Patent Regimes, Host Country Policies and the Nature of MNE Activities
Usha Nair-Reichert
Georgia Institute of Technology
and
Roderick Duncan
CharlesSturtUniversity
Abstract
This paper provides useful insights in the debate regarding the relationship between stronger patent rights, host country policies and multinational enterprise (MNE) activity using panel data from US MNEs. It analyzes the impact of stronger patent protection on the exports, local affiliate sales and licensing activities by explicitly modeling the joint nature of the MNE’s decision making process in servicing a foreign market. The key findings support the idea that the policy environment in the host country influences the impact of stronger IPRs on US MNE activities during the period 1992 to 2000. A risky environment in the host country appears on average to have a negative and significant impact on unaffiliated exports and affiliate sales. Increased patent protection in high-risk countries on average appears to reduce licensing, and increase unaffiliated exports, suggesting a dominant monopoly effect of stronger IPRs in the former case and a dominant market expansion effect in the latter case.
JEL Classification: Q10, Q17, F23, K33, O34.
Key words: patent protection, intellectual property rights, imitation, innovation, exports, affiliate sales, licensing, R&D.
Address for correspondence: Usha Nair, School of Economics, Georgia Institute of Technology, Atlanta, GA 30332-0615; Tel: (404) 894-4903; Fax: (404) 894-1890; E-mail:
Roderick Duncan; School of Marketing and Management, CharlesSturtUniversity, Bathurst, NSW 2795, Australia; Phone: (02) 6338-4982; Fax: (02) 6338-4769; Email:
We would like to thank Janusz Mrozek, Amy Glass and the participants at the 2006 Western Economics Association meetings, the 2006 EEFS conference, and the Emory Trade Workshop for their useful feedback.
- Introduction
In recent years, intellectual property rights (IPRs) have become increasingly significant in many types of international transactions, and strengthening IPRs has become a key issue in global trade negotiations.[1] The goal of this paper is to address the following question: to what extent does the host country policy environment matter when thinking about the effects of stronger IPR protection? A review of the literature suggests that previous empirical work on the impact of IPRs has focused largely on the monopoly and market expansion effects of stronger IPRs and analyzed how the imitative capabilities of the country influence the impact of stronger IPRs. The literature does not provide unconditional arguments regarding the impact of stronger patent rights on trade, foreign direct investment (FDI), and licensing. In particular, host-country specific collateral factors, which often affect the impact and economic value of patent rights, have been not been studied much in the empirical literature on IPRs.[2] Maskus (1998) notes that it is inadequate to analyze the implications of IPR systems without also considering their position in the general regulatory structure.
There are a large number of theoretical models that assess the impact of intellectual property protection on MNE activities in the host country, but the results of these models vary a great deal depending upon the modeling assumptions (for example Lai (1998) and Glass and Saggi (1999, 2002a, 2002b, 2002c), Yang and Maskus (2001), Branstetter et al. (2007)). The theoretical models that consider the interactions between the host country environment and intellectual property protection in assessing the impact of stronger IPRs on MNE activities in the host country have largely focused on human capital and openness. Vishwasrao (1994) uses a strategic partial equilibrium model to argue that lack of patent protection may encourage FDI relative to licensing where there are country-specific fixed costs of establishing a subsidiary in the host country. For example, subsidiary production faces certain inherent disadvantages such as the fear of expropriation and other political and economic risks. Branstetter et al. (2007) analyze how the strengthening intellectual property rights in developing countries impacts the level and composition of industrial development. They conclude that IPR reform in the South leads to increased FDI from the North, as Northern firms shift production to Southern affiliates.[3] Rodrik (1991) suggests that “even moderate amounts of policy uncertainty can act as a hefty tax on investment, and that otherwise sensible reform may prove damaging if they induce doubts as to their permanence.”[4]
Empirical evidence on the impact of IPRs is also rather diverse.[5] Despite the importance of the host country policy environment in explaining the impact of patent rights on trade, investment and technology flows, there is limited empirical evidence in this regard. Most of the existing work has focused either on the interaction between IPRs and human capital / technological capability or the openness of the economy to trade. Among the papers analyzing the impact of IPRs the most relevant to the current work are Maskus (1998) and Smith (2001). Maskus (1998) uses a simultaneous equation framework to analyze the impact of IPRs on different types of commercial flows, namely patent applications, sales, exports, and level of affiliate investment assets using data from 1989-1992. He finds that exports to affiliates are highly positively affected by patent strength in developing economies. While the average patent strength has limited effect on affiliate sales, the impact is significantly positive in developing countries. Smith (2001) uses cross-sectional data for 1989 to analyze the impact of stronger IPRs on US unaffiliated MNE exports, local subsidiary sales and licensing receipts of subsidiaries from unaffiliated firms and indicates that strong patent rights increase US affiliate sales and licensing, particularly in countries with strong imitative abilities. Related research also suggests that the level of IP protection itself varies, depending on the technological capabilities of the host countries. In short, there has been relatively little evaluation of patent rights in conjunction with other international and domestic policies to assess their full impact on a multinational’s mode of servicing a host country market.[6]
Saggi (2002) has pointed out that most of the literature discussed suffers from another fundamental problem; either FDI or licensing is the only channel through which the Northern firm can produce in the South. A more sophisticated analysis of the consequences of strengthening patent rights in the South requires that we consider all of the MNE’s options (trade, FDI and licensing) jointly. Most of the existing empirical studies, with the exception of Maskus (1998), have not adequately captured the joint effects of MNE activities over time, as they are based on single equation studies, cross sectional analysis, or time-averaged data.
This paper builds on the Maskus (1998) and Smith (2001) papers. It develops a theoretical framework and empirically evaluates the interactive effects of multiple institutions and polices, including IPRs on MNE activities. It differs from the earlier papers in that it focuses on the role of the host country policy environment and its impact on the influence of stronger IPRs in facilitating the process of internationalization of technology transfer, investment, and trade. The analysis also uses panel data for the years 1992-2000 and a simultaneous equation framework to account for the fact that the decision by a firm to export a patentable product is jointly determined with decisions to service the market through licensing or FDI, as was done in Horstmann and Markusen (1987) and also Maskus (1998). The key findings indicate that the policy environment in the host country influences the impact of stronger IPRs on US MNE activities during the period 1992 to 2000.
The rest of this paper proceeds as follows. Section 2 presents the basic model and the propositions that are being tested. Section 3 presents the empirical analysis and section 4 summarizes the key conclusions and policy implications.
2.Theoretical Framework
This section develops a framework to examine how the host country policy environment and patent protection jointly affect the MNE’s mode of entry into a foreign market through exports, FDI or technology licensing to the host country.[7] The assumptions of the model are similar to those in a standard North-South framework. An MNE in the North invents a new technology or good and patents it. Production for the Southern market can take place in the North or the South. The Northern firm has three choices or modes of production: it can produce at home and export to the South, it can form an affiliate in the South to service that market, or it can license its technology to a Southern firm. Although firms typically produce for both their domestic and export markets, we assume in this model that all production in the North is for export and that all production in the South is for the local market.[8] This is a static one-period model and does not take into account the post-imitation rivalry between firms and the strategic interactions among the rival firms.[9]
The decision of a Northern firm about the mode of production will depend on several factors in the host country. A unit of production, x, requires one unit of labor, at a cost of w in the North and ws in the South. Southern labor is assumed low cost, w > ws, and production in the South avoids the per unit tariff, t.[10] However production in the South will incur either a given fixed cost, f, for setting up a local affiliate or must share profits with a Southern licensor at some rate, 1-.
We assume technological sophistication of local firms and labor is related to the level of research and development conducted by local firms, R. Technological sophistication is assumed to improve productivity of Southern labor and so lower the per unit cost of production, ws(R)/R < 0. Technological sophistication is also assumed to increase the imitative ability for new technologies by local firms. As in Fosfuri (2000) an increase in the imitative ability of local firms raises the bargaining power for Southern firms in a licensing agreement with Northern firms. This increase in bargaining position will lower the share of profits in the licensing agreement which can be captured by the Northern firm, /R <0.
The Northern firms face an inverse demand function for their output in the Southern market, p = a-bx, where p is the price and x represents quantity. xe, xf and xl represent the quantities under each of the three modes of servicing the market: exports, FDI, or licensing. The Northern firm makes a profit of e by exporting, a profit of f from opening an affiliate in the host country, and l by licensing its technology to the foreign firm. We assume that the Northern firm will choose the mode that produces the highest profit.[11]
In the case of exports, the Northern firm’s profits will be:
Differentiating with respect xe, and solving for the equilibrium profits from exports, we get,
In the case of FDI, the Northern firm’s profit function is:
The equilibrium profits from FDI are
The third approach to servicing the Southern market is by licensing the technology to local firms in the host country. The division of rents between the Northern firm and the Southern firm depends on several factors such as the relative bargaining powers of the contracting parties, degree of market competition and government intervention (Contractor and Sagafi-Nejad, 1981). Several papers (Yang and Maskus (2001b) and Taylor (1994) and Markusen (2001)) indicate that stronger IPRs lead to a reduction in the licensors’ monitoring and contacting costs and raise the return on transferring technology. There is also evidence in the literature that stronger IPRs increase the licensor’s share of rents. Gallini and Wright (1990) show that where information is asymmetric and imitation is possible, the licensor sacrifices some of the rent, although the licensor’s share increases with the increase in imitation costs.
In our model, we follow Yang and Maskus (2001b) and Fosfuri (2000) and assume that the licensor’s rent share is a positive function of Southern IPRs. Stronger IPRs make it more difficult to imitate, and hence the licensee commits not to imitate at a lower rent share, thus increasing the licensor’s share.
Let be the Northern firm’s share of rents, such that 0<<1, =(k, R), and /k >0, where k measures the strength of the Southern IPR regime and R the level of domestic research and development. Hence the Northern firm makes a profit of (k, R)sl by licensing technology to the Southern firm, where sl is the profits of the Southern firm.
The Southern firm’s profit function can be written as follows:
The equilibrium profits of the Southern firm are:
Hence, the equilibrium profits of the Northern firms from licensing its technology to the Southern firm is
The profitability of each mode of operation for the Northern firm is not observable. However we can observe indicators of the level of operation of each mode by all Northern firms in a host country. The relative profitability of the three modes of MNE operation given by Equations 2, 4 and 7 will determine the aggregate level of operation of each mode in a host country. Holding all else constant, an increase in the profitability of one mode should raise the level of operation of that mode and lower the level of operation of the other two modes. For a higher level of tariffs for example, we would expect to see a lower level of exports and higher levels of FDI and licensing.
Applying the logic from the previous paragraph to Equations 2, 4 and 7, we derive the expected signs of the changes in the aggregate level of operation of exports, FDI and licensing for the parameters of interest from our model: k, ws, t, f and R. These expected signs are reported in Table 1. We can find expected signs for all of the coefficients of the parameters for all three modes of operation, except in the cases of ws and R for licensing. In these cases the coefficients are ambiguously signed due to the interactions between , R and ws, as well as the substitution away or towards that mode of operation from other modes. In the case of a rise in ws the profitability of FDI falls, while the profitability of licensing also falls, but not by as much. Some firms will switch away from licensing to exports, but other firms will switch to licensing from FDI. The net impact in ambiguous.
3.Empirical Analysis
In this section we develop an empirical framework based on the gravity model to test the propositions on the impact of patent protection and country specific fixed costs on the MNC’s mode of entry that we derived from theory.
i) Data Considerations
The data for this study covers the activities of majority-owned US MNEs from 1992 to 2000, and includes aggregate bilateral data for each type of MNE activity: unaffiliated exports, local affiliate sales and unaffiliated licensing receipts (lMNA in equation 12 below represents real values of three different types of US MNE activity in logs). This data is different from that used in earlier work in that it covers a later time period, and has both time series and cross-sectional dimensions. Before we proceed with the analysis it is useful to discuss the main characteristics and limitations of the data.
Data on exports by US parents to unaffiliated firms is not available for the non-benchmark survey years. So, we use data on exports from the US to unaffiliated firms, which is calculated as the difference between bilateral exports from the US to the host country, and the exports of US parents to their affiliates in the host country. We also use local sales by affiliates in the host country obtained from the BEA, instead of FDI flows, since our analysis focuses on the outputs of MNE activity rather than inputs. Licensing receipts of US parents from unaffiliated firms are from the Survey of Current Business.[12] While interpreting these variables it is worth noting that in the case of local affiliate sales, US MNEs may already have assets in the host country and may vary the degree of utilization of these assets to service the local market.[13]
The host country policy environment variables in our theoretical model include the wage rate, ws, patent protection, k, tariffs, t, the fixed costs of doing business in the host country, f, and the host country’s level of research and development, R. Data availability considerations compel us to be parsimonious with the use of variables included in our regressions. The home country’s wage rate, ws, is proxied by the compensation (Wages), the patent protection, k, by the Ginarte and Parks index of patent protection (PR), tariffs, t, by the Sachs-Werner openness index and trade to GDP ratio (Open), and the fixed costs of doing business in the host country, f, by the Euromoney country risk index and the ICRG risk indices (Risk).[14] The ICRG Country Risk Index is a composite of political, economic and financial risk indices. Our measure of country risk is defined as (100- Country Risk Index).[15] Our transformation allows higher index values to be associated with higher levels of risk, and hence higher fixed costs, and facilitates interpretation of the results.
The real wages in the host country are calculated from the total compensation and the number of employees of MNE subsidiaries available from the BEA deflated by the appropriate GDP deflator.[16] The Sachs and Werner (1995) openness index, Open, which takes values of either 0 or 1, measures the trade distortions in the host country.[17] If Open has a value of 1, it is comparable to having low tariffs in the theoretical model. The expected sign of Open from our model in Section 2 is the opposite of the expected sign of tariff in Table 1. The index of patent protection (PR) is based on Ginarte and Parks (1998), and Park (2002) work analyzing national patent law. The overall composite index ranges between zero and five with higher values signifying greater patent protection.[18] The literature also suggests that the impact of increased patent protection in a country depends on its ability to imitate. Hence we include the host country’s R&D as a proportion of GNP to proxy for the host country’s ability to imitate (R&D), and the interaction between the patent index and R&D is a measure of the threat of imitation. The classification of countries according to regions,[19] the real GDP per capita (gdppc), population (pop), are from the World Development Indicators (2004); the other gravity variables such as distance (dist) proxied by the geographical distance between the US and its trading partners and language (langdum) a dummy variable that equals 1 if the official language of the country is English, and zero otherwise are from Jon Haveman’s website.[20] The GDP per capita, distance, population, and wages are all expressed in logs. We note that MNEs can typically choose both the country in which to invest in as well as the mode of entry, and if some country specific factors change, in addition to shifting the mode of entry, they can also chose to invest in another country. Our data however does not provide information for this type of analysis, and hence we focus only on the mode of entry into a given market.