Lecture notes: James R. Markusen Department of Economics University of Colorado Boulder CO 80309-0256 james.markusen@colorado.edu Multinational Firms 1. Review of empirical evidence 2. Dunning's OLI, joint inputs, firm versus plant-level scale economies 3. A model with endogenous multinationals 4. Pattern of trade in goods and services 5. Motives for internalization 6. A model of internalization MULTINATIONAL FIRMS AND THE NEW TRADE THEORY Firm and Industry Characteristics (1) Multinationals are associated with high ratios of R&D relative to sales. (2) Multinationals employ large numbers of scientific, technical, and other "white collar" workers as a percentages of their work forces. (3) Multinationals tend to have a high value of "intangible assets"; roughly, market value minus the value of tangible assets such as plant and equipment. (4) Multinationals are associated with new and/or technically complex products. (5) Evidence suggests that multinationality is negatively associated with plant-level scale economies. (6) Multinationals are associated with product-differentiation variables, such as advertising to sales ratios. (7) A minimum or "threshold" level of firm size seems to be important for a firm to be a multinational, but above that level firm size is of minimal importance. (8) Multinationals tend to be older, more established firms. Country Characteristics (1) The high-income developed countries are not only the major source of direct investment, they are also the major recipients. Most direct investment seems to be horizontal. (2) There has been a major boom of direct investment into the developing countries in the 1990s, but most of it has gone to the more advanced LCDs and to China. Little goes to the least developed countries. (3) Direct investment stocks have grown significantly faster than trade flows over the last two decades, even though trade barriers have fallen dramatically. (4) High volumes of direct investment are associated with similarities among countries in terms of relative factor endowments and per capita incomes, not differences. (5) A high volume of outward direct investment is positively related to a country's endowment of skilled labor and insignificantly or negatively related to its physical capital endowment. (6) There is little evidence that direct investment is primarily motivated by tariff avoidance or measurable transport costs, (7) There is mixed evidence that tax avoidance and/or risk diversification are important motives for direct investment. Some evidence does suggest that political risk discourages inward investment. (8) Infrastructure, skill levels, and a minimum threshold level of per capita income seem to be very important determinants of direct investment. (9) There is evidence that agglomeration effects are important in direct investment. But it is admittedly difficult to distinguish agglomeration effects from firms being drawn to the same (unobserved) site-specific resources. Table 1 Annual growth rate (%), all countries 1986-1990 1991-1996 FDI inflows FDI stocks Sales of foreign affiliates Royalties and fee receipts 24.7 19.8 17.4 21.8 14.4 11.4 6.6 12.9 GDP at factor cost Exports of goods and non-factor services 10.8 14.3 6.4 7.4 Table 2 FDI inflows and outflow, share in total Year Developed In out Developing in out CEE in out 1983-1987 1988-1992 1993 1994 1995 76 78 62 59 65 24 21 35 39 32 0 1 3 3 4 95 93 85 83 85 5 7 15 17 15 0 0 0 0 0 Share of inward FDI to/from developing countries going to: China Least developed inward (48 countries) inward LDC share of Hong Kong 1984-89 1990-92 1993 1994 1995 10 15 38 39 38 2 2 2 1 1 world GDP inward 4 5 4 4 4 outward 24 28 54 56 51 UNCTAD world investment report, 1996. Zhang and Markusen, 1996. Correlation between Foreign Ownership Shares and Characteristics of Industries (Andrew Dick, JIE 1993) (1982 and 1987 data) Variable Pearson Correlation Coefficient Net Income divided by Assets (proxy for intangible assets) 0.357 Advertising Intensity 0.283 % of Total Employees in R&D 0.228 Median Plant Shipments/ Industry Average Shipments (economies of scale?) 0.610 Capital to Value Added Ratio 0.175 Patent Effectiveness process R&D product R&D 0.814 0.574 Research Lead Time process R&D product R&D 0.785 0.615 Point of Departure for Theory: Firms incur significant costs of doing business abroad relative to domestic firms in those countries. Therefore, for a firm to become a multinational, it must have offsetting advantages. Dunning (OLI): There are three necessary conditions for firms to be willing to undertake investments abroad Ownership Advantage: the firm must have a product or a production process such that the firm enjoys some market power advantage in foreign markets. Location Advantage: the firm must have a reason to want to locate production abroad rather than concentrate it in the home country, especially if there are scale economies at the plant level. Internalization Advantage: the firm must have a reason to want to exploit its ownership advantage internally, rather than license or sell its product/process to a foreign firm. Ownership Advantages, Firm-Specific Assets, and Knowledge Capital Multinationality related to R&D, marketing, scientific and technical workers, product newness and complexity, product differentiation. MNEs intensive in knowledge capital, knowledge-based assets (1) services of knowledge capital easily transported to distant plants (2) joint input or "public goods" nature of knowledge capital. Physical capital intensity by itself should not give rise to multinationality. What is being traded? Multinationals are exports of the services of knowledge-based assets: managerial and engineering services, financial services, reputations and trademarks. Location advantages. Horizontal multinationals producing the same goods and services in each location: Large markets and high trade costs. Vertical multinationals geographically fragmenting the production process by stages: factor-price differences across countries are linked to the factor intensities of different stages, low trade costs. Internalization advantages. The same joint-input, public-goods property of knowledge that makes it easily transferred to foreign locations makes it easily dissipated. Firms transfer knowledge internally in order to maintain the value of assets and prevent dissipation. Elements of a simple model. Firm's technology. F - firm-specific fixed costs G - plant-specific fixed costs c - constant marginal cost per unit t - shipping or tariff or other cost unit Multi-plant economies of scale. F is a joint-input across plants. Fixed costs of a two-plant firm is less than the fixed costs of a one-plant national firm. (F + 2G) < (F +G) + (F + G) = 2F + 2G. But, this does not immediately imply that all firms will be multinationals. That will depend on the size of G, the fixed costs of a second plant, versus t, the unit trade cost. A vertical dimension is added by allowing F, G, and c to have different factor intensities; e.g., skilled-labor intensity is higher for F than for G which is in turn higher than for c. Markusen and Venables, "Multinational Firms and the New Trade Theory", Journal of International Economics, 1998 1. Two Goods, X and Y 2. Y - competitive, constant returns to scale 3 Two Factors, L and R. R specific to Y sector, L mobile between sectors 4. X is homogeneous, produced from labor only. Technology in X F G c ô - firm-specific fixed cost in units of labor - plant-specific fixed cost in units of labor - constant marginal cost in units of labor - unit shipping cost in units of labor 5. X is imperfectly competitive, firms are Cournot competitors. Markets are segmented. 6. There are two countries, country h and country f. Factors L and R are immobile between countries. 7. There are four firm types, and there is free entry and exit into and out of firm types. Type mh - multinational firms which maintain plants in both countries, firms-specific capital is located in country h. Type mf - multinational firms which maintain plants in both countries, firms-specific capital is located in country f. Type nh - national firms that maintain a single plant in country h. Type h firms may or may not export to country f. Type nf - national firms that maintain a single plant in country f. Type f firms may or may not export to country h. We will model this as a complementarity problem; that is, at set of inequalities with associated non-negative variables. But it is analogous to a game where there is a continuum of players. A firm decides whether or not to enter, and if entry, what type. After entry firms play a Cournot output game. An equilibrium is charaterize by no active or inactive firm wishing to change is strategy. Production function for Y á 1&á Yi ' Liy R i i ' h, f. , (1) Factor prices wi ' á ( Liy /R i )á & 1 , ri ' (1 & á) ( Liy /Ri )á i ' h, f. (2) Labor demand in country i by type-ni firms n n cXii % (c % ô)Xij % G % F, i, j ' h, f, i … j. (3) Labor demand in country i by type-mi firms m c X ii % G % F, i ' h, f. (4) Labor demand in country i by type-mj firms m c X ij % G, i, j ' h, f. (5) Labor market adding up condition n n L̄i ' Liy % ni ( cXii % ( c % ô) Xij % G % F ) % m mi ( c Xii % G % F) % m mj ( cXji (6) % G) Income Mi ' wi L i % ri Ri i ' h, f. (7) Utility â 1& â Ui ' Xic Yic , n n m m Xic ' ni Xii % nj Xji % m i X ii % m j X ji (8) Y ic ' (1&â)Mi. (9) Demand Xic ' â M i /pi, Price equals marginal cost conditions n n pi ( 1 & eii ) # wi c (10) ( Xii ) n n pj ( 1 & eij ) # wi (c % ô) m pi ( 1 & eii ) # wi c m pj ( 1 & eij ) # wj c ( Xij ) (11) ( Xii ) m (12) m (13) ( Xij ) Cournot markup formulas k k eij ' X ij k ' Xjc pj X ij k ' n, m, âMj i, j ' h, f. (14) "Supply" functions n Xii $ â M i n Xij $ â M j m Xii m $ â Mi Xij $ â M j p i & wi c 2 pi pj & wi (c % ô) 2 pj p i & w ic 2 pi p j & w jc 2 pj (15) . , (16) (17) (18) Free entry conditions n n n n ( n h) (19) n n n n ( nf ) (20) m m m m m m m m ph ehh Xhh % pf ehf Xhf # wh ( G % F ) pf eff Xff % ph efh Xfh # wf ( G % F ) ph ehh X hh % pf ehf X hf # wh ( G % F ) % wf G (21) ( mh ) pf eff X ff % ph efh X fh # wf ( G % F ) % wh G (22) ( mf ) If outputs are positive, then using (14)-(18), these free entry conditions can be expressed as: â Mh â Mh â Mh â Mh p h & wh c 2 % Mf ph ph & wf (c % ô) ph p h & wh c 2 % Mf ph p h & wh c ph 2 % Mf pf & wh (c % ô) 2 pf 2 % Mf pf &wf c p f & wf c pf pf ( nh ) (23) 2 # wf ( G % F ) , ( nf ) (24) 2 pf p f & wf c # wh ( G % F ), # w h ( G % F ) % wf G , ( mh ) (25) # wf ( G % F ) % wh G , ( mf ) (26) 2 These can be expressed more simply as (a > b): n (27) n (28) m (29) m (30) Ðh ' a h M h % b f M f & d h # 0 Ðf ' b h M h % a f M f & d f # 0 Ðh ' a h M h % a f M f & d h & e f # 0 Ðf ' a h M h % a f M f & d f & e h # 0 Change in Total Income: dMh = dMf > 0. m m n n d Ðh ' d Ðf > d Ðh ' d Ðf $ 0 Change in the Distribution of Income: dMh = - dMf > 0 n m m n d Ðh > d Ðh ' d Ðf ' 0 > d Ðf Change in w: dwf = - dwh> 0 n m m n d Ðh > d Ðh > 0 > d Ðf > d Ðf Change in Firm versus Plant Cost Ratio: dF = - dG > 0 m m n n d Ðh ' d Ðf > 0 ' d Ðh ' d Ðf Change in Transport Costs: dô > 0 m m n n d Ðh ' d Ðf ' 0 > d Ðh ' d Ðf Now let us summarize these results. Multinational firms will have an advantage relative to type-nh and/or nf firms when: (1) The overall market is large. (2) The markets are of similar size. (3) Labor costs are similar. (4) Firm-level scale economies are large relative to plant-level scale economies. (The added fixed costs of becoming a multinational firm are low.) (5) Transport costs are high. Figures 1, 2, 3, 4 Extension of the model to six firm types. This permits “vertical” firms which have a headquarters and single plant in different countries. Characteristics of US transnational Corporations in Manufacturing, 1989 R&D expenditures as a percentage of Sales Parents Affiliates Developed countries Developing countries 3.33 1.12 1.27 0.30 R&D employment as a percentage of total employment Parents 5.46 Affiliates 2.42 Skill level (compensation per employee $000) Parents Affiliates Developed countries Developing countries 38.9 25.2 33.3 9.5 Physical capital intensity ($000 per employee) Parents 47 Affiliates by industry of parent 31 by industry of affiliate 31 Assets per employee Parents Affiliates Developed countries Developing countries 186 114 147 52 Ratio of ratios (Slaughter) (Parents non-production workers/ parents production workers)/ (affiliates non-production workers/ affiliates production workers) 1.626 Table 1: Selected statistics, unweighted averages Skilled worker % of labor force Seven Developed Countries 26.0 Ratio of skilled to unskilled 1.81 All Developed Fifteen Developing Countries 11.0 All Developing Sources: 3.55 Inward FDI % of GDP Outward FDI % of GDP 13.4 16.5 9.1 11.5 19.6 5.2 15.4 4.5 UNCTAD world investment report, 1996 and 1997. GTAP data set, 1995. Developed countries are: USA, Canada, Japan, Denmark, Germany, Great Britain, Sweden. Developing countries are: Mexico, Korea, Singapore, Phillipines, Malaysia, Thailand, Indonesia, China, Countries, Brazil, Chile, Turkey, Venezuela, Columbia, South Africa, Sri Lanka Are MNEs predominantly horizontal or vertical? Characteristics of US Transnational Corporations in Manufacturing, 1989 Trade as a % of Sales Parents Total Exports 11.6 Exports to affiliates 5.3 Imports from affiliates 4.2 Affiliates Total Exports 37.8 Sales to US parents in the US Industry of affiliate Industry of parent 11.2 9.0 Imports from parents Industry of affiliate Industry of parent 13.2 12.0 Source: World Investment Report 1993, UNCTAD, page 64. Vertical six-firm-type model 1. Two Homogeneous Goods, X and Y; Two Countries, h and f. Two Factors, unskilled labor: L skilled labor: S. 2. Y - competitive, constant returns to scale, L intensive 3. X - imperfectly competitive, increasing returns to scale, S intensive overall. "headquarters and "plant" may be geographically separated. a firms may have plants in one or both countries 4. There are six firm types, with free entry and exit into and out of firm types. Regime denotes a set of firm types active in equilibrium. Type mh - horizontal multinationals which maintain plants in both countries, headquarters is located in country h. Type mf - horizontal multinationals which maintain plants in both countries, headquarters is located in country f. Type nh - national firms that maintain a single plant and headquarters in country h. Type h firms may or may not export to country f. Type nf - national firms that maintain a single plant and headquarters in country f. Type f firms may or may not export to country h. Type vh - vertical multinationals that maintain a single plant in country f, headquarters in country h. Type vh firms may or may not export to country h. Type vf - vertical multinationals that maintain a single plant in country h, headquarters in country h. Type vf firms may or may not export to country f. Factor-intensity assumptions: ranked from most skilled-labor-intensive to least skilled-labor-intensive Activities [headquarters only] > [integrated X] > [plant only] > [Y] Firm Types [type-m firms] > [type-v and type-n firms] Operations of a given firm within a Country [local type-v firm] > [local type-m firm] > [local type-n firm] > [plant of foreign type-m or type-v firm] > [Y] What is Traded: X, Y, and headquarters' services, the latter produced with skilled labor. Factors are not directly tradable. What is Technology Transfer: Headquarters' services, "blueprints". They may be complements or substitutes for host-country skilled labor. This depends on whether branch plants draw skilled labor from the host-country Y sector or an existing X sector. Why is Investment Liberalization different from Trade Liberalization: Imports of X embody technology, but they can only be substitutes for host-country X production, and thus substitutes for substitutes for hostcountry skilled labor. Determininants of the equilibrium regime Total Market Size: Type-m firms have an advantage when the total market size is large. Trade costs: Type-m firms have an advantage when ô is large. Firm versus plant level scale economies: Type-m firms have an advantage when firm-level costs are large relative to plant-level costs. Differences in Market Size: Type-m firms have a disadvantage when the markets are very different in size. A type-n firm headquartered in the large country or a type-v firm in the other country has the advantage. Differences in factor prices: Type-m firms and type-n firms have a disadvantage when factor prices are very different across countries. Type-v firms headquartered in the skilled-labor-abundant country with their plant in the unskilled-labor-abundant country have the advantage. Figures 5, 6, 7 Specific Case: country f small, unskilled-labor abundant. focus on country f Motivated by: Mexico liberalizing with the US Eastern Europe liberalizing with the EU Parameterization: Country h three times as large (S/L)h > 2.5*(S/L)f HP IL TL FF - high pportection: high trade costs, no FDI allowed investment liberalization: high trade costs, yes FDI trade liberalization: no trade costs, no FDI free, free: investment and trade liberalization Production Regime HP IL TL Types nh, nf Types mh, nh Type nh FF Types vh, nh Trade Regime of country f HP IL TL FF* export Y import X export Y import X import S export Y import X import Y export X import S Results for a Specific Case: Variable, country f HP IL TL FF Welfare 0.90 0.94 0.99 1.00 X production 0.16 0.49 0.00 1.00 z 0.87 1.79 0.65 1.00 w 0.90 0.92 1.00 1.00 z/w 0.97 1.95 0.64 1.00 zf/zh 0.91 1.75 0.65 1.00 wf/wh 0.91 0.92 1.00 1.00 Variable, country h HP IL TL FF Welfare 1.00 0.99 1.00 1.00 X production 1.26 1.14 1.39 1.00 z 0.96 0.99 0.97 1.00 w 1.00 0.99 1.00 1.00 z/w 0.96 1.00 0.97 1.00 Country f, Policy Criterion Welfare FF > TL > IL > HP X production FF > IL > HP > TL z IL > FF > HP > TL w FF > TL > IL > HP z/w IL > FF HP > TL zf/zh FF = TL > IL > HP wf/wh IL > FF HP > TL > > General comment Trade liberalization and investment liberalization are not substitutes. 1. Trade liberalization and investment liberalization have opposite effects on some variables such as factor prices. 2. Neither trade liberalization nor investment liberalization along exhausts gains from trade (in the sense of achieving FPE). 3. For a small, skilled-labor-scarce country, trade and investment liberalization are "complements", in a welfare sense and often in a volume-of-trade sense. Empirical Results (Carr, Markusen, Maskus, AER 2000) RSALES Production by affiliates of i parents in country SUMGPD Sum if country i's plus country j's GDP GPDDIFF Difference between country i and country j's GPD. GDPDIFSQ Sq diff between country i and country j's GDP. SKDIFF Skilled labor abundance of country i relative to j. SKDIFSQ SKDIFF squared (GDPDIFF*SKDIFF) INVCJ Cost of investing in country j TCJ Cost of exporting to country j (TCJ*SKDIFSQ) TCI Cost of exporting to country i RSALES = Other Variables B1(SUMGDP) (+) B2(GDPDIFSQ) (-) B3(SKDIFF) (+) B4(GDPDIFF*SKDIFF) (-) B5(INVJ) (-) B6(TCJ) (+) B7(TCJ)(SKDIFSQ) (-) B8(TCI) (-) DIST Distance from i to j Table 1. Results for Real Sales Volume of Affiliates: WLS and Tobit estmination Variable WLS Estimate (t stat) Sign as Predicted? (marginal significance) Tobit Estimate (Chi-sq) Sign as Predicted? (marginal significance) SUMGDP 13.92 (9.80) Yes (0.0001) 15.04 (105.5) Yes (0.0001) GDPDIFSQ -0.0014 (-8.90) Yes (0.0001) -0.0010 (34.67) Yes (0.0001) SKDIFF 31044 (4.01) Yes (0.0001) 61700 (52.98) Yes (0.0001) GDPDIFF *SKDIFF -4.27 (-2.12) Yes (0.035) -10.20 (18.81) Yes (0.0001) INVCJ -455.6 (-3.92) Yes (0.0001) -378.6 (7.98) Yes (0.005) TCJ 190.6 (2.20) Yes (0.029) 156.2 (2.28) Yes (0.131) TCJ*SKDIFSQ -569.9 (-0.41) Yes (0.683) -1264 Yes (0.57) (0.573) TCI -93.3 (-1.14) Yes (0.256) -122.0 (2.13) Yes (0.144) DIST -1.34 (-6.63) ? (0.0001) -1.48 (41.92) ? (0.0001) INTERCEPT -5381 (-0.42) (0.676) -23282 (2.59) (0.108) Observations Adjusted R2 Log Likelihood 509 628 0.60 -5755 Result 1: An increase in the host-country's trade costs will raise production by affiliates of parent country firms. Result 2: A bilateral-increase in parent and host-country trade costs: (A) Weighted least squares: decreases affiliate production, so trade and investment are "complements". (B) Tobit: generally decreases affiliate production when the non-US country is a developing country ("complements") but increases affiliate production when the non-US country is another high-income country ("substitutes"). Result 3: A convergence in income (GDP) between the United States and country j (holding the sum their incomes constant) increases affiliate sales in both directions. Result 4: (A) When the US is parent, an increase in country j's skilledlabor abundance increases US affiliate production in country j (production by US affiliates is attracted to skilled-labor abundant host countries). (B) When the US is host, an increase in country j's skilledlabor abundance increases country j's affiliate production in the US. Result 5: Affiliate production is income elastic: a bilateral increase in parent and host country incomes increases affiliate production by a greater proportion. Markusen and Maskus 2000b RSALESL RSALESE Local sales of affiliates of i parents in country j Export sales of affiliates of i parents in country j RSALESUS RSALESF Export sales back to the US (US outward sample only) Export sales to third countries (US outward sample only) SUMGPD GDPDIFF GDPDIFSQ Sum if country i's plus country j's GDP Difference between country i and country j's GPD. Squared difference between country i and country j's GDP. Skilled labor abundance of country i relative to j. SKDIFF (GDPDIFF*SKDIFF) INVCJ TCJ TCI Cost of investing in country j Cost of exporting to country j Cost of exporting to country i RSALES = B1(SUMGDP) (+) (> for local) B2(GDPDIFSQ) (-) B3(SKDIFF) (+) B4(GDPDIFF*SKDIFF) (-) B5(INVCJ) (-) (> for export) B6(TCJ) (+) (- for export) B7(TCI) (-) (> for export) Other Variables DIST Distance from i to j Table 2 Results 1. The theoretical model receives support in the empirical estimation. Both horizontal and vertical motives for foreign production seem to be present in the data. 2. Empirical results confirm the hypothesis that the local market size is more important for production for local sales than for export sales. 3. Empirical results confirm the hypothesis that production shifts relatively in favor of production for local sales when the host is more skilled-labor abundant and relatively in favor of exports when the host is skilled-labor scarce. 4. Empirical results confirm the hypothesis that the ratio of production for export to production for local sales is negatively related to host-country investment and trade costs. 5. There is a quantitative difference between the two-way and USoutward-only samples with respect to the host-country skilledlabor share. For the two-way sample, production for export is attracted by host-country unskilled-labor abundance. A hypothesis that US affiliate production for export is attracted by host-country unskilled-labor abundance is not supported. Markusen and Maskus 1999 Explicitly runs a horse face between three models: An unrestricted model and two restricted versions. KK model (KK1) There are firm-level as well as plant-level scale economies (KK2) A single-plant firm may geographically separate headquarters and plant (KK3) Firm-level fixed costs are skilled-labor intensive relative to plant-level fixed costs and the marginal costs of production. VER model (VER1) There are no firm-level scale economies (KK2) A single-plant firm may geographically separate headquarters and plant (KK3) Firm-level fixed costs are skilled-labor intensive relative to plant-level fixed costs and the marginal costs of production. HOR model (KK1) There are firm-level as well as plant-level scale economies (HOR2) A single-plant firm may not geographically separate headquarters and plant (HOR3) Firm-level fixed costs, plant-level fixed costs and the marginal costs of production all use factors in the same proportion. KK model (1) Both type-H and type-V multinationals can exist. (2) Multinational are important when countries are similar in size and in relative endowments, and trade costs are moderate to high (type-H multinationals). (3) Multinationals are important when countries differ in relative endowments, particularly if the skilled-labor abundant country is small. VER model (1) Only type-V multinationals can exist. (2) Multinationals are important when countries differ in relative endowments. (3) Multinationals do not arise between identical countries. HOR model (1) Only type-H multinationals can exist. (2) Multinational are important when countries are similar in size and in relative endowments, and trade costs are moderate to high. Figures 8, 9 Tables 3, 4, 5 Results The econometric results give strong support to the HOR model, and strongly reject the VER model. The coefficient estimates in the HOR model have the right signs and high statistical significance. An F-test cannot reject the HOR model's restrictions. These results tell us what researchers have long believed from casual empiricism: direct investment is most important between countries that are similar both in size and in relative endowments. It is the "hill" of Figures 1c and 2c rather than the "valley" of Figures 1b and 2b that best describes the world. The VER model clearly should not be taken seriously as a general description of the world. Internalization General Idea: some of the same properties of knowledge capital that create ownership advantages create internalization advantages. These arise from the jointness property of knowledge along with moral hazard, asymmetric information, and the infeasibility of complete and/or enforceable contracts. Some internalization models involving the stylized facts on knowledge capital, product newness and complexity. (1) A firm is reluctant to reveal its product or process to a licensee, who may reject the proposal, but now has the knowledge. But the potential licensee is not going to sign an agreement without knowing what it is buying. (2) The licensee knows that the firm may not have an incentive to truthfully reveal the product's quality. (3) The newness of the product may create an informational asymmetry in the opposite direction: the potential licensee may have a much better idea of how the product will sell in its local market, while the MNE does not. The licensee extracts rent to reveal the information. (4) Bi-lateral uncertainty over stat-up problems, worker productivity and learning rates. (5) Knowledge is easily learned by new employees. The licensee may be able to defect, starting a new firm in competition with the MNE. (6) Product quality is an intangible asset. A licensee may have an incentive to reduce quality, capturing a short-run gain at the expense of losing the contract. (7) Difficulties in choosing between costly monitoring and suffering the costs of moral hazard when employing licensees. (8) Parties must make relation-specific investments (implies investments are sunk and cannot be used for other uses). (9) Differences in objectives and goals between the firm and the licensee. Elements of the Model (Markusen, JIE 2000) (1) The MNE introduces (or attempts to introduce) a new product every second time periods. Two periods are referred to as a "product cycle". A product is economically obsolete at the end of the second period (end of the product cycle). (2) The probability of the MNE successfully developing a new product in the next cycle is 1/(1+r) if there is a product in the current cycle, zero otherwise (i.e., once the firm fails to develop a new product, it is out of the game). The probability of having a product in the third cycle is 1/(1+r)2 etc. Ignore discounting.1 (3) The MNE can serve a foreign market by exporting, or by creating a subsidiary to produce in the foreign market. (4) Because of the costs of exporting, producing in the foreign country generates the most potential rents. (5) But any local manager learns the technology in the first period of a cycle and can quit (defect) to start a rival firm in the second period. Similarly, the MNE can defect, dismissing the manager and hiring a new one in the second period. The (defecting) manager can only imitate, not innovate and compete in the next product cycle. 1 Alternatively, r is the discount rate between product cycles, with discounting within between periods of a cycle ignored (not elegant I know, but it avoid very messy present value expressions). (6) Initially, no binding contracts can be written to prevent either partner from undertaking such a defection. (7) Initially, I will assume that the MNE either offers a selfenforcing contract or exports. The possibility that defection occurs as an equilibrium is allowed later in the paper. (8) Notation is as follows. R- Total per period licensing rents from the foreign country. E- Total per period exporting rents (E < R). F- Fixed cost of transferring the technology to a foreign partner. These include physical capital costs, training of the local manager, etc. T- Training costs of a new manager that the MNE incurs if it dismisses the first one (i.e., if the MNE defects). G- Fixed cost that the manager must incur if he/she defects. This could include costs of physical capital, etc. Li- Licensing or royalty fee charged to the subsidiary in period i (i = 1,2). V- Rents earned by the manager in one product cycle: V = (R-L1) + (R-L2). V/r- Present value of rents to the manager of maintaining the relationship. The manager ("a" for agent) has an "individual rationality" constraint (IR): the manager must earn non-negative rents. The manager also has an incentive-compatibility constraint: the manager must not want to defect in the second period. (1) (R - L1) + (R - L2) $ 0 IRa (2) (R- L2) + V/r $ (R - G) ICa where V = (R- L1) + (R - L2) V/r is the present value to the manager of the future rents, if there are any. (R - G) is the payoff to unilaterally defecting. The MNE similarly has an "individual rationality" constraint (IR): the MNE must earn non-negative rents. The MNE also has an "incentive-compatibility constraint: the MNE must not want to defect (fire the manager) in the second period. (3) L1 + L2 - F $ 2E IRm (4) L2 $ R - T Icm Combine the IC constraints. (5) R - T # L2 # G + V/r Firm's objective is to minimize V subject to this incentive compatibility. Solving this problem yields: (6) 2R - L1 - L2 = V = r(R - T - G) > 0 (rent share to the manager) Result 1: If R # G + T, the MNE captures all rents in a product cycle, henceforth referred to as a rent-capture (RC) contract. This situation occurs when (1) The market is relatively small. (2) Defection costs for the MNE (T) are high. (3) Defection costs for the manager (G) are high. If R > T + G, there is no single-product fee schedule that will not cause one party to defect. Now consider the case where the manager's IR constraint does not hold; that is, the MNE shares rents with the manager. Combine (2) and (3). (7) R - T # L2 # G + V/r It is clear that the multinational should minimize V. V is therefore chosen such that (7) holds with equalities, L2 = R - T. Write (7) as an equality and substitute for V. (8) R - T - G = V/r = (2R - L1 - L2)/r = (2R - L1 - (R - T))/r (9) r(R - T - G) = 2R - R + T - L1 (10) L1 = R + T - r(R - T - G) Adding together L1 and L2 = R - T and subtracting F, we get the earnings of the MNE (11) L1 + L2 - F = 2R - F - r(R - T - G) > 0 The MNE has to share some rents with the subsidiary. The subsidiary earns (12) 2R - L1 - L2 = V = r(R - T - G) > 0. The subsidiary earns positive rents, and no potential rents are dissipated. Result 2: If R > G + T, the MNE can credibly offer a long-term commitment, but must share rents with the subsidiary. This is henceforth referred to as a rent-sharing (RS) contract. The one-period rents earned by the subsidiary are smaller as (1) (2) (3) (4) r is small (future rents are more valuable) G is large (the incentive to defect is smaller) T is larger (the MNEs incentive to defect is smaller). R is larger (the subsidiary's share increase faster than R). The one-period rents earned by the MNE are larger when (1), (2), (3) and (4) hold. Increases in F reduce profits (but do not dissipate the rents 2R) and make the MNE worse off and leave the subsidiary unaffected. Increases in G shift rents toward the MNE (G is never incurred in equilibrium). The boundary between the area in which a rent-capture contract is possible and the area in which a rent-sharing contract is needed is given by SC in Figure 1. RC occurs to the right, RS to the left of SC. MNE chooses exporting over the RC contract if 2R - F < 2E. Indifference between exporting and the RC subsidiary is given by the horizontal line in Figure 1, F = 2R - 2E. We label this line EC. At point A in Figure 1, exporting and a RS subsidiary also yield the same profits (by transitivity). Beginning at this point, differentiate the right-hand side of (16) holding it equal to zero, in order to derive the locus of indifference between exporting and a subsidiary. (13) dF - rdG = 0 dF/dG = r The locus of points giving the same profits from a subsidiary as point A in Figure 1 is given by SE in Figure 1. This line intersects the vertical axis (G=0) at F = 2R - 2E - r(R-T). Result 3 Figures 10, 11 (1) Above SE and EC in Figure 1, the MNE chooses exporting (2) Below EC and right of SC, the MNE chooses a subsidiary and captures all rents. (3) Left of SC and below SE, the MNE chooses a subsidiary and shares rents. The exporting region of Figure 1 below EC and above (left of) SE is a region in which the MNE prefers to dissipate some rents in exporting rather than share rents with a subsidiary. Extension: Property rights and contract enforcement can be modeled by adding a defection penalty. Results shown in Figure 2. A few summary comments. The theory of foreign direct investment and multinational location decisions should be viewed as largely distinct from the theory of portfolio capital flows. The theory of foreign direct investment is based on the existence of firm-specific assets which give firms ownership advantages necessary to overcome the costs of doing business abroad. First-specific advantages are often located in knowledge-based assets that create firm-level scale economies (multi-plant economies of scale) that encourage horizontal expansion into foreign markets. Vertical geographic fragmentation of activities is encouraged by differences in factor intensities among stages of production, combined with differences in factor prices across countries.2 The same properties of knowledge-based assets that create ownership advantages, often create internalization advantage, making it profitable to maintain firm ownership across geographically dispersed plants rather than use arm’s-length markets. Note some confusion in terminology. The activities are vertically integrated within the firm by definition of being a MNE, but geographically disintegrated. 2 Figure 1: Equilibrium Regimes, τ = .15 Of .95 .90 .85 .80 Endowment of Resources .75 .70 .65 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 Oh .05 .10 .15 .20 .25 .30 .35 .40 .45 .50 .55 .60 .65 .70 .75 .80 .85 .90 .95 Endowment of Labor Multinational firms only Mixed regimes of national and multinational firms National firms only Figure 2: Compatative Statics Of .95 .90 .80 Countries Smaller: 40% Endowment of Resources .85 .75 .70 .65 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 Oh .05 .10 .15 .20 .25 .30 .35 .40 .45 .50 .55 .60 .65 .70 .75 .80 .85 .90 .95 Endowment of Labor Of .95 .90 Trade Costs Smaller: 10% Endowment of Resources .85 .80 .75 .70 .65 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 Oh .05 .10 .15 .20 .25 .30 .35 .40 .45 .50 .55 .60 .65 .70 .75 .80 .85 .90 .95 Endowment of Labor Of .95 MNE fixed costs higher: 1.75 Endowment of Resources .90 .85 .80 .75 .70 .65 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 Oh .05 .10 .15 .20 .25 .30 .35 .40 .45 .50 .55 .60 .65 Endowment of Labor Multinational firms only Mixed regimes of national and multinational firms National firms only .70 .75 .80 .85 .90 .95 0.25 0.25 0.20 0.20 0.15 0.15 0.10 0.10 0.05 0.05 0.00 0.00 -0.05 -0.05 -0.10 -0.10 -0.15 -0.15 -0.20 -0.20 -0.25 Production minus labor share in country h without MNEs (diamonds) Production minus labor share in country h with MNEs (squares) Figure 3: Effect of Multinational Entry on the Location of Production, Welfare, and the Volume of Trade: NW-SE Diagonal -0.25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 NW-SE Diagonal (countries differ in relative endowments) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 NW-SE Diagonal (countries differ in relative endowments) 18 19 Effect of multinational entry on the welfare of country h 1.20% 1.00% 0.80% 0.60% 0.40% 0.20% 0.00% 18 19 Volume of trade and affiliate sales 450 400 350 300 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 NW-SE Diagonal (countries differ in relative endowments) Volume of trade - NE model Volume of trade - MNE model Affiliate sales - MNE model Figure 4: Effect of Multinational Entry on the Location of Production, Welfare and the Volume of Trade: SW-NE Diagonal 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0.00 0.00 -0.02 -0.02 -0.04 -0.04 -0.06 -0.06 -0.08 -0.08 Effect of multinational entry on the welfare of country h 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% -0.50% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 200 450 180 400 160 350 140 300 120 250 100 200 80 150 60 40 100 20 50 0 Volume of affiliate sales Volume of trade in X SW-NE Diagonal (countries differ in size) 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 SW-NE Diagonal (countries differ in size) Voume of trade - NE model Volume of trade - MNE model Affiliate sales - MNE model Figure 5: Volume of Affliate Production: 25% trade costs Affiliate Production 500 450 400 350 300 250 200 150 7 0 3 Skilled Labor 50 11 19 15 100 1 2 3 4 5 6 7 8 9 15 16 17 18 19 10 11 12 13 14 Unskilled Labor Origin for Country h Figure 6: Affiliate Production by h-owned plants in country f Affiliate Production 500 450 400 350 300 250 200 150 19 100 15 50 11 0 7 3 Skilled Labor 1 Origin for Country h 2 3 4 5 6 7 8 9 18 19 15 16 17 14 13 10 11 12 Unskilled Labor Figure 7: Change in Affiliate Production, h plants in f: 25% trade costs minus 5% trade costs in both directions Change in Affiliate Production 200 150 100 50 0 -50 -100 -200 Skilled Labor 3 7 11 15 19 -150 1 2 Oriigin for Country h 3 4 5 6 7 8 9 15 16 17 18 19 10 11 12 13 14 Unskilled Labor Table 2: Effects of Host-Country Size and Skilled-Labor Abundance on Foreign Affiliate Production for Local Sale and Export (derviatives evaluated at the mean of independent variables) Effect on: $1 billon increase in country j's GDP One percentage point increase in SKLJ* One point increase in INVCJ One point increase in TCJ Local sales of country i affiiates in country j (inward and outward data) $16.7 million ε = 1.558 ** - $12.1 million ε = -0.017 - $633.4 mill $366.6 mill Export sales of country i affiliates in j to all countries (inward and outward data) $4.7 million ε = 1.118 - $189.0 million ε = -0.681 - $277.2 mill $52.1 mill Local sales of US affiliates in country j (US outward data only) $22.7 million ε = 1.044 $280.5 million ε = 0.599 - $517.8 mill $314.8 mill Export sales of US affliliates in j to the US (US outward data only) $1.3 million ε = 0.213 $8.3 million ε = 0.064 - $346.6 mill $207.6 mill - 211.3 mill - $8.6 mill Export sales of US affliliates in j to other (US outward data only) εw = 0.681 $9.0 million ε = 0.989 εw = 0.062 $12.0 million ε = 0.061 * By a one-percentage-point increase in SKLJ we mean, for example, an increase from 15% to 16%, not an increase from 15% to 15.15% ** ε denotes elasticity Figure 8a: Affliate Production, 25% trade costs, KK model 500 400 300 200 19 16 100 origin for country h 19 17 15 13 11 9 7 5 3 1 1 4 7 10 13 0 Skilled Labor Unskilled Labor Figure 8b: Affiliate Production, 25% Trade Costs, VER model 500 400 300 200 19 17 15 13 11 9 7 5 3 1 1 4 7 10 Skilled Labor 0 13 19 16 100 Unskilled Labor Figure 8c: Affilliate Production, 25% Trade Costs, HOR model 100 80 60 40 Unskilled Labor 19 17 15 13 11 9 7 5 3 1 1 4 7 10 Skilled Labor 0 13 19 16 20 Figure 9a: Affiliate production by h-owned plants in f: KK model 1 0.8 0.6 0.4 7 Unskilled Labor 1 4 10 Skilled Labor 0 13 19 16 0.2 origin for country h Figure 9b: Affiliate production by h-owned plants in f: VER model 500 400 300 200 19 17 15 13 11 9 7 5 3 1 1 4 7 Skilled Labor 0 10 19 16 13 100 Unskilled Labor Figure 9c: Affiliate production of h-owned plants in f: HOR model 70 60 50 40 30 20 Unskilled Labor 19 17 15 13 11 9 7 5 3 1 1 4 7 Skilled Labor 0 10 19 16 13 10 Table 3. Theoretical Predictions for Three Models ( i = parent, j = host) Variable KK HOR VER SUMGDP + + 0 GDPDIFSQ - - 0 D2*SKDGDPD - 0 0 D2*SKDSUMG + - + D1*SKDSUMG 0 - 0 DISTANCE ? ? ? INVCJ - - - TCJ + + + TCI - - - SUMGDP = GPDi + GDPj GDPDIFF = (GDPi - GDPj) SKDGDPD = SKDIFF*GDPDIFF = (SKi - SKj)*GDPDIFF SKDSUMG = SKDIFF*SUMGDP = (SKi - SKj)*SUMGDP D1 = -1 if SKDIFF = SKi - SKj < 0 =0 D2 GDPDIFSQ = (GDPi - GDPj)2 if SKDIFF = SKi - SKj > 0 = 1 if SKDIFF = SKi - SKj > 0 = 0 if SKDIFF = SKi - SKj < 0 D2 is non-zero if the parent country is skilled-labor abundant, D1 is non-zero if the host country is skilled-labor abundant Table 4. WLS Estimation Including Investment and Trade Costs (509 obs) KK HOR SUMGDP 15.04 (0.0001) Yes 15.01 (0.0001) Yes GDPDIFSQ -0.001 (0.0001) Yes -0.001 (0.0001) Yes VER D2SKDGDPD 5.39 (0.12) No D2SKDSUMG -6.85 (0.04) ? -2.94 (0.21) Yes D1SKDSUMG -12.83 (0.0001) Yes -12.53 (0.0001) Yes DISTANCE -1.31 (0.0001) ? -1.23 (0.0001) ? -2.21 (0.0001) ? -436.6 (0.005) Yes -367.9 (0.002) Yes -639.3 (0.0001) Yes TCJ 173.9 (0.03) Yes 149.9 (0.05) Yes 321.6 (0.001) Yes TCI -90.69 (0.26) Yes -85.7 (0.29) Yes -85.2 (0.40) Yes INTERCEPT -7597 (0.55) -10592 (0.40) 43337 (0.0001) 0.61 0.61 6.60 6.65 0.37 87.41 3.34 INVCJ Adjusted R2 F-Test Critical F 99% 3.79 (0.16) Yes Marginal significance levels are in parentheses: "Yes" and "No" indicate conformity to predictions. Variable Table 5. Tobit Estimation Including Investment and Trade Costs (628 obs) KK HOR SUMGDP 17.4 (0.0001) Yes 17.23 (0.0001) Yes GDPDIFSQ -0.001 (0.0001) Yes -0.001 (0.0001) Yes VER D2SKDGDPD 11.83 (0.003) No D2SKDSUMG -15.05 (0.0001) ? -6.93 (0.001) Yes D1SKDSUMG -24.51 (0.0001) Yes -23.97 (0.0001) Yes DISTANCE -1.48 (0.0001) ? -1.36 (0.0001) ? -2.63 (0.0001) ? -386.9 (0.008) Yes -227.2 (0.09) Yes -616.3 (0.0004) Yes TCJ 134.1 (0.15) Yes 76.3 (0.40) Yes 261.2 (0.03) Yes TCI -136.7 (0.10) Yes -125.1 (0.13) Yes -243.4 (0.02) Yes INTERCEPT -26271 (0.07) -32791 (0.02) 46477 (0.0001) -5747 -5751 8.0 7.88 -5933 372.0 14.86 INVCJ Log-Likelihood LR Test Critical Chisq 13.17 (0.0001) Yes Marginal significance levels are in parentheses: "Yes" and "No" indicate conformity to predictions. Variable Figure 10: Values of F and G supporting alternative Modes SC F SE E A 2R - 2E EC RC 2R - 2E - r(R-T-2P) RS R - T - 2P G Figure 11: Elimination of Contract Enforcement SC F SC' SE' E SE 2R - 2E EC RC to E RS to E 2R - 2E - r(R-T) RC RC to RS RS R - T - 2P R-T G