Intellectual Property Rights, MNFs and Technology Transfers

Transcription

1 Intellectual Property Rights, MNFs and Technology Transfers Sara Biancini and Pamela Bombarda July 2016: VERY PRELIMINARY AND INCOMPLETE Abstract We build a theoretical model in which MNFs based in developed countries (the North) offshore production of an intermediate good in a developing country (the South). MNFs can open a subsidiary in the country (vertical integration), or rely on independent contractors (outsourcing), through a licensing contract based on a Northern blueprint. The environment in the destination country is characterized by incomplete contracts and an imitation risk related to weak IPR enforcement. Our model shows that under reasonable assumptions on model parameters, a reinforcement of IPR in the South increases the relative share of imports from vertically integrated suppliers (and decrease the share of ousourcing to independent suppliers). The model predictions are tested using data on intra-firm trade of US multinationals provided by the U.S. Related-Party Trade database. JEL classification: F23, F12. Keywords: Intellectual Property Rights, MNFs, FDI, outsourcing, international trade. We would like to thank... Université de Caen Normandie, CREM, sara.biancini@unicaen.fr. Université de Cergy-Pontoise, THEMA, pamela.bombarda@u-cergy.fr. 1

2 1 Introduction The paper studies the impact of Intellectual Property Rights (IPR) on the location of production by multinational firms (MNFs). Several empirical papers show that increasing IPR enforcement has a positive effect on bilateral trade flows (Maskus and Penubarti, 1995, Smith, 1999, Maskus and Fink, 2005). Others highlight the positive impact of IPRs on licensing activity (Yang and Maskus, 2001). In general, the empirical literature often considers the impact of IPRs on either export, licensing or FDI and does not consider how IPRs influence the preliminary choice between these production modes. In the present paper, we concentrate on the choice between FDI and outsourcing. This issues has been studied by Antràs (2003) and Antràs and Helpman (2004) in a context of contract incompleteness, which is naturally related to the presence of weak property rights. The specific role of IPRs has not received much attention in the literature. We build a theoretical model in which MNFs based in developed countries (the North) offshore production of an intermediate good in a developing country (the South). MNF can open a subsidiary in the country (vertical integration), or rely on independent contractors (outsourcing). Outsourcing requires transferring technological knowledge through a Northern blueprint. The environment in the destination country is characterized by incomplete contracts, as in Antràs and Helpman (2004). In addition to contract incompleteness, production in the South exposes the MNF to a risk of imitation, depending on the level of enforcement of IPR in the South. In case of a contractual breach, if intellectual property rights are not well protected, the Southern manufacture can operate the technology on the side because it has learnt the Northern technology. We assume that this can occur both in the case of vertical integration and outsourcing, but the payoffs of the multinational firm in the incomplete contract framework are affected differently depending on the chosen ownership structure. Stronger IPR increase the profitability of the MNF under both offshoring modes. However, we show that a reinforcement of IPR in the South, under reasonable assumptions on model parameters, is likely to increase the relative share of imports from vertically integrated suppliers (i.e. intrafirm imports) and decrease imports from independent suppliers. We then test the predictions of the model using data on intra-firm trade of US multinationals provided by the U.S. Related-Party Trade database (US Census). 2

3 2 The model We adopt a standard model with two countries, North and South, and a factor of production (labor). World consumers have identical preferences represented by: U = x µ J X µ j, 0 < µ < 1 j=1 Where x 0 is a homogeneous good, X j is aggregate consumption of differentiated varieties j and µ is a parameter. The differentiated sector produces a continuum of horizontallydifferentiated varieties, x (v), from two intermediate goods, z and m. Both z and m are produced using one unit of labour. Each variety produced in the differentiated sector is then supplied by a Dixit-Stiglitz monopolistically-competitive firm which produces under increasing returns to scale arising from a fixed cost. World aggregate consumption in sector j is given by: [ X j = ] 1 x j (i) α di α, 0 < α < 1 Where 1/(1 α) is the elasticity of substitution between two varieties. As a consequence, the inverse demand for each variety writes: p j = A j x j (i) α 1, 0 < α < 1 (1) where A = X µ α j. New varieties can be only invented in the North. To start producing a new variety, a Northern firm has to bear a fixed cost f E unit of Northern labor. Upon paying this fixed cost, productivity level θ is drawn. Production of the final-good variety requires the combination of two factors, z j (i) and m j (i), associated with headquarter services and manufacturing. Output follows a Cobb-Douglas production function: [ ] ηj [ ] 1 ηj zj (i) mj (i) X j = θ, 0 < η j < 1 (2) η j 1 η j The parameter η j is sector specific and it is the higher the more the sector is intensive in headquarter services. Intermediate output m i can be produced in the North or in the South, while we assume that final production by multinational takes place in the North. Transportation costs are set to zero for simplicity. We assume that, when deciding to produce the intermediate m overseas, multinational 3

4 firms (henceforth MNFs) can choose two alternative modes of production: vertical integration or outsourcing. Under vertical integration, intermediate m is produced in the South by a foreign affiliate and then shipped to the North headquarter where final assembling takes place. Under outsourcing, intermediate m is produced in the South by an independent manufacturer and then exported to the North MNF. In both cases production requires transferring technological knowledge to the Southern supplier. Fixed organisational costs are higher under vertical integration than under outsourcing: f V > f O (3) In case of outsourcing, the MNF transfers the technology to the independent supplier, which produces the intermediate. If IPR are poorly enforced in the South, multinationals are exposed to a risk of imitation and expropriation of the technology developed by the headquarter. We assume that both intermediate and final goods are protected by patents. The location of the production of m and the mode of ownership (V or O) are chosen ex-ante by the multinational. Each MNF offers a contract to a manufacturer in exchange for an upfront fixed fee. Dropping the index j for the sake of notation, we can write the potential revenue from the sale of each differentiated good as R(i) = p(i)x(i). If IPR are protected (i.e. with probability λ), the price for firm i corresponds to the inverse market demand. Replacing (1) and (2): [ ] αη [ ] α(1 η) z(i) m(i) R(i) λ = Aθ α (4) η α(1 η) 3 The incomplete contracts problem Following the property right approach, we assume that both under vertical integration and outsourcing, contracts are totally incomplete (see Antràs and Yeaple, 2014). Due to ex post bargaining, the parties share the surplus from the agreement (i.e., revenues net of the sum of the parties outside options). We denote β and 1 β, with β [0, 1], the bargaining weights of the headquarter and the South manufacturer respectively. Both vertical integration and outsourcing feature an ex-ante contracting stage t 0, an investment stage t 1, and an ex-post bargaining stage t 2. At t 2, the outside options available to the headquarter (H) and the manufacturer (M) will now be a function of the ownership decision at t 0. For simplicity, we model imperfect enforcing by assuming that IPR are enforced only 4

5 with some probability λ, which represents the strength of IPR enforcement in the South. In our setup, the protection of intellectual property rights matters both under outsourcing and vertical integration. We assume that the probability of IPR enforcement λ does not depend on the offshoring mode and it is the same under ousourcing and vertical integration (we can relax this assumption later). 3.1 Vertical Integration In case of vertical integration, we follow the property right approach, by assuming that in case of contractual breach the MNF can fire the manufacturer and find an alternative way to assemble the final good. In fact, H holds property rights over the input m produced by M, and thus H has the ability to fire M if he has refused to agree on a transfer price, while still being able to capture a fraction of the potential revenue generated by combining z and m. This share δ is assumed to be lower than one, 0 < δ < 1, which reflects the intuitive idea that MNF cannot use the input m as effectively as it can with the cooperation of the South intermediate producer M. However, we also assume that the potential realize revenue δr is fully realized only if IPR are enforced, i.e. with probability λ. With probability 1 λ the manufacturer, who has learnt the Northern technology, can decide to operate it on the side, paying a small fixed cost to develop a concurrent variety. Doing so, the manufacturer can divert a share δφ of the revenue. For simplicity, we assume that the headquarter keeps the residual share δ φ V of the revenue in case of contractual breach. 1. Thus, the expected revenue of the MNF if there is no contractual breach is R. In case of contractual breach if is λδr + (1 λ)(δ φ V )R. As a consequence, the H s share of the revenue under vertical integration writes: Similarly, M s share of revenue writes: β(1 δ) + δ (1 λ)φ V β V [λ] (5) (1 β)(1 δ) + (1 λ)φ V 1 β V [λ] (6) The operating profit of the MNF under vertical integration: 1 This assumption is similar to the one in Antràs and Yeaple, 2014, chapter 2. 5

6 where β V Π MNF V = β V [λ]r zw N (7) is the MNF s share of total revenue under vertical integration, R represents the MNF revenue from the sale of the final good, f V is the fixed cost, and w N represents the wage rate in the North (which includes transportation costs to the South). Similarly, the operating profit of the integrated manufacturer writes: Π M V = (1 β V [λ])r mw s (8) 3.2 Outsourcing Under outsourcing, in the absence of an agreement at t = 2, H is left with a zero payoff. H has no time to find an alternative supplier for the provision of m. On the other hand, if there is a contractual breach, the manufacturer can operate the technology on its own and realize a share of potential revenue φ O R. We also assume that φ O R is lower than 1 because the manufacturer is less efficient when attempting to realize the output after breaking the initial collaboration with H. In addition, the potential income only realizes if intellectual property rights are not protected. Otherwise, the manufacturer is not allowed to operate the foreign technology. As a results, the H s share of revenue under outsourcing is as follows, Similarly, M s share of revenue writes: β(1 (1 λ)φ O ) β O [λ] (9) (1 β)(1 (1 λ)φ O ) + (1 λ)φ O 1 β O [λ] (10) The operating profit of the MNF under outsourcing writes: Π MNF O = β O [λ]r zw N (11) Similarly, the operating profit of the independent manufacturer writes: Π M O = (1 β O [λ])r mw s (12) In line with the literature, we assume that β V [λ] > β O [λ] which requires that δ(1 β)(1 6

7 λ) > φ V βφ O. This captures the idea that H captures a higher share of the surplus under vertical integration than under outsourcing. 4 Equilibrium H and M maximise (7) and (8) respectively under vertical integration, and (11) and (12) respectively under outsourcing. From the first order condition of these programs and using (4) we can write the total operating profits as: where π k = A 1/(1 α) θ α/(1 α) Ψ k [λ] w N f k, k = {V, O} (13) 1 α (β k [λ]η + (1 β k [λ])(1 η)) Ψ k [λ] =, k = {V, O} (14) ((1/α)(w n /β k [λ]) η (w s /(1 β k [λ])) 1 η α/(1 α) ) We first note that Ψ k [λ] is not necessarily increasing in λ. When λ increases, the share of revenue retained by H, β k increases. However, due to incomplete contracts, this does not necessary increases profits. When headquarter intensity η is close to one, higher values of β k yield higher operating profits. Conversely, when η is close to zero, lower values of β k yield higher operating profits. As in Antràs and Helpman (2004), due to incomplete contracting, efficient allocation of property rights requires that ownership is allocated to party that contributes more to the value of the relationship, to limit underinvestment. Because we have assumed that the fixed costs of vertical integration is higher than the one under outsourcing, and that β V [λ] > β O [λ], then vertical integration can arise only if η is high. In this case, if β V [λ]/ λ > β O [λ]/ λ, increasing λ increases the relative profitability of vertical integration. This realizes in our model if φ V > βφ O. For this to be satisfied, φ O has to be not too large (although it can be higher than φ O, because β < 1). Intuitively, if under weak property rights outsourcing increases disproportionately the outside option of the manufacturer, then increasing IPR unambigously increases the attractivity of this production mode. However, vertical integration also require technology transfer and a risk of expropriation (possibly lower but non negligible). Because under vertical integration the headquarter have property rights on production facilities, this can help to respond to the risk of contractual breaches and allows to reap a higher part of the surplus, which is even larger when intellectual property rights are strongly enforced. 7

8 We now consider firm heterogeneity. We suppose that the productivity parameter θ is randomly drawn at the ex-ante stage t O. Because f V > f O and Ψ V [λ] > Ψ O [λ], less efficient firms choose outsourcing and more efficient firms select into vertical integration. More precisely, the critical threshold above which firms outsources is given by the free-entry condition: π O = A 1 1 α θ α 1 α ΨO [λ] f O w n 0 (15) From which we deduce that firms outsource in the South if their productivity level is higher than the threshold θ O [λ]: θ O [λ] = ( ) α fo w 1 α n AΨ O [λ] (16) In addition, sorting into vertical integration is given by the relation π V π O > 0: π V π O = (A 1 1 α θ α 1 α ΨV [λ] f V w n ) (A 1 1 α θ α 1 α ΨO [λ] f O w n ) > 0 (17) So that vertical integration is preferred to outsourcing by MNFs having productivity above the threshold θ V [λ]: θ V [λ] = ( (fv f O )w n ) α 1 α A(Ψ V [λ] Ψ O [λ]) (18) Under our assumptions, an increase in λ move both thresholds to the left. Relative less efficient firms begin to offshore, and among them the more efficient ones choose vertical integration. decreases. The threshold above which vertical integration is preferred to outsourcing also To compute the share of share of manufacturing inputs that are transacted within multinational firm boundaries, we assume foreign inputs are priced such that these input expenditures constitute the same multiple of operating profits under all organizational form. Integrating over firm s types and taking the ratio, we have that the share of imports transacted within firm boundaries is given by: σ V [λ] = A 1 α θ V 1 α θ 1 α ΨV [λ] dg(θ) θ V θ O A 1 α 1 α θ 1 α ΨO [λ]dg(θ) + A 1 α θ V 1 α θ 1 α ΨV [λ] dg(θ) If we assume that θ follows a Pareto distribution (with parameter κ > 1/(1 α) 1), (19) 8

9 using thresholds (16) and (18) we obtain: σ V [λ] = [ ( θv [λ] θ O [λ] Ψ V [λ] Ψ O [λ] ) κ ( 1 ] 1 α 1) 1 + Ψ V [λ] Ψ O [λ] (20) where: [ θ V [λ] θ O [λ] = f O f V f O ( )] α ΨV [λ] Ψ O [λ] 1 1 α (21) We note that the relative share of intrafirm imports σ V [λ] is increasing in the ratiosψ V [λ]/ψ O [λ]. This ratio can be written: Ψ V [λ] Ψ O [λ] = [ (1 βv [λ] 1 β O [λ] ) 1 η ( ) ] α η βv [λ] β O [λ] 1 α ( 1 α[βv [λ]η + (1 β V [λ])(1 η)] 1 α[β O [λ]η + (1 β O [λ])(1 η)] ) (22) Equation (22) is a complex function of λ. However, we can establish that for η and α sufficiently large the ratio in (22) behave as the ratio β v [λ]/β o [λ]. For more general values of α, we can show that the same occurs as long as φ V βφ O not too small. If these conditions are satisfied, an increase in IPR enforcement increase the share of intrafirm trade. 5 Empirical evidence The goal of this section is quantify empirically the role of IPR in affecting the global sourcing decisions of firms. Since our model describes firm organizational decisions, firm-level data could seem more appropriate. Nevertheless, the alternative firm level dataset available was not well suited for this study. In fact, the French firm-level data from EIIG (Echanges Internationaux Intra-Groupe), has different limitations. First it is a survey available only for 1999, and so it misses a time dimension. Second, it is restricted to firms that are owned by manufacturing groups that control at least fifty percent of the equity capital of an affiliate based outside France. 2 Therefore, our attempt to test the relationship between intrafirm imports and Intellectual Property Right (IPR), exploits the U.S. cross-industry data provided by Antràs (2016). 3 2 For a discussion about firm-level data sets see Antràs (2013). 3 See Antràs web page 9

10 5.1 Data description The share of U.S. intrafirm imports comes from U.S. Related Party Trade database made available in Antràs web page. To control for headquarter intensity of an industry, we follow Nunn and Trefler (2013) and Antràs and Chor (2013) in using physical capital per worker, skill and R&D intensities of U.S. manufacturing firms. The on line dataset also includes additional industry level covariates, such as within industry size dispersion, capital equipment etc. All these are U.S. industry controls that varies by NAICS and time. Nevertheless, we follow Antràs dataset and use only average cross-industry variables. These averages capture the type of inputs imported by the headquarter. U.S. import tariffs correspond to applied tariffs from the World Integrated Trade Solution (WITS) database, averaged across exporting countries and over all years in For additional information on these data see section 4.3 in Antràs and Chor (2013) and Chapter 5 in Antràs (2016). The measure of IPR protection is taken from Park (2008a), who built an IPR index ranging between 0 and 5 for both the origin and the destination country. This measure of IPR updates the index of patent protection published in Ginarte and Park (1997) and is available for 122 countries for (it is calculated in periods of 5 years). Additional country level controls, like rule of law or being a WTO member come from World Development Indicators (WB) and CEPII dataset. To isolate the intermediate input component of U.S. imports, we combine U.S. Related Party Trade with BEC classification using the concordance provided by Pierce and Schott (2009). 5.2 Empirical Specification In this section we exploit the exogenous variation in IPR across country to analyze the organizational decisions of U.S. multinational firm. We will then try to explain the intrafirm import share using the following equation: S ict = β 1 + β 2 IP R ct + β 3 X it + β 4 Z ct + µ c + γ t + ɛ ict (23) where S ict is the share of intrafirm import from country c at time t, IP R ct is the IPR index at the country level. X it are a set of U.S. industry controls to headquarter intensity. Z ct are destination country controls. The specification also controls for country and year fixed effect, µ c and γ t. We cluster at the product level to avoid the artificially small standard errors. Since 10

11 our model predicts that IPR should encourage intra-firm import, our coefficient of interest is β Results The results are in line with the model predictions: higher IPR foster intra-firm import. This is particularly true for intermediates Results on all samples. In this section we estimate the relationship between U.S. intrafirm imports and IPR on all NAICS products. Table 1: IPR and Intrafirm Trade Shares Dependent variable Intrafirm Import Share (1) (2) (3) (4) IPR 0.019*** 0.021*** 0.020*** 0.019*** (0.006) (0.006) (0.007) (0.007) rule of law *** *** *** (0.011) (0.011) (0.011) GATT/WTO member (0.008) (0.008) log R&D 0.035*** (0.004) log skill (0.015) Country FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Observations 39,744 39,744 39,744 39,744 R-squared Notes: The regressions are OLS estimations of equation (23). The dependent variable is the U.S. intrafirm import share in industry i in year t. Only manufacturing sectors. Heteroskedasticity-robust standard errors clustered by sectors are reported in parentheses.,, and indicate significance at the 1, 5 and 10 percent levels respectively. 11

12 5.3.2 Results with interaction. To account for intermediate input we use the BEC classification. Since BEC uses HS6 classification, we use the concordance between NAICS industries and U.S. Harmonized system codes provided by Pierce and Schott (2009). Table 2: IPR and Intrafirm Trade Shares (Interactions) Dependent variable Intrafirm Import Share (1) (2) (3) (4) IPR (0.009) (0.009) (0.009) (0.009) interm*ipr 0.019*** 0.019*** 0.019*** 0.011*** (0.004) (0.004) (0.004) (0.004) final*ipr (0.004) (0.004) (0.004) (0.004) rule of law *** *** *** (0.012) (0.012) (0.012) GATT/WTO member (0.011) (0.011) log R&D 0.031*** (0.001) log skill 0.022*** (0.005) Country FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Observations 39,744 39,744 39,744 39,744 R-squared Notes: The regressions are OLS estimations of equation (23). The dependent variable is the U.S. intrafirm import share in industry i in year t. Only manufacturing sectors. Heteroskedasticity-robust standard errors clustered by sectors are reported in parentheses.,, and indicate significance at the 1, 5 and 10 percent levels respectively. 12

13 5.3.3 Results only for intermediates. Finally, in this section we restrict the analysis to intermediates only. Table 3: IPR and Intrafirm Trade Shares (Intermediates) Dependent variable Intrafirm Import Share (1) (2) (3) (4) IPR 0.024*** 0.025*** 0.026*** 0.025** (0.009) (0.009) (0.010) (0.010) rule of law (0.016) (0.016) (0.016) GATT/WTO member (0.014) (0.013) log R&D 0.032*** (0.002) log skill 0.030*** (0.006) Country FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Observations 28,037 28,037 28,037 28,037 R-squared Notes: The regressions are OLS estimations of equation (23). The dependent variable is the U.S. intrafirm import share in industry i in year t. Only manufacturing sectors. Heteroskedasticity-robust standard errors clustered by sectors are reported in parentheses.,, and indicate significance at the 1, 5 and 10 percent levels respectively. 13

14 6 Conclusion... 14

15 References Antràs, P., Firms, Contracts, and Trade Structure. Quarterly Journal of Economics, 118(4), Antràs, P., Global Production: Firms, Contracts, and Trade Structure. Princeton University Press, 336 pp. Antràs, P., Grossman-Hart (1986) Goes Global: Incomplete Contracts, Property Rights, and the International Organization of Production. Journal of Law, Economics and Organization, 30(1), Antràs, P., Helpman, E., Global Sourcing. Journal of Political Economy, 112(3), Antràs, P., Chor, D., Organizing the Global Value Chain. Econometrica, 81(6), Antràs, P., Yeaple, S., Multinational Firms and the Structure of International Trade. Handbook of International Economics, 4, Bertrand, M., Duflo, E., Mullainathan, S., How Much Should We Trust Differences-in- Differences Estimates?. The Quarterly Journal of Economics, 119(1), Branstetter, L., Saggi, K., Intellectual property rights, foreign direct investment and industrial development. The Economic Journal, 121(55), Helpman, E., Melitz, M.J., Yeaple, S.R., Exporting and FDI with Heterogeneous Firms. American Economic Review, 94(1), Keller, W., Yeaple, S.R., Gravity in the Knowledge Economy. American Economic Review, forthcoming, 103(4), 55. Maskus, K. E., Fink, C., Intellectual property and development: lessons from recent economic research. World Bank. Maskus, K. E., Penubarti, M., How trade-related are intellectual property rights?. Journal of International economics. 39(3),

16 Melitz, M.J., The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity. Econometrica, 71(6), Nunn, N., Trefler, D., Incomplete Contracts and the Boundaries of the Multinational Firm. Journal of Economic Behavior & Organization, 94(1), Pierce, J.R., Schott, P., A Concordance Between Ten-Digit U.S. Harmonized System Codes and SIC/NAICS Product Classes and Industries. NBER Working Paper, Ottaviano, G., Turrini, A., Distance and foreign direct investment when contracts are incomplete. Journal of the European Economic Association, 5(4), Smith, Pamela J., Are weak patent rights a barrier to US exports?. Journal of International Economics, 48(1), Yang, G., and Maskus, K. E., Intellectual property rights, licensing, and innovation in an endogenous product-cycle model. Journal of International Economics, 53(1),