# Expropriation of multinational firms: the role of domestic market conditions and domestic rivalries.

EXPROPRIATION OF MULTINATIONAL FIRMS: THE ROLE OF DOMESTIC MARKET
CONDITIONS AND DOMESTIC RIVALRIES

HAMID MOHTADI(*)

The causes of expropriation of Multinational Firms (MNFs) by their host governments are investigated, based on social welfare analysis and also on public choice theory. A key feature is the presence of Cournot-Nash rivalry with domestic firms. Thus, the likelihood of expropriation differs according to whether or not a domestic rival exists. The likelihood also depends on the MNF's initial technological superiority over domestic technology, the host country's demand structure, the type of control to be exercised over the expopriated firm (private vs. state) and the strategic behavior of the MNF, ex anti, when threat of expropriation.

I. INTRODUCTION

Evidence over the past thirty years points to widespread expropriation of multinational firms (MNFs), particularly by the less developed countries (LDCs).(1) Yet, expropriation has received far less attention than impediments to commodity trade (tariffs, quotas, etc.). One exception is Eaton and Gersovitz [1984] (EG), who focus primarily on the impact of an expropriation threat on international capital mobility and the response of foreign capital to such a threat. However, even allowing for a strategic response to a threat of expropriation, the issue of why governments engage in the actual expropriation of existing firms deserves attention.(2) Explaining expropriation of MNFs already in the host country is the task of this paper.

While the host government's motives for expropriation may vary, as suggested by Akinsanya [1980], this paper shall focus on economic motives, based on social welfare analysis, and on political-economic motives, based on public choice (lobbying) theory. A key feature will be an emphasis on domestic vs. foreign firm rivalry, captured by a Cournot-Nash duopoly mode, which is aimed at highlighting the potential conflict of interest between the two firms and its effects on the host government's expropriation decision.(3) To highlight the role of rivalry (section 3), a simple monopoly model of expropriation will be developed in section 2.

A second key feature will be to distinguish between the host government's expropriation and ownership decisions, regarding the expropriated firm, since chances of expropriation may depend on how the ownership issue is settled. Two cases will be distinguished: (a) the expropriated firm is turned over to domestic residents and the firm continues to maximize profits (sections 2 and 3); and (b) the state controls the expropriated firm (i.e. nationalizes it) (section 4).(4) In this case, the government is likely to pursue an objective different from profit maximization (see Aharoni [1986, ch. 4]).

Following Caves [1982, chs. 1,7], the MNF will be assumed to possess technology superior to that domestically available because of access to intangible assets (organizational and managerial know-how, specific technical knowledge, etc.), not available to domestic firms.(5) Unlike physical capital, intangible assets may be withdrawn in the event of an expropriation threat, as in EG's model, or actual expropriation, as in the present model, resulting in increased production costs to the expropriated firm.(6)

In this paper, expropriation in the monopoly model will be motivated by the host country's desire to capture MNF profits, otherwise repatriated, to increase welfare. But, this welfare gain may be offset by a loss to consumers, via the impact on costs and prices of the MNF's withdrawal of its technology. (Under state ownership some of these results will be altered.) In the duopoly model, the cost increase to the expropriated firm will raise the domestic rival's profits, providing it with an incentive to lobby for expropriation. This raises the opportunity cost of not lobbying and thus the incentive to lobby by other competing groups (MNFs and consumers), suggesting that expropriation may be the outcome of lobbying forces, rather than of pure welfare considerations.

II. EXPROPRIATION OF A MONOPOLY UNDER PRIVATE-DOMESTIC CONTROL

Consider a monopolistic and locally profit-maximizing MNF producing in and for the host country market.(7) Before expropriation, social welfare equals consumer surplus, as MNF profits are repatriated. The motive for expropriation is to raise social welfare by internalizing the MNF's profits.(8) But expropriation causes the MNF to withdraw its intangible assets, increasing production costs of the expropriated firm. Assuming the firm's ownership is passed to the domestic private sector (section 4 analyzes expropriation under state control), the firm remains a profit maximizer, and the cost increase translates into a price increase, reducing output and consumer surplus.

Figure 1 illustrates the trade-off between expropriated profits-gain and the loss of consumer surplus. This is done by comparing the net welfare gain (loss) (a) under a steeper and a flatter demand, D and d respectively (where d is pivoted through the intersection of [MC.sub.0] and D for expositional purposes), but subject to the same marginal cost shift, [MC.sub.0] to [MC.sub.1], caused by expropriation and, (b) under two different size marginal cost shifts [MC.sub.0] to [MC.sub.1] vs. [MC.sub.0'] to [MC.sub.1], upon expropriation (reflecting the MNF's different initial technological advantage, [MC.sub.0] vs. [MC.sub.0'], but subject to the same demand, D. The figure yields the following intuitive result:

RESULT 1. The welfare gains of expropriation vary inversely with the flatness of demand as well as the initial technological advantage of the MNF. Thus the likelihood of expropriation increases as either factor declines.

To explain, the profits-gain of expropriation are larger, and the drop in consumer surplus is smaller, the steeper is the demand, given a cost increase, or the smaller is the cost increase, given a demand. (Compare, as shown, the non-overlapping parts of expropriated profits [the rectangles] and the change of consumer surplus [the triangles].) Thus, monopolistic MNFs facing steep demand and operating in industries with little technological advantage over domestic technology are the more likely candidates for expropriation.

Strategic Behavior Under Expropriation Threat

If expropriation is not a complete surprise, the MNF may be able to influence (reduce) the chances of its occurrence, ex anti, by altering its output to reduce the host government's potential welfare gains, behaving strategically as a Stackelberg leader vis-a-vis the host country government. Let [PI.sub.0] and [PI.sup.s.sub.0] be the MNF's pre-expropriation profits in the cases of non-strategic and strategic behavior, respectively (O = pre-expropriation status). Then

[PI.sub.0] = P([x.sub.0]) [x.sub.0] - [C.sub.0] ([x.sub.0]) (1) and

[PI.sup.s.sub.0] = ([rho] [PI.sub.0]) = (1- [rho]) [PI.sub.0]

where [chi.sub.0] = pre-expropriation output, P([chi.sub.0)] = inverse demand, [C.sub.0] = pre-expropriation cost, and [rho] = probability of being expropriated. Thus [rho][PI.sub.0] is the expected loss from expropriation, appearing as an extra cost to the MNF. Expropriation is more likely the higher are its potential welfare gains, [DELTA]W, relative to the initial welfare, [W.sub.0], i.e., [rho] = [rho] ([DELTAW/W.sub.0]) with [rho'](*) > 0.(9) In general, [DELTAW/W.sub.0] falls with a rise in [chi.sub.0]. This is because [W.sub.0], which equals pre-expropriation consumer surplus, [S.sub.0], rises with higher [chi.sub.0], so that [DELTA]W = [W.sub.1] - [W.sub.0] (where 1 = post-expropriation status), and from there [DELTAW/W.sub.0] both fall with higher [chi.sub.0]. Hence, [Mathematical Expression Omitted]. A strategically acting MNF no longer maximizes [PI.sub.0] but [Mathematical Expression Omitted]. Denoting the relevant output by [Mathematical Expression Omitted], the first-order condition yields (from the equation for [Mathematical Expression Omitted] and after, rearrangement):

[Mathematical Expression Omitted]

In equation 3, since [Mathematical Expression Omitted] is the product of a positive and a negative term, it is < 0. Moreover, since [rho] < 1, it follows that [Mathematical Expression Omitted], implying that [Mathematical Expression Omitted] is located on the downward portion of the strictly concave profit function [PI.sub.0]([x.sub.0]), i.e. past the optimum output that would have maximized [PI.sub.0]. Denoting the latter output by [Mathematical Expression Omitted] (ns = non-strategic), [Mathematical Expression Omitted] and thus [Mathematical Expression Omitted]. Moreover, at [Mathematical Expression Omitted] the second partial of [Mathematical Expression Omitted] is negative so that the second-order condition is also satisfied for [Mathematical Expression Omitted] Hence:

RESULT 1a. Under an expropriation threat strategic behavior causes a monopolistic MNF to produce more than it would in the absence of such a threat. This reduces the host country's potential welfare gains from expropriation and thus the chance of its occurrence.

III. EXPROPRIATION OF A DUOPOLY UNDER PRIVATE-DOMESTIC CONTROL

By introducing a domestic rival, this section analyzes expropriation in the context of Cournot-duopoly rivalry.(11) Figure 2 shows the reaction functions, iso-profit curves, and equilibrium outputs of the domestic (foreign) firm at the initial (pre-expropriation) state by [R.sub.D(F)], [PI.sub.D(F) and [x.sub.D0(F0)], respectively, with equilibrium at [EPSILON.0]. Expropriation raises marginal costs to the expropriated MNF as it withdraws its intangibles, thus equalizing foreign and domestic marginal costs. In Figure 2 the MNF's iso-profit curve then shift right (to [PI.sub.F']) and its reaction function shifts down (to [R.sub.F']); for its domestic rival the iso-profit curve shifts down (to [PI.sub.D']). Thus the expropriated firm's market share falls and that of the domestic rival rises.(12) It is easy to show that the domestic rival's profits depend positively on the cost level of its foreign rival,(13) and since the latter rises under expropriation, the former must also rise, i.e.:

[Delta PI.sub.D] = [PI.sub.D1] - [PI.sub.D0] > 0.

(4)

RESULT 2. Expropriation of a MNF which is a Cournot-Duopoly rival to a local firm increases the latter's profits.

Interestingly, a comparable result by Davidson et al. [1985] shows that the imposition of local content requirements on a MNF raises its cost and therefore its domestic rival's profits. One may thus think of expropriation as a particular form of state intervention in a continuum in which a local content requirement is another.

This potential increase in the domestic firm's profits may cause it to lobby in support of expropriation. This increases, relative to the monopoly model, the likelihood of the MNF lobbying against expropriation, as the opportunity cost of not doing so is now higher. Also, consumers facing higher prices may lobby against expropriation, to protect the consumer surplus, and in favor of it to gain part of expropriated profits (see below). Since lobbying is more likely than in a monopoly model, an expropriation decision will be analyzed as an outcome of interest group lobbying rather than of social welfare considerations. The approach followed here is similar to Cassing and Hillman's [1985] analysis of trade policies where a political objective function of producer and consumer lobbying is maximized by the government.(14) Two cases will be considered: first, lobbying occurs only among domestic interest groups (consumers and the domestic firm); and second, the MNF also lobbies.

Case a: Lobbying Domestic Consumers and Producers Only

In this case the tradeoff occurs between the loss of consumer surplus ([DELTA]S < 0) and the rise in the domestic firm's profits ([DELTA PI.sub.D] > 0). To counter the consumers' loss (and thus their opposition to expropriation), the government, which is here assumed to turn the expropriated firm over to the domestic private sector, transfers part of the expropriated profits to consumers (e.g., by taxing its new owners). Denote such profits by [PI.sub.F1] and their allocation to consumers and producers by [R.sub.1] and [R.sub.2]. Then [PI.sub.F1] = [R.sub.1] + [R.sub.2]. Producer groups as a whole lobby in favor of expropriation to the extent of the increase in own domestic profits (result 2) plus their share of the expropriated firm's profits, [DELTA PI.sub.D] + [R.sub.2]; consumers lobby against expropriation to the extent of their net loss if [DELTA]S + [R.sub.1] < 0, or in favor of it to the extent of their net gain if [DELTA]S + [R.sub.1] > . The government chooses [R.sub.2] chooses [R.sub.1] and [R.sub.2] to maximize its political support function among the two groups, subject to the above constraint: [Mathematical Expression Omitted]

where M is the political support function. Assuming that M is strictly quasi-concave, as in Cassing and Hillman [1985], the first-order condition [Mathematical Expression Omitted] yields optimum R, say R(*), where [M.sub.1](R*) = [M.sub.2](R*). In Figure 3, the horizontal axis depicts consumer lobbying for or against expropriation as [DELTA]S + [R.sub.1] > 0 (first quadrant) or <0 (second quadrant); the vertical axis depicts producer lobbying for expropriation, proportional to [DELTA PI.sub.D] + [R.sub.2]. Maximization yields the tangency of the M indifference curves and the revenue constraint lines [R.sub.1] + [R.sub.2] = [PI.sub.F1], which in the [DELTA]S] + [R.sub.1], [DELTA][PI.sub.D] + [R.sub.2] space are re-expressed as ([DELTA]S] + [R.sub.1]) + ([DELTA PI.sub.D] + [R.sub.2]) space as re-expressed as

([DELTA]S + [R.sub.1]) + ([DELTA PI.sub.D] = [DELTA]S] + [DELTA PI.sub.D] + [PI.sub.F1].

Both the M contours and the revenue constraints may of course cross into the second quadrant as [DELTA]S] + [R.sub.1] may be > 0 or < 0. The special M=0 contour, however, passes through the origin since M(0,0)=0. Points to the northeast of this contour imply a positive likelihood of expropriation as M>0 in this region.

To determine a specific revenue constraint line and its relevant tangency in Figure 3, [DELTA]S], [DELTA PI.sub.D] and [PI.sub.F1] in the government's maximization decision (eqs. 5 and 6) must be known. These in turn depend on the extent of the cost increase to the expropriated MNF. Assume a linear demand, P=a-bx (a,b= constant), and constant marginal costs, [C.sub.D] and [C.sub.F] (D= domestic, F= foreign), and let the MNF's cost advantage before expropriation be [epsilon] such that [C.sub.DO] = [epsilon] [C.sub.FO] where [epsilon] >1. Then the removal of this cost advantage after expropriation means equality of the two costs, [C.sub.F1] = [C.sub.D1] (0 (1)= pre (post-expropriation values). Once equilibrium duopoly profits for each firm and consumer surplus before and after expropriation are determined, one can easily show that:

[Mathematical Expression Omitted]

Intuitively, the higher the MNF's technological edge, the greater is the cost increase upon its expropriation, and the larger the are the consumers' loss and domestic firm's profit gain. As for [PI.sub.F1], higher costs to the expropriated firm raise the price, reducing total market size, and thus its output and profits. Figure 4 shows that [DELTA PI.sub.D] and [PI.sub.F1] are positive and convex while [DELTA] S] is negative and concave in [epsilon],(16) all starting at [epsilon] = 1 (as the MNF is at least as efficient as its local rival) and reaching [epsilon.sub.max] (not shown here).(17) When [epsilon] = 1, no consumer losses nor increased domestic profits occur and [DELTA][PI.sub.D] = [DELTA]S = 0; but the expropriated firm's profits [PI.sub.F1] will be >0. In Figure 4 the four curves, [DELTA PI.sub.D], [DELTA]S, [DELTA PI.sub.D] + [PI.sub.F1], and [DELTA]S + [PHI.SUB.F1], are labeled P, S, PP and SS, respectively.

To assess the impact of lobbying on the likelihood of expropriation, information from Figures 3 and 4 are combined. For each given value of [epsilon] the height of curves S, P, SS and PP in Figure 4 identify a specific revenue constraint line in Figure 3. At the upper-left end-point of such a line all expropriated profits accrue to producers and none to consumers ([R.sub.1] = 0 and [R.sub.2] = [PI.sub.F1]), yielding the coordinate values of [DELTA]S and [DELTA PI] + [PI.sub.F1]. Thus to find this point the height of the S and PP curves (for a given [epsilon]) are traced from Figure 4 over to Figure 3. Similarly, at the lower-right end-point of this line all expropriated profits accrue to consumers and none to producers ([R.sub.1] = [PI.sub.F1] and [R.sub.2] = 0), yielding the coordinate values of [DELTA]S + [PI.sub.F1] and [DELTA PI.sub.D]. These are then traced from the curves SS and P in Figure 4.

In Figure 3 four revenue constraint lines and four tangencies are shown, based on arbitrary values of [epsilon.sub.0] = 1, [epsilon.sub.1], [epsilon.sub.2] and [epsilon.sub.3] in increasing order. For a typical family of indifference curves, as [epsilon] increases the expansion path connects tangencies of higher indifference curves, suggesting higher probability of expropriation. (Note that all tangencies are in the M > 0 region.) Though one may visualize how an extremely steep family of indifference curves might yield tangencies that move in the opposite direction as [epsilon] increases (with a tangency at first for [M.sub.0], but then changing to corner solutions that touch the lower-right end-points of the revenue constraint lines for other contours), such an outcome is unlikely, as it requires that consumer lobbying be far more effective than producer lobbying ([M.sub.1] >> [M.sub.2]) which is not usually the case.(18) To summarize:

RESULT 3. Excluding the unusual case of a much greater effectiveness of consumer lobbying relative to producers, expropriation by the host government is more likely, the larger is the initial technological advantage of the MNF.

Thus, in contrast to result 1, when a domestic rival exists whose lobbying power exceeds that of consumers, a higher technological superiority of the MNF implies greater chances of expropriation, rather than less. This reversal of result 1 arises as the higher technological advantage of the MNF implies higher potential domestic profits that would accrue from its expropriation, and thus more intensive lobbying by the domestic firm in favor of such an action.

Case b. Lobbying by the MNF

If the MNF lobbies, this lobbying would be against expropriation, and proportional to the size of its initial profits, [PI.sub.F0], lost upon expropriation. Then the function M (eq. 5) is modified as:

Max M = M([DELTA]S + [R.sub.1], [DELTA][PI.sub.D] + [R.sub.2], - [PI.sub.F0])

([M.sub.1],[M.sub.2],[M.sub.3] > 0),

[R.sub.1], [R.sub.2]

subject to [R.sub.1] + [R.sub.2] = [PI.sub.F1].

Note that [M.sub.3] > 0, but that [Mathematical Expression Omitted], which indicates that lobbying occurs against expropriation. With foreign lobbying the M = 0 contours shifts out, starting now at a positive vertical intercept, rather than at the origin. This is because M = 0 is now consistent with no consumer lobbying but with positive domestic firm lobbying just counteracting the negative MNF lobbying [i.e. M(0, [DELTA PI.sub.D] + [R.sub.2], - [PI.sub.F0]) = 0, for some [DELTA PI.sub.D] and [PI.sub.F0]]. Since the region of M > 0 would then contract and that of M < 0 expand (not shown here), lobbying by MNF reduces chances of expropriation, ceteris paribus, as expected.

RESULT 3a. Lobbying by the MNF against expropriation reduces the chances of expropriation, ceteris paribus.

Welfare Comparison of Monopoly and Duopoly Results

For a linear demand and constant costs this comparison shows that

[Mathematical Expression Omitted]

and

[Mathematical Expression Omitted]

Thus when [epsilon] is large,

[Mathematical Expression Omitted]

RESULT 4. Expropriation of a MNF with a high initial technological advantage (large [epsilon]) yields a greater welfare gain (smaller loss) when there is a domestic rival than when it operates as a monopoly in the host country.

To explain, when expropriation raises costs, duopolistic rivalry prevents output from falling and prices from rising by as much as they would under monopoly. Thus consumer surplus falls by less. Moreover, although expropriated profits are smaller in the presence of a rival, this effect is more than offset by the rise in the domestic firm's own profits when the expropriated firm suffers a large cost increase.(20)

Finally, comparing the welfare effects of expropriation and local content requirements as in Davidson et al. [1985], one finds that for an equivalent cost increase faced by the MNF under both schemes, expropriation yields a larger welfare gain (smaller loss) due to the added term, [PI.sub.F1] (the local firm's profit increase and the consumers' loss will remain the same in both cases).

Strategic Behavior Under Expropriation Threat

In the monopoly model, strategic behavior under an expropriation threat implied overproduction by the MNF to reduce chances of expropriation, via increasing consumer surplus, and thus increasing consumers' potential loss if expropriation occurred. Under duopoly, overproduction by the MNF also increases total output and thus consumer surplus. Additionally, however, it would reduce the domestic rival's pre-expropriation market shares and profits, increasing the latter's potential gain ([DELTA PI.sub.D] ) by expropriation. Weighing the two impacts, the outcome is ambiguous to the MNF, since overproduction now intensifies both the domestic firm's lobbying in support of and consumer lobbying against expropriation. To resolve the ambiguity, the MNF must know, a priori, the relative lobbying power of each group. If, as is often the case, local producers from a stronger lobby than consumers, then the MNF strategy would call for reduced pre-expropriation output, since the avoidance of a fall in the local firm's pre-expropriation profits ([PI.sub.D0]) would be more important to the MNF than inducing a larger pre-expropriation consumer surplus. Otherwise the reverse holds.

IV. EXPROPRIATION UNDER STATE CONTROL

Expropriation is frequently observed to accompany state ownership, i.e., nationalization, as discussed by Sigmund [1980, ch. 7]. Assessing the chances of expropriation when nationalization is an option requires a comparison with the case of expropriation under private-domestic ownership. For simplicity this comparison is made for a monopolistic MNF.(22) State ownership raises two new issues. First, the objective function of a state-run firm differs from that of a private firm. While several possibilities exist,(23) the more conventional assumption that the nationalized firm maximizes social welfare is pursued here. Second, there may exist an additional cost increase due to the state operation of the firm. A survey by Aharoni [1986] shows that this may only (but not necessarily) occur in the absence of competition from similar private firms, implying that in the present monopoly model state-run firms may experience costs either higher or the same as their private counterparts, a point which will be utilized shortly.

In Figure 5, with linear demand and generalized cost curves, D is demand, [MC.sub.0] and [MC.sub.n] are the pre- and post-nationalization marginal costs, and [AC.sub.n] is the post-nationalization average cost. Since social welfare is consumer surplus and profits, its maximum, when the firm is nationalized, occurs when output is [x.sub.n] and price is [P.sub.n] (= minimum of [AC.sub.n]). Maximum welfare is then the entire triangle [ARP.sub.n], yielding zero profits, a finding loosely corroborating the evidence that state-run enterprises operate at non-positive profits.(24)

Further, compared with an equivalent cost increase under the alternative scheme of expropriation and private ownership (also shown by [AC.sub.n] and [MC.sub.n]), nationalization yields larger optimum output ([x.sub.n]) than private ownership ([x.sub.1]), and thus also higher social welfare (area of [ARP.sub.n]) than private ownership (area of ABGC), as the usual "dead-weight loss" vanishes under nationalization.

However, cost increase need not be the same, as argued above. With an added cost increase owing to state operation, post-nationalization costs exceed [AC.sub.n] and [MC.sub.n]. There then exists a threshold level of such costs, i.e. [AC.sub.n]' and [MC.sub.n]' in Figure 5, which gives the same welfare as expropriation under private ownership. It follows that the areas of AR'P.sub.n'] and ABGC must equal, and thus the added welfare loss of the higher-cost state-run firm should just match the gain due to no deadweight loss. For smaller cost increases nationalization yields more welfare than expropriation with private ownership, and thus more likely; for larger cost increases the reverse holds. Thus:

RESULT 5. If state and privately-run firms are equally efficient, then welfare analysis would favor nationalization (where the nationalized firm maximizes social welfare); if state-run firms are less efficient, a threshold efficiency level exists below which nationalization (in the sense above), and above which expropriation under domestic-private ownership, would be favored.

Finally, since the pre-expropriation welfare level in this case is the same as the consumer surplus, strategic behavior by the MNF would yield the same outcome as in the monopoly case, requiring overproduction, ex anti, as in Result 1a.

V. SUMMARY AND CONCLUSION

Although a subject of many studies in international law, the issue of expropriation has received only scant attention in the economic literature. One exception is Eaton and Gersovitz [1984], though their study primarily focuses on the impact of an expropriation threat on international capital mobility. However, even allowing for a strategic response by the MNF to a threat of expropriation, the issue of why governments engage in the actual expropriation of existing firms deserves attention.

Expropriation occurs either to increase social welfare or because of lobbying influences, when such lobbying is presumed likely to exist. The chances of expropriation, in either scheme, are found to depend on the MNF's initial technological edge over domestic technology, the host country market's demand structure, the presence or absence of domestic rivals, the type of control over the expropriated firm (private vs. state), and the strategic behavior of MNF, ex anti, when under the threat of expropriation.

Further extensions might include extractive export-oriented MNFs, as the evidence suggests frequent past expropriation of such firms (see Caves [1982, 122-123]). Owing to its distinct market and organizational features, however, modeling expropriation in this case deserves separate treatment.(25) (*) Assistant Professor University of Wisconsin-Milwaukee and Graduate School of International Studies, University of Denver. Earlier versions of this paper were presented at the University of Wisconsin-Madison, University of Wisconsin-Milwaukee, University of Colorado-Boulder, Atlantic Economic Association and Western Economic Association. I am most indebted to an anonymous referee for his insightful comments and to Satya Das, John Heywood, William Kaempfer, and Randall Crane for their feedback on the earlier drafts.

(1.) Williams [1975] reports that during 1956-72, 25 percent of all foreign investments in LDCs were expropriated. By 1974, according to Sigmund [1980,7], U.S. estimates showed 106 investment disputes in 39 countries with $3.5 billion of a total of $25 billion.

(2.) The entry of a rational MNF in the face of an expropriation threat may occur either in the absence of full information, i.e. if expropriation was a "surprise", or if the expectation of its occurrence, when taken into account by the MNF, yielded a higher expected profit than was attainable elsewhere.

(3.) The use of a duopolistic framework also suggests greater realism, as in Knickerbocker [1973], and generality than the monopolistic or perfectly competitive models, and follows the broadening of trade models to oligopolistic behavior as done by Dixit [1983].

(4.) For much evidence on this type of expropriation see Sigmund [1980, ch. 8].

(5.) Differential access stems from the absence of markets, as transaction costs drive a wedge between the buyer's and the seller's price (see Williamson [1975]). Caves [1982, 4-5] outlines such costs for the MNFs.

(6.) The host government may restore the technological superiority of the MNF by licensing that technology. To the owners of the intangibles, licensing is an alternative to the "arms-length" markets of the MNF, to capture the Ricardian rents from their ownership. Such rents are not entirely bid away (owing to the market failure discussed above) and may be captured by a nonzero licensing fee. Any positive licensing fee would raise production costs, when defined exclusive of any quasi rent gains captured from expropriation, to the expropriated firm. These quasi rents are then treated separately. I wish to thank an anonymous referee for clarifying this issue. (For further discussion of licensing costs see Buckley and Casson [1985, 104-105].)

(7.) Production for exports is ignored, but an extensive analysis of expropriation of export-oriented MNFs is available in Mohtadi [1986]. (See also footnote 25.)

(8.) There may also be onetime compensation payments to the parent company, usually based on book or other contractual value of the expropriated assets, as discussed by Sigmund [1980, 265-66]. As fixed costs, their inclusion will not affect the analysis, and is thus ignored.

(9.) Note that dividing by [W.sub.0] is only for normalization convenience (allowing the argument [DELTA W/W.sub.0 and its function [rho] to both be expressed in pure numbers), and the analysis would be identical, though more cumbersome, with [rho(Delta W.sub.0)] instead.

(10.) From eq. 2 we get:

[Mathematical Expression Omitted]

Since the non-strategic profit function is concave, the first term on the right hand side is negative. As for the second term, both [Mathematical Expression Omitted] and [Mathematical Expression Omitted] are negative as discussed above. Finally, a [Mathematical Expression Omitted] must be [greater than] 0 (i.e. [rho] must be in [chi.sub.0]), if [rho] is a well-behaved probability function, starting at 1 and gradually approaching 0, as [chi.sub.0] increases. Letting [DELTA]W/W.sub.0 = [omega], this is satisfied

[Mathematical Expression Omitted]

(11.) Alternatively, a Bertrand-Duopoly model is inconsistent when products are homogeneous and costs are different, as discussed by Friedman [1983, 48], since it yields a discontinuous output for the MNF at the lowest possible price, say [rho*], of its rival (from zero at slightly below [rho*] to the entire market at slightly above [rho*]. Thus the MNF can force its domestic rival out of the market by underselling below the latter's cost, violating the coexistence assumption.

(12.) However, overall output must fall. Thus the drop in MNF output exceeds the rise in domestic output, in conformity with Figure 2.

(13.) To show this, first write the Cournot profit equations:

[Mathematical Expression Omitted]

where D = Domestic and F = Foreign. Assume, for simplicity, but without loss of generality, constant marginal costs, [c.sub.i], so that, [C.sub.i] (x.sub.i) = c.sub.i][x.sub.i, (i = D, F). Since expropriation raises cf, we differentiate the domestic firm's maximized profit function, [PI(*).sub.D], in cf and use the Envelope Theorem:

[Mathematical Expression Omitted]

The negative sign of the first term follows from differentiating (a) (written for the domestic firm) with respect to [chi.sub.F]; and that of the second term, from the MNF's optimum output,

[Mathematical Expression Omitted].

(14.) See also Kaempfer, Willett and McClure [1989], for extensions of Cassing and Hillman. For other analyses of lobbying as a determinant of trade policy outcomes, see Baldwin [1982], Findlay & Wellisz [1982; 1986] and Mayer [1984].

(15.) For the above cost and linear demand parameters:

[Mathematical Expression Omitted], and

[Mathematical Expression Omitted]. From these it follows that: [Mathematical Expression Omitted], and [Mathematical Expression Omitted]. Now, positive output for both firms requires that a>(2.[epsilon] - 1)cfo (see footnote 17). Using this constraint it follows that [Mathematical Expression Omitted] and [Mathematical Expression Omitted].

(16.) From footnote 15, [[Mathematical Expression Omitted] ([epsilon][epsilon] implies 2nd derivatives in [epsilon].) The latter two results imply convexity while the first result implies concavity, because the function is itself negative.

(17.) Maximum [epsilon] occurs when the MNF's pre-expropriation cost advantage is so high as to drive its local rival out of the market (yielding the latter's output of zero). Since both firms must coexist by assumption, [epsilon] must be < [epsilon.sub.max]. For the resent model [Mathematical Expression Omitted]. Setting [Mathematical Expression Omitted].

(18.) Also observe that for very flat M contours, the tangencies (or corner solutions) may first occur in the first quadrant, gradually moving to the second quadrant as [epsilon] rises, suggesting that when the producer lobby is very effective, consumers are initially net gainers but for higher [epsilon] values may eventually become net losers.

(19.) While [Mathematical Expression Omitted] is initially < 0 at [epsilon] = 1, monotonically increases with [epsilon]. Thus it must become positive for large enough [epsilon].

(20.) Inclusion of real resource cost to pursue rent-seeking activities (lawyers, accountants, etc.), as discussed in the DUP literature, would reduce the gain in welfare (increase welfare loss). However, since [[DELTA PI.sub.D]] + [Mathematical Expression Omitted]b.F1\duopoly] rises with [epsilon] without bound (note 19), and since the resource cost of lobbying is bound by each group's potential gains, result 4 still holds for large enough [epsilon].

(21.) The MNF's increased output in this scheme would be less than offset by the fall in the domestic firm's output, raising overall output.

(22.) Though one may analyze nationalization under duopoly, the specification of the objective function in this case poses unique problems and ambiguities. First, should the nationalized firm's social objective function include the private rival's profits as well? Secondly, if competition with a private rival equalizes costs (as suggested above) then the Cournot results would yield zero profits to both firms and zero output to the domestic firm, regardless of either form of the objective function, thus forcing the latter out of the market, an implausible outcome. Alternative formulations, to yield more realistic results, were tried, but did not yield closed form solutions, were not defensible on grounds of causal empiricism, or both.

(23.) Aharoni [1986, 123] outlines three stated goals of which maximization of social welfare is one.

(24.) See Aharoni [1986, 174-190]. Also Baumol and Bradford [1970] show that state-run firms are less profitable than their private counterparts (even when costs are the same), owing to the inclusion of consumer surplus in the firm's objective function, corroborating the present finding.

(25.) In a working paper, Mohtadi [1986] has analyzed certain aspects of this issue. One of the key findings is that expropriation of an export-oriented MNF under duopolistic rivalry may actually reduce social welfare, though the impact of lobbying was not included in that analysis.

REFERENCES

Aharoni, Yair. The Evolution and Management of State Enterprises. Cambridge: Ballinger, 1986.

Akinsanya, Adeoye A. The Expropriation of Multinational Property in the Third World. New

York: Praeger, 1980.

Baumol, William J. and David F. Bradford. "Optimal Departures from Marginal-Cost

Pricing." American Economic Review, June 1970, 265-83.

Baldwin, Robert E. "The Political Economy of Protection," in Import Competition and

Response, edited by Jagdish Bhagwati and T.N. Srinivansan. Chicago: University of

Chicago Press, 1982.

Buckley, Peter J. and Mark Casson. The Economic Theory of the Multinational Enterprise.

London: Macmillan, 1985.

Cassing, James H. and Arye L. Hillman. "Political Influence Motives and the Choice

Between Tariffs and Quotas." Journal of International Economics, May 1985, 279-90.

Caves, Richard E. Multinational Enterprise and Economic Analysis. Cambridge: Cambridge

University Press, 1982.

Davidson, Carl, Steven J. Matusz and Mordechai E. Kreinin. "Analysis of Performance

Standards for Direct Foreign Investments." Canadian Journal of Economics, November

1985, 876-90.

Dixit, Avinash K. "International Trade Policy to Oligopolistic Industries." Economic Journal,

Supplement 1983, 1-16.

Eaton, Jonathan and Mark Gersovitz. "A Theory of Expropriation and Deviation from

Perfect Capital Mobility." Economic Journal, March 1984, 16-40.

Findlay, Ronald E. and Stanislaw H. Wellisz. "Endogenous Tariffs, the Political Economy

of Trade Restrictions and Welfare," in Import Competition and Response, edited by

Jagdish Bhagwati and T.N. Srinivansan. Chicago: University of Chicago Press, 1982.

_____."Tariffs, Quotas, and Domestic Content Protection: Some Political Economy

Considerations." Public Choice, 50, 1986, 221-42.

Friedman, James W. Oligopoly Theory. Cambridge: Cambridge University Press, 1983.

Knickerbocker, Frederick, T. Oligopolistic Reaction and Enterprise. Boston: Harvard School

of Business, 1973.

Kaempfer, William H., J. Harold McClure and Thomas D. Willett. "Incremental Protection

and Efficient Political Choice Between Tariffs and Quotas." Canadian Journal of

Economics, May 1989, 228-36.

Mayer, Wolfgang. "Endogenous Tariff Formation." American Economic Review, December

1984, 970-85.

Mohtadi, Hamid. "Expropriation of Export-Oriented Multinationals." Working Paper,

University of Wisconsin-Milwaukee, 1986.

Sigmund, Paul E. Multinationals in Latin America: The Politics of Nationalization. Madison:

University of Wisconsin Press, 1980.

Williams, M. L. "The Extent and Significance of the Nationalization of Foreign-Owned

Assets in Developing Countries 1956-72." Oxford Economic Papers, July 1975, 260-73.

Williamson, Oliver. Markets and Hierarchies: Analysis and Antitrust Implications. London: The

Free Press, 1975.

PHOTO : FIGURE 2 Duopolistic Firm Interaction Before and After Expropriation

PHOTO : FIGURE 3 Key: rc lines are revenue-constraint lines for given values of [epsilon] in Figure 4. They satisfy the condition [R.sub.1 + R.sub.2 = PHI.sub.F1.

HAMID MOHTADI(*)

The causes of expropriation of Multinational Firms (MNFs) by their host governments are investigated, based on social welfare analysis and also on public choice theory. A key feature is the presence of Cournot-Nash rivalry with domestic firms. Thus, the likelihood of expropriation differs according to whether or not a domestic rival exists. The likelihood also depends on the MNF's initial technological superiority over domestic technology, the host country's demand structure, the type of control to be exercised over the expopriated firm (private vs. state) and the strategic behavior of the MNF, ex anti, when threat of expropriation.

I. INTRODUCTION

Evidence over the past thirty years points to widespread expropriation of multinational firms (MNFs), particularly by the less developed countries (LDCs).(1) Yet, expropriation has received far less attention than impediments to commodity trade (tariffs, quotas, etc.). One exception is Eaton and Gersovitz [1984] (EG), who focus primarily on the impact of an expropriation threat on international capital mobility and the response of foreign capital to such a threat. However, even allowing for a strategic response to a threat of expropriation, the issue of why governments engage in the actual expropriation of existing firms deserves attention.(2) Explaining expropriation of MNFs already in the host country is the task of this paper.

While the host government's motives for expropriation may vary, as suggested by Akinsanya [1980], this paper shall focus on economic motives, based on social welfare analysis, and on political-economic motives, based on public choice (lobbying) theory. A key feature will be an emphasis on domestic vs. foreign firm rivalry, captured by a Cournot-Nash duopoly mode, which is aimed at highlighting the potential conflict of interest between the two firms and its effects on the host government's expropriation decision.(3) To highlight the role of rivalry (section 3), a simple monopoly model of expropriation will be developed in section 2.

A second key feature will be to distinguish between the host government's expropriation and ownership decisions, regarding the expropriated firm, since chances of expropriation may depend on how the ownership issue is settled. Two cases will be distinguished: (a) the expropriated firm is turned over to domestic residents and the firm continues to maximize profits (sections 2 and 3); and (b) the state controls the expropriated firm (i.e. nationalizes it) (section 4).(4) In this case, the government is likely to pursue an objective different from profit maximization (see Aharoni [1986, ch. 4]).

Following Caves [1982, chs. 1,7], the MNF will be assumed to possess technology superior to that domestically available because of access to intangible assets (organizational and managerial know-how, specific technical knowledge, etc.), not available to domestic firms.(5) Unlike physical capital, intangible assets may be withdrawn in the event of an expropriation threat, as in EG's model, or actual expropriation, as in the present model, resulting in increased production costs to the expropriated firm.(6)

In this paper, expropriation in the monopoly model will be motivated by the host country's desire to capture MNF profits, otherwise repatriated, to increase welfare. But, this welfare gain may be offset by a loss to consumers, via the impact on costs and prices of the MNF's withdrawal of its technology. (Under state ownership some of these results will be altered.) In the duopoly model, the cost increase to the expropriated firm will raise the domestic rival's profits, providing it with an incentive to lobby for expropriation. This raises the opportunity cost of not lobbying and thus the incentive to lobby by other competing groups (MNFs and consumers), suggesting that expropriation may be the outcome of lobbying forces, rather than of pure welfare considerations.

II. EXPROPRIATION OF A MONOPOLY UNDER PRIVATE-DOMESTIC CONTROL

Consider a monopolistic and locally profit-maximizing MNF producing in and for the host country market.(7) Before expropriation, social welfare equals consumer surplus, as MNF profits are repatriated. The motive for expropriation is to raise social welfare by internalizing the MNF's profits.(8) But expropriation causes the MNF to withdraw its intangible assets, increasing production costs of the expropriated firm. Assuming the firm's ownership is passed to the domestic private sector (section 4 analyzes expropriation under state control), the firm remains a profit maximizer, and the cost increase translates into a price increase, reducing output and consumer surplus.

Figure 1 illustrates the trade-off between expropriated profits-gain and the loss of consumer surplus. This is done by comparing the net welfare gain (loss) (a) under a steeper and a flatter demand, D and d respectively (where d is pivoted through the intersection of [MC.sub.0] and D for expositional purposes), but subject to the same marginal cost shift, [MC.sub.0] to [MC.sub.1], caused by expropriation and, (b) under two different size marginal cost shifts [MC.sub.0] to [MC.sub.1] vs. [MC.sub.0'] to [MC.sub.1], upon expropriation (reflecting the MNF's different initial technological advantage, [MC.sub.0] vs. [MC.sub.0'], but subject to the same demand, D. The figure yields the following intuitive result:

RESULT 1. The welfare gains of expropriation vary inversely with the flatness of demand as well as the initial technological advantage of the MNF. Thus the likelihood of expropriation increases as either factor declines.

To explain, the profits-gain of expropriation are larger, and the drop in consumer surplus is smaller, the steeper is the demand, given a cost increase, or the smaller is the cost increase, given a demand. (Compare, as shown, the non-overlapping parts of expropriated profits [the rectangles] and the change of consumer surplus [the triangles].) Thus, monopolistic MNFs facing steep demand and operating in industries with little technological advantage over domestic technology are the more likely candidates for expropriation.

Strategic Behavior Under Expropriation Threat

If expropriation is not a complete surprise, the MNF may be able to influence (reduce) the chances of its occurrence, ex anti, by altering its output to reduce the host government's potential welfare gains, behaving strategically as a Stackelberg leader vis-a-vis the host country government. Let [PI.sub.0] and [PI.sup.s.sub.0] be the MNF's pre-expropriation profits in the cases of non-strategic and strategic behavior, respectively (O = pre-expropriation status). Then

[PI.sub.0] = P([x.sub.0]) [x.sub.0] - [C.sub.0] ([x.sub.0]) (1) and

[PI.sup.s.sub.0] = ([rho] [PI.sub.0]) = (1- [rho]) [PI.sub.0]

where [chi.sub.0] = pre-expropriation output, P([chi.sub.0)] = inverse demand, [C.sub.0] = pre-expropriation cost, and [rho] = probability of being expropriated. Thus [rho][PI.sub.0] is the expected loss from expropriation, appearing as an extra cost to the MNF. Expropriation is more likely the higher are its potential welfare gains, [DELTA]W, relative to the initial welfare, [W.sub.0], i.e., [rho] = [rho] ([DELTAW/W.sub.0]) with [rho'](*) > 0.(9) In general, [DELTAW/W.sub.0] falls with a rise in [chi.sub.0]. This is because [W.sub.0], which equals pre-expropriation consumer surplus, [S.sub.0], rises with higher [chi.sub.0], so that [DELTA]W = [W.sub.1] - [W.sub.0] (where 1 = post-expropriation status), and from there [DELTAW/W.sub.0] both fall with higher [chi.sub.0]. Hence, [Mathematical Expression Omitted]. A strategically acting MNF no longer maximizes [PI.sub.0] but [Mathematical Expression Omitted]. Denoting the relevant output by [Mathematical Expression Omitted], the first-order condition yields (from the equation for [Mathematical Expression Omitted] and after, rearrangement):

[Mathematical Expression Omitted]

In equation 3, since [Mathematical Expression Omitted] is the product of a positive and a negative term, it is < 0. Moreover, since [rho] < 1, it follows that [Mathematical Expression Omitted], implying that [Mathematical Expression Omitted] is located on the downward portion of the strictly concave profit function [PI.sub.0]([x.sub.0]), i.e. past the optimum output that would have maximized [PI.sub.0]. Denoting the latter output by [Mathematical Expression Omitted] (ns = non-strategic), [Mathematical Expression Omitted] and thus [Mathematical Expression Omitted]. Moreover, at [Mathematical Expression Omitted] the second partial of [Mathematical Expression Omitted] is negative so that the second-order condition is also satisfied for [Mathematical Expression Omitted] Hence:

RESULT 1a. Under an expropriation threat strategic behavior causes a monopolistic MNF to produce more than it would in the absence of such a threat. This reduces the host country's potential welfare gains from expropriation and thus the chance of its occurrence.

III. EXPROPRIATION OF A DUOPOLY UNDER PRIVATE-DOMESTIC CONTROL

By introducing a domestic rival, this section analyzes expropriation in the context of Cournot-duopoly rivalry.(11) Figure 2 shows the reaction functions, iso-profit curves, and equilibrium outputs of the domestic (foreign) firm at the initial (pre-expropriation) state by [R.sub.D(F)], [PI.sub.D(F) and [x.sub.D0(F0)], respectively, with equilibrium at [EPSILON.0]. Expropriation raises marginal costs to the expropriated MNF as it withdraws its intangibles, thus equalizing foreign and domestic marginal costs. In Figure 2 the MNF's iso-profit curve then shift right (to [PI.sub.F']) and its reaction function shifts down (to [R.sub.F']); for its domestic rival the iso-profit curve shifts down (to [PI.sub.D']). Thus the expropriated firm's market share falls and that of the domestic rival rises.(12) It is easy to show that the domestic rival's profits depend positively on the cost level of its foreign rival,(13) and since the latter rises under expropriation, the former must also rise, i.e.:

[Delta PI.sub.D] = [PI.sub.D1] - [PI.sub.D0] > 0.

(4)

RESULT 2. Expropriation of a MNF which is a Cournot-Duopoly rival to a local firm increases the latter's profits.

Interestingly, a comparable result by Davidson et al. [1985] shows that the imposition of local content requirements on a MNF raises its cost and therefore its domestic rival's profits. One may thus think of expropriation as a particular form of state intervention in a continuum in which a local content requirement is another.

This potential increase in the domestic firm's profits may cause it to lobby in support of expropriation. This increases, relative to the monopoly model, the likelihood of the MNF lobbying against expropriation, as the opportunity cost of not doing so is now higher. Also, consumers facing higher prices may lobby against expropriation, to protect the consumer surplus, and in favor of it to gain part of expropriated profits (see below). Since lobbying is more likely than in a monopoly model, an expropriation decision will be analyzed as an outcome of interest group lobbying rather than of social welfare considerations. The approach followed here is similar to Cassing and Hillman's [1985] analysis of trade policies where a political objective function of producer and consumer lobbying is maximized by the government.(14) Two cases will be considered: first, lobbying occurs only among domestic interest groups (consumers and the domestic firm); and second, the MNF also lobbies.

Case a: Lobbying Domestic Consumers and Producers Only

In this case the tradeoff occurs between the loss of consumer surplus ([DELTA]S < 0) and the rise in the domestic firm's profits ([DELTA PI.sub.D] > 0). To counter the consumers' loss (and thus their opposition to expropriation), the government, which is here assumed to turn the expropriated firm over to the domestic private sector, transfers part of the expropriated profits to consumers (e.g., by taxing its new owners). Denote such profits by [PI.sub.F1] and their allocation to consumers and producers by [R.sub.1] and [R.sub.2]. Then [PI.sub.F1] = [R.sub.1] + [R.sub.2]. Producer groups as a whole lobby in favor of expropriation to the extent of the increase in own domestic profits (result 2) plus their share of the expropriated firm's profits, [DELTA PI.sub.D] + [R.sub.2]; consumers lobby against expropriation to the extent of their net loss if [DELTA]S + [R.sub.1] < 0, or in favor of it to the extent of their net gain if [DELTA]S + [R.sub.1] > . The government chooses [R.sub.2] chooses [R.sub.1] and [R.sub.2] to maximize its political support function among the two groups, subject to the above constraint: [Mathematical Expression Omitted]

where M is the political support function. Assuming that M is strictly quasi-concave, as in Cassing and Hillman [1985], the first-order condition [Mathematical Expression Omitted] yields optimum R, say R(*), where [M.sub.1](R*) = [M.sub.2](R*). In Figure 3, the horizontal axis depicts consumer lobbying for or against expropriation as [DELTA]S + [R.sub.1] > 0 (first quadrant) or <0 (second quadrant); the vertical axis depicts producer lobbying for expropriation, proportional to [DELTA PI.sub.D] + [R.sub.2]. Maximization yields the tangency of the M indifference curves and the revenue constraint lines [R.sub.1] + [R.sub.2] = [PI.sub.F1], which in the [DELTA]S] + [R.sub.1], [DELTA][PI.sub.D] + [R.sub.2] space are re-expressed as ([DELTA]S] + [R.sub.1]) + ([DELTA PI.sub.D] + [R.sub.2]) space as re-expressed as

([DELTA]S + [R.sub.1]) + ([DELTA PI.sub.D] = [DELTA]S] + [DELTA PI.sub.D] + [PI.sub.F1].

Both the M contours and the revenue constraints may of course cross into the second quadrant as [DELTA]S] + [R.sub.1] may be > 0 or < 0. The special M=0 contour, however, passes through the origin since M(0,0)=0. Points to the northeast of this contour imply a positive likelihood of expropriation as M>0 in this region.

To determine a specific revenue constraint line and its relevant tangency in Figure 3, [DELTA]S], [DELTA PI.sub.D] and [PI.sub.F1] in the government's maximization decision (eqs. 5 and 6) must be known. These in turn depend on the extent of the cost increase to the expropriated MNF. Assume a linear demand, P=a-bx (a,b= constant), and constant marginal costs, [C.sub.D] and [C.sub.F] (D= domestic, F= foreign), and let the MNF's cost advantage before expropriation be [epsilon] such that [C.sub.DO] = [epsilon] [C.sub.FO] where [epsilon] >1. Then the removal of this cost advantage after expropriation means equality of the two costs, [C.sub.F1] = [C.sub.D1] (0 (1)= pre (post-expropriation values). Once equilibrium duopoly profits for each firm and consumer surplus before and after expropriation are determined, one can easily show that:

[Mathematical Expression Omitted]

Intuitively, the higher the MNF's technological edge, the greater is the cost increase upon its expropriation, and the larger the are the consumers' loss and domestic firm's profit gain. As for [PI.sub.F1], higher costs to the expropriated firm raise the price, reducing total market size, and thus its output and profits. Figure 4 shows that [DELTA PI.sub.D] and [PI.sub.F1] are positive and convex while [DELTA] S] is negative and concave in [epsilon],(16) all starting at [epsilon] = 1 (as the MNF is at least as efficient as its local rival) and reaching [epsilon.sub.max] (not shown here).(17) When [epsilon] = 1, no consumer losses nor increased domestic profits occur and [DELTA][PI.sub.D] = [DELTA]S = 0; but the expropriated firm's profits [PI.sub.F1] will be >0. In Figure 4 the four curves, [DELTA PI.sub.D], [DELTA]S, [DELTA PI.sub.D] + [PI.sub.F1], and [DELTA]S + [PHI.SUB.F1], are labeled P, S, PP and SS, respectively.

To assess the impact of lobbying on the likelihood of expropriation, information from Figures 3 and 4 are combined. For each given value of [epsilon] the height of curves S, P, SS and PP in Figure 4 identify a specific revenue constraint line in Figure 3. At the upper-left end-point of such a line all expropriated profits accrue to producers and none to consumers ([R.sub.1] = 0 and [R.sub.2] = [PI.sub.F1]), yielding the coordinate values of [DELTA]S and [DELTA PI] + [PI.sub.F1]. Thus to find this point the height of the S and PP curves (for a given [epsilon]) are traced from Figure 4 over to Figure 3. Similarly, at the lower-right end-point of this line all expropriated profits accrue to consumers and none to producers ([R.sub.1] = [PI.sub.F1] and [R.sub.2] = 0), yielding the coordinate values of [DELTA]S + [PI.sub.F1] and [DELTA PI.sub.D]. These are then traced from the curves SS and P in Figure 4.

In Figure 3 four revenue constraint lines and four tangencies are shown, based on arbitrary values of [epsilon.sub.0] = 1, [epsilon.sub.1], [epsilon.sub.2] and [epsilon.sub.3] in increasing order. For a typical family of indifference curves, as [epsilon] increases the expansion path connects tangencies of higher indifference curves, suggesting higher probability of expropriation. (Note that all tangencies are in the M > 0 region.) Though one may visualize how an extremely steep family of indifference curves might yield tangencies that move in the opposite direction as [epsilon] increases (with a tangency at first for [M.sub.0], but then changing to corner solutions that touch the lower-right end-points of the revenue constraint lines for other contours), such an outcome is unlikely, as it requires that consumer lobbying be far more effective than producer lobbying ([M.sub.1] >> [M.sub.2]) which is not usually the case.(18) To summarize:

RESULT 3. Excluding the unusual case of a much greater effectiveness of consumer lobbying relative to producers, expropriation by the host government is more likely, the larger is the initial technological advantage of the MNF.

Thus, in contrast to result 1, when a domestic rival exists whose lobbying power exceeds that of consumers, a higher technological superiority of the MNF implies greater chances of expropriation, rather than less. This reversal of result 1 arises as the higher technological advantage of the MNF implies higher potential domestic profits that would accrue from its expropriation, and thus more intensive lobbying by the domestic firm in favor of such an action.

Case b. Lobbying by the MNF

If the MNF lobbies, this lobbying would be against expropriation, and proportional to the size of its initial profits, [PI.sub.F0], lost upon expropriation. Then the function M (eq. 5) is modified as:

Max M = M([DELTA]S + [R.sub.1], [DELTA][PI.sub.D] + [R.sub.2], - [PI.sub.F0])

([M.sub.1],[M.sub.2],[M.sub.3] > 0),

[R.sub.1], [R.sub.2]

subject to [R.sub.1] + [R.sub.2] = [PI.sub.F1].

Note that [M.sub.3] > 0, but that [Mathematical Expression Omitted], which indicates that lobbying occurs against expropriation. With foreign lobbying the M = 0 contours shifts out, starting now at a positive vertical intercept, rather than at the origin. This is because M = 0 is now consistent with no consumer lobbying but with positive domestic firm lobbying just counteracting the negative MNF lobbying [i.e. M(0, [DELTA PI.sub.D] + [R.sub.2], - [PI.sub.F0]) = 0, for some [DELTA PI.sub.D] and [PI.sub.F0]]. Since the region of M > 0 would then contract and that of M < 0 expand (not shown here), lobbying by MNF reduces chances of expropriation, ceteris paribus, as expected.

RESULT 3a. Lobbying by the MNF against expropriation reduces the chances of expropriation, ceteris paribus.

Welfare Comparison of Monopoly and Duopoly Results

For a linear demand and constant costs this comparison shows that

[Mathematical Expression Omitted]

and

[Mathematical Expression Omitted]

Thus when [epsilon] is large,

[Mathematical Expression Omitted]

RESULT 4. Expropriation of a MNF with a high initial technological advantage (large [epsilon]) yields a greater welfare gain (smaller loss) when there is a domestic rival than when it operates as a monopoly in the host country.

To explain, when expropriation raises costs, duopolistic rivalry prevents output from falling and prices from rising by as much as they would under monopoly. Thus consumer surplus falls by less. Moreover, although expropriated profits are smaller in the presence of a rival, this effect is more than offset by the rise in the domestic firm's own profits when the expropriated firm suffers a large cost increase.(20)

Finally, comparing the welfare effects of expropriation and local content requirements as in Davidson et al. [1985], one finds that for an equivalent cost increase faced by the MNF under both schemes, expropriation yields a larger welfare gain (smaller loss) due to the added term, [PI.sub.F1] (the local firm's profit increase and the consumers' loss will remain the same in both cases).

Strategic Behavior Under Expropriation Threat

In the monopoly model, strategic behavior under an expropriation threat implied overproduction by the MNF to reduce chances of expropriation, via increasing consumer surplus, and thus increasing consumers' potential loss if expropriation occurred. Under duopoly, overproduction by the MNF also increases total output and thus consumer surplus. Additionally, however, it would reduce the domestic rival's pre-expropriation market shares and profits, increasing the latter's potential gain ([DELTA PI.sub.D] ) by expropriation. Weighing the two impacts, the outcome is ambiguous to the MNF, since overproduction now intensifies both the domestic firm's lobbying in support of and consumer lobbying against expropriation. To resolve the ambiguity, the MNF must know, a priori, the relative lobbying power of each group. If, as is often the case, local producers from a stronger lobby than consumers, then the MNF strategy would call for reduced pre-expropriation output, since the avoidance of a fall in the local firm's pre-expropriation profits ([PI.sub.D0]) would be more important to the MNF than inducing a larger pre-expropriation consumer surplus. Otherwise the reverse holds.

IV. EXPROPRIATION UNDER STATE CONTROL

Expropriation is frequently observed to accompany state ownership, i.e., nationalization, as discussed by Sigmund [1980, ch. 7]. Assessing the chances of expropriation when nationalization is an option requires a comparison with the case of expropriation under private-domestic ownership. For simplicity this comparison is made for a monopolistic MNF.(22) State ownership raises two new issues. First, the objective function of a state-run firm differs from that of a private firm. While several possibilities exist,(23) the more conventional assumption that the nationalized firm maximizes social welfare is pursued here. Second, there may exist an additional cost increase due to the state operation of the firm. A survey by Aharoni [1986] shows that this may only (but not necessarily) occur in the absence of competition from similar private firms, implying that in the present monopoly model state-run firms may experience costs either higher or the same as their private counterparts, a point which will be utilized shortly.

In Figure 5, with linear demand and generalized cost curves, D is demand, [MC.sub.0] and [MC.sub.n] are the pre- and post-nationalization marginal costs, and [AC.sub.n] is the post-nationalization average cost. Since social welfare is consumer surplus and profits, its maximum, when the firm is nationalized, occurs when output is [x.sub.n] and price is [P.sub.n] (= minimum of [AC.sub.n]). Maximum welfare is then the entire triangle [ARP.sub.n], yielding zero profits, a finding loosely corroborating the evidence that state-run enterprises operate at non-positive profits.(24)

Further, compared with an equivalent cost increase under the alternative scheme of expropriation and private ownership (also shown by [AC.sub.n] and [MC.sub.n]), nationalization yields larger optimum output ([x.sub.n]) than private ownership ([x.sub.1]), and thus also higher social welfare (area of [ARP.sub.n]) than private ownership (area of ABGC), as the usual "dead-weight loss" vanishes under nationalization.

However, cost increase need not be the same, as argued above. With an added cost increase owing to state operation, post-nationalization costs exceed [AC.sub.n] and [MC.sub.n]. There then exists a threshold level of such costs, i.e. [AC.sub.n]' and [MC.sub.n]' in Figure 5, which gives the same welfare as expropriation under private ownership. It follows that the areas of AR'P.sub.n'] and ABGC must equal, and thus the added welfare loss of the higher-cost state-run firm should just match the gain due to no deadweight loss. For smaller cost increases nationalization yields more welfare than expropriation with private ownership, and thus more likely; for larger cost increases the reverse holds. Thus:

RESULT 5. If state and privately-run firms are equally efficient, then welfare analysis would favor nationalization (where the nationalized firm maximizes social welfare); if state-run firms are less efficient, a threshold efficiency level exists below which nationalization (in the sense above), and above which expropriation under domestic-private ownership, would be favored.

Finally, since the pre-expropriation welfare level in this case is the same as the consumer surplus, strategic behavior by the MNF would yield the same outcome as in the monopoly case, requiring overproduction, ex anti, as in Result 1a.

V. SUMMARY AND CONCLUSION

Although a subject of many studies in international law, the issue of expropriation has received only scant attention in the economic literature. One exception is Eaton and Gersovitz [1984], though their study primarily focuses on the impact of an expropriation threat on international capital mobility. However, even allowing for a strategic response by the MNF to a threat of expropriation, the issue of why governments engage in the actual expropriation of existing firms deserves attention.

Expropriation occurs either to increase social welfare or because of lobbying influences, when such lobbying is presumed likely to exist. The chances of expropriation, in either scheme, are found to depend on the MNF's initial technological edge over domestic technology, the host country market's demand structure, the presence or absence of domestic rivals, the type of control over the expropriated firm (private vs. state), and the strategic behavior of MNF, ex anti, when under the threat of expropriation.

Further extensions might include extractive export-oriented MNFs, as the evidence suggests frequent past expropriation of such firms (see Caves [1982, 122-123]). Owing to its distinct market and organizational features, however, modeling expropriation in this case deserves separate treatment.(25) (*) Assistant Professor University of Wisconsin-Milwaukee and Graduate School of International Studies, University of Denver. Earlier versions of this paper were presented at the University of Wisconsin-Madison, University of Wisconsin-Milwaukee, University of Colorado-Boulder, Atlantic Economic Association and Western Economic Association. I am most indebted to an anonymous referee for his insightful comments and to Satya Das, John Heywood, William Kaempfer, and Randall Crane for their feedback on the earlier drafts.

(1.) Williams [1975] reports that during 1956-72, 25 percent of all foreign investments in LDCs were expropriated. By 1974, according to Sigmund [1980,7], U.S. estimates showed 106 investment disputes in 39 countries with $3.5 billion of a total of $25 billion.

(2.) The entry of a rational MNF in the face of an expropriation threat may occur either in the absence of full information, i.e. if expropriation was a "surprise", or if the expectation of its occurrence, when taken into account by the MNF, yielded a higher expected profit than was attainable elsewhere.

(3.) The use of a duopolistic framework also suggests greater realism, as in Knickerbocker [1973], and generality than the monopolistic or perfectly competitive models, and follows the broadening of trade models to oligopolistic behavior as done by Dixit [1983].

(4.) For much evidence on this type of expropriation see Sigmund [1980, ch. 8].

(5.) Differential access stems from the absence of markets, as transaction costs drive a wedge between the buyer's and the seller's price (see Williamson [1975]). Caves [1982, 4-5] outlines such costs for the MNFs.

(6.) The host government may restore the technological superiority of the MNF by licensing that technology. To the owners of the intangibles, licensing is an alternative to the "arms-length" markets of the MNF, to capture the Ricardian rents from their ownership. Such rents are not entirely bid away (owing to the market failure discussed above) and may be captured by a nonzero licensing fee. Any positive licensing fee would raise production costs, when defined exclusive of any quasi rent gains captured from expropriation, to the expropriated firm. These quasi rents are then treated separately. I wish to thank an anonymous referee for clarifying this issue. (For further discussion of licensing costs see Buckley and Casson [1985, 104-105].)

(7.) Production for exports is ignored, but an extensive analysis of expropriation of export-oriented MNFs is available in Mohtadi [1986]. (See also footnote 25.)

(8.) There may also be onetime compensation payments to the parent company, usually based on book or other contractual value of the expropriated assets, as discussed by Sigmund [1980, 265-66]. As fixed costs, their inclusion will not affect the analysis, and is thus ignored.

(9.) Note that dividing by [W.sub.0] is only for normalization convenience (allowing the argument [DELTA W/W.sub.0 and its function [rho] to both be expressed in pure numbers), and the analysis would be identical, though more cumbersome, with [rho(Delta W.sub.0)] instead.

(10.) From eq. 2 we get:

[Mathematical Expression Omitted]

Since the non-strategic profit function is concave, the first term on the right hand side is negative. As for the second term, both [Mathematical Expression Omitted] and [Mathematical Expression Omitted] are negative as discussed above. Finally, a [Mathematical Expression Omitted] must be [greater than] 0 (i.e. [rho] must be in [chi.sub.0]), if [rho] is a well-behaved probability function, starting at 1 and gradually approaching 0, as [chi.sub.0] increases. Letting [DELTA]W/W.sub.0 = [omega], this is satisfied

[Mathematical Expression Omitted]

(11.) Alternatively, a Bertrand-Duopoly model is inconsistent when products are homogeneous and costs are different, as discussed by Friedman [1983, 48], since it yields a discontinuous output for the MNF at the lowest possible price, say [rho*], of its rival (from zero at slightly below [rho*] to the entire market at slightly above [rho*]. Thus the MNF can force its domestic rival out of the market by underselling below the latter's cost, violating the coexistence assumption.

(12.) However, overall output must fall. Thus the drop in MNF output exceeds the rise in domestic output, in conformity with Figure 2.

(13.) To show this, first write the Cournot profit equations:

[Mathematical Expression Omitted]

where D = Domestic and F = Foreign. Assume, for simplicity, but without loss of generality, constant marginal costs, [c.sub.i], so that, [C.sub.i] (x.sub.i) = c.sub.i][x.sub.i, (i = D, F). Since expropriation raises cf, we differentiate the domestic firm's maximized profit function, [PI(*).sub.D], in cf and use the Envelope Theorem:

[Mathematical Expression Omitted]

The negative sign of the first term follows from differentiating (a) (written for the domestic firm) with respect to [chi.sub.F]; and that of the second term, from the MNF's optimum output,

[Mathematical Expression Omitted].

(14.) See also Kaempfer, Willett and McClure [1989], for extensions of Cassing and Hillman. For other analyses of lobbying as a determinant of trade policy outcomes, see Baldwin [1982], Findlay & Wellisz [1982; 1986] and Mayer [1984].

(15.) For the above cost and linear demand parameters:

[Mathematical Expression Omitted], and

[Mathematical Expression Omitted]. From these it follows that: [Mathematical Expression Omitted], and [Mathematical Expression Omitted]. Now, positive output for both firms requires that a>(2.[epsilon] - 1)cfo (see footnote 17). Using this constraint it follows that [Mathematical Expression Omitted] and [Mathematical Expression Omitted].

(16.) From footnote 15, [[Mathematical Expression Omitted] ([epsilon][epsilon] implies 2nd derivatives in [epsilon].) The latter two results imply convexity while the first result implies concavity, because the function is itself negative.

(17.) Maximum [epsilon] occurs when the MNF's pre-expropriation cost advantage is so high as to drive its local rival out of the market (yielding the latter's output of zero). Since both firms must coexist by assumption, [epsilon] must be < [epsilon.sub.max]. For the resent model [Mathematical Expression Omitted]. Setting [Mathematical Expression Omitted].

(18.) Also observe that for very flat M contours, the tangencies (or corner solutions) may first occur in the first quadrant, gradually moving to the second quadrant as [epsilon] rises, suggesting that when the producer lobby is very effective, consumers are initially net gainers but for higher [epsilon] values may eventually become net losers.

(19.) While [Mathematical Expression Omitted] is initially < 0 at [epsilon] = 1, monotonically increases with [epsilon]. Thus it must become positive for large enough [epsilon].

(20.) Inclusion of real resource cost to pursue rent-seeking activities (lawyers, accountants, etc.), as discussed in the DUP literature, would reduce the gain in welfare (increase welfare loss). However, since [[DELTA PI.sub.D]] + [Mathematical Expression Omitted]b.F1\duopoly] rises with [epsilon] without bound (note 19), and since the resource cost of lobbying is bound by each group's potential gains, result 4 still holds for large enough [epsilon].

(21.) The MNF's increased output in this scheme would be less than offset by the fall in the domestic firm's output, raising overall output.

(22.) Though one may analyze nationalization under duopoly, the specification of the objective function in this case poses unique problems and ambiguities. First, should the nationalized firm's social objective function include the private rival's profits as well? Secondly, if competition with a private rival equalizes costs (as suggested above) then the Cournot results would yield zero profits to both firms and zero output to the domestic firm, regardless of either form of the objective function, thus forcing the latter out of the market, an implausible outcome. Alternative formulations, to yield more realistic results, were tried, but did not yield closed form solutions, were not defensible on grounds of causal empiricism, or both.

(23.) Aharoni [1986, 123] outlines three stated goals of which maximization of social welfare is one.

(24.) See Aharoni [1986, 174-190]. Also Baumol and Bradford [1970] show that state-run firms are less profitable than their private counterparts (even when costs are the same), owing to the inclusion of consumer surplus in the firm's objective function, corroborating the present finding.

(25.) In a working paper, Mohtadi [1986] has analyzed certain aspects of this issue. One of the key findings is that expropriation of an export-oriented MNF under duopolistic rivalry may actually reduce social welfare, though the impact of lobbying was not included in that analysis.

REFERENCES

Aharoni, Yair. The Evolution and Management of State Enterprises. Cambridge: Ballinger, 1986.

Akinsanya, Adeoye A. The Expropriation of Multinational Property in the Third World. New

York: Praeger, 1980.

Baumol, William J. and David F. Bradford. "Optimal Departures from Marginal-Cost

Pricing." American Economic Review, June 1970, 265-83.

Baldwin, Robert E. "The Political Economy of Protection," in Import Competition and

Response, edited by Jagdish Bhagwati and T.N. Srinivansan. Chicago: University of

Chicago Press, 1982.

Buckley, Peter J. and Mark Casson. The Economic Theory of the Multinational Enterprise.

London: Macmillan, 1985.

Cassing, James H. and Arye L. Hillman. "Political Influence Motives and the Choice

Between Tariffs and Quotas." Journal of International Economics, May 1985, 279-90.

Caves, Richard E. Multinational Enterprise and Economic Analysis. Cambridge: Cambridge

University Press, 1982.

Davidson, Carl, Steven J. Matusz and Mordechai E. Kreinin. "Analysis of Performance

Standards for Direct Foreign Investments." Canadian Journal of Economics, November

1985, 876-90.

Dixit, Avinash K. "International Trade Policy to Oligopolistic Industries." Economic Journal,

Supplement 1983, 1-16.

Eaton, Jonathan and Mark Gersovitz. "A Theory of Expropriation and Deviation from

Perfect Capital Mobility." Economic Journal, March 1984, 16-40.

Findlay, Ronald E. and Stanislaw H. Wellisz. "Endogenous Tariffs, the Political Economy

of Trade Restrictions and Welfare," in Import Competition and Response, edited by

Jagdish Bhagwati and T.N. Srinivansan. Chicago: University of Chicago Press, 1982.

_____."Tariffs, Quotas, and Domestic Content Protection: Some Political Economy

Considerations." Public Choice, 50, 1986, 221-42.

Friedman, James W. Oligopoly Theory. Cambridge: Cambridge University Press, 1983.

Knickerbocker, Frederick, T. Oligopolistic Reaction and Enterprise. Boston: Harvard School

of Business, 1973.

Kaempfer, William H., J. Harold McClure and Thomas D. Willett. "Incremental Protection

and Efficient Political Choice Between Tariffs and Quotas." Canadian Journal of

Economics, May 1989, 228-36.

Mayer, Wolfgang. "Endogenous Tariff Formation." American Economic Review, December

1984, 970-85.

Mohtadi, Hamid. "Expropriation of Export-Oriented Multinationals." Working Paper,

University of Wisconsin-Milwaukee, 1986.

Sigmund, Paul E. Multinationals in Latin America: The Politics of Nationalization. Madison:

University of Wisconsin Press, 1980.

Williams, M. L. "The Extent and Significance of the Nationalization of Foreign-Owned

Assets in Developing Countries 1956-72." Oxford Economic Papers, July 1975, 260-73.

Williamson, Oliver. Markets and Hierarchies: Analysis and Antitrust Implications. London: The

Free Press, 1975.

PHOTO : FIGURE 2 Duopolistic Firm Interaction Before and After Expropriation

PHOTO : FIGURE 3 Key: rc lines are revenue-constraint lines for given values of [epsilon] in Figure 4. They satisfy the condition [R.sub.1 + R.sub.2 = PHI.sub.F1.

Printer friendly Cite/link Email Feedback | |

Author: | Mohtadi, Hamid |
---|---|

Publication: | Economic Inquiry |

Date: | Oct 1, 1990 |

Words: | 6081 |

Previous Article: | The voluntary provision of public goods? The turnpike companies of early America. |

Next Article: | Financial innovation, price smoothing, and monetary policy. |

Topics: |