Firm-specific advantages, multinational joint ventures, and host country tariff policy.

I. Introduction

In recent years multinational firms have shown a tendency to vertically disintegrate the production process into a number of plants across different countries. These firms have also shown an increasing willingness to form joint ventures with local firms. For example, some U.S.-Japanese joint ventures now assemble U.S. and Japanese autoparts in the U.S. to sell the final products in the North American market. The phenomenon is especially important in less developed countries (LDCS) where, in some cases, as much as 80 percent of total foreign capital may be in the form of joint ventures [9].

The literature on vertically integrated multinationals is voluminous (see Rugman and Eden [12] for a survey), but very few authors have explored the implications of vertically disintegrated joint ventures between an upstream input supplier and a downstream final goods producer. While Hirshleifer [8] analyzed the transfer price implications of a vertically integrated industry, Monteverde and Teece [11] were the first to point out that quasi-vertical integration may be a stable organizational form when appropriable quasi-rents exist between the downstream and the upstream firms.(1) Recently, Contractor and Lorange [4], Beamish [2], and Harrigan [7] have pointed out that quasi-vertical integration has evolved as an organizational form because both firms have some firm-specific advantages which make joint ventures mutually beneficial.(2) In the Japanese market, for example, some U.S. firms tend to collaborate with local Japanese partners simply because the Japanese firms have much better knowledge of the local distribution networks [9].(3)

This paper constructs a model of multinational operation which has the choice of forming a joint venture with a downstream firm in another country. I assume that the downstream "local" firm is interested in establishing an alliance with the multinational because the multinational has some firm specific advantages. More specifically, the multinational can produce an input in its upstream plant that is not available to the local firm. Similarly, the multinational also finds it advantageous to form an alliance with the local firm because the local firm has unique entrepreneurial knowledge of local conditions, offers cheaper inputs and has better ties to the government and the important buyers. Welfare implications of government tariff policy are explored and it is shown that under certain circumstances, an import-restricting tariff may indeed increase the welfare of the domestic consumers.

II. The Model

Assume that a multinational firm has an upstream plant located in Country 1. This plant produces a technologically complex input X with the help of skilled labor [L.sup.s] and other inputs not available in Country 2. With regard to Country 2, the firm now faces two choices:

(a) Produce good [Q.sub.m] in a facility in Country I and export [Q.sub.m] to Country 2. Country 2 has a high tariff barrier against all final goods imports.

(b) Produce [Q.sub.h] in a location inside Country 2 and sell it in Country 2's domestic market. In this case the multinational will have to export X to Country 2 to produce [Q.sub.h] We assume that X has a lower tariff than [Q.sub.m].

(c) If the firm chooses option (b), it must also decide whether or not to engage in a joint venture with a local firm. We assume that the government does not impose any restrictions on foreign equity holdings.

A crucial assumption of the model is that in all cases, the high technology intermediate inputs are produced by the multinational in a location outside Country 2's territory. Clearly, options (a) and (b) are not mutually exclusive. Assume that [Q.sub.m] and [Q.sub.h] are closely related on the production side (i.e., [Q.sub.m] stands for luxury cars and [Q.sub.h] stands for economy cars: both use similar intermediate inputs), but [Q.sub.m] has a less elastic demand compared to the demand for [Q.sub.h] in Country 2's domestic market.

In the case of (c), the possibility of a joint venture raises some conceptual issues about bargaining power within the joint venture. First of all, the joint venture must resolve the degree of ownership between the two firms. Existing literature does not shed much light on this issue. Most authors assume that the ownership decision is made by the government of the host country which sets limits to maximum ownership and implicitly assume that this constraint is binding. This need not necessarily be so. The multinational's decision to form a joint venture will stem from the costs and benefits of the project. In many cases the multinationals decide on minority, or 50-50 joint ventures even if there if no pre-set limit on the degree of ownership [9].(4)

A second, related, question concerns the joint venture's price-output decisions. The joint venture must decide on the price to set, the quantity [Q.sub.h] to produce, and determine the transfer price of the high technology intermediate input X. In a more general model these variables would be decided in a duopoly game; but for simplicity, this paper does not follow this route. I assume instead that the multinational decides the optimal level of joint venture ownership and the transfer price of the high technology intermediate input, while the local firm decides the price of final output.

This assumption is probably not too unrealistic because the local firm can never have effective control over the ownership decision. Suppose the local firm decides to own a higher share of the joint venture. Since the upstream multinational controls the transfer price, it can always manipulate the transfer price to control the accounting profits of the joint venture. The local firm thus may have higher nominal ownership, but total profits of the local firm are always determined by the transfer price set by the upstream firm. The local firm, however, is likely to have more intimate knowledge of the domestic market and would be able to more accurately judge the domestic demand and cost conditions and would be able to set the optimal output price.(5)

The formal model, therefore, is as follows [1]. Let the demand for [Q.sub.m] in Country 2's domestic market be

[Q.sub.m] = [P.sup.-[epsilon].sub.m], [epsilon] > 1. (1)

[Q.sub.m] is the quantity demanded of the importable good, [P.sub.m] the price of [Q.sub.m] and [epsilon] the elasticity of demand.

The demand for [Q.sub.h] in Country 2's domestic market is

[Q.sub.h] = [P.sup.-[eta].sub.h], [eta] > [epsilon] > 1. (2)

[Q.sub.h] is the quantity demanded of the domestic good produced by the joint venture, [P.sub.h] the price of [Q.sub.h] and [eta] the elasticity of demand. In Country 1, the high technology intermediate input X and other primary inputs [L.sup.*] are assumed to be combined in fixed proportions to produce [Q.sub.m]:

[Q.sub.m] = min{X,[L.sup.*]}. (3)

Similarly, [Q.sub.h] is produced with X and Country 2's primary input L:

[Q.sub.h] = min{X,L}. (4)

X is assumed to be a function of sector-specific skilled labor ([L.sup.s]) and other factors in Country 1:

X = f([L.sup.s],[multiplied by]). (5)

Let the prices of [L.sup.*] and L be denoted by and [W.sup.*] and W respectively, W < [W.sup.*]. Let [C.sub.[chi]] be the unit cost of inputs required in the production of X.

The multinational enjoys certain locational and organizational advantages if it decides to engage in a joint ienture [15]. If the multinational decides to produce in Country 2, it not only produces [Q.sub.h] with cheaper inputs, it can also reduce the cost of production of [Q.sub.m], through franchising and various other cost saving techniques.(6) Assume that these country-specific benefits increase as the multinational increases its ownership in industry [Q.sub.h]. Denoting the multinational ownership of the joint venture by [sigma], 0 [less than or equal to] [sigma] [less than or equal to] 1, the benefit function accruing to the multinational's production in industry [Q.sub.m] is defined as [B.sub.m]([sigma]), where

[Mathematical Expressions Omitted]

(6)

The joint venture with the local firm also generates firm-specific advantages,(7) but these advantages erode as the multinational assumes progressively more ownership. Denoting these benefits by [B.sub.h]([sigma])

[Mathematical Expressions Omitted]

(7)

Assume now that the government imposes unit taxes on imports of intermediate and final goods given by [T.sub.[chi]] and [T.sub.m] respectively. In view of(1)-(47), the multinational's global profit function is

[[pi].sub.g] = {[P.sub.m][Q.sub.m] - ([C.sub.[chi]] + [W.sup.*] - [B.sub.m] + [T.sum.m])[Q.sub.m]} + ([P.sub.[chi]] - [C.sub.[chi]])X + [sigma]{[P.sub.h][Q.sub.h] - ([P.sub.[chi]] + W + [T.sub.[chi]])[Q.sub.h] + [B.sub.h]/[sigma]} (8)

where [[pi].sub.g] is the global profit of the multinational and [P.sub.[chi] is the transfer price of X. The multinational's problem is to maximize (8) subject to (1) through (5). The first-order conditions imply

[Q.sub.m] = [{[epsilon]/([epsilon] - 1)}([C.sub.[chi]] + [W.sup.*] - [B.sub.m] + [T.sub.m])].sup.-[epsilon] (9)

[Mathematical Expressions Omitted]

(10)

[Mathematical Expressions Omitted]

(11) where [[pi].sub.j] is the profit of the joint venture. Notice that the multinational does not treat [Q.sub.h] as constant. It realizes that demand for X is a derived demand and uses the derivative property of Shephard's lemma to derive (10).(8) The local firm has the following profit function

(1 - [sigma]){[P.sub.h][Q.sub.h] - ([P.sub.[chi]] + W + [T.sub.[chi]])[Q.sub.h]}. (12)

The local firm maximizes (12) subject to (2). Its first-order condition is

[Q.sub.h] = [{[eta]/([eta] - 1)}([P.sub.[chi]] + W + [T.sub.[chi]])].sup.-[eta] (13)

Using (2), (10), (12) and (13) the transfer price [P.sub.[chi]] can be eliminated to derive the equilibrium values in terms of true costs of production

[Q.sub.h] = [{[[eta].sup.2]([C.sub.[chi]] + W + [T.sub.[chi]])}/{([eta] - 1)([eta] + [sigma] - 1)}].sup.-[eta] (14) [[pi].sub.j] = {1/([eta] - 1)}{[eta]/([eta] - 1)}.sup.-[eta][{[eta]([C.sub.[chi]] + W + [T.sub.[chi]])/([eta] + [sigma] - 1)}].sup.1-[eta] (15) Given the wage rates, the unit taxes and the elasticities, the system is completely determined. Note that [sigma] is now implicitly determined from equation (11). In general [sigma] need not take an extreme value of 1. The multinational will weigh the marginal benefits and costs of increasing [sigma] and decide on its optimal ownership level of the joint venture.

PROPOSITION 1. An increase in multinational ownership of joint venture increases total profit of the joint venture.

Proof. Differentiating equation (15) we get

d[[pi].sub.j]/d[sigma] = {[eta]/([eta] - 1)}.sup.-[eta][[eta]([C.sub.[chi]] + W + [T.sub.[chi]])].sup.1-[eta]([eta] + [sigma] - 1).sup.[eta]-2 (16)

which is positive in view of [eta] > 1. Q.E.D.

Proposition 1 is surprising. An intuitive explanation is that if the multinational increases its ownership of the joint venture, it is now in its interest to reduce the transfer price [P.sub.[chi]] and it is easy to show that [[ ]P.sub.[chi]]/[ ][sigma] < 0. Reduction in transfer price then increases the total profit of the joint venture [1].

III. Host Country Tariff Policy

An interesting implication of the model is that the government may now use tariff policy to increase Country 2's domestic welfare. Define welfare as the sum of consumer's surplus and producer's profits.(9) Assuming that the government makes a lump sum transfer of all tariff proceeds to domestic consumers, welfare is then

[Mathematical Expressions Omitted]

(17)

Country 2's welfare in the market for [Q.sub.h] is

[Mathematical Expressions Omitted]

(18)

Notice that only domestic retained profit of the joint venture has entered the welfare calculations.

PROPOSITION 2. An increase in [T.sub.m] - [T.sub.[chi]] will increase the multinational's ownership of the joint venture.

Proof. Totally differentiating equation (11), using (1) and (2), and rearranging terms we get(10)

[Mathematical Expressions Omitted]

(19)

In view of (1), (6), (7) and Proposition 1, (19) is positive. Q.E.D.

We can now explore the effect of an increase in [T.sub.m] - [T.sub.[chi]] on welfare. From (17) and (18) it can be shown that

d([[omega].sub.m] + [[omega].sub.h]])/[dT.sub.m] = d[[omega].sub.m]]/[dT.sub.m] + {[[eta].sup.2]/([[eta] - 1)}.sup.1-[eta]]([[C.sub.[chi]] + W + [T.sub.[chi]]).sup.1-[eta]([[eta] + [sigma] = 1)].sup.[eta]-2(1 - [sigma]d[sigma]/[dT.sub.m]. (20)

The first term may be negative, but from proposition 2, the second term is always positive. Thus the expression above is positive if the beneficial effect of further vertical integration outweighs the possible negative welfare effects in the first market,[[ ][omega].sub.m]/[[ ]T.sub.m]. This is likely to happen at a certain range of the [B.sub.[iota]([sigma]) functions. We have thus shown that a country may gain by creating tariff barriers for importable final goods. Tariff barriers now force the multinational to reduce the transfer price, which tends to increase Country 2's welfare.(11) From (19) and (20) one can implicity derive an optimal tariff rate for Country 2.

IV. Concluding Comments

The model shows that firm-specific and country-specific advantages may determine the extent of foreign ownership in a joint venture firm. This may explain why the same multinational may form joint ventures with different degrees of ownerships across different countries and over time. The model also shows the possibility of a beneficial tariff policy that increases domestic welfare by constructing "tariff-walls" that encourage the multinationals to increase ownership and reduce the transfer price. Formation of Customs Unions such as the ones in Europe and North America may also have the same effect. Since the multinationals are known to charge a high transfer price, a country may impose tariff simply to lower the transfer price. The model is especially relevant for LDCs where the local firms are unable to produce complex high-technology inputs, but have an excellent knowledge of local conditions to market the jointly produced commodity.

References

[1.] Bardhan, Pranab K., "Imports, Domestic Production and Transnational Vertical Integration: A Theoretical Note." Journal of Political Economy, October 1982, 1020-34. [2.] Beamish, Paul W., "Joint Ventures in LDCs: Partner Selection and Performance." Management International Review, First Quarter 1987, 23-37. [3.] Brander, James A. and Barbara J. Spencer, "Tariffs and the Extraction of Foreign Monopoly Rents Under Potential Entry." Canadian Journal of Economics, August 1981, 371-89. [4.] Contractor, Farok J. and Peter Lorange, "Competition vs. Cooperation: A Benefit/Cost Framework for Choosing Between Fully-Owned Investments and Cooperative Relationships." Management international Review, Special Issue, 1988, 5-18. [5.] Falvey, Rodney E., and Harold O. Fried, "National Ownership Requirements and Transfer Pricing." Journal of Development Economics, December 1986, 249-54. [6.] Grossman, Gene M., "The Theory of Domestic Content Protection and Content Preference." The Quarterly Journal of Economics, November 1981, 583-603. [7.] Harrigan, Kathryn R., "Strategic Alliances and Partner Asymmetries." Managenwnt International Review, Special Issue, 1988, 53-72. [8.] Hirshleifer, Jack, "On the Economics of Transfer Pricing." Journal of Business, July 1956, 172-84. [9.] Korbin, Stephen J., "Trends in Ownerships of American Manufacturing Subsidiaries in Developing Countries: An Inter-Industry Analysis." Management International Review, Special Issue, 1988. [10.] McConnell, John J. and Timothy J. Nantell, "Corporate Combinations and Common Stock Returns: The Case of Joint Ventures." Journal of Finance, June 1985, 519-36. [11.] Monteverde, Kirk and David J. Teece. "Appropriable Rents and Quasi-Vertical Integration." Journal of Law and Economics, October, 1982, 321-28. [12.] Rugman, Alan M. and Lorraine Eden, eds. Multinationals and Transfer Pricing. New York: St. Martin's Press, 1985. [13.] Silva-Echenique, Julio, "Quasi-Vertical Integration and Rate-of-Return Regulation." Canadian Journal of Economics. November 1989, 852-66. [14.] Svejnar, Jan and Stephen Smith, The Economics of Joint Ventures in Less Developed Countries." The Quarterly Journal of Economics, February 1984, 149-67. [15.] Tabeta, Naoki. "A Model of Quasi-Vertical Integration in the Japanese Automobile Industry." Working Paper, Takachiho Commercial College, Tokyo, Japan, June 1991.

(*) I would like to thank an anonymous referee for extremely useful comments on an earlier version of the paper. (1.) Silva-Echenique [13] discusses the issues related to rate of return regulation for upstream and downstream firms, but assumes a predetermined level of indigenous ownership. (2.) Tabeta [15] uses a quasi-vertical model to explain the Japanese subcontracting system in the auto industry. Svejnar and Smith [14] use a Nash bargaining framework to explain joint ventures but ignore the issue of firm specific costs and benefits. (3.) McConnell and Nantell [10] show that for a sample of 210 firms listed in New York and American Stock Exchanges, firms with joint ventures had significant increases in their share values. (4.) In some cases, of course, government-stipulated maximum-ownership limits or the domestic content laws may be a binding constraint [6; 5]. Ownership or content laws can be treated as a special case of our model. (5.) Svejnar and Smith [14] assume that decisions are made jointly; but their approach does not shed light on the determinants of bargaining power. I assume that the bargaining power is based on firm-specific advantages. Note that in this model, the domestic firm has no other means of obtaining X except from the foreign firm. Since the domestic firm can not enter the market without foreign collaboration, it has considerably less bargaining power. Firm-specific advantages also enable the local firm to retain some of its profits. If firm-specific advantages do not exist for the local firm, the upstream firm will wholly own the downstream firm and appropriate all profits [13; 15]. (6.) These benefits reduce "transaction costs" as well as production costs in Country 1. Lower wage, for example, may be a reason why a multinational may want to operate in an LDC [4]. (7.) In a survey of multinational executives, Beamish [2] found that the local partner's knowledge of local business, economy, politics and customs were the most significant perceived benefits of collaboration. (8.) This has been pointed out by Bardhan [1]. (9.) Bardhan [1] uses the same measure for a licensed firm. (10.) Equations (19) and (20) are derived by indexing [T.sub.m] to zero. We assume that [[ ]B.sub.m]/[ ][sigma] is sufficiently large.