries.

4.1. Impacts of Free Trade on Developing Countries

Before the topic of intellectual property rights gets examined, one should recall that becoming a member of the WTO is more than just implementing the rules of the TRIPS. Most of all, the WTO is an organization that monitors the regulation of free trade, which represents substantial changes for most developing countries. As previously described in the historical consideration of TRIPS, the agreement was a necessary undertaking for developed countries due to the rising economic interrelations with emergent economies. Furthermore, technological change and the shift to more knowledge-dependent industries increase these countries’ demand for protection.

Before the WTO, several developing countries followed strategies to protect their domestic markets from foreign firms (Dornbusch, 1992). A common way to follow this strategy was to implement import substitution to develop the domestic industry behind protective barriers such as tariffs and import quotas. Adhering to these strategies was meant to avoid problems of impacting the terms of trades in a negative way. Nevertheless, protecting the domestic market from imports affects a country’s terms of trade. These impacts occur by overvaluing the domestic currency when a country does not demand foreign exchanges. In turn, protecting a country from imports results consequently in a ‘tax’ on exports. Besides many different effects, which depend on the particular circumstances13_{, the most crucial effects, at least with} respect to our research topic, are those causing the deterioration of competition. This distortion of competition leads to less variety as well as less quality of products which consequently leads to a loss of welfare. Furthermore, countries´ access to new technologies gets restricted. Hence, opening an economy to free trade avoids these distortions. However,

more free trade is not a panacea for economies in general. Essentially, the resource allocation within a country is affected, it creates a more competitive environment, diffusion of new technologies and access to intermediate products.

Particularly, free trade partly led to a higher fragmentation (unbundling of tasks) of the Global-Value-Chain (GVC)14. This is and was a big opportunity for developing countries to engage in the global trade system by carrying out production steps of MNEs, thus leading to benefits from the inclusion in the production process (Kowalski et al., 2015). These benefits arise from specializing in specific tasks to exploit comparative advantages along the GVC which can increase a country’s level of productivity. This can lead, in turn, to upgrades in terms of economic activities15_{. Hence, raising the prosperity of a country can happen faster} and easier than in previous decades and consequently encounter poverty.

Recent findings in the literature indicate a decisive role of technological change on economic development. As previously discussed, technological change can occur through domestic innovations and diffusion of foreign knowledge.

According to a study of Bassanini and Scarpetta (2002), domestic R&D activity is a significant driver of real GDP growth in OECD countries. These findings are in line with the empirical results by Park (1995) who identifies positive impacts of R&D on productivity improvements, which in turn stimulate economic growth. Moreover, the author emphasizes the important role of R&D activities, carried out by the private sector since they impact productivity more than public R&D activities. Jones (1995), as well as Ulku (2004), also identify stimulating impacts of R&D activities on economic growth. Moreover, both authors suggest finite effects of R&D on growth. However, since Ulku (2004) use patent data as well as R&D expenditures to capture innovative outcomes, the author does not regard these results as a rejection of the endogenous growth theory16_{. These results rather suggest that patents do} not capture every innovative outcome of an economy.

Especially in the case of developing countries, diffusion of knowledge is a crucial driver of economic development. Borensztein et al. (1995) investigate the effect of technology diffusion in terms of FDI on economic growth in developing countries. Their results suggest that, in these countries, technology diffusion contributes more to economic growth than domestic investments do. Though, they condition these results on the level of human capital which reflects absorptive capacities within a country. Peña-Lopez (2008) adds to these findings that in developing countries, technological development basically depends on prior existing technologies which get adopted. Xu and Wang (2000) examine the role of FDI and trade with respect to technology diffusion. Their results are consistent with previous findings

14 _{Reducing tariffs is only one part of the story; The ICT revolution evolved the fragmentation} of tasks even further (Baldwin and Evenett, 2012).

15 _{Referring to Humphrey and Schmitz (2002) there exist four different kinds of upgrading:} Process upgrading, product upgrading, functional upgrading and chain upgrading.

that foreign R&D spillovers contribute to domestic growth, however more in terms of capital goods trade17_.

The empirical investigation of Ciruelos and Wang (2005) finds results, in line with those of Borensztein et al. (1995). Technology diffusion which arises by FDI from developed to developing countries only affects technology improvements in developing countries when they have a certain amount of human capital. Moreover, their results contribute to the findings of Xu and Wang (2000) as well which suggest exports from developed to developing countries to affect productivity improvements more than FDI.

4.2 Impact of Patent Protection on Innovations and Economic Development

With respect to the in section 3 explained theory of patent protection, the TRIPS agreement ought to encourage innovations in developing countries through the inherent incentive for economic agents to innovate. In the following part, the results from the currently relevant literature are presented to prove whether and under which conditions this causality holds true.

Chin and Grossman (1988) present a global welfare analysis regarding the effects of a global patent system. In their model, they consider the ‘North’ as the global innovator and attribute the ability to invest in R&D only to northern firms. The ‘South’ does not invest in R&D and is consequently the global imitator in the absence of patent protection. They conclude that the South does not benefit from enforcing patent protection for foreign firms contrarily to the North, who benefits from it. They suggest that intellectual property protection starts to be effective when a country passes the middle-income threshold. Thus, to be efficient from the welfare point of view, the South should not implement patent protection until they gained sufficient knowledge from imitating northern technologies.

According to Braga et al. (2000), before the foundation of the WTO, the fewest developing and least developed countries had an intact system of intellectual property rights or just weak protection which was accepted by policy makers of the industrialized countries. The predominant opinion is that these country groups have a lack of capabilities of creating new knowledge and thus would not benefit from IP protection, which is in line with the analysis by Chin and Grossman (1988). Moreover, the author emphasizes the connection between the effectiveness of a patent system and investments in R&D. Since developing countries spent much less on R&D than developed countries, the efficacy of the spurring effects of TRIPS can be questioned. Within their report, the World Bank (2001) reveals that especially the least developed economies devote almost no resources to the generation of new knowledge. In addition, they do not have a stock of intellectual property which they could protect.

Others, such as Chen and Puttipanun (2004) assert a positive effect of stronger IPR´s in developing countries on innovativeness. However, these effects depend on the economic development of a country and its coherent innovative abilities. Ginarte and Park (1997)

suggest that the strength of a patent system does not only depend on economic development but also on the level of R&D expenditures, market freedom and how open the economy is. Additionally, Chen and Puttipanun (2004) find positive impacts of patents on innovations but recommend to industrialized economies to assist developing countries with improving their innovative abilities. However, others like Sharma and Saxena (2012) present overall negative results of strong patent protection in developing countries. Furthermore, they condition the economic growth of developing countries on the possibility to generate knowledge via imitating foreign technologies which get hampered by strong patent protection. In line with these findings are the empirical results by Thompson and Rushing (1996), which identify the relation between the stage of development of a country and the possibilities to benefit from stronger patent protection. Referring to their results, countries first need to pass a certain threshold of economic development before they can benefit from stronger protection. This finding is also emphasized by Chin and Grossman (1998), The World Bank (2001), Falvey Foster and Greenway (2006), and Allred and Park (2007). Their common argument is, middle-income economies are still imitating countries and not likely to be innovators. Patent protection merely impedes them to imitate, and thus hinders further benefit from knowledge assimilation as well as increase the prices for protected goods.

Moreover, Sharma and Saxena (2012) show that universally imposed patent protection, as it is the case under TRIPS, is not likely to enhance growth in developing countries, which did not apply patent protection previously. Hassan, Yaqup and Diepeveen (2010) present similar results and stress the dependency of economic development on the effectiveness of patent protection, too. Yet, they emphasize the necessity of adopting intellectual property protection at a certain level of economic development since the potential of catching up decreases for countries when they come closer to the frontier of existing knowledge. To further stimulate economic growth, it is necessary to engage in innovating instead of imitative activities.

The terminal finding interesting for this field of research, revealed by Forero-Pineda (2006), is that R&D investments in a certain area are only spurred when there was no interest of foreign investments in this specific area. In other words, when foreign firms were not engaged in a certain industry.

4.3 Protection and Diffusion of Knowledge

Higher patent protection restricts diffusion in terms of imitation at least through non-market channels such as reverse engineering, which has been an economic driver for the now developed countries in the past (Forray, 2009). This causes the dependency on market-channels solutions for developing countries nowadays like FDI, trade, and licensing. This dependency occurs from the previously mentioned results of the North/South trade model by Chin and Grossman that identified the South as the global importer of technologies. Firms from industrialized countries have an allegedly higher incentive to engage in these types of market transactions when patent protection in developing countries is higher since they face a smaller threat of imitation. Yet, differences in strength of patent protection make firms switch between the mentioned market-channels. For instance, exporting can entail a threat of imitation by reverse engineering, FDI by employee movements, and licensing when the contract partner does not comply with the contract conditions. In the following part, the impact of patent protection on the diffusion of innovations, in terms of technology transfer, is revealed. This is important since it is a fixed obligation in the TRIPS agreement.Referring to article 66 in the TRIPS, the industrialized members of the WTO are required to assist developing countries and to promote their technological progress by transferring technologies. The conditions and effects from transferring technologies are controversially debated in the current and past literature, but results remain ambiguous. Braga et al. (2000) state that most technologies are introduced in developing countries through diffusion instead of domestic innovations. Thus, if stronger patent rights lead to a higher transfer of technologies, this could be beneficial for developing countries. Park and Lippoldt (2008) oppose that developing countries can only gain from imports as a source of technological spillovers when they have sufficient absorptive capabilities. Falvey et al. (2006) emphasize ambiguous effects of a stronger patent protection on the different channels of technology transfer. The selection of how to engage in foreign markets is mostly affected by the level of economic development of the receiving country, which is reflected by imitative or innovative capabilities, respectively.

4.3.1 Trade

former arises due to the possibility to enlarge the sales volume in a foreign market whereas the latter occurs through the monopoly power of the patent owner. Hence, less of a product is sold at a higher price and others are excluded from selling substitutes18. These effects vary for large developing countries which benefit more in terms of import attraction than small developing countries. Moreover, Maskus und Penubarti (1995) show that these effects differ among different industries depending on their particular patent sensitivity. Smith (1999) adds the impact of the threat of imitation to these findings. Thus, the market power effect arises mainly in small developing countries that represent a small threat of imitation whereas the market expansion effect mostly occurs in large developing countries with more competitive domestic firms with higher imitative capabilities. This is not valid for countries with low imitative capacities because their lack of abilities already acts as a protection against imitation. Therefore, strong patent protection can have negative net impacts on imports in countries with weak imitative capabilities (Smith, 2001). This is caused by the higher market power effect that outweighs the market expansion effect. Park and Lippoldt (2008) empirically identify positive impacts of patent protection on imports in developing as well as in developed countries. However, the quantitative outcome in developed countries is much higher than in the other country groups. Smith´s (2001) empirical results show that strong patent protection does not affect exports at all. Neither in countries with high imitative capabilities nor low imitative capabilities. Hassan et al. (2010) show that, especially for high-technology products, increasing patent protection has virtually no impact on trade flows. This is explained by either the weak imitative capacities of developing countries in this sector or, more likely, that firms prefer to transfer this kind of knowledge via licensing or FDI.

4.3.2 Foreign Direct Investment and Licensing

Important issues when debating about firms’ engagement in FDI with respect to patent protection primarily are, which circumstances affect the decision for firms to engage in FDI in general. Secondly, it is important to consider how different levels of protection and economic development affect the composition of FDI. Park and Lippoldt (2008) show positive empirical effects of strong patent protection on inward FDI flows for developed, developing, and least developed countries. Yet, their identified effects become stronger the more developed a country is. Firms have different reasons to engage in FDI. For instance, in some industries, the level of protection is not crucial for the firms’ decision, when a firm only uses the affiliate for distributive or administrative tasks instead of R&D or manufacturing related activities. This suggestion is in line with Lee and Mansfield (1996) who argue that weak protection could make firms shifting their FDI´s from R&D related activities towards i.e. sales and distribution. Braga and Fink (1998) oppose that stronger patent protection can make firms to switch from FDI to licensing. Park and Lippoldt (2003) however, state that stronger patent

protection increases inward as well as outward FDI in developing and least developed countries, nevertheless. Moreover, they find in their empirical analysis that developing countries which are a member of the WTO face higher inflows of FDI than non-members. However, for least developed countries being a member of the WTO does not reveal any effects. Nicholson (2007) finds evidence for both; higher patent protection leads to more FDI and more licensing. Firms favor FDI if costs of capital are high paired with low patent protection, which consequently leads to firms maintaining control over their assets. If the receiving country has strong patent protection and the firm faces high capital costs they are more likely to license to avoid additional costs for an affiliate. Furthermore, firms with high R&D expenditures prefer licensing when the technology receiving country has a highly protective system. Strong patent protection is associated with higher affiliate sales plus an even higher effect on licensing in countries with high imitative capabilities (Smith, 2001). Furthermore, if countries with high imitative abilities don’t provide strong protection the test shows negative effects on affiliate sales and even stronger effects on licensing.

The overall positive effects on exports which were found in Smith (1999) are now insignificant due to the simultaneous consideration of all three market-channel solutions. Furthermore, the author found that knowledge flows towards the affiliate in technology receiving country are positive and significant with countries that have a highly protective system and high imitative capacity (Smith, 2001). Countries which provide patent protection and have weak imitative capacities face negative effects regarding knowledge flows. Therefore, it is doubtful whether especially the least developed countries should gain from the TRIPS agreement when they are neither allowed to imitate nor able to absorb the transferred knowledge by firms.

4.4 Gap in the Literature

whether the claimed effects may arise with a time lag. We expect if the effect is positive, to see, higher domestic patent application with respect to higher protection. A supporting argument for the time lag theory is that patents deal with the static losses and dynamic benefits. The static effects arise immediately whereas the dynamic benefits which are supposed to occur over a non-defined time span. Especially the papers by Allred and Park (2007) and Chen and Puttipanun (2004) encourage our motivation to examine that topic more detailed. Both papers, although aiming just slightly different research questions, come to different conclusions referring to a patent system. Allred and Park (2007) reveal all in all no positive impacts from patent protection in developing countries. On the contrary, Chen and Puttipanun (2004) present positive impacts of patent protection on innovative activities in developing countries. However, these positive impacts depend on a ‘perfect’ patent system which respects the countries level of economic development. Thus, their results imply that developing countries should choose an individual level of protection, too. That means in turn, not to adopt a universal patent system.

Recent empirical analysis conduct analyses about the impact of patent protection on innovative activities with mixed country groups or divided by their country status19.

What distinguishes our study from previous ones is that we only include countries that implemented the TRIPS at the same time.

The research question this thesis examines is the following:

How have the impacts of certain variables on innovative outcomes in developing countries changed after increasing national patent protection with respect to the TRIPS agreement? Particularly, how protection itself explains innovativeness in these countries.

Innovativeness is examined under the following measures:

1) Domestic patent application by residents and non-residents. 2) Patents granted to residents and non-residents.

5 Methodology and Data Description

To measure the impact of higher patent protection on the innovativeness of our selected countries, two different approaches with different levels of intensity are used.

The first measure of innovative activities is patent applications (Equation 1). It represents the least preferred measurements of innovative activities since not every filed patent gets granted. Some inventions are not patentable20 and some innovators consciously decide not to file a patent.

The more preferred measurement of innovative activity is patents which were granted (Equation 2). It is a more powerful test of innovative activity since it means that the new technology or process meets the requirements of the international standards of patent protection. Especially for developing countries it is a substantial step to meet these requirements since many of them did not have such strict protection before signing TRIPS21. With respect to the relevant literature as well as growth theory, we select the independent variables that are supposed to explain innovative activities. In our model, patent applications as well as granted patents, are explained by several explanatory variables. These variables are R&D expenditures as a share of GDP, level of patent protection, population size, GDP per capita, level of economic freedom, trade openness, share of highly educated people in the population and FDI inflows. We further distinguish between residents and non-residents because we want to test for the particular impacts after aligning to the TRIPS on the explanatory variables. Our main interests are the effects of protection on innovative outcomes. Patent application, as well as granted patents by residents, are regarded as domestic innovative outcomes.

Patent applications by non-residents can be both: A measurement of innovation (when a patent is filed for the first time in a foreign country) or diffusion because it introduces a new product or process to a foreign market. We use it in this thesis as a proxy for technology diffusion because first filings mostly occur in the innovator’s home country and not in a foreign country.

Patents granted to non-residents is our strictest measurement of diffusion even though most developed countries already applied strict protection22 before TRIPS commenced.

20 _{As we mentioned in previous parts of this thesis especially developing countries tend to} produce incremental innovations which don´t get granted under international standards. 21 _{One should be aware of the fact that activities by residents are restricted to residents of our} country set. However, data for non-residents comprise activities of all countries which have access to a market (thus, also every industrialized country).

5.1 Model Specification

1) $% &'()%( *++,-.'(-/%0₁₂

= 45+ $% 47(9&; <=+)%>-(?@)0)12+ 4A$% (B&C1DEF)12+ 4G$% (&/+)12 + 4_H$% (B;&)₁₂+ 4_I$% (<J)₁₂+ 4_K L&₁₂ + 4_M$% <;N ₁₂

+ 4O$% J;P 12+ QRLSTUUV+ W12

2) $% &'()%( B@'%(0₁₂

= 4₅+ $% 4₇(9&; <=+)%>-(?@)0)₁₂+ 4_A$% (B&_C1DEF)₁₂+ 4_G$% (&/+)₁₂ + 4_H$% (B;&)₁₂+ 4_I$% (<J)₁₂+ 4_K(L&)₁₂+ 4_M$% <;N ₁₂

+ 4_O$% J;P ₁₂+ QRL_STUUV+ W₁₂

Referring to the research question, whether aligning the domestic patent legislation with the TRIPS regulations stimulates innovativeness in developing countries, we impose two equations which represent innovative outcomes on the left-hand side. Since we consider two different time periods, before and after the implementation of TRIPS, we expect different effects arising in each period. That is because the TRIPS regulations raised the level of patent protection of the selected countries substantially. The tested dependent variables are Patent Applications (Equation 1) as well as Patent Grants (Equation 2). Both equations entail the same set of explanatory variables: R&D Expenditures which are expressed in R&D expenditures as a share of GDP, GP-Index which captures different levels of provided patent protection, Pop represents the population size of a country, GDP is GDP per capita, EF shows the level of economic freedom, OP represents the level of trade openness, EDU is the share of highly educated people withing a country, FDI are the FDI inflows and WTO is a dummy for the time after a country became a member of the WTO and had to implement the TRIPS agreement. Since we use panel data, n= 1…N indexes countries and t= 1…T time periods. W₁₂ capture the errors.

literature review, foreign firms tend to use their affiliates for non-R&D related activities. Hence, our expectations are not clear. Since we consider a country set which contains almost exclusively developing countries we assume the impacts of the GP-Index to be insignificant. That´s because developing countries traditionally don’t file patents. Thus, they have other incentives to either engage in innovative activities, or they just don´t have enough innovative capacities to innovate.

In a sense, economic freedom would lead to more innovative activities, but we do not assume economic freedom to show significant results. That is due to the low level of institutions in developing countries.

For non-residents, we assume that domestic R&D expenditures have a negative impact on granted patents since it represents a high innovative environment. The GP-Index is supposed to be positive because higher protection is an incentive for foreign firms to file patents. Since innovations in developed countries are mostly drastic, they are likely to meet the requirements of the TRIPS. Moreover, non-residential patents are our proxy for technology diffusion and thus are assumed not to be the first filing, it is convincing that we assume that these patents are granted again. The effects of GDP per capita can be positive as well as negative. A filed patent can be refused in a richer country because it does not represent a novelty (richer countries tend to have more innovations and therefore more substitutes). However, richer countries attract more innovators since the market is more promising. Our index of economic freedom should have positive impacts on granted patents since it represents a well-working environment for economic agents. Though, it could be that the effects of economic freedom and the ones of the GP-Index are overlapping. The effects of trade openness are assumed to be ambiguous again, but the effects of FDI inflows are supposed to be positive. Education levels can have a negative impact on granted patents for non-residents23.

5.2 Data Sources Dependent variables

Measuring of innovations or innovative behavior is a difficult task since they are intangible. None of the following indicators perfectly measures the innovativeness of countries24. However, they are indicative of the direction a country is moving in. Therefore, the following two different measurements are used to minimize the risk of receiving biased results: Patent applications by residents and residents as well as patents granted to residents and non-residents.

We obtain our data for granted patents as well as patent application by non-residents and residents from the WIPO Intellectual Property Statistics Center and the World Bank, World Development Indicators. The time span we examine covers each year from 1995 until 201025.

23 _{Same explanation as for patent applications by non-residents.}

Independent variables

R&D expenditures as a share of GDP is an input to produce innovations because investments in R&D are usually a necessary step to obtain innovations. We obtain this variable from the UNESCO Institute for Statistics (2017).

We use the index of economic freedom from the Heritage Foundation (2017). The index is a weighted average of the four key contributors to economic freedom, to wit rule of law, government size, regulatory efficiency and market openness. Having more economic freedom means that it is more beneficial for entrepreneurs to invest in innovative activities. That is since governments grant their innovators higher ownership- as well as judicial protection. The level of economic freedom can be summarized as, at least in the case related to this thesis, the easiness to make business in a country. The regulation system of a country supports private economic activity. Especially developing countries can largely suffer from a bad economic environment that affects the action of private economic agents. The index emphasizes rather an individualistic point of view instead of a collectivistic one (Miller and Kim, 2017).

Ginarte and Park Index (1997)26: The index incorporates five different categories:

1) Coverage (which contents can be patented); 2) Duration of the patent (the time span until it expires); 3) Enforcement (judicial possibilities for the owner of the patent); 4) Membership in international patent treaties (TRIPS, Paris Convention etc.) 5) Limitations which countries provide (i.e. compulsory licensing). The overall score of the index ranges from zero, which is the lowest score, until 5, which in turn is the highest possible score. The index is an equally weighted score thus there is no emphasis on a particular category. We are aware of the possible problem of endogeneity that can arise since the TRIPS agreement is part of the index as well as other components as for instance the Paris Convention. However, the idea to include the index was to observe the differences in the extent of protection in the entire dataset. Furthermore, it is useful to include a quantitative measurement of patent protection. Since the GP index is a quinquennial index we supposed a linear relationship between the intervals and adjusted the missing values by the moving average method.

Gross Domestic Product per capita (PPP, constant 2011 international $): In order to compare differences between countries with respect to their economic levels the most eligible indicator is GDP per capita converted into international dollars with purchasing power parity rates. Even though, GDP per capita is not a perfect measure of economic development because it is just an average measurement it provides information about the economic power of a country and it is useful for classifications. We obtained the data from the World Bank, World Development Indicators (2017).

Foreign Direct Investment Inflows (FDI as a % of GDP) and Openness to trade (Trade as a % of GDP): Both measurements represent how much a country is involved in the global trade system. We obtain our data from the World Development Indicator database from the World Bank, World Development Indicators (2017).

26 _{We use an updated version which provides data from 1960-2010.}

Education: Education is our proxy for human capital. As a measure of education, we use the Barro-Lee Index (2013) with tertiary education completion rates which covers the time span between 1950-2010. Since the Barro-Lee index is a quinquennial index, we supposed a linear relationship between the intervals and adjusted the missing values with a moving average method.

Population: The data comes from the World Bank, World Development Indicators (2017). Although we already use a variable that captures the size of an economy we introduce population size as well because some countries have a high GDP per capita but represent a small population like Hong-Kong. Thus, we hope to capture the market size much better by including the population size of a country.

Table 1 - Summary Statistics

Variable Observations Mean Standard

Deviation Minimum Maximum

6 Empirical Results

We started in the beginning with all developing countries that could use the transition period before fully implementing TRIPS (1995-2000). This was a country group of 70 countries. However, after checking the review of implementing legislation27 which was conducted by the WTO between 2000-2001 we had to find out that a lot of countries were not capable meeting the required deadline. Furthermore, countries like China only joined the WTO in 2001. Thus, we agreed to the implementation year 2001 even though some countries did not implement the entire TRIPS legislation and some countries already did before 2001. This procedure reduced the country set to 44 countries. After gathering all the data for the analysis, we had to drop 16 countries due to lack of data. Finally, we use a data set which covers 28 countries. We conduct our analysis for all measurements of innovative activities with this country set. To test our hypotheses more precisely we conduct a robustness test afterward without the richest countries in terms of GDP per capita28. Therefore, we excluded each country which belongs to the 90th percentile of mean GDP per capita. Hence, we excluded every country from the data set which has a higher mean GDP per capita than 31.336.63 international dollars.

Hausman test provided different suggestions for both time periods (before and after the implementation of TRIPS). For the first period between 1995-2000, the test suggested using a random effects model. However, for the second period between 2001-2010, the test suggests using a fixed effects model. We allowed for having unobserved heterogeneity between different countries and decided to use a fixed effects model since the second period entails more observations as well as more periods29. Furthermore, since we suspect heteroscedasticity, as well as serial correlation in our data set, we provide robust and clustered standard errors. As one can observe from the summary statistics Table 1, we are dealing with large parameters. Because we are aware of having non-linear parameters we transform our data and test our research question in a log-log30 _{model. Moreover, the log-log specification} allows for straightforward interpretations of our results. In order to ensure that the results are not disturbed by the implementation year of the TRIPS, we conduct the tests again for the year 2003 and 2005.

As can be observed from the summary statistics, Table 1, there is an increase of innovative activity when one compares both selected periods. All indicators of innovative activities such as R&D expenditures, patent applications as well as patent awards show higher numbers in the after TRIPS implementation period than before. The goal of our tests is to find out which

27 _{https://www.wto.org/english/tratop_e/trips_e/intel8_e.htm} 28 _{Both country lists can be found in the Appendix.}

29 _{We´re discussing the outcomes of the Random-Effects model briefly to show the} differences and similarities.

variables explain the increase of innovative activity and to which extent. Especially, the importance of protection (GP-Index) will be examined.

Firstly, before the main test is conducted, we introduce a regression analysis about the entire period (1995-2010) with an included WTO dummy. As one can observe from Table 1 in the appendix, there are no significant coefficients for the WTO dummy. These results suggest that there are no impacts of TRIPS on the innovativeness of developing countries. Since we want to further analysis which variables contribute to innovative outcomes in each period, we split the sample in a pre-TRIPS and after TRIPS implementation period.

Table 2 (a) shows the results of our first test, the patent application by residents, before as

well as after the implementation of the TRIPS. As we expected, in the case of developing countries, higher patent protection did not spur domestic innovative activities.

Before TRIPS has been implemented, applications were mainly driven by R&D expenditures and education. R&D expenditures consistently show a coefficient about 0.5 which are significant at 5 % level of significance. When including population, results represent varying coefficients for having a high share of highly educated people (EDU) of 0.7-0.9 that are significant to 10 % level of significance in model 1. In model 2, which omits the population variable31 education is significant at 5 % level and in model 3, which includes FDI also at a 5 % level of significance. For model 2 and 3, we observe positive and significant results for the GP-Index which suggest a positive relative importance of protection in developing countries for patent applications by residents. Furthermore, GDP per capita, economic freedom, trade openness, FDI inflows, and population size do not show any significant effects.

After the implementation of TRIPS, we do not obtain significant coefficients for the GP-Index anymore. The significant results for R&D expenditures become insignificant as well which suggests that the relative importance of R&D expenditures has declined after TRIPS was implemented. Only model 3 shows a positive coefficient of 0.5 at a level of significance of 10% when FDI is included.

Moreover, each test shows positive and results at 1% level of significance, for GDP per capita and the education. This suggests that the relative importance of being a rich country (or having a more promising market size) increases the number of patents filed by residents in. In every test, a 1 per cent rise in GDP per capita leads to 1.5 per cent rise in patent application. Moreover, having relatively more highly educated people raises the domestic patent applications, too. In each model, a 1 percent increase in education leads to about 0.14 percent more patent applications. In model 3, we do not obtain significant impacts of FDI inflows. However, this test shows significant results for R&D expenditures but only at 10% level of significance. This can be explained by the more competitive environment which occurs when foreign firms enter the domestic market.

The robustness test (Appendix: Table A4 a) is in line with our main results. In the pre- TRIPS period, positive and significant impacts of patent protection as well as R&D expenditures on applications by residents can be observed. Suggesting that aligning the domestic legislations to the requirements of TRIPS is less beneficial for developing countries with respect to protection. For the implementation period, as in our main test, only results for GDP per capita and education levels are positive and significant. Thus, we conclude that the economic more powerful countries do not distort results of the examination.

The Random Effects model confirms the impact of GDP per capita and Education on increasing patent applications. Furthermore, the Random Effects model finds positive and significant coefficients for the GP-Index and R&D Expenditures for the pre-TRIPS and after TRIPS period (Appendix: Table A2 a). However, since the Hausman test gives a clear proposal for the TRIPS implementation period, we stick to the results we obtain from the Fixed Effects model.

Table 2 b shows the results for the same test, however for patent applications by

non-residents, which we defined as diffusion of foreign technologies. As expected, contrarily to impacts on domestic applications, higher patent protection has a relatively higher importance for non-residents to file patents in developing countries after TRIPS got implemented.

Before the implementation of TRIPS, the choice of filing a patent by a non-resident was basically driven by high economic freedom (model 1-3). As one can observe from the table, this coefficient is highly significant in each model. According to our results, a rise in economic freedom by 1 per cent leads to a rise in patent applications by 3.6, 2.8, and 3 percent, respectively. The GP-Index is not significant in any model, suggesting that before TRIPS was implemented, protection was relatively unimportant for non-residents’ decision to file a patent. Moreover, model 1 shows positive and significant coefficients for population size and GDP per capita which suggests a positive impact resulting from a relatively bigger market. For that model, we also obtain highly significant, at 5 per cent, and negative coefficients for education. This would mean that a relatively higher domestic level of education decreases the proneness of non-residents to file a patent. Though this result occurs only in one specification. Thus, we do not attend to it.

After countries started adhering to the regulations of TRIPS, our results present a clear pattern. In each test, only the coefficient GP-Index is significant to 5 % level of significance and positive. This finding suggests a relative importance of higher protection levels on the diffusion of foreign technologies. In each model increasing the GP-Index by 1% leads to a rise of patent applications about 4%.

robustness test confirms these results again. As in the last test, we cannot assert any distortions by the outlying countries (Appendix: Table A4 b).

The basic findings from the Random Effects model are also in line with our main findings. Higher protection has a relatively high importance of diffusion of foreign knowledge which gets shown by significant and positive coefficients of the GP-Index (Appendix: Table A2 b).

Table 3 a shows the results of our more preferred test, patents that were granted to residents,

before and after implementing the regulations of TRIPS. As expected, in the case of developing countries, there are no significant impacts of patent protection.

Before implementing TRIPS, the control variables show different effects. First off, we do not obtain significant and positive coefficients of the GP-Index. In model 1, we obtain a negative and significant impact of population size at 10 % level of significance. This would imply that relatively larger countries have comparatively fewer patent holders than smaller countries. However, we observe this finding only in one specification. The impact of education is significant at a 1 % level of significance and positive for each model (model 1-3). This is in line with the theory as well as our presumptions since more highly educated people raise the production of innovations. The coefficients suggest a rise in granted patents up to 3.3 percent after increasing the share of highly educated people by 1 percent.

After the full implementation of TRIPS, we do not obtain significant results for the GP-Index, as expected. However, this result is not surprising since the requirements of getting a patent awarded are much higher under the TRIPS legislations. In addition, the implementation of TRIPS has further impacts of our previous empirical results. For each model, GDP per capita is significant at a 1 % level and positive (model 1-3). A rise of GDP per capita of 1 percent leads to an increase in granted patents of more than 2 percent in each model. This finding suggests that after TRIPS was implemented, the importance of a relatively higher GDP per capita increases patents which are granted. The results for education completely change which we did not expect. Only in model 3, which includes FDI, education is significant at a 10% level. Furthermore, we do not observe any impact of FDI inflows on patents awarded to residents. A result from our test, which is very surprising is that R&D expenditures show no relative importance for innovative outcomes. However, this is probably due to the incremental innovations which developing countries tend to produce and do not meet the high requirements of the TRIPS regulations.

These results are confirmed by our robustness check which does not present any deviations (Appendix: Table A5 a).

Applying the Random Effects model produces different results. The biggest difference is observable for the after TRIPS period because R&D expenditures as well as the GP-Index present significant and positive coefficients (Appendix: Table A3 a).

Table 3 b shows the results for our more preferred model for diffusion of foreign

Before the implementation of TRIPS, we find for each model, negative and at 5% level of significance coefficients for domestic R&D expenditures. This suggests that domestic R&D expenditures, or better a more innovative environment, have a relatively negative impact of patents granted to non-residents. The GP-Index shows as well as R&D expenditures consistently significant and positive coefficients at a 5% level of significance. After omitting population in model 2, economic freedom starts to be significant, too. In model 1, we obtain a relatively negative impact of population size on diffusion. Moreover, the impact of education is significant and positive. In model 3, which includes FDI inflows, our previous results keep the same, but interestingly our measurement for trade openness is negative and significant at a 10% level. The strongest conclusion we can extract from the pre-TRIPS period is that economic freedom and an already existing level of patent protection led to patents that were awarded to non-residents and hence impacted diffusion of foreign innovations.

After the implementation of TRIPS, the results change. Though, the general impact of domestic R&D expenditures and the GP-Index keep the same. Domestic R&D expenditures still have a negative and higher protection a positive impact on patents that were granted to non-residents. However, the coefficients for the GP-Index substantially increase. In model 2 and model 3, a 1 per cent increase of protection leads to 3.2 as well as 3.4 percent more patents awarded to non-residents, respectively.

However, the previously significant results of economic freedom become insignificant which probably arises due to overlapping effects between the GP-Index and economic freedom. From model 1, we observe significant and negative impacts from a higher population again. The same is valid for GDP per capita, which coefficient is positive at a 1% level of significance. These are individual cases which do not occur in the other models. In model 3, we do not observe impacts of patent grants with respect of FDI inflows.

Again, excluding the three richest countries from the data set does not distort the previous results. Hence, we can conclude that our results are robust (Appendix: Table A5 b).

Using a Random Effects model (Appendix: Table A3 b) instead of Fixed effects do not change the results with respect to the GP-Index. Protection is a highly significant and positive contributor to patents which are awarded to non-residents. The only difference is that domestic R&D expenditures do not have an influence on the patent awards to non-residents as the Fixed Effects model asserts.

Since the full implementation of the TRIPS regulations differed among countries of our data set, we agreed on using the year 2001 as the ‘common’ year of implementation. We are aware of possible distortions by using this year and conducted the same analysis with 2003 and 2005 as the year of implementation. These specifications do not change our results with respect to the GP-Index.

Table 2 (a) – Ln Patent Applications by Residents

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Model 1 Model 1 Model 2 Model 2 Model 3 Model 3

VARIABLES Before TRIPS After TRIPS Before TRIPS After TRIPS Before TRIPS After TRIPS

Ln_RD_SHARE_GDP 0.519** 0.335 0.490** 0.328 0.492** 0.520* (0.244) (0.324) (0.230) (0.335) (0.232) (0.266) LN_GP_Index 1.228 1.103 0.982* 1.076 0.967** 0.977 (0.815) (0.989) (0.470) (0.901) (0.446) (0.790) Ln_Pop -1.083 -0.214 --- --- --- --- (2.191) (1.281) --- --- --- --- LnGDPCAP -1.250 1.479** -1.103 1.444*** -1.104 1.385*** (1.298) (0.535) (1.198) (0.512) (1.204) (0.454) Ln_EF 0.315 1.034 0.459 1.034 0.482 1.193 (1.141) (0.936) (1.048) (0.928) (1.006) (0.837) Ln_TRADE_OP -0.122 -0.0576 -0.112 -0.0488 -0.145 0.157 (0.429) (0.298) (0.431) (0.292) (0.422) (0.292) Ln_EDU 0.900* 0.136*** 0.771** 0.136*** 0.768** 0.140*** (0.444) (0.0411) (0.345) (0.0429) (0.342) (0.0349) Ln_FDI --- --- --- --- 0.0164 -0.0563 --- --- --- --- (0.109) (0.0525) Constant 33.33 -9.910 12.99 -13.25* 13.05 -13.90** (44.16) (22.00) (10.76) (6.822) (10.90) (6.121) Observations 70 189 70 189 70 187 R-squared 0.333 0.423 0.330 0.422 0.331 0.483 Number of id 18 25 18 25 18 25