INRA-LORIA

5 downloads 0 Views 427KB Size Report
Nous donnons également à la multinationale la possibilité de former une partie des ... la multinationale à y investir n'est pas systématiquement monotone. ... (1996), shows that no proxy of the human capital is a significant determinant to. FDI.
INRA-LORIA Laboratoire d'Organisation Industrielle Agro-alimentaire

Multinationals and the Training of workers in the host country : the Tariff-jumping argument revisited

Hakim Hammoudi Wadii Hatit

Octobre 2004

Cahier n° 2004-05

Laboratoire d'Organisation Industrielle Agro-Alimentaire 65 Bd Brandebourg 94205 IVRY Cedex (33) 1 49 59 69 87 http://.fr/ mailto:[email protected]

Multinationals and the Training of workers in the host - country : the Tariff-jumping argument revisited

Hakim Hammoudi1 Wadii Hatit2 Octobre 2004 Cahier n° 2004-05

Résumé:

Nous analysons la décision "investissement direct v. exportation" d'une multinationale en concurrence avec un entrant potentiel local sur un marché hôte. Nous considérons une asymétrie au niveau de la qualification de la main d'oeuvre entre le pays d'origine de la multinationale et le pays hôte. Nous donnons également à la multinationale la possibilité de former une partie des salariés embauchés en cas d'investissement. Nous montrons que la relation entre le niveau de qualification dans le pays hôte et l'incitation de la multinationale à y investir n'est pas systématiquement monotone. Nous montrons également comment la stratégie de contournement tarifaire peut améliorer le bien-être du pays hôte même si elle provoque l'exclusion de la firme locale du marché.

Abstract:

We analyse the "direct investment v. export" decision of a multinational firm in competition with a potential entrant in a host country. We consider a workers' skills asymetry between the host and the multinational home countries. We also give to the multinational the possibility to choose and train a part of the hired workers when investing. We show that an improvement in the workers' skills in the host country does not increase systematically the multinational incentive to invest. We also demonstrate that under the multinational's training assumption, the tariff-jumping investment can always be welfare improving even if it excludes the local firm from the market.

Mots clés :

Investissement direct étranger, firmes multinationales, politique commerciale, formation, bien-être

Key Words :

Foreign direct investment, multinational firms, trade policy, training, welfare

Classification JEL: F12, J24, O12

1 2

ERMES, Université Panthéon-Assas et INRA-LORIA. E.mail. [email protected] ERMES, Université Panthéon-Assas et INRA-LORIA. E.mail. [email protected]

La version électronique du cahier est disponible à l’adresse : www.inra.fr/Internet/Departements/ESR/UR/ivry/PDF/LORIA2004-05.pdf

1

Introduction

The theoretical literature dealing with the determinants of the multinationalisation of …rms (especially Foreign direct investment (FDI) vs. Export to a host market) was renewed with a serial of works based on the imperfect competition hypothesis and the strategic interaction between …rms [Smith (1987), Horstmann and Markusen (1987,1991), Motta (1992,1994)]. Early contributions to this strand of literature propose models in which a foreign …rm, already established in its own country and facing the "Exportation vs. direct investment" dilemma, competes with a local potential entrant in a single market. These works aim at determining the conditions under which the foreign …rm adopts the direct investment as a tari¤-jumping strategy and the impact of this strategy on the host-country welfare. The authors deal with these questions by introducing several variables (market size, exportation costs, sunk costs...) which have a priori an impact on the …rms’strategies. For the sake of simplicity, these contributions suppose that the return of the multinational’s technology is independent of its implementation country. They don’t take into account the fact that, when investing abroad, the multinational leaves its workers which are already experienced and familiar with its technology to operate with the host-country workers who have not necessary the required skills to be e¢ cient. In fact, if we consider the issue of the foreign investment in less developed countries, the workforce’s under-quali…cation, increases the investment’s cost and in‡uences consequently the multinational’s tradeo¤ between Exportation and FDI. A large number of empirical works have considered the latter point. If we focus on the cross-country studies, we distinguish two groups. The …rst one, which includes Root and Ahmed (1979), Schneider and Frey (1985) and Hanson (1996), shows that no proxy of the human capital is a signi…cant determinant to FDI. The second is composed of Noorbakhsh et al. (2001) and Nunnenkamp and Spatz (2002). They show a positive relation between human capital and FDI. In spite of this empirical controversy, the few theoretical works dealing with this question lead to the following consensual result : the quali…cation level has a positive impact on the foreign investment. For instance, Lucas (1990) conjectures that the lack of human capital discourages the foreign investment. Zhang and Markusen (1999) show that the availability of skilled workers in the hostcountry makes it more attractive to foreign investors. Dunning (1988) suggests that the skill and the education level of labor can in‡uence both the volume of the foreign investment and the type of the activities that multinationals can undertake in a country. If we reconsider the relation between the human capital and the FDI under the potential competition assumption, the standard result of the theoretical literature might change. An increase of the workers’quali…cation in a country improves naturally the return of the FDI. However, it also incites local investors to enter the market. The intensi…cation of the competition after the human capital’s raise makes the relation between human capital and FDI ambiguous. This imposes a reconsideration of the question under an endogenous market 1

structure of the host-country. A part from the exogeneity of the market stucture, the previous theoretical works have another common characteristic. All of them consider that the human capital level in the host-country is constant. They exclude de facto several strategies that multinationals can undertake in order to improve the quali…cation level in the host-country, such as "training". Nevertheless, the involvement of multinationals in the development of human ressources’skills where they invest is supported by a large number of empirical works and study cases. OECD (2002, 2003) and UNCTAD (1994) give many examples showing that multinationals invest substantially in both general and speci…c training. They also provide a considerable support to upgrade the formal education in host countries. Otherwise, many empirical studies show that multinationals train more that local …rms [See Tan and Batra (1996), Tan and Lopez-Acevedo (2003), Miyamoto and Todo (2003), and Almeida (2003)]. Hence, the human capital level is not an irreversible variable. The multinational can be incited to improve it in order to increase its return. This is notably the issue of the FDI in less developed countries where the multinationals are not able to reproduce e¢ ciently their technologies without incuring an additional cost of training1 . Such a scenario is not inherent to the North-South trade relation. The Japanese FDI in USA is another illustration of this point. After the excessive raise of trade barriers imposed by the USA, the Japanese …rms persued di¤erent types of global strategies such as FDI to supply the US market. However, the relatively low quali…cation of US workers leads the Japanese investors to undertake considerable e¤orts in training [See Urata (1998) and UNCTAD (1994)]. In our paper, we analyze the investment decision of a multinational considering explicitly this potential commitment in training. Furthermore, we consider on the one hand the interaction between …rms’ strategies and, on the other hand, the interaction between …rms’decisions and the trade policy of the hostcountry. Our paper can be seen as an extention of Smith (1987) and Motta (1992). Contrary to these works, we introduce the hypothesis of under quali…cation of workers in the host-country. Moreover, we give to the multinational the strategic possibility to train all or a part of the hired workers. Three options are o¤ered to the multinational. It can put up with the low host-country workers quali…cation or invest with an involvement in workers training. It can also choose the exportation instead of the investment in order to bene…t by the better quali…cation in its home country. The consideration of the multinational’s ability to train brings new elements to deal with standard questions studied by the previous literature: (i) the private incentive to FDI which may be higher with the training strategy, (ii) the host-country welfare which may be positively a¤ected by the human capital externality generated by the FDI. 1 In India, "Simens" provides a three-year training programme for 140 young workers. "Intel" adopts a training strategy in all countries where it invests : Argentina, Brasil, Costa Rica, China, Malaysia, Russia, Poland etc.

2

Concerning the …rst question, we show that the multinational has always incentive to train a part of the hired workers when investing in the host-country. Owing to the training strategy, the multinational is more likely to adopt FDI. We also show that an improvement in the human capital in the host-country does not encourage systematically the multinational to FDI. Concerning the second question, we show that the multinational’s investment can exclude the local …rm due to the training. However, the host-country bene…ts of the higher multinational’s return. The trade-o¤ between these two opposite e¤ects leads to new results concerning the tari¤-jumping impact on the host-country. In fact, contrary to the intuition and previous works [See Smith (1987) and Motta (1992)], we explain how can an anti-competitive tari¤, inducing the multinational’s investment, improve the host-country welfare. The remainder of this paper is set out as follows. Section 2 presents the model assumptions. In the section 3, we discuss the issue of the game when the multinational is not able to train. In section 4, we remove this assumption. In section 5, we analyze the e¤ect of the tari¤-jumping investment on the hostcountry welfare. We conclude in the section 6.

2

Model

2.1

The assumptions

We suppose that the world economy in composed of a foreign and a host countries. In the foreign country, there is a multinational …rm (M )2 already established. In the host-country, there is a potential local entrant (L). The two …rms produce a homogenous good and supply only the home country market. The demand function in this market is given by : P (Q) =

Q

(1)

Where P denotes the price of the good in the host market, represents the size of its market and Q is the total quantity sold in the same market. The production process requires only the labor input. We suppose that the workforce’s productivity is higher in the foreign country. More precisely, a worker in the foreign country produces one unit of the good (which is the best return that we can get) while a worker of the home country produces g units (g < 1). We also assume that the number of workers in the two countries is very large to satisfy the work’s supply of …rms. We denote by w the wage in the two countries3 . To serve the host market, the multinational has to choose between the exportation and the direct investment. Using the …rst strategy, the multinational makes use of its production facilities and quali…ed workers. However, it has to 2M

will be called multinational even if it’s a potential one. assumption is kept to simplify the analysis. Our main results are qualitatively unchanged when we suppose that the host-country wage is lower than the foreign-country one. 3 This

3

incur a per unit transportation cost s. With the second strategy, the multinational avoids all the exportation costs but has to incur, as the local …rm, a …xed cost K. When investing, we suppose that only the multinational has the possibility to train all (or a part of) the hired workers4 . The aim of the training action is to improve the productivity of workers from g to 1. We suppose also that the training is speci…c and the multinational incurs all its cost. Thus, the trained workers don’t get a higher wages after the skills acquisition. The training cost of N workers, having initially g as productivity, is given by the following expression : 1 (1 g):N 2 (2) 2 This function re‡ects diseconomies of scale in the training process because the …rms run into constraints of capital equipments as the number of the trainees increases5 . Moreover, the training costs are lineray decreasing in g : the higher the initial workers productivity, the lower the costs of the training. (N; g) =

2.2

The game

We assume that …rms’ decisions on the entry mode are taken in a sequential way. The local …rm is supposed to move …rst because of its proximity from the host market. We analyse the game described in the …gure 1 : In the …rst stage, the local …rm decides whether to enter (E) or not (NE). In the second, the multinational decides whether to export (EX), invest (INV) or not to serve the host market (?). Finally, in a third stage, a cournot game occurs to determine the number of workers to hire which determines the quantities produced at the equilibrium. L

E

NE

M

INV

M

EX

Ø

INV

COURNOT

EX

Ø

COURNOT

Figure 1 4 The

assumption of asymetry between local and multinational …rms with respect to their ability to train is supported by many empirical works (See the introduction). 5 See Booth and Chatterji (1998).

4

We assume that the decisions at earlier stages are made in the knowledge of the equilibria in the later stages. The equilibrium is solved by backward induction. Below, we give the notations of …rms’ payo¤s conditionally to the strategies adopted in the earlier stages. LnM

EX

IN V

E

(

ex m =e ;

NE

(

ex m =ne ; 0)

e l =ex )

?

(

inv m =e ;

e l =inv )

(

inv m =ne ; 0)

(0;

e l =? )

(0; 0)

In the two following sections, we give the …rms’ payo¤s in the di¤erent con…gurations and discuss the issue of the game relatively to the host-country characteristics (workers’skills and market size) and the multinational ability to train.

3

FDI and host-country workers’quali…cation

All multinationals do not invest in the training of their workers. In the literature, there is some evidence to identify why this is the case6 . In this section, we start to analyse a benchmark case where the multinational has not the ability to train for exogenous raisons. Let’s determine the payo¤s of the last game’s stage. After deciding on the entry mode, …rms that have chosen to sell in the hostcountry maximize their pro…ts by determining the number of workers to hire in a simultaneous way. Below, we give the pro…ts of both …rms at the equilibrium according to the strategies chosen in the earlier stages7 . When the multinational exports and the local …rm decides not to enter, an exporting monopoly emerges in the host-country. The local …rm gets a nil pro…t and the pro…t of the multinational is as follows :

ex m =ne

=

(

(

0

s w)2 4

if if

>s+w 2g(s+w) ex 9g 2 g (4) = = m e w 0 if < 2g(s+w) g 6 See 7 For

OECD (2003) for a discussion of this point. the full derivation of results see the appendixes.

5

e l =ex

=

(

[g( +w+s) 2w]2 9g 2 ( g w)2 K 4g 2

K

if

>

if


> >

1 2 3

, , ,

e l =inv > 0 e l =ex > 0 e l =? > 0

(8) (9) (10)

The curves 4 and 5 represent the minimum levels which yield to the multinational a positive pro…t given the local …rm’s strategy. > >

4 5

, ,

ex m =e > 0 ex m =ne > 0

(11) (12)

Finally, the locci 6 et 7 are the multinational’s indi¤erence curves between the investment and the exportation according to whether the local …rm enters or not at the equilibrium. > >

6 7

, ,

inv ex m =e > m =e inv ex m =ne > m =ne

(13) (14)

The …gure 2 shows how the game’s issue changes when the host-country workers’skill increases. We distinguish three cases according to the market size. When is very low, the raise of g has no e¤ect on the equilibrium : No …rm invests in the host-country. For medium values of , a …rst increase of g induces the entry of the local …rm without modifying the multinational’s strategy (EX). In fact, if the di¤erence in wokers’ skills between the two countries remains considerable after the increase of g, the multinational keeps exporting to bene…t of the higher quali…cation in its home country. A further increase in g may discourage the multinational to export and incites it not to serve the hostcountry without any increase in the transportation cost. The local …rm keeps always its entry decision. In this con…guration, a raise in the skill level has the same qualitative e¤ects on the two …rms as an increase in the trade cost. At high values of , the relationship between the …rms’incentive to invest and the quali…cation level is a monotonical one : The higher the quali…cation level, the higher the incentive to invest. This result is in agreement with those of the theoretical literature9 . If we focus on the e¤ect of the size, our analysis leads to an other intuitive result : a higher market size incites …rms to invest. Althrough this result is embodied in the literature10 , it was be challenged by some contributions. In fact, Motta (1992) shows that the interaction between the multinational and the local 9 See 1 0 See

the introduction. Horstmann and Markusen (1987 and 1992), and Motta (1994).

7

…rm in a simultaneous game can alter the monotony of the relation. Horstmann and Markusen (1996) exlpain how a multinational, which is not well informed about the host-country market characteristics, can …nd it more pro…table to contract with a local agent instead of investing directly when the market size is large enough to support the FDI. However, all these contributions agree that, in a large market, at least one …rm decides to invest. The consideration of the asymetry in workers’skills between the two countries challenges this result. It gives rise to an equilibrium con…guration not revealed by the previous works : The host-country market is large and no …rm invests. This is illustrated by the region (1b) of the …gure 2. In this region, g is very low and the market is large. The multinational decides to export to bene…t of the higher skills in his country. Disadvantaged by the low skills of its workers, the local …rm responds by not entering.

4

FDI and host-country workers’training

In this section, we consider the multinational ability to train. Note that when the multinational exports or doesn’t serve the host market, the payo¤s of both …rms do not change with respect to the benchmark case11 . When investing, the multinational decision on hiring workers changes relatively to the previous case. Here, the multinational has to determine the number of workers to be trained and the number of workers to be a¤ected directly to the production. In the case of entry, the local …rm determines the number of workers to be hired. Remember that all decisions on the workers’ numbers are taken in a simultaneous way. These payo¤s are given by the following expressions12 . When the multinational invests and the local …rm decides to enter, the two …rms get the pro…ts :

inv m =e

e l =inv

=

(

(3 g)[ g+(1 2g)w]2 2g 2 (5 2g)2 2 ( g w)2 + (1 2gg)w 2 9g 2

=

(

[ g(2 g) (3 2g)w]2 g 2 (5 2g)2 ( g w)2 K 9g 2

K

if




K

if




4w g 4w g

(15)

4w g 4w g

(16)

When the multinational invests while its rival does not enter, it gets the following pro…t : ( ( w)2 3w K if inv 6 2g g (17) m =ne = ( g w)2 (1 g)w2 3w + 2g2 K if 4g 2 g 1 1 See 1 2 For

expressions (3), (4) and (7). the full derivation of results see the appendixes.

8

The multinational has the choice between three strategies : investment without training, investment with training, and exportation. Note that the multinational gets higher investment pro…t when it trains. Thus, the multinational’s strategic space is reduced to investment with training and exportation13 . The …gure 3 presents the sub-game perfect equilibria in ( ; g) space.

µ µ1 (E, INV)

µ4 (NE, INV)

µ2 µ8

(E, EXP) U

X

V

µ6

Z

(NE, INV)

µ7

(NE, INV)

(E, Ø)

µ3

µ9

(NE, EXP)

µ5 (NE, Ø)

g

Figure 3 Before discussing the results, remark that the investment doesn’t yield the same pro…t for both …rms. We keep 1 and 3 to designate the minimum level of the investment pro…tability for the local …rm and we denote by 8 and 9 the minimum levels of which insure the multinational investment pro…tability given that the local …rm has chosen INV or NE. > >

8 9

, ,

inv m =e > 0 inv m =ne > 0

(18) (19)

As the …gure 3 shows, the multinational’s training changes considerably the game’s issue especially when the workers skills are low in the host-country. Remember that, in the benchmark case, the multinational is incited to export when g is low14 . With the training strategy, the multinational can improve its return. Furthermore, if the market is large enough to cover the …xed costs, the multinational’s strategy switches to the investment even if g is very low. Note that under the latter condition, the non entry constitutes a dominant strategy for the local …rm. This analysis presents the training as a crucial condition that gives rise to the multinational investment. In fact, if we focus on the FDI in 1 3 In the reminder, (INV) will be used to qualify the strategy "investment with training" when (M ) is able to train. 1 4 See the region (1b) of the …gure 2.

9

developing countries, we can see that it is often accompanied by training e¤orts. Many stylised facts support this result15 . The …gure 3 leads to another unexpected result : the relation between the “export vs direct investment” decision and the workers skills’ level in the host-country is not systematically monotone. At the point U , the host-country is endowed with a low skilled workers. The local …rm decides not to enter the market regardless its rival strategy. The multinational chooses between exporting and the direct investment by comparing their respective costs. At the considered point, the best response of the multinational is to invest. As g rises, which corresponds to a move from the point U to the point V , the local …rm …nds it pro…table to enter if the multinatinational doesn’t invest16 . At the equilibrium, the local …rm enters and the multinational responds by exporting. If g increases further (till the point X), the market size of the host-country is still allowing only one …rm to invest. If the local …rm enters, the multinational decides not to invest. Furthermore, the exportation quantity is nil that’s why the multinational decides not to serve the market. Finally, a further increase in g to the point Z, gives to both …rms the incentive to invest.

5

Tari¤-jumping investment and welfare

There is a large literature dealing with the impact of trade policy on the investment decision in a country and its consequences on its welfare. The …rst contributions consider the market structure as exogenous. They disregard the interactions between local and foreign producers. Their main result is that an inducing-investment tari¤ is always welfare improving. Recent contributions have reconsidered the same question with more formal analysis based on concepts of the game theory. Thus, Smith (1987) analyses the strategic aspect of the direct investment and the relation between the trade policy and the incentives to invest in a host-country. He shows that the tarif does not induce systematically the investment of the multinational. His analysis reveals two important cases. In the …rst, a tari¤ increase allows the local …rm to enter. As a reaction, the multinational deviates from the investment to the exportation. In the second, the tari¤ induces the invesment of the multinational which excludes the local …rm from the market. Thus, the tari¤ can increase or decrease the competition that makes its e¤ect on the welfare ambiguous. Motta (1992) keeps the same framework as Smith (1987) and proposes a more formalised contribution. He concludes that the tari¤-jumping investment improves the host-country welfare only when the local …rm would not have entered the market under free trade. Horstmann and Markusen (1992) adopt a more general framework with two countries. In each country, there is a …rm already established. Both …rms have the possibility to produce locally or to a set up a plant abroad. They show that a low tari¤ variation can cause a considerable welfare variation (eventually a decrease). In fact, the market structure endogeneisation creates a discontinuity 1 5 For 1 6 the

instance see OECD (2003). host market is not large enough to hold the investment of both …rms.

10

in the welfare function which limits the applicability of the optimal tari¤ theory. The latter contribution was extended by numerous works. Venables and Markusen (1998) introduce the free entry hypothesis and consider an asymety on the market size and the labor endowment between the two countries. They rely on numerical solutions to shows that the FDI is often bene…cial to the two countries. De Santis and Stähler (2004) consider the market entry and develop a way to solve analytically the model. They investigate the impact of an asymetry in the headquaters costs on the …rm’s decisions and the countries’welfare. They show that, when the headquarters costs are large in the foreign country, the FDI improves the foreign country welfare and reduces the domestic country one. However, when these costs are symetric, they show that the FDI is bene…cial to the two countries irrespective of market structure e¤ects. As we mentionned in the introduction, all these contributions do not consider the possibility of the foreign …rm(s) to train the hired workers when investing in the host-country. This strategy a¤ects the quantities produced and the payo¤s earned by both …rms17 . Thus, it’s legitimate to reconsider the e¤ects of the tari¤-jumping investment on the market structure and the welfare of the host country under this assumption. To this end, we suppose that the host-country adopts a per-unit speci…c tari¤ t on the multinational’s exports18 . We de…ne the host-country welfare as the sum of consumers surplus, local …rm’s pro…t and the tari¤ revenue. We denote by Q(i;j) and W(i;j) respectively the total quantity consumed and the welfare of the host-country when the equilibrium (i; j) occurs. In this section, we study the e¤ect of the host-country tari¤ on the game’s issue and its consequences on the welfare. We limit the welfare analysis to cases where the tari¤ induces the FDI. We deal with two cases according as the multinational trains or not.

5.1

Non training Case

After the host-country protection, the per-unit exportation cost raises from s to s + t. When the multinational exports, the pro…ts of both …rms change when we move from free trade to protection. As a result, many changes occur in the …gure 1 : the curves 4 and 5 move upwards while the curves 2 and 6 move 1 7 See 1 8 The

section 2. tari¤ is considered as exogenous. We take t = 0:05, for all the following …gures.

11

downwards. These changes are illustrated in the …gure 4.

VI

V IV

III II I w w+s+t

w w+s

g Figure 4 The …gure 4 shows that the tari¤ a¤ects the game’s issue in many ways. In the region (II), the tari¤ does not change the game’s issue. In fact, the local …rm keeps not entering the market because g is very low. For the same reason, the multinational continues with exporting. In the regions (I) and (III), the tari¤ compells the multinational to give up serving the market with no changes on the local …rm’s strategy. In the region (IV), we …nd the opposite e¤ect : the multinational is still exporting and the local …rm’s strategy switches to the entry. The regions (V) and (VI) constitute the only situations where the tari¤ induces the investment of the multinational. Let’s move to study the welfare variation in the latter case. Under the free trade, the multinational exports and the local …rm enters. The host-country welfare is given by the following expression :

W(E;EX) =

[w(1 + g)

s)]2 + [w(g 18g 2

g(2

2) + g( + s)]2

K

(20)

After the tari¤ adoption and the multinational’s move to the investment, the host-country welfare becomes : 3K) + w2 2 gw (21) 3g 2 Using these two expressions, the host-country welfare variation is given by the following expression : W(E;IN V ) =

W(E;IN V )

g2 (

2

W(E;EX) = 12

[w

g(w + s)]2 6g 2

0

(22)

Note that this variation is negative. Thus, the host-country is always worseo¤ after the tari¤ adoption. Motta (1992) …nds the same result. Furthermore, he shows that the multinational’s investment improves the consumer surplus and reduces the local …rm pro…t with respect to the free trade. In our model, this is not always true even if the resultant e¤ect is the same. We can easily verify that the following conditions are satis…ed : w w+s w , g>( ) w+s

(23)

Q(E;IN V ) > ( ) Q(E;EX) , g > ( ) e l =inv

w+s …nd the same results as Motta (1992).

5.2

Training Case

In this section, we analyse the impact of the tari¤ on the game’s issue when the multinational is able to train its hired workers. The …gure 5 illustrates all these changes.

µ

µQ

µW II c II b

II a III

Ia Ib

g Figure 5 13

This …gure shows that the multinational adopts the tari¤-jumping investment in the two regions (I), and (II)19 . In the …rst, the local …rm maintains the non entry strategy. Thus, the equilibrium moves from a monopoly where the multinational exports to a monopoly with the multinational as an investor [(N E; EX) ) (N E; IN V )]. In the second, the local …rm is excluded from the market and the equilibrium moves from a duopoly with export to a monopoly with investment [(E; EX) ) (N E; IN V )]. In the previous section, we have shown that the tari¤ induces the multinational’s investment only if the market size and the workers’ quali…cation are su¢ ciently high to provok the entry of the local …rm in the free trade situation. When we consider the multinational’s ability to train, the tari¤ becomes more e¤ective to induce the foreign investment. As shown in the regions (I), the tari¤-jumping investment occurs even if the the host-country market size and the workers quali…cation are very low. In the following, we move to study the welfare variation in the tari¤-jumping investment cases already exposed. In the region (I), the total quantity consumed in the host-country under free trade is given by : w s (25) 2 After the tari¤ adoption, this quantity is given by the following expression : ( w if < 3w (3 g) g Q(N E;IN V ) = (26) g w 3w if 2g g Q(N E;EX) =

The variation analysis of the host-country welfare, which is reduced to the consumers surplus, reveals two cases : 3w Case 1 : g , this corresponds to the region (1b) in the …gure 5. We verify that the following condition holds :

Q(N E;IN V ) > ( ) Q(N E;EX) , g > ( )

w w+s

(27)

The …gure 5 shows that, for all points of the region (1b), the condition g > is satis…ed. Thus, we conclude that :

w w+s

Q(N E;IN V ) > Q(N E;EX)

(28)

Case 2 : < 3w g , this corresponds to the region (1a) in the …gure 5. We verify that the following condition holds : Q(N E;IN V ) > ( )Q(N E;EX) , 1 9 (I)

= (Ia) [ (Ib) and (II) = (IIa) [ (IIb) [ (IIc)

14

s

. Thus, for all points of the region (1a), the property (29) already proved in the region (1b) is satis…ed. Then, we conclude that the tari¤-jumping investment is always bene…cial to the host-country when the local …rm has chosen not to enter under the free trade regim20 . Now, we focus on the region (II). Using the two expressions (21) and (27), we plot the host-country’s indi¤erence curve (W(E;EX) = W(N E;IN V ) ) and the consumers’ indi¤erence curve (Q(E;EX) = Q(N E;IN V ) )21 . We denote respectively these two curves by W and Q . We verify that : W(E;EX) Q(E;EX)

< ( ) W(N E;IN V ) , < ( ) Q(N E;IN V ) ,

2g(s+w) 3g g nL = (A10) 2 w ( g w) if < 2g(s+w) 4g 2 g The pro…ts at the equilibrium are as follows : ( [ g+w 2g(s+w)]2 if ex 9g 2 m =e = 0 if ( [g( +w+s) 2w]2 K if e 9g 2 l =ex = ( g w)2 K if 4g 2 17

>


2g(s+w) w g 2g(s+w) w g

(A12)


0 holds only if > 3w g . In this case, the equilibrium M ’s pro…t is : (1 g)w2 ( g w)2 inv + K (A23) m =ne = 2 4g 2g 2 When < as follows :

3w g ,

M trains all the hired workers. In fact, nM 2 = 0 and nM 1 is

( w) 6 2g This yields the following equilibrium pro…t : nM 1 =

inv m =ne

7.7

(

=

6

w)2 2g

(A24)

K

(A25)

Case 7 : L enters and M invests with traning.

In this case, M has to determin the number of workers to train (nM 1 ) and the numbers of workers to a¤ect directly to production without a preliminary training (nM 2 ).L has to determin the number of workers hired (nL ). The return of both …rms is : q M = nM 1 + g nM 2 q L = g nL

(A26)

The expressions of the ex ante pro…ts are : e l =inv inv m =e

= =

( (

qL qL

qM )qL

wnL

qM )qM

K

w(nM 1 + nM 2 )

(A27) 1 (1 2

g):n2M 1

K(A28)

The simultaneous maximisation of both (A27) and (A28) with respect to nL , nM 1 and nM 2 gives the following solutions :

nL

=

nM 1

=

nM 2

=

( g w) 3g 2 w g ( g 4w) 3g 2 19

(A29) (A30) (A31)

Under the assumption > 4w g , we have nL > 0, nM 1 > 0.and nM 2 > 0. Hence, at the equilibrium the pro…ts are given by : e l =inv

=

inv m =e

=

( g w)2 K 9g 2 (1 g)w2 ( g w)2 + 9g 2 2g 2

(A32) K

(A33)

4w If g which corresponds to nM 2 = 0, M and L respond by choosing nL and nM 1 in the following manner :

nL

=

nM 1

=

g(2 ( g

g) (3 2g)w (5 2g)g 2 w) + 2w(1 g) (5 2g)g

Note that nM 1 > 0. However, nL > 0 holds only if this condition, the pro…ts of both …rms are : e l =inv

=

inv m =e

=

g) (3 2g)w]2 g 2 (5 2g)2 g)[ g + (1 2g)w]2 2g 2 (5 2g)2

[ g(2 (3

(A34) (A35) >

(3 2g)w (2 g)g .

Thus under

K

(A36)

K

(A37)

If < (3(2 2g)w g)g , then nL = 0, nM 2 = 0. The best reponse of M is given by the expression (A24). Thus, its pro…t is given by (A25).

8

References

Almeida. R. (2003). The e¤ects of foreign owned …rms on the labor market. IZA Discussion Paper 785. Batra. G and Tan. H. (2002). Upgrading Work Force Skills to create HighPerforming Firms. In building Competitive Firms : Incentives and capabilities, in Nabi. I and Luthria. M (eds.), World Bank. Washington. D.C. De Santis R.A and Stähler F. (2004) Endogenous market structures and the gains from foreign direct investment. Journal of International Economics (article in press). Dunning. J. (1988). Explaining international production. London : Unwin Hyman. Fosfuri. A, Motta. M and Ronde. T. (2001). Foreign direct investment and spillovers through workers’ mobility. Journal of International Economics 53, 205-222. 20

Hanson. J.R. (1996) Human Capital and Direct Investment in Poor Countries, Explorations in Economic History 33, 86-102. Horstmamm. I.G and J.R Markusen. (1987) Strategic investments and the development of multinationals. International Economic Review 28,109-121. Horstmamm. I.G and J.R Markusen. (1991) Endogenous market structures in international trade. Journal of international Economics 32,109-129. Lucas. R. (1990) Why doesn’t capital ‡ow from rich to poor countries. American Economic Review 80, 92-96. Motta. M. (1992). Multinational …rms and the tari¤-Jumping argument. European Economic Review 36, 1557-1571. Motta. M. (1994). International trade and investments in a vertically di¤erentiated industry. International Journal of Industrial Organization 12, 179-196. Miyamoto. K and Todo. Y. (2002) Knowledge Di¤usion From Multinational Enterprises : The role of Domestic and foreign Knowledge-enhancing Activities, Technical paper 196, OECD Development Centre. Paris. Miyamoto. K (2003) Human capital formation and foreign direct investment in developing countries, Technical paper 211. OECD Development Centre. Paris. Nunnekamp.P. and Spatz. J. (2002) Determinants of FDI in Developing Countries : Has Globalization Changer the Rules of the Game, Unctad vol 2, n 2, August 2002. Nourbakhsh. F., Paloni A. and Youssef. A. (2001) Human capital and FDI to developing Countries : New empirical evidence, World Development 29, issue 9, 1593-1610. Root. F. and Ahmed.A. (1979) Empirical Determinants of Manufacturing Direct Foreign Investment in Developing Countries, Economic Development and Cultural Changes, 27, 751-767. Schneider. F. and Frey.B.(1985) Economic and political Determinants of Foreign Direct Investment, World Development13,161-175. Smith. A. (1987) Strategic investment, multinational corporations and trade policy. European Economic Review 31, 89-96. Zhang. K and Markusen. J. (1999) Vertical multinationals and host-country characteristics. Journal of Development Economics 59, 233-252.

21

CAHIERS DU LORIA Année 2002 2002-01

C. Chambolle et E.Giraud-Héraud

Certification de la qualité par une AOC : un modèle d’analyse.

2002-02

E. Giraud-Héraud, L.G. Soler et H. Tanguy

Concurrence internationale dans le secteur viticole : Quel avenir au modèle d’Appellation d’Origine Contrôlée ?

2002-03

L.G. Soler et H. Tanguy

Le fonctionnement des marchés entre vignoble et négoce dans le secteur des vins AOC : trois études régionales.

2002-04

C. Caron et J. Laye

Equilibre de Cournot-Nash avec contraintes de capacité : une formule explicite.

2002-05

O. Saulpic et H. Tanguy

Influence de la structure financière sur les choix stratégiques : étude de cas dans l’industrie du vin.

2002-06

C. Arnaud, E. Giraud-Héraud et A.Hammoudi

Harmonisation des taxes à la consommation : le cas des boissons alcoolisées.

2002-07

M. Mainsant et F. Porin

Les effets des politiques de promotion sur la fiabilité des outils d’évaluation des prix de détail.

2002-08

R. Green et L. Pierbattisti

Principales tendencias del mercado mundial de vinos.

2002-09

E. Giraud-Héraud, H. Hammoudi et M. Mokrane

Multiproduct Firm Behavior in a Differentiated Market

2002-10

M.L. Allain et C. Chambolle

Les relations entre la grande distribution et ses fournisseurs : bilan et limites de trente ans de régulation

2002-11

C. Chambolle

Stratégies de revente à perte et réglementation

2002-12

C. Chambolle

Faut-il interdire la revente à perte ?

2002-13

E. Giraud-Héraud, L. Rouached, L.G. Soler

Standards de qualité minimum et marques de distributeurs : un modèle d'analyse

2003-01

R. Green, M. Hy

Sécurité alimentaire et traçabilité

2003-02

R. Green, M. Rodriguez Zuniga et Leandro Pierbattisti

Global Market changes and business behavior in the wine sector

2003-03

E. Giraud-Héraud, J. Mathurin, L.G. Soler

Quelle légitimité à des mécanismes de régulation de l’offre dans les Appellations d’Origine Protégée ?

Année 2003

2003-04

C. Chambolle, E. GiraudHéraud

Certification of Origin as a Non-Tariff Barrier

2003-05

S. Poret

A Note on the Location of the Anti-Drug Law Enforcement Policy

2003-06

C. Chambolle, E. GiraudHéraud

Do Vertical Extension and Horizontal Mergers Interact

2003-07

P. Sans, G. de Fontguyon

L’industrie de transformation de la viande bovine en France : Une approche historique (1950-2003)

2003-08

J.A. Laye

Bilan économique du stockage en réserve qualitative de vin de Champagne

2003-09

C. Chambolle, L. Muniesa M.A Ravon

Concentrations Horizontales et Puissance d’Achat

2004-01

M.L. Allain, C. Chambolle

Below cost pricing laws as vertical restraints

2004-02

M.L. Allain, C. Chambolle

Forbidding Resale at a Loss : a Strategic Inflationary Mechanismv

2004-03

S. Poret

Les politiques publiques des drogues : vers une analyse économique

2004-04

S. Poret

L’échec de la guerre à la drogue : une revue de la littérature théorique

2004-05

H. Hammoudi, W. Hatit

Multinationals and the training of workers in the host – country : the tariff – jumping argument revisited

Année 2004

INRA – LORIA 65, Bd de Brandebourg 94205 IVRY Cedex (33) 1 49 59 69 87 [email protected]