Vertical FDI versus Outsourcing: A Welfare Comparison from the Perspective of the Host Country Arti Grover [email protected] [email protected] Delhi School of Economics/Princeton University

Abstract: Internalization theory is most often viewed through the lens of a sourcing firm, especially so in a vertical relationship. In this paper we recognize the importance of the mode of organization of fragmented production for the host country. Whether production sharing arrangement is internalized by the parent firm in the form of a vertical foreign direct investment (VFDI) relationship or transacted externally through outsourcing contracts, affects the welfare of the host country in a significant way. Using a GrossmanHelpman quality ladders kind of set up, we compare the welfare derived from the two alternative regimes of production sharing arrangements in the host country. The ability to maximize welfare in the alternative regimes is found to be contingent on the absorptive capacity of the host country. If the host country's absorptive capacity is above a threshold level, outsourcing is more welfare enhancing vis-à-vis VFDI; while even with an absorptive capacity lower than this threshold, outsourcing being welfare improving over VFDI cannot be ruled out. Keywords: outsourcing, foreign direct investment, absorptive capacity JEL Classification: F12, F23

Acknowledgements: I am extremely grateful to Partha Sen for his guidance and motivation throughout the course of this project. The paper has benefited immensely from discussions with Gene Grossman, Arghya Ghosh, Mausami Das and Abhijit Banerji. I am also thankful to the participants of the Global Competitiveness conference at the Indian Institute of Management, Bangalore and the International Conference on Economic Growth and Development at the Indian Statistical Institute at Delhi for their useful comments.

I.

BACKGROUND

Offshoring is inevitable; firms no longer contemplate on whether to offshore or not, but strategize on the organization of offshore production. There has been a proliferation of papers addressing the tradeoff between vertical foreign direct investment (VFDI) and outsourcing from the firm’s perspective, see for example, Antràs (2003, 2005), Grossman and Helpman (2002, 2003, 2004). On one hand, offshoring raises the overall productivity (See for example, Amiti, and Wei, 2006, Görg and Hanley, 2005, Egger and Egger, 2006, and Calabrese and Erbetta, 2005) in the source country, expanding global production and thereby raising consumer welfare. On the other hand, offshore production benefits the host country through technology transfer, worker training, skill upgradation, job creation, and higher wages1. However, which mode of organization of fragmented production – intra-firm or outsourcing – is more welfare enhancing from the perspective of the host country is not known. In this paper, we attempt to explore this question. Internalization literature is silent on the welfare consequences of the organizational mode of vertical production transfer. Correspondingly, the literature that deals with the welfare impact of offshoring does not make a distinction between its different modes of organization. For example, Glass and Saggi (2001) focus on the welfare implications of international fragmentation in the home country but do not distinguish between the organizational forms. On the same note, none of the existing models which differentiate between VFDI and outsourcing discuss their effect on welfare in the host country. For example Antràs (2003, 2005), Grossman and Helpman (2002, 2003, and 2004) determine the tradeoff between VFDI and outsourcing but do not deal with their relative welfare implications on the host country2. Further, in all these models of vertical production transfer, the issue of internalization has always been viewed through the lens of the sourcing firm and the home country with limited attention to the host country3. The mode of organization of production sharing arrangement is crucial because it can potentially determine the skill formation, technology transferred, growth and welfare in the host country. Thus there is a pressing need to link the two strands of literature and specifically so from the host country’s perspective. This is what we intend to do in this paper. This paper compares the effect of VFDI with that of outsourcing on the host country factor prices and real Gross Domestic Product (GDP), which is our metric for welfare. We formally distinguish between VFDI4 and outsourcing in terms of their technique of production in a general equilibrium framework to contrast their effects on the host country welfare. Using evidence from existing literature, we hypothesize that an unaffiliated supplier employs a higher skill intensive technique of production vis-à-vis a subsidiary. Production is transferred to a low wage country either internally through a VFDI contract or externally through an outsourcing relationship in a Grossman-Helpman quality ladder framework. Assuming that “all pervasive” VFDI constitutes the initial equilibrium, we analyze the impact of a comparative static exercise of a change in production technique. Since the only difference between VFDI and outsourcing is found in its production methodology, an increase in skill intensity of production would therefore imply a regime switch from VFDI to outsourcing. Increase in skill intensity of production increases the relative demand for skilled labor and hence their relative payoffs. A regime switch from VFDI to outsourcing thus changes real GDP and welfare of the host country5. An important insight of our model is that the ability to maximize welfare from any regime of international production sharing is found to be contingent on the absorptive capacity of the host country. Outsourcing leads to higher welfare vis-à-vis VFDI if the absorptive capacity of the host country is higher than a critical threshold. However, even if the absorptive capacity is lower than this critical threshold, outsourcing being welfare improving over VFDI cannot be ruled out. The results are in conformity with our intuition. Outsourcing The existing literature on welfare effects of international fragmentation in the host country yields ambiguous results which shall be discussed in section II 2 Antràs (2003) model does present the impact of internalization on welfare of the home country but not the host country. Similarly, in the horizontal production transfer literature, there are many papers that compare the impact of horizontal FDI with licensing on welfare and growth but not so in the vertical production transfer. See for example, Saggi (1996, 1999), Glass and Saggi (2002). 3 Grossman and Helpman (2003) in some sense highlight that the legal framework of the host country influences the decision to outsource, but do not refer to welfare at all. 4 The term VFDI may sound a bit misleading since our model does not involve any capital formation. However, we stick to this term because it is commonly used to indicate internal production transfer. 5 In our model, the source country variables are exogenous, implying that we do not deal with why sourcing firms internalize but explore as to which form of organization is better for the host country. 1

1

being more skill-intensive creates greater demand for skilled labor thereby raising its skill premium. A high skill premium combined with a higher proportion of skilled labor (which is our measure of absorptive capacity), implies higher GDP with outsourcing vis-à-vis VFDI. On the other hand, a lower proportion of skilled labor in the host country, can still generate higher aggregate real income in the outsourcing regime depending on certain parameters. The result indicates a clear space for policy formulation. If the host country finds, say, outsourcing to have higher welfare impact relative to VFDI, then it can design policy incentives to bend the sourcing firms’ decision to offshore production through external contractors vis-à-vis internal transfer. For example, if higher intellectual property rights (IPR) enforcement encourages outsourcing, as suggested in Grossman and Helpman (2003), then, a tighter IPR regime is called for by the host government to encourage outsourcing. The paper beyond this point is organized in the following way: Section II discusses the related literature on VFDI and outsourcing. Section III describes the model for comparing the welfare impact of VFDI relative to outsourcing; section IV describes equilibrium in the host country and the conditions under which one form of foreign sourcing scores over the other in terms of welfare. Section V concludes the paper. II.

RELATED LITERATURE

This paper benefits from the existing literature on internalization theory and the welfare effects of international production sharing. Glass and Saggi (1999) and Balcão Reis (2001) model offshoring but are cautious of its widely cited economic prospects on the host country welfare. Glass and Saggi (1999) discuss the impact of production sharing in a general equilibrium model with oligopolistic industries. Offshoring shifts labor demand to a low wage host country, reducing the cost of production and thereby increasing the sourcing firm’s profits. This however comes at the cost of domestic (host country) firm’s profits because a rise in host country’s labor demand also pushes up domestic wages. Welfare in the host country may thus decline if the fall in the domestic entrepreneur’s profit is high. Balcão Reis (2001) considers a similar tradeoff between innovation and the quality available to the consumers due to offshoring. Production sharing with a low wage host country enhances the speed of product development for the sourcing firms and enables quicker access to higher quality products for all consumers but at the same time, it lowers the rate of innovation and profit of domestic entrepreneurs. Hence welfare may fall in the host country if the utility loss from latter is more than the benefit from higher quality consumption. Different approaches, viz., factor endowment (Grossman and Helpman, 2003, 2005, Antràs, 2003), contract theory (Antràs, 2003, 2005, Antràs and Helpman, 2004) and incentive system (Grossman and Helpman, 2004) - have been used to model the organizational forms of production sharing which have brought to fore many other differences between VFDI and outsourcing. The prevalence of the two alternative forms of organization is determined by the tradeoff between the governance cost of complex vertical organizations and the costs emerging from incomplete contracts due to the lack of ex-post verifiability with independent suppliers. Vertical integration has high costs fixed operational costs (Antràs and Helpman, 2004) due to more complex governance and limited specialization while outsourcing to unaffiliated suppliers have additional costs of search (Grossman and Helpman, 2002, 2003, 2005) and contractual incompleteness. Vertical integration with the offshore supplier is preferred if the independent suppliers in the host country are too little or more costly which makes search and matching probability low; product differentiation is high and contribution by the sourcing firm in the total value of the relationship is high. The share in the value of total surplus is positively related to the capital intensity of the offshored good (Antràs, 2003), and negatively related to its degree of standardization (Antràs, 2005). We analyze the production structure of the supplier and differentiate between the two organizational modes in terms of the offshored input’s production technique. The technique applied to produce an intermediate good by a sourcing firm’s subsidiary is different from that used by an unaffiliated supplier. We propose that the difference in production technique between a subsidiary and a third party vendor is due to the differences in skill intensity of production. We are inclined to assume that an unaffiliated supplier employs a higher skill-intensity in production vis-à-vis an affiliated supplier because of the following reasons. First, empirical evidence by Feenstra and Hanson (1996a, b), Sachs and Shatz (1994), Slaughter and Swagel (1997), Slaughter (2000) indicates that outsourcing leads to a greater divergence between the skilled and unskilled workers’ wages, while FDI does not. Greater divergence of skilled and unskilled labor factor prices implies that outsourcing creates higher demand for skilled labor. Similarly, using intra-firm US multinational’s data, Hanson et al. (2005) estimate that the demand for imported intermediate inputs for further processing in 2

affiliates is negatively related to lower-skill wages (absolute and relative) implying a greater usage of low-skill labor in affiliate operation. Our assumption can also be justified by the results obtained in Borga and Zeile (2004). They regress the volume of intra-firm trade on a number of parent firm related factors, host country characteristics and affiliate related variables. Affiliate R&D intensity (which is a good measure of its skill intensity) is found to be negatively related to the volume of intra-firm trade. This implies that the volume of intra-firm trade or the extent of VFDI falls when the skill intensity (or R&D intensity) of the affiliate rises. The second important result of Borga and Zeile that supports our paper’s conclusion (and hence justifies the use of our assumption) relates to the education standards of the host country and its income. Their results confirm that the volume of intra-firm trade falls if the host country has higher levels of education or income. As the absorptive capacity of the host country increases in our model, outsourcing certainly yields higher welfare enhancing vis-à-vis VFDI, which is what Borga and Zeile observe in their empirical analysis. Second, Mansfield and Romeo (1980) document that the technology transferred to a subsidiary is on an average 9.8 years old while that for an arm’s length transaction is 13.1 years old. Empirical evidence in Pack and Saggi (1997) also indicates that the best technologies are transferred internally through FDI while older or obsolete technologies are transferred through arm’s length transaction like licensing6. Assuming that skilled labor can substitute for technology7, an outsourcing partner will have to employ more skilled labor to match a subsidiary’s quality level8. A recent internalization model by Stähler (2007) assumes that in an internal production transfer, the supplier does not undertake the costs of training of host labor vis-à-vis an outsourcing relationship. His model supports our assumption that unskilled labor of the host country together with the technology of the multinational can be substituted by the skilled labor of the host country to produce the same fragmented piece of the final good. Therefore, it is clear in his model that to produce an input in the host country, a sourcing firm has two alternatives. If the firm wants to avoid spillovers it must transfer production internally by making a fixed investment, employing its own pool of skilled labor or technology and hiring the host country unskilled labor. On the other hand, if the firm does not fear spillovers, it can transfer production externally. Third, since technology closer to the frontier is transferred to a subsidiary, the desire to protect their technology from leaking to other firms may also induce them to hire less skilled workers because skilled workers can be potential carriers of the MNC industrial secrets. Since more recent technology is transferred to a subsidiary vis-à-vis an arm’s length agent, hence a multinational has a greater incentive to prevent dissipation of this technology from a subsidiary and hence lesser employment of skilled labor. We build a standard quality-ladder type model and initialize the equilibrium with all pervasive VFDI. From the VFDI equilibrium, we carry out a comparative static exercise of an increase in skill intensity of the supplier in the host country. This comparative static exercise is indicative of a regime switch from VFDI to international outsourcing because our distinction between VFDI and outsourcing entails only one exogenous change, that is, the change in the technique of production (via the exogenous change in skill intensity). Our model is different from Feenstra and Hanson (1996b) and Glass and Saggi (2001) as we compare the welfare and factor price effects of international outsourcing with that of VFDI while these two models measure wages and welfare Outsourcing may seem to share similarities with licensing but in essence it is different. First, outsourcing usually involves fragmentation of the production process while licensing does not. Outsourcing partners in the host country provide one of the inputs for the final good produced by the multinational company whereas, the licensee produces the entire product herself and directly sells it to the market. Second, in literature, licensee search has been modeled more like an auction (Casson and Buckley, 1981 and Casson, 1979) while searching for outsourcing partners have been modeled using probability theory (Grossman and Helpman, 2002). Third, an outsourcing partner contributes to an input in the final good produced by the MNC; therefore, appraisal and quality check of the partner’s product is inevitable. Fourth, if a licensor makes effort to internalize, assimilate and further improve the technology, it is purely his gain, while if the outsourcing partner builds on the technology given by the multinational, both agents gain. The differences between horizontal FDI and vertical FDI could perhaps come close to the differences between licensing and outsourcing. 7 High-end technology is surely a complement to skilled labor; however, the technology that is transferred by a foreign firm to a subsidiary is never a high-end technology. Yeaple (2003) finds that the U.S. outward FDI to LDCs is concentrated mostly in low to medium skilled industries and least of all in the high skill industry. In such a situation it is possible to substitute technology for skilled labor. An example of this can be seen in a CA firm which fills and analyses the tax liability of its client (using tax software). Suppose this process is offshored. In the case where the process is handled by a subsidiary, the CA firm can transfer the knowledge of the software that will help the employees in the subsidiary perform the job. In this case, the subsidiary need not employ skilled CA professionals for the work, while in case of an unaffiliated supplier, the CA firm will not transfer the software or the knowledge, then, the vendor must employ skilled CA to perform the same job. 8 The fact that we mention technology transfer is only to build the rationale of our assumption. Our model, however does not explain the mechanism of how outsourcing will be relatively skill intensive when technology transfer occurs. 6

3

in production fragmentation with no distinction between the different modes of organization. We also allow for substitutability between skilled and unskilled labor in manufacturing, a feature, which is missing in Glass and Saggi (2001). The model’s framework shares similarity with Balcão Reis (2001), however, he focuses on VFDI only. Most of the research in this area focuses on the home country and therefore the factors that determine internalization decision of the firm is well explored. For example, Antràs (2003, 2005), Antràs and Helpman (2004), Grossman and Helpman (2003, 2004) which discuss the sourcing firm’s profits in the two alternative regimes while Glass and Saggi (2001) discuss the source country’s welfare and the multinational firm’s profits, taking the host country as given. Exceptions to this are found in Feenstra and Hanson (1996b) which discusses the home and the host while Balcão Reis (2001) focuses only on the host country. We address the problem from the perspective of a host country. We take the source country and the sourcing firm’s internalization decision – the choice between VFDI and outsourcing as given, and analyze the host country’s perspective to the problem. This paper is about the preferred mode of organization of foreign sourcing in the host country taking the home country as given. Keeping the basic premise of this paper in mind, that is, the supplier technique of production is contingent on their affiliation status; we model the impact of outsourcing relative to VFDI on the welfare of the host country. III.

MODEL

This section builds a north-south framework where the source country is referred to as “north” while the low wage host country is called “south”. Each country is endowed with fixed supplies of two types of labor, the skilled and the unskilled labor. Labor is the consumer of final goods and derives utility from consumption of two kinds of goods. One of the consumption goods is the homogeneous agricultural good, produced only in the south under perfect competition. The other good, is a vertically differentiated manufacturing good. The production of the basic stage of the vertically differentiated good can be shifted to the low wage south either via intra-firm (VFDI) or external (outsourcing) transfer. VFDI and outsourcing are modeled as mutually exclusive forms of foreign sourcing. This is unlike Grossman and Helpman (2003) where VFDI and outsourcing co-exist in the industry equilibrium. Even though the assumption is less realistic, it is essential to evaluate the independent impact of the two forms of foreign sourcing on host country welfare. Our approach is similar to Ottaviano and Turrini (2007), where they model exports, VFDI and outsourcing as mutually exclusive ways of capturing the southern market. Antràs (2003, 2005) share this feature of “all pervasive” VFDI or outsourcing for one product line, that is, a good cannot be offshored using two different governance strategies at the same time. In the model presented below, we do not discuss the internalization decision of the sourcing firm but focus on the preferred mode of organizing foreign sourcing by the host country. To keep the algebra clean, we assume that the host country is relatively small9 to affect any variable of the north. Thus, variables like northern wages and cost of production of northern goods are exogenous for the southern equilibrium. At the same time, the sourcing firms10 are large enough to impinge their impact on the host country labor markets hence affect its wages and welfare. Household Behavior

Consumption

Consumers’ problem is modeled in the spirit of Grossman and Helpman (1991) quality ladders model. Consumers live in one of the two countries, h ∈ {N, S}, North or South respectively, belongs to one of the two

We make this assumption based on research, for example, by the McKinsey Global Institute which shows that US will likely lose to offshoring (all host countries taken together) no more than 300,000 jobs each year, an insignificant number when set against normal job turnover (4.7 m in 2005) in the economy. Therefore, our assumption holds true in the current context, but the future of offshoring may perhaps be different. 10 Antràs (2005) point out that a sourcing firm is a multinational only if it vertically integrates with its input supplier not otherwise. Therefore, we prefer to call the final good producer firm a sourcing firm. 9

4

labor types, l ∈ {1, 2}, unskilled or skilled labor respectively. Consumers take market variables as given and maximize their utility11. The utility function of a representative consumer is given by: (1) U = [ Χ]γ [ y ]1−γ Where y is the homogeneous perfectly competitive agricultural good and is chosen to be the numeraire while X, the aggregate vertically differentiated manufacturing good is represented as: 1

⎡

⎤

Χ = ∫ log ⎢∑ ξ m x m ( j )⎥ dj 0

⎣

m

⎦

ξ m is the assessment of quality level m and x m ( j ) is the consumption of quality level m of product j. For consumption of X, consumers choose from a continuum of manufacturing products indexed by j ∈ [0, 1] available in discrete quality levels indexed by m. Quality level12 m of product j provides quality q m ( j ) ≡ ξ m . All consumers value higher quality of manufacturing goods, that is, ξ > 1. Where ξ denotes the innovation size or magnitude of the quality jump. A consumer utility maximization problem is broken in three stages. In stage one, they decide on the proportion of expenditure to be spent on X and y. In stage two, they allocate the available expenditure for each product. Since the elasticity of substitution between various products of the manufacturing sector is assumed to be unity, the consumer evenly spreads spending across the unit measure of all products, that is, γ E ( j ) = γ E . ~ ( j ) of product j and no other units of Aggregate Consumers demand x m~ ( j ) = γ E p m~ ( j ) units of quality level m other quality level of that product. In the final stage, the consumer allocates spending for each product to the ~ ( j ) offering the lowest quality adjusted price. Thus, in equilibrium, the consumers choose only one quality level m quality of a product that has lowest quality adjusted price. A representative consumer in either country maximizes utility as given by the utility function (1) subject to the budget constraint: (2) y + PX X = E Where E = E N + E S = E N + w 1 L 1 + w 2 L 2 E is the aggregate income of the world. Eh is the income in country h . As mentioned before, south being small cannot influence E N . Ll denotes the stock of type l labor in south and w l the corresponding wage. PX is the composite price of the manufacturing good. Maximizing (1) subject to (2), we get the aggregate demands for y and X as: y = (1 − γ ) E

Px X = γ E

Producers

Production Structure

There are two kinds of firms in the north that can potentially produce X – the leader and the followers. The leader firm or the northern national firm has the ability to produce a quality of X closer to the frontier. We assume that the frontier quality of good X can be produced only by the leader firm while the standardized quality of this good is produced by many follower firms. The quality of X produced by the follower firms is one level lower than the one produced by the leader firm. These follower firms offshore the basic stage of good X production to a low cost nation (either through VFDI or outsourcing) to lower its production cost. In the south, there are two types of firms, the domestic sector or the southern national firms which produce the homogeneous good y and the foreign sector or supplier firms, which provide the basic stage of production of good X. The production structure is depicted in figure 1. We separate figure 1 in two parts by a dotted horizontal line. The variables of all entities lying above the line are taken to be exogenous13 by the host country. The paper focuses on the host country and therefore the southern national firms and the supplier – subsidiary or the outsourcing partner are endogenous in our model. We assume that a country can trade the good it does not produce, at zero cost. Since this is a static model we look at one discrete level jump in quality that generates a product cycle. 13 It may seem that the decision to internalize should be endogenous as the supplier is in the south. However, the decision to offshore and to internalize is made by the sourcing firm, which is essentially in the north. 11 12

5

Figure 1 depicts the two types of firms in the North – the northern national firm (or the leader) which produces the higher quality of the vertically differentiated good and the sourcing firms (or the followers) who produce one quality level below the frontier quality. This standardized good can be fragmented into two stages, the basic and the advanced stage. The basic stage of production can be offshored to the low wage south either through an affiliated supplier (VFDI) or an unaffiliated supplier (outsourcing). On the other hand, the southern national firms produce a homogeneous good y. Leader/ Northern National Firms: Frontier quality of X

Standardization

Follower/ Sourcing Firm: One quality level lower Advanced stage of X– one quality level lower than leader)

Subsidiary/ OS partner Search

Domestic Sector/

Foreign Sector/

Southern National Firms:

Supplier firms:

y Production

Basic Stage of X

Figure 1: Production Structure of the World Economy

Manufacturing Technique

We assume that the production of a sourcing firm can be separated into two stages – the basic stage of production and the advanced stage. We assume that to produce one unit of final good, a sourcing firm must combine 1α units of output from basic stage of production with 1 ( 1−α ) units of output from advanced stage of production produced in the north. This can be envisioned as a fragmented production structure whereby advanced production involves the manufacturing of sophisticated intermediate inputs and basic production involves the bundling of these intermediate inputs to produce the final good. Assuming that the low wage south lacks the efficiency to produce the advanced stage, only the basic stage of production can be offshored. Glass and Saggi (2001) assume that one unit of labor produces one unit of output, which makes their production strategy rigid as labor requirements do not respond to factor prices. We do away with this assumption in our model by introducing two factors of production, the skilled and the unskilled labor, which can be substituted for each other. We assume that the techniques of production for the advanced and the basic stage of the standardized good X is given by neoclassical function, f A (.) and f B (L 1 , L 2 ) respectively. We represent the production technology for final good X produced by the sourcing firm as:

(

X M = Min (1 − α ) f

A

(.), α

)

f B (L1 , L 2 )

Out of the total manufacturing output produced by the sourcing firm, a proportion (1 − α ) , that is, the advanced stage, is produced using the sourcing firm’s technique in the north, f A (.) and a proportion α , that is, the basic stage, is produced using the supplier (subsidiary or the outsourcing partner) technique, f B (L 1 , L 2 ) in the south. It should be noted that the production technique of the basic stage of X is different for an affiliated supplier (VFDI subsidiary) vis-à-vis an unaffiliated supplier (outsourcing partner) and thus all variables like wages, marginal costs, aggregate expenditure are affected by mode of organization of foreign sourcing.

6

With this production technology in hand, the marginal cost (MC) of production of the sourcing firm can be written as a weighted sum of MC B , the basic stage MC incurred by the supplier (affiliated or unaffiliated) in the south, and MC A the advanced stage MC incurred by the sourcing firm in the north. MC

M

Where MC B

= (1 − α ) MC =

[a

B 2

A

+ α MC B

( ω )w 2 + a 1B ( ω )w 1

]

(3)

w2 is wages of the skilled labor relative to unskilled labor in the south and a lk (ω ) is the marginal w1 requirement14 of type l southern labor, by the firm type k ∈ {D, B}, where D and B corresponds to the domestic sector firms and the foreign sector firms in the host country respectively. For example, a 1B ( ω ) is the unskilled

ω=

labor requirement by a firm in the foreign sector (or the supplier firm). The domestic sector in the south produces the homogeneous good y under perfect competition with a CRS production function, f D ( L 1 , L 2 ) . The marginal cost of producing good y in a southern firm is given by: (4) MC D = a 2D (ω )w 2 + a 1D (ω )w 1

Pricing Decisions of the firms

By assumption, consumers prefer higher quality goods and therefore they are willing to pay a premium of

ξ for a single jump in the quality of the product. In the north, the northern national firm or the leader competes with the sourcing firms or the followers in a Bertrand way which determines their price. The good produced by the leader is one quality level higher than the standardized one produced by the follower firms. The leader can charge, at best, a quality markup over the cost of production of the follower firm. This is because even if the follower firm prices its good at its marginal cost, the leader firm can still sell its output at a quality mark-up price (as its quality adjusted price would be the same as the standardized good price) 15. (5a) P N = ξ [(1 − α ) MC A + α MC B ] On the other hand, the standardized quality of good X is produced by many firms under perfectly competition16 implying that its price equals their marginal cost of production. (5b) P M = [(1 − α ) MC A + α MC B ] In the host country, southern domestic producers face perfect competition in good y, which drives its price to equal the marginal cost. Since, good y is chosen to be the numeraire, the price of good y equals 1. (5c) P D = MC iD = 1

Resource Constraints Southern skilled labor required by southern firms for good y production is a 2D (ω ) [(1 − γ ) E ] . The derived demand for skilled labor17 by the intermediate good of X is α a B (ω ) ⎛⎜ γ E ⎞⎟ . Thus, skilled labor market 2 M ⎝ MC

⎠

equilibrium in south is represented by an equality of supply with demand.

14 It is well known that for a neo-classical, constant returns to scale production function, the unit labor requirements is equal to the marginal labor requirement and that these coefficients of labor requirements can be expressed as a function of the relative wage. To obtain the unit/marginal labor requirements, we consider the implied cost function and use Shephard’s lemma along with homogenous of degree one property of the production function. 15 Even though northern firm’s price depends on the supplier’s MC and therefore on southern wages, however, by assuming the south to be a relatively small open economy, we can preclude any effect on northern variables or even repercussion effect and focus only on the host country effects. 16 In equilibrium, both quality levels can sell because they have the same quality adjusted price. Each quality can capture some market share that sums to the total market share for good X. Then, using definition of good X, and the price equation (5a) and (5b) the total output of good X is be given by γ E .

MC M

7

⎛ γ E ⎞ L 2 = a 2D (ω ) [(1 − γ ) E ] = + α a 2B (ω ) ⎜ ⎟ M ⎝ MC ⎠

(6)

Similarly, in equilibrium, unskilled labor demand in south is equal to the given unskilled labor supply: ⎛ γ E ⎞ L1 = a1D (ω ) [(1 − γ ) E ] + α a1B, (ω ) ⎜ M ⎟ ⎝ MC ⎠

(7)

This completes the formulation of the model. We can now turn to the next section that compares welfare under the two alternative regimes of international sourcing in south. IV. EQUILIBRIUM IN SOUTH AND WELFARE COMPARISON In our model, the two organizational modes of international production sharing, VFDI and outsourcing differ only in the skill intensity employed by the supplier in the host country. Therefore, we can analyze the regime switch from one mode of governance to the other as a comparative static exercise of a change in skill intensity of the basic stage of production in the X sector. To execute this exercise, we need to specify one of the two modes of foreign sourcing as the initial equilibrium. WLOG, we assume that in the initial equilibrium, outsourcing to an unaffiliated supplier is not a viable way for international production sharing, say because independent suppliers are not available. Thus, VFDI constitutes the initial equilibrium in the host country. Suppose after some time periods, the market is thick with suppliers, then we ask if it is worthwhile to shift all production sharing contract to these unaffiliated suppliers. If this regime switch from VFDI to outsourcing can increase the real GDP18 of the host country then certainly outsourcing is called for. We qualify the conditions for which a mode change from intra-firm production transfer to external transaction raises the welfare of the host country. We also find conditions in which the aggregate real wage earnings fall after regime shift from VFDI to outsourcing, that is, the host experiences a welfare loss. As we switch from VFDI equilibrium to the outsourcing steady state, there is an exogenous increase in the skill intensity of the basic stage of production. The unit requirement of labor, a lk , depends on exogenous factors and endogenously on host country wages. The Mode of organization of foreign sourcing is one such exogenous factor. Let the exogenous rate of change in marginal labor requirements of l type of labor with regime switch in south be represented by da B ( exogenous ) l = 1, 2 uˆ = l l

a lB

Our assumption of a subsidiary being less skill intensive implies that skill intensity increases after regime switch from VFDI to outsourcing, that is: uˆ 2 − uˆ1 > 0 Moreover, as the regime changes from VFDI to outsourcing, wages also change, which impact the skill requirement endogenously after the regime shift. da k (ω ) l = 1, 2 and k = D, B aˆ lk = l k al Both exogenous and endogenous changes in marginal labor requirement entail a change in marginal cost of production with a shift from VFDI equilibrium to outsourcing equilibrium. Totally differentiating MC B , equation (3), to get the change in marginal cost of production of the southern unit of the sourcing firm (as we move from VFDI equilibrium to outsourcing steady state): (8) MCˆ B = θ 2B wˆ 2 + θ 1B wˆ 1 + θ 2B uˆ 2 + θ 1B uˆ 1 17 Ideally the derived demand for labor in the foreign sector should be determined by the share of market captured by the follower firms. Without loss of generality, we work with the case where it is equal to 1. 18 It should be noted that the change in utility will be equivalent to deriving the change in real GDP because the demand for X and y are proportional to the expenditure and also that all income in spend on consumption.

Uˆ = γXˆ + (1 − γ ) ˆy = Eˆ − γPˆx = Eˆ − Pˆ = change in real GDP

8

dw l wl Totally differentiating the southern domestic firm’s marginal cost, equation (4), and using (5c), we get: (9) MCˆ D = θ 2D wˆ 2D + θ1D wˆ 1D = 0

where wˆ l =

⇒ wˆ 1 = −

θ 2D wˆ 2D θ1D k

In equation (8) and (9) θ k = a l . wl is the distributive share of type l labor in type k firm in the host country. l

MC k

Using the above expression we get that, k = D, B (10) aˆ 1k = − θ 2k σ k ( wˆ 1 − wˆ 2 ) and aˆ 2k = θ 1k σ k ( wˆ 1 − wˆ 2 ) Where σ k is the elasticity of substitution between the two factors of production. Totally differentiating the southern resource constraints, given by equation (6) and (7), and substituting for aˆ lk from (10) and using Eˆ = wˆ 1 e 1 + wˆ 2 e 2 we get: wˆ 2 λ1B λ B2 = uˆ 2 − uˆ 1 ⎡σ D D D B σB B B θ 2D ⎞⎤ D D B B B ⎛ ⎢ D (θ 1 λ 2 λ1 + θ 2 λ1 λ 2 ) + D (λ 2 λ1 ) + (λ 2 − λ1 )⎜⎜ e 2 − D e 1 ⎟⎟⎥ θ1 θ 1 ⎠⎦⎥ ⎝ ⎣⎢ θ 1 Where e l = w l L l is the income share of the type l labor of the host country in global GDP and E

proportional labor shares. For example, λB = α a lB (ω ) ⎛⎜ l

Ll

γ E ⎞⎟

⎜ MC M ⎟ ⎝ ⎠

λ s are the

is the fraction of type l labor of south employed by the

foreign sector in the host country. D It should be noted that (λB2 − λ1B ) and ⎛⎜ e − θ 2 e ⎞⎟ are always of the same sign such that their product s 2 D 1⎟ ⎜ ⎝

always positive19. Thus, wˆ 2 > 0 uˆ 2 − uˆ 1

θ1

⎠

. Intuitively, when the skill intensity of production increases, we would expect

the skill premium to increase and wages of unskilled labor falls. To make welfare comparisons across the two alternative regimes of foreign sourcing, we need to look at real wage effect. The price of good X depends on the MC of the supplier which changes with regime change. Since the price of both quality levels (leader and follower) depends on the marginal cost of production of the supplier, price index changes in proportion to the consumption share of good X.

(

pˆ = γ α MCˆ B

)

Using (8) and (9) we get: ⎤ ⎡ MCˆ B wˆ 2S ⎢θ 2B θ 2S ⎥ θ 2B uˆ 2 + θ1B uˆ1 B = θ1 ⎢ − ⎥+ uˆ − uˆ θ1B θ1S ⎥ uˆ 2 − uˆ1 uˆ 2 − uˆ1 12231 ⎢ 1 424 3 14 4244 3 + − ⎣⎢ + / − ⎦⎥ B B Let Kˆ = θ 2 uˆ2 +θ1 uˆ1

uˆ2 − uˆ1

CRIS INFAC (2005) questions the sustainability of captives in the Indian BPO industry due to their high employee and training costs. Their research clearly implies a lower marginal production costs for the unaffiliated supplier. Grossman and Helpman (2003) also assume that the marginal cost of production of a subsidiary is To ensure full employment, the labor capital ratio of the host country must always lie between the skill intensities of the two sector, that D is, the domestic and the foreign sector. This ensures that if λB2 > λ1B , then e > θ 2 e 2 1 θ 1D 19

9

always higher than an outsourcing partner. In their model, the sourcing firms face a trade-off between higher fixed cost of searching and contracting with suitable outsourcing partners on one hand, and a lower marginal cost of production on the other. VFDI in their model does not entail costly search and contractual costs and hence their marginal costs are higher vis-à-vis an unaffiliated supplier. The rationale behind this assumption can be explained by the familiarity of domestic producers with the host country market and economies of scale. There are fixed cost of offshoring and the economies of scale generated by this fixed cost can be exploited if the supplier can sell its product to several users. The unaffiliated supplier can then offer the input at a price below the average cost of in-house production. If we go by Grossman and Helpman (2003) assumption and CRIS INFAC (2005) we get the result that a switch from VFDI equilibrium to the outsourcing steady state, lowers the MC of the supplier falls, that is: MCˆ B < 0

⇒ Kˆ < 0 .

However, our results do not require such a strong assumption, all we require is that Kˆ ≤ 0 . A switch from VFDI to outsourcing will produce a higher magnitude of welfare if the real GDP after the regime shift rises, that is,: wˆ 2 − pˆ wˆ − pˆ L2 + 1 L1 > 0 uˆ 2 − uˆ 1 uˆ 2 − uˆ 1

⎧⎪ ⎛⎜ wˆ 2 L 1 ⎨ ⎜⎜ ζ uˆ 2 − uˆ 1 ⎪⎩ ⎜⎝

or,

⎛ B ⎞ ⎫ ⎜ θ θD ⎟ ⎟ − γ α θB⎜ 2 − 2 ⎟ ζ + 1 ⎪ ⎬ ⎟ ⎟ 1 ⎜ B θD ⎟ ⎟ ⎜ θ ⎪⎭ ⎠ ⎝ 1 1 ⎠ ⎞

θD ⎟

2 − θD 1

(

) − γ α (ζ + 1) Kˆ > 0

S

Where ζ = L 2 S L1

Taking Kˆ ≤ 0 , a sufficient condition for outsourcing to generate higher welfares is:

ζ>

⎛ B ⎞ ⎜ θ ⎟ θD ⎟ +γ αθB⎜ 2 − 2 ⎜ ⎟ 1 θB θD ⎜ ⎟ ⎝ 1 ⎠ 1 ⎞ ⎛ B ⎜ θ θD ⎟ 1 − γ α θ B ⎜⎜ 2 − 2 ⎟⎟ 1 θB θD ⎟ ⎜ ⎝ 1 1 ⎠

⎛ D ⎜θ ⎜ 2 ⎜ D ⎜θ ⎝ 1

⎞ ⎟ ⎟ ⎟ ⎟ ⎠

=ς

Thus, if the absorptive capacity of the host country is above this threshold, outsourcing certainly leads to higher welfare. It should be noted that, if the foreign sector, B, is skill intensive vis-à-vis the domestic sector, D, that is, θB

θD 2 2 > B θ θD 1 1

, then the absorptive capacity required to generate higher welfare with outsourcing is higher. The result is

intuitive because higher skill intensity of the foreign sector leads to a greater increase in demand for skilled labor after regime switch. This in turn implies a greater rise in skill premium and correspondingly a greater decline in the wages of the unskilled labor. Since change in wages are the only source of welfare gain or loss, we can conclude that unless the relative supply of skilled labor is high in the host country, rise in skill premium due to outsourcing may not benefit a large proportion of the population and hence may lead to an overall welfare loss. A high proportion of X consumption also warrants a higher threshold absorptive capacity. The intuition can again be explained by the fact that a higher proportion of X consumption implies greater demand for the basic input and hence a greater increase in skill premium. Assuming Kˆ ≤ 0 , we can conclude that a necessary condition for VFDI to generate higher welfares is: ζ <ς Thus, VFDI may lead to a higher welfare relative to outsourcing if the host country has a lower absorptive capacity relative. To see the intuition behind this result, we note that, with VFDI, the demand for skilled labor is not high. So, if VFDI is matched with high ζ , the low demand for skills pushes the wages of skilled labor to lower levels. This reduces welfare because a greater proportion of (skilled) labor earns this lower level of skill premium. On the other hand, outsourcing may still lead to higher welfare vis-à-vis VFDI even if it is matched

10

with lower skill abundance in the host country provided the rise in skill premium due to outsourcing is not too high to crash the unskilled labor wages. This result indicates some lesson for the developing countries that compete blindly for VFDI by offering subsidies and attractive package incentives to the sourcing firms. It makes sense for the host country to attract VFDI only if they have low level of skills relative to domestic sector absorption in the host country. However, the results also indicate that even if the host absorptive capacity, ζ , is low, then under certain conditions, VFDI may still not lead to higher welfare. On the other hand, if the domestic absorptive capacity is above the threshold defined, then, the host country certainly gains from outsourcing contracts rather than VFDI. Our result also mirrors Antràs (2005) result on organizational cycles. In his model, VFDI always precedes outsourcing because the degree of standardization of a product increases with time as its high-tech intensity falls. A lesser standardized product is more compatible with an intra-firm production contract because the sourcing firm contributes more to the value of the relationship when the headquarter intensity of the product is high. As the degree of standardization of the product increases with time, the product’s governance moves from internal production arrangement to market transaction thereby exhibiting organizational cycles. Our results show that this kind of organizational cycle is also what the host country desires. At low levels of development of a host country, when the availability of human capital is also short, it benefits the host country to have VFDI. With time, as its absorptive capacity grows then it is welfare improving to host outsourcing contracts. V. CONCLUSIONS In this paper foreign sourcing in the form of either VFDI or outsourcing is taken as an exogenous event and the focus of interest lies in comparing their relative affect on welfare in the host country. The approach that we chose runs as follows. We develop a quality ladder model that focuses specifically on the events in the host country. The empirical and theoretical differences between VFDI and outsourcing helps us support our assumption that a subsidiary activity is relatively less skill intensive vis-à-vis an outsourcing partner’s production technique. Our model argues that under certain conditions, depending on the absorptive capacity of the host country, outsourcing leads to a higher level of real GDP and welfare. Specifically, if the absorptive capacity of the host is higher than a given threshold (which depends on the proportion of manufacturing good consumption, relative skill intensity of the foreign sector in the host country) then, outsourcing certainly leads to higher welfare. However, if the absorptive capacity of the host country is below this derived threshold, then, VFDI may lead to higher welfare. Even in this case, outsourcing being welfare improving over VFDI is not ruled out. This result should raise alarm for countries blindly trying to attract VFDI by giving incentives especially in the form of subsidies and tax relaxation. Going further, it may be interesting to accommodate the differential impact of VFDI and outsourcing on skill formation in the host country by endogenizing the skill acquisition decision of workers as in Sayek and Sener (2001) and Beaulieu et al. (2004). In such a setting, VFDI will induce skill formation by getting the unskilled labor in close contact with the new technology of the north, while under outsourcing, individuals respond to increased skill premium by undertaking training and becoming skilled. REFERENCES Amiti, M. and S. Wei, “Service Offshoring and Productivity: Evidence from the United States”, NBER Working Paper, No. 11926, 2006. Antràs, P., “Firms, Contracts, and Trade Structure,” Quarterly Journal of Economics, 118(4), (2003), pp. 1375-1418. Antràs, P., “Incomplete Contracts and the Product Cycle,” American Economic Review, 95(4), (2005), pp. 1077-1091. Antràs, P., and Helpman, E., “Global Sourcing,” Journal of Political Economy, 112, (2004): 552-580 Buckley, P. and Casson, M., “The Optimal Timing of a Foreign Direct Investment,” The Economic Journal, 91(361), (1981), pp. 75-87. Balcão Reis, A., “On the Welfare Effects of Foreign Investment,” Journal of international Economics, 54 (2), (2001), pp. 411-427. Beaulieu, E., Benarroch, M. And Gaisford J., “Trade Barriers and Wage Inequality in a North-South Model with Technology- Driven Intra-Industry Trade,” Journal of Development Economics, 75(1), (2004), pp. 113-136

11

Calabrese, G. and F. Erbetta, “Outsourcing and Firm Performance: Evidence from Italian Automotive Suppliers”, International Journal of Automotive Technology and Management , 5(4), (2005), pp. 461 - 479. Casson, M. Alternatives to the Multinational Enterprise, (1979): Macmillan, London CRIS INFAC, “IT-enabled services, Third party vendor model the better bet”, February, (2005) Egger, H. and P. Egger, “International Outsourcing and the Productivity of Low-skilled Labour in the EU”, Economic Inquiry, 44(1), (2006), pp 98-108. Feenstra, R. and Hanson, G, “Foreign Investment, Outsourcing and Relative Wages”, in R.C. Feenstra, G.M. Grossman & D.A. Irwin, eds, “Political Economy of Trade Policy: Essays in Honor of Jagdish Bhagwati,” MIT Press, Cambridge, Mass. (1996a), pp. 89-127. Feenstra R. and Hanson, G., “Globalization, Outsourcing and wage Inequality”, American Economic Review, 86(2), (1996b), pp. 240-245. Glass, A.J. and Saggi, K., “FDI Policies under Shared Factor Markets,” Journal of International Economics, 49(2), (1999), pp. 309-332. Glass, A.J. and Saggi, K., “Innovation and Wage effects of International Outsourcing,” European Economic Review, 45(1), (2001), pp. 67-86. Glass, A.J. and Kamal Saggi, “Licensing versus direct investment: implications for economic growth,” Journal of International Economics, 56(1), (2002), pp. 131-153. Grossman, S. and Hart, O., “The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration,” Journal of Political Economy, 94(4), (1986), pp. 691-719. Grossman, G. and Helpman, E., “Integration versus Outsourcing in Industry Equilibrium,” Quarterly Journal of Economics, 117(1), (2002), pp. 85-120. Grossman, G. and Helpman, E., “Outsourcing versus FDI in Industry Equilibrium,” Journal of the European Economic Association, 1(2-3), (2003), pp. 317-327. Grossman, G. and Helpman, E., “Managerial Incentives and International organization of production”, Journal of International Economics, 63(2), (2004), pp. 237-262. Grossman, G. and Helpman, E., “Outsourcing in a Global Economy,” Review of Economic Studies, 72(1), (2005), pp. 135-160. Görg, H. and A. Hanley, “International Outsourcing and Productivity: Evidence from the Irish Electronics Industry”, North American Journal of Economics and Finance, 16(2), (2005), pp. 2005. Hanson, G., Mataloni, R.J. and Slaugter, M. “Vertical Specialization in Multinational Firms,” Review of Economics and Statistics, 87, (2005), pp. 285-293. Hart, O. and Moore, J., “Property Rights and the Nature of the Firm,” Journal of Political Economy, 98(6), (1990), pp. 1119-1158. Mansfield, E. and Romeo, A. “Technology Transfer to Overseas Subsidiaries by U.S. Based Firms,” Quarterly Journal of Economics 95(4), (1980), pp. 737-749. Ottaviano, G. I. P. and Turrini, A. “Distance and Foreign Direct Investment When Contracts and Incomplete” Journal of the European Economic Association, 5(4), (2007), pp. 796-822 Pack, H. and Saggi, K., “Inflows of Foreign Technology and Indigenous Technological Development,” Review of Development Economics, 1(1), (1997), pp. 81-98 Sachs, J. & Shatz, H., “Trade and jobs in U.S. manufacturing,” Brookings Papers on Economic Activity, Vol. 1994(1), (1994), pp. 1-84. Saggi, K., “Entry into a Foreign Market: Foreign Direct Investment versus Licensing,” Review of International Economics 4(1), (1996), pp. 99-104. Saggi, K., “Foreign Direct Investment, Licensing, and Incentives for Innovation,” Review of International Economics, 7(4), (1999), pp. 699-714. Sayek, S. and Sener, F. “Dynamic Effects of Outsourcing on Wage Inequality and Skill Formation,” Union College, Schenectady, NY mimeo. (2001) http://www1.union.edu/senerm/Research/Outsourcing_wages_10_01.pdf Slaughter, M., “Production Transfer Within Multinational Enterprises and American Wages” Journal of International Economics, 50(2), (2000), pp. 449-472. Slaughter, M. and Swagel, P., “The effects of globalization on wages in the advanced economies,” IMF Working Paper WP/97/43 (1997).

12

Stähler, F. “A Model of Outsourcing and Foreign Direct Investment,” Review of Development Economics, 11(2), (2007), pp. 321-332, Yeaple, S.R., “The Complex Integration Strategies of Multinationals and Cross Country Dependencies in the Structure of Foreign Direct Investment.” Journal of International Economics, 60, (2003), pp. 293-314.

13