Ownership and control in multinational joint ventures

Ownership and Control in Multinational Joint Ventures

Rafael Bautista

School of Management

Universidad de los Andes

Cra. 1E No. 18A-10, Bogotá, Colombia

This version: October, 2011

Abstract

In international joint ventures, where one of the partners is a multinational enterprise

(MNE) and the other is a local firm that possesses some significant advantage in its

market, there are sometimes issues of control (who is in charge of what) that may be

reflected in the financial structure of the venture. In particular, it may be the case that

the structure of equity is a signal of whether or not the distribution of control among

partners has been efficiently achieved. This possibility seems to go against the intuition

that equity, and capital structure should be irrelevant.

JEL codes: D01, D21, F21, F23, G32.

Key words: joint venture, equity structure, capital constraints, control

Introduction

In the last thirty years international cooperative ventures have become a

significant component of world business (Froot 1993). Such ventures now represent at

least 20% of the revenue of the 1,000 largest U.S. firms (Economist 1999). These joint

ventures (JV), at least in the case of many developing economies, are preceded, and

accompanied, by frequent conflicts and renegotiations. As the result of such

experiences, some common elements of the negotiation process have become a

Electronic copy available at: http://ssrn.com/abstract=1952681

structural component of the JV experience. Within the frame of political economy, De

la Torre (1981) discusses several of these elements, including the crucial aspects related

to capital restrictions. The factors determining why a multinational enterprise (MNE)

and a local government, or a domestic firm embedded in a foreign legal environment,

decide to enter in a JV seem to vary from country to country (Contractor 1990). This

fact presents the challenge to build a theoretical frame that helps to account for at least

some of the basic observations. Kogut (1988) classifies theories as belonging to one of

two main streams: explanations based on transaction costs as derived from the work

given by Williamson (1975, 1985), or strategic. The transaction costs approach benefits

from a wide range of perspectives, going from the costs of technology transfer (e.g.

Teece 1977) to costs related to inefficient internal markets (Hennart 1988). The strategic

frame is more a collection of approaches that share a common preoccupation with

competitive positioning and its impact on profitability (e.g. Lecraw 1984). Harrigan

(1988) proposes a general frame for competitive strategies through the use of joint

ventures and similar business arrangements. More recently there is a stream of literature

that systematizes many of the separate streams coming from both the transaction cost

approach as well as some of the ideas in the strategic approach, though still clinging

closer to the ideas in I/O (see Markusen, 2002).

Ownership structure can depend on many factors, which have been qualitatively

studied by a number of authors (e.g. Fagre and Wells 1982; Sercu and Uppal 1995).

Gomes-Casseres (1990) makes a statistical study comparing predictions from both the

transaction costs and strategic perspectives in their predictions about ownership

structure. At least some of these factors can be explained by the existence of capital

Electronic copy available at: http://ssrn.com/abstract=1952681

restriction rules. These may explain, for example, why it is that many JVs have minority

participation by the MNE (Hladik 1985).

International ventures are a fertile ground where to investigate the effects of

different arrangements of bargaining power and of private information between the

partner firms. These effects will reflect on different types of contracts, structured

according to an optimization program that incorporates all constraints resulting from

limited liability, reservation value constraints, participation constraints and varied

assortments of incentive constraints. This line of research into the structure of joint

ventures is relatively less explored 1 than has been the motivations for the formation of

joint ventures (Gomes-Casseres 1990). In a recent paper (Noe, Rebello and Shrikhande

2002, from now on NRS) the authors build the problem of structuring international

ventures along four separate instances, depending on who has the information and who

the bargaining power. In their work the basic scenario consists of a multinational that

enters a joint venture with a foreign domestic firm. Different investment strategies result

from each of the four possible combinations. Overinvestment occurs when the

multinational has both the bargaining power and the information about the true

prospects of the venture; this outcome assumes that there are no legal rules imposing a

capital participation restriction on the multinational. If the equity participation of the

multinational is restricted, then depending on the stringency of this constraint, the

investment level that optimizes the utility of the multinational varies considerably. The

main aim of this paper is to find out the effects of capital restriction policies for this

particular case. A second issue that is addressed in both NRS and the present work

concerns the quality of the private information possessed by the informed party. With

1
See the observations to this point made by D. Yoffie, 1993.

some frequency venture contracts between heterogeneous partners attempt to specify an

optimal schedule of transfer payments (Darrough and Stoughton 1989 and also

Stoughton and Talmor 1994). This kind of solution may indeed be optimal only if there

is no residual uncertainty left on the future cash flows, conditional on private

information. In Stoughton and Talmor (1994) the informed party has perfect

information and they focus on a mechanism for the transfer of payments, with the

multinational facing differential income taxes. Imperfect information begets the

necessity for contract design. Given limited liability, contracts have to be state

dependent if private information is imperfect, this mechanism solves for the optimal

contract form, instead of the mechanism of a simple transfer payment used if all

information were known to the informed part. NRS suggest that this reasoning follows

the same intuition underlying the models for takeover when the shareholders of the

target firm have private information (Eckbo, Giammarino and Heinkel, 1990; Fishman

1989; Hansen, 1987). This gives the problem characteristics similar to those found in

optimal security design and optimal mechanism design under asymmetric information,

when investment policy is the main issue (see, for instance, Green 1984). Income tax

considerations are left out of the present work, since the purpose of this work is to focus

on the effect of capital restrictions on the structuring of the venture contract. Indeed,

capital participation constraints can be seen as a form of regulatory taxation in the

classic sense.

A review of previous work

Fagre and Wells (1982) report results from an empirical study made using

information on the experiences of several U.S.-based MNEs in Latin American

countries on the relationship between certain qualities of an MNE and its ability to

obtain a certain level of ownership in a JV. This paper makes the central assumption

that the determinants of the percentage of equity participation in a JV are best explained

by the concept of relative bargaining power. Relative bargaining power is defined in

terms of a set of characteristics that distinguish both the MNE and the host country. The

authors propose that much of the relative bargaining power that a MNE musters reflects

in the proportion of ownership struck in the bargain with the local government.

Therefore, they see ownership as an important measure of the MNE's bargaining power.

They duly observe that there are other measures, such as control and allocation of

economic benefits that might be of importance. Economically, they point out, there are

some ways in which one could see ownership as the least important of the measures that

have been mentioned, due to the existence of other mechanisms, such as taxation and

rights to name board members, in which governments can have effective control and

redistribute allocations. ''Nevertheless, many governments have generally considered

local ownership as an important goal when they negotiate with foreign investors. Their

concerns presumably arise from political motivations;...Whatever the facts about the

relationship of ownership to control and economic benefits, both parties to negotiations

seem to view the distribution of shares as an important outcome in it own right.'' 2 In

brief, the authors see ownership linked in practice, though in an imperfect way, to

control. The paper devotes its main efforts to investigate how some of the MNE

resources relate to its overall bargaining power, as measured by equity participation,

vis-à-vis the foreign government.

Three different measures of ownership are introduced in the paper, each at an

increasing degree of refinement. Measure I takes the naive approach of assuming that a

2
Op. cit. page 10.

larger equity participation by the MNE corresponds to a larger measure of bargaining

power. The second measure recognizes the fact that there is no such thing as a universal

attitude toward ownership that all multinational enterprises follow faithfully. Some may

voluntarily place limits to their share of ownership, even in the absence of external

pressures from any particular government. Measure II therefore is constructed in order

to account for such idiosyncratic differences among MNEs. First, for each MNE in the

sample an average value for ownership was calculated for its holdings in its subsidiaries

in Europe. This geographical area was chosen because there ownership participation by

the MNEs in their subsidiaries is largely left to their own reckoning. Therefore, the

proposed average is considered to be a true measure of attitude of the MNE toward

ownership. Measure II is then defined as the difference between the ownership share in

the Latin American subsidiary and its European average. The third measure tries to

account for differences in attitude toward ownership adopted by the different countries

in Latin America. The third measure is defined as the difference between the parent's

ownership share in the subsidiary and the average U.S. ownership of subsidiaries for the

particular country.

The model proposed in the paper attempts to establish the effect of five

economic characteristics of the MNE on each of the measures defined above. According

to the authors, these measures are meant to represent the impact of these economic

characteristics on the outcome of negotiations as a whole, though they concede that to a

more skeptical reader those measures only represent the impact on equity alone. These

characteristics are: the level of technology, the degree of product differentiation through

advertising, to access provided to export markets, the amount of capital, and the

diversity of the firm's product line. These five characteristics act as independent

variables and each separate measure of ownership as defined earlier act as the

dependent variable.

The level of technology is proxied by the percentage of sales revenues spent by

the parent on R&D during 1974. This choice is open to some criticism, since not all

firms are equally efficient at making the R&D budget to translate into tangible results.

The justification for this choice resides in the fact that R&D expenditures are an

imperfect measure of the rate of technological innovation produced and put into practice

by the MNE. The higher the R&D expenditures as a percentage of sales, the more likely

is that innovation and technological uniqueness are at the heart of the strategic

commitments of the firm. Most developing countries are at a considerable disadvantage

when negotiating with a firm that deals in markets with participants that compete

through a fast paced product improvement race. These conditions will not be met by

those countries without the involvement of a foreign investor that has the necessary

R&D experience and the required material and human resources. By comparing studies

where each of the three measures of ownership is taken as dependent of this single

variable, the results are all consistent with each other, and point to a not too simple

dependence: If the percentage of R&D expenditures is less than about 5%, then there is

no clear effect, as seen through any of the alternative dependent variables. However, if

it is higher than 5%, then all of the measures show a strong effect, where all MNEs

within this category take on average nearly 100% of ownership, regardless of industry,

number of competitors in the particular industry, or other similar criteria.

Fagre and Wells suggest that product differentiation seems to be one of the main

bargaining advantages for those MNEs that have strong marketing skills. ''Product

differentiation by multinational enterprises presents a formidable barrier to potential

local enterprises in developing countries. Although the production technology needed to

manufacture satisfactory ball point pens and carbonated beverages is hardly secret,

Parker pens and Coca Cola are still the preferred products in many developing

countries.'' 3 This independent variable is proxied by the percentage of sales that the

MNE spends on advertising. Advertising technology is also a rather perfected art;

therefore, in this case, there is a much higher confidence that the proxy represents more

closely the associated concept than in the case for the level of technology. The authors

find that advertising seems to be related to bargaining power at nearly all levels of

expenditure. As with the previous criterion, they seem to distinguish a threshold value

of about 7% of total sales revenues, above which MNEs in this group attain on average

a level of ownership of 98% (see their TABLE 3.) The measures corrected for firm

attitude toward ownership and for country attitude show this same tendency as well.

This result must be taken with some caution, since the mix of MNEs at the top of the

scale are to some extend of the same characteristics as those that have a strong

technological background, therefore there is the possibility for confounding effects. The

authors argue that since expenditures in advertising show significance for nearly the

entire range of analysis, this variable is important on its own. To strengthen this

argument, they point out that no other variable nearly approaches the power to explain

the bargaining success of pharmaceutical and cosmetics multinational enterprises.

A high capacity for exporting the product of the subsidiary is found by the

authors to be another important factor in the bargaining power of a MNE. ''Fundamental

to the strategy of some multinational enterprises is an ability to rationalize production

3
Op. cit. page 12

on a global scale and a capacity to acquire and utilize sophisticated knowledge of

foreign markets.'' 4 In particular, if a large portion of the product from the subsidiary is

for export to another affiliate of the same multinational, this will result in considerable

negotiating strength for the MNE, although in general if the greatest part of production

is for exports, then Fagre and Wells find that this translates into high relative bargaining

power. The proxy variables for these concepts are the percent of total production

transferred and the percent of total production exported, respectively. They find that

especially the third measure: the country corrected ownership, shows a significant

increase when the proportion of exports in either category are in excess of 50% of the

total product. Again there is here the ever present risk of confounding, caused by the

fact that the multinational enterprises used in the sample may be also either strong in

technological level or marketing. The authors justify the independent relevance of this

variable by looking at MNEs in the electronic part sector, where product differentiation

or technology are not of particular importance, and find a marked difference between

the ownership measures for those subsidiaries that export most of its product, and those

that do not.

The case for capital as a source of bargaining power is less clear. From the

sample used to make this particular study, there is no clear evidence for this to be a

factor that favors the MNE. Only when the 1975 investments involved assets over 100

million, there was a visible effect in equity participation. But then, the authors disclose

the fact that subsidiaries with assets of this size were usually located in countries that

have lenient policies toward foreign ownership.

4
Op. cit. page 13

The last of the independent variables, product diversity, is an additional factor

not found, according to the authors, in the previous literature on the subject. ''When

subsidiaries are classified by the number of products (3-digit SIC) they produce, a quite

strong relationship to ownership appears. The greater the number of products, the

greater the apparent bargaining of the firm...'' 5 They find somewhat puzzling though

why this variable turns out to be very significant in their study. This is more striking

after they confirm the fact that variable is not correlated to any of the others already

discussed that are of particular interest to the host country. Differently from the other

four variables, there doesn't seem to be a straightforward reason as to why this one

seems to weaken the relative bargaining position of the host country in Latin America.

To be sure this is not a general sort of happening, the authors check the European case,

and find that the effect there is almost the opposite. The authors attempt to understand

the reasons by concentrating in the single case of Mexico, where the numbers are large

enough to produce some significant statistics. When parent firms are classified

according to the number of affiliates, and each affiliate is associated to a certain group,

the authors find the following effect: ''Enterprises that have only one affiliate in Mexico

use it as a base for several product lines in 55 percent of the cases, roughly double the

similar figure for corporations owning 5 or more affiliates in Mexico. Parent firms using

a single affiliate strategy also wholly owned their affiliate in 64 percent of the cases,

again roughly double the similar figure for corporations having 5 or more affiliates.'' 6

The possible reasons that they conjecture for these observations relate to higher

management skills, preferences by the host country of larger, multi-product,

investments over those for a single product, there may be a concentration of specially

sensitive industries which use a single affiliate with several products, and the one

5
Op. cit. page 17.
6
Op. cit. page 18.

favored by the authors, that the observed relationship is the consequence of government

regulations in the host country.

Fagre and Wells not only consider enterprise related factors in order to measure

its relative bargaining power, they also looked at the degree of competition that the

multinational could face within its industry. They classified the parents into industrial

sectors, using the 3-digit SIC classification and within each industry counted the

number of MNEs acting in each country. As it turns out, the number of competitors has

a very significant effect on bargaining power, as measured by the average level of

ownership, using any of the three measures. The authors report that ''16 of the 18

industries in which all foreign investment were wholly owned contained only one

corporation, and the average parent ownership of subsidiaries for the 24 single-

corporation industries was 95 percent.'' 7 They obtain a consistently negative relationship

between ownership and degree of competition, as long as they exclude from the study

three industries that have a special meaning to most governments in Latin America:

Pharmaceuticals (SIC 283), petroleum refining (SIC 291) and office machinery (SIC

357); they discuss in some detail why these three are to be left out of the study. The

reasons they offer for the case of office machinery are a bit dated, and it might be worth

to repeat this same study with modern data.

The paper closes with a brief discussion of multivariate analysis, where the

different measures of equity participation are modeled using linear regression taking all

five independent variables at once. Their results show all coefficients but the one for

size (assets) of affiliate to be significant at least to the 0.05 level. The R2s are for all

7
Op. cit. page 19.

three measures in the neighborhood of 13%, with a total of 648 degrees of freedom.

These results point to problems due to imperfection in the measures defined for both

dependent and independent variables, as well a to inadequacies due to the application of

a linear model.

The main general concern left unanswered by this study is whether actually the

main assumption made in the study; that equity participation is a sufficient measure of

bargaining power, can survive further inspection. This is one of the central

considerations, and the starting point, for some of the more recent studies.

In LeCraw (1984) the problem posed by the relationship between bargaining

power and ownership in the JV is further analyzed, beyond the findings already

discussed by Fagre and Wells (1982). In Fagre and Wells (1982) the driving concept is

to define relative bargaining power through a set of five characteristics of the MNE, and

then to measure it via the MNE's percent of equity ownership participation. While

recognizing the important contribution of the bargaining power framework provided in

their paper, LeCraw, as well as others, is skeptical of the assumption made there that

equity participation alone is enough measure. LeCraw (1984) is a paper mostly

empirical in its content. It has three distinguishable components. In the first, LeCraw

extends the study of Fagre and Wells (1982) to include countries in the ASEAN group:

Thailand, Malaysia, Singapore, Indonesia and the Philippines. The second part studies

the connection between the relative bargaining powers of a MNE vis-à-vis the local

government, as proxied by a chosen set of characteristics, to the degree of control the

MNE exercises over its subsidiary in the host country. This particular dimension of

bargaining power is missing in Fagre and Wells (1982). Quoting from LeCraw (1984, p.

27): ''Poynter [1982] has shown that a TNC [MNE] may find it advantageous to bargain

not for increased equity ownership, but for control over the variables critical to the

success of the subsidiary from the TNC's point of view.'' Control provides means other

than equity participation by which the MNE may appropriate the return on its

investment: ''...licensing and management fees paid by the subsidiary, sale of inputs to

the subsidiary, sale of outputs to other units of the TNC [MNE] or on world markets,

and interest on intra-company debt. The TNC may use its bargaining power not to

increase its equity ownership, but to secure some other means by which to appropriate

this return, possibly by manipulating the transfer price of these other payments.'' 8 In the

third part of the study, LeCraw explores how the exertion of control over a specific list

of important operations and functions in the MNE's subsidiary (marketing, finance,

technology, production, imports, exports, etc.) affects the overall success of the JV, as

perceived by the sampled MNEs. The paper draws its conclusions from a sample of 153

subsidiaries operating in the five countries of the ASEAN community. These countries

differ widely in the characteristics of their respective policies toward foreign direct

investment. The MNEs operated in six different manufacturing sectors. The MNEs

subject of the study were based in the United States, Europe, Japan, and several LDCs.

The care in the construction of the sample encourages the author to believe that ''the

sample then may give a good basis on which to reach generalizations concerning the

determinants of ownership and control of the subsidiaries of TNCs [MNEs] in LDCs,

and concerning the effects of ownership and control on the success of these

investments.'' 9

8
LeCraw (1984) page 30.
9
Op. cit. page 28.

The study proceeds along methodological lines that are similar to those

discussed above for Fagre and Wells (1982). This work concentrates its attention in

three dependent variables: Actual equity ownership, MNE bargaining success, and

effective control. These measures are related through multivariate linear regression to

the following set of independent variables: Technological leadership, Advertising

intensity, Subsidiary assets, Capital/output, Export/sales, MNE assets, MNE-subsidiary

linkages, Host-country attractiveness, Potential MNE investors, Time (1960=1),

Dummy-Japanese MNE, Dummy-LDC MNE, and European MNE.

The measure of actual equity participation is immediate. What the author calls

bargaining success is defined thus

EO − DE HC
S MNE =
DE MNE − EO

Where EO is the actual equity ownership, DEHC is the resulting equity

ownership of the MNE, if the host country gets its desired level of equity ownership,

and DEMNE is the desired level of equity ownership of the MNE. This measure is

approximately equivalent to the ratio of the country-corrected to the company-corrected

measures devised by Fagre and Wells. There are three observations that come to mind

about the introduction and intended use of this measure. First, the author does not state

all the arguments for why this particular construct is a good representative of bargaining

success. Second, inspecting the regression results that he presents in TABLE 2 of the

paper, there doesn't seem to be any new insight gained from this variable that it is not

already included in the results for actual equity ownership. Third, if the idea is that the

measure is an increasing function of bargaining power, then it is worth noticing that for

fixed values of DEHC and DEMNE, the measure is monotonically increasing with actual

equity ownership. This seems, at least superficially, somewhat at odds with the J-shaped

relationship between equity ownership and the company and country corrected success

presented in Figure 3 of the same paper.

The third measure is effective control. This measure is constructed using

managerial evaluations. Each manager of the 153 subsidiaries rated the importance of

control over each of a list of 18 factors: output pricing, output volume, output quality,

technology transfer, technology control, capital expenditures, financing source,

financing cost, financing amount, dividends timing, dividends amount, fees paid to the

MNE, advertising and marketing expenditures, channels of distribution, import price,

import source, import volume, export price, export destination, export volume and

overall management. The rating system used a scale from 1(none) to 10(critical). The

managers were also asked to rate the degree of control that the MNE had over each

factor from 1(no control) to 10(complete control). These data were then used to

construct the measure of effective control by averaging the scores for the degree of

control over the 18 factors, weighed with the scores for the importance. The author

interprets this measure by stating that ''...Effective Control measured the degree of

control over the critical success variables retained within the TNC [MNE] compared to

the control lost to those outside the TNC, such as, local partners or the host

government.'' 10 He finds the correlation between effective control and equity ownership

to be 0.57, much less than a 1 to 1 correspondence.

Next, using multiple regression, the connection is established between the three

dependent variables and the set of factors associated with relative bargaining strength.

The independent variables used to describe bargaining power were: technological

10
Op. cit. page 37.

leadership, advertising intensity, subsidiary assets, capital/output, export/sales, MNE

assets, MNE-subsidiary linkages, host country attractiveness, potential MNE investors

(this refers to the number of potential competitors), time (1960=1), dummy for Japanese

MNE, dummy for LDC-based MNE and dummy for European MNE. Now we provide a

brief description of the measures for each independent variable, and place in

parentheses the respective sign for each dependent measure (Actual equity, Bargaining

success and Effective control) in the results for the regression. The measure of

technological leadership (intensity) (+ all) 11 was made by asking the managers of the

153 subsidiaries to rate from 1 to 10 some aspects that are usually associated with this

characteristic. The measure includes ratings for the technology that was initially

transferred and the potential for further transfers in the future. ‘Advertising intensity’ (+

all) was represented with the advertising to sales ratio of the subsidiary relative to other

firms in the industry; capital intensity (+ all) and capital requirements (+ all) were

measured using total assets/output and total assets of the subsidiary, respectively, but

since these two are correlated, when put together in the regression the first shows

significance below the ten percent level and the second is significant to only ten percent

level. ‘Export intensity’ (+ all) was measured as exports/sales; this was significant at the

one percent level in all three regressions. The total assets of the parent MNE (+ all)

relative to others in the same industry were included to test the hypothesis that smaller

MNEs would take a minority equity position, this variable showed high significance in

both the first and third dependent measures, it was not significant for ‘Bargaining

success’. The ‘Linkage effects’ (-, -, +) is proxied by the ratio of total flow of resources

between the parent and the subsidiary over sales, the flows considered included ''inputs,

interests on loans and intrafirm suppliers' credit, intrafirm sales, management and

11
This means that all signs in the regression positive. The same interpretation applies if the sign in
parentheses is negative.

technical service fees, and imputed rental value on machinery and equipment supplied

by the MNE.'' 12 Notice the opposite sign between ‘Actual equity’ and ‘Effective

control’, this confirms the hypothesis that if linkages are strong, then the MNE will be

less interested in equity and more interested in having control over critical operations

and functions; the coefficients for these two measures were significant at least at the

five percent level, the coefficient for ‘Bargaining success’ was not significant. For

attractiveness of the host country (- all), the managers of the subsidiaries in the sample

were asked to rank the country on a 1 to 10 scale; the results show that the more

attractive the country the lower the relative bargaining power of the MNE vis-à-vis the

host country, the coefficient for ‘Effective control’ was not significant. Potential MNE

investors (- all) represents the degree of competition found by the MNE at its arrival.

The number of MNEs that had already undertaken FDI in the same industry in the

particular country was used as the measure; all coefficients for the three regressions are

significant to the five percent level or better. An absolute reference for Time (1960=1)

(- all) was introduced to show the learning effect of the host countries, which got better

in their bargaining ability over time; all coefficients are significant at the five percent

level. The coefficient of the dummy for a Japanese MNE showed a somewhat

unexpected result, it was negative for Actual equity and positive for effective control,

with both significant to the five percent level. The apparent interpretation is that the

Japanese bargain less for equity participation while keeping a relatively higher degree of

control over operations. The LDC dummy (-, +, -) shows that these MNEs have

relatively less bargaining power; this is stressed by the fact that the coefficient for

Bargaining success is not significant. The European dummy was not significant to any

12
Op. cit. page 35.

2
of the dependent measures. The R for each of the three regressions were 0.63 for

‘Actual equity’, 0.47 for ‘Bargaining success’ and 0.55 for ‘Effective control’.

LeCraw also analyses in more direct ways the relationship between equity

ownership, effective control and the overall success of the JV, from the MNE's point of

view. ''...TNCs [MNEs] may bargain for increased equity participation in order to

increase their control over the operations of their subsidiary, to try to ensure that the

internalization advantages are in fact realized. The link between the level of equity

participation and the TNC’s control over its subsidiary, however, may not be

straightforward. Depending on type of technology transferred, the capabilities of the

local partners, and the host government policies, a TNC may be able to control the

operations of its subsidiary that are critical from its viewpoint without a majority

ownership, or, conversely, may have little control over these operations despite majority

(or even complete) ownership. A TNC may therefore be willing to trade reduced equity

ownership for increased control of variables crucial to the success of the venture from

its point of view...The link, therefore, between the bargaining power of the TNC, the

level of its equity participation, its control of the subsidiary, and its perception of the

success of the investment is complex and may be difficult to trace.'' 13 In an attempt to

map the relationships among these factors introduces three measures of success: the

profitability of the subsidiary, the success of the subsidiary as rated by the MNE (on a

scale from 1 to 10); and a ''country and industry corrected'' success (CICS) measure.

This way of measuring success was introduced to account for the perception that

profitability was not the only measure of success and for the fact that profitability

reports from subsidiaries of MNEs have sometimes been found to differ from actual

13
Op. cit. pages 30-31.

profitability. The CICS was plotted against Effective control, using the latter as

independent variable. The plot is shown as FIGURE 2 in the paper, and it displays a

strong linear dependence. In FIGURE 3 CICS is plotted against the percent of actual

equity ownership; it shows a J-shaped diagram, with its lowest points drawn around the

equity region where ownership is roughly equally distributed between the partners.

Meaning that in the sample studied, ventures were the least successful when ownership

was split roughly equally between the partners, condition that possibly reflects a hard

bargaining process where both parties saw a close tie between ownership and control,

resulting in a poor managerial structure. This result is consistent with the findings in

Killing (1982).

On relative bargaining power and project control

There are several reasons given by different governments for the existence of

equity participation restrictions: better access to information, control of payments for

technology transfer and management fees, control of pricing of output and intra-

company trade, reinvestment and remittance of capital. These reasons sometimes do not

provide a coherent information set on which to draw conclusions, much less to predict

the outcome of a particular negotiation. Also, it is common to find equity participation

restrictions related to politically sensitive issues to the host government, quite apart

from economic considerations.

The main purpose of this section is to propose a starting point for the

formalization of the ideas contained in Poynter (1982) and LeCraw (1984). It may be

useful first to summarize briefly the relevant aspects of what has been found and

presented in the previous section. In Fagre and Wells (1982) the plausible assumption is

made that there is a close relationship between the bargaining power of the MNE and

the level of equity participation that it can negotiate with local firms and governments.

Their study shows results where this assumption finds empirical support. Nevertheless,

Poynter (1982) makes the observation that such bargaining power will not necessarily

be focused on equity participation alone, but that often it is used in order for the MNE

to keep decision control over operations that are critical to the success of the venture,

even if the MNE has minority participation. After additional empirical research LeCraw

(1984) finds that the implied additional assumption made by in Fagre and Wells (1982):

that increased equity participation goes hand in hand with increased control of critical

functions, is not necessarily true. Control of critical operations and percentage of equity

participation are not, in general, perfectly correlated. These empirical results have a

bearing on the theoretical side of JV studies. If an important part of the contribution

from the MNE to the JV is related to managerial skills, then it is not adequate to

describe models for JV only in terms of equity participation, but control must also be

included as part of the model. In terms of bargaining power, it is difficult not to see as

somewhat odd the fact that the MNE has the bargaining power, and yet it has no say in a

negotiation process with a local government which is perhaps acting under its own

pressures to seek outside investment, especially for the increase of the nation's export

capacity. The main items over which there is something to bargain for are equity

participation and project control, therefore, if such power is on the side of the MNE, it

would have to reflect in at least the second of these two components. A more careful

inspection of this argument though would hint at a stronger characterization of the

concept of bargaining power itself. The standing literature limits its scope of bargaining

power to setting the bargaining problem such that the party having the power maximizes

its own utility, subject to some rationality constraint given by the counterpart. This

maximization is done with respect to a set of variables that specify the terms of the

problem; these variables represent the main items that are subject to negotiation at the

''bargaining table''. However, in order for this maximization exercise to make sense,

there are sometimes conditionality restrictions that the set of variables most satisfy. In

the present literature, the treatment of ownership and governance variables, even when

the latter is directly treated at all, does not seem to follow clear rules of conditionality.

This does not seem to be an entirely satisfactory scheme. Consider an MNE that does

indeed wield a ''big stick'' in terms of, for instance, marketing power. If it faces a

negotiation process with a government that is lenient in its capital participation policy,

then it is to be expected that the MNE will must certainly have full control of the main

components of the project as well. Assume on the other hand that the same MNE faces

negotiations in a country with tough policies regarding foreign investment in certain

industries. Then it may be the case that capital participation restrictions could act as a

barrier for the MNE to have access to majority participation. However, if it is true that

the MNE carries weight in the negotiation, it still should be able to keep the full control

of critical parts of the project. This can always be the case, since project control is not

an aspect that is easily visible to outsiders, as opposed to what happens with equity

participation. On the other hand, it is not easy to conceive of a situation in which the

result of negotiations between a local government and an MNE were that the latter gets

to keep, say, 95% of equity, but it has no effective control of operations. In other words,

a clear sign that the MNE has any true bargaining power is that it keeps control of at

least those components of the project that are critical for a successful outcome, from the

MNE's point of view. We will refer to having control, or having effective control to a

situation in which managerial control is not shared to the extend where decisions result

in actions that are inconsistent among each other, and with the well being of the project.

Some arrangements of shared management in JVs result in each of the parties

controlling some vital components of the project, leading almost always to conflicts that

cause the JV to fail (Killing, 1982).

An example may help to clarify the arguments given above. Consider the

following simple model: suppose that a JV project produces a certain amount L if things

do not turn out well. This amount can be seen as a riskless cash flow that the partners

get for the very fact of undertaking the project. If all comes out perfect, then the final

cash flow for the project is L+ where is a ''premium'' that defines the degree of
,

success over failure. Suppose moreover that the probability of success for the project is

some concave, increasing function of initial investment I , and that it also depends on

the degree of control c exerted by the MNE over critical parts of the project. Let's

denote this probability by p(∆ | c, I ) . Then the expected utility for a risk-neutral MNE

will be given by U (c, β , γ , I ) = γL + β∆p(∆ | c, I ) − I . Here γ corresponds to the share the

MNE gets of the riskless part of the cash flow contributed by the mere undertaking of

the project, and β is its equity participation. This formula assumes that all initial

investment comes from the MNE. This assumption makes sense at least in the scenario

when the MNE has both the bargaining power and privileged information, because then

one can not expect the local firm to have an incentive to put cash up front as part of the

initial investment in the project. To simplify matters, let's assume that c ∈ {0,1} , where

c = 0 if the MNE has no control and c = 1 if it has total control. The probability function

is assumed to be such that p(∆ | c = 1, I ) ≥ p(∆ | c = 0, I ) and

DI p(∆ | c = 1, I ) ≥ DI p(∆ | c = 0, I ) for all feasible I . In what follows we shall incur in a

small abuse of notation and rename these two probabilities simply as p1 (I ) and p0 (I ) ,

respectively. The stated conditions are meant to ensure that the project with full control

will be comparatively more successful than the project with no control. Let's analyze a

hypothetical situation in which the MNE has to choose between having a higher than

50% equity participation β H , but with little or no control, and having a minority

participation β L , but with the possibility of having full control of critical operations.

Then, all else been equal, the MNE will have to decide the best use of its bargaining

power on the basis of which of the two products, β H p0 (I ) or β L p1 (I ) is higher. Given

that there exist some indifference point of investment level between these two options,

then the MNE will have an incentive to put a larger investment I into the second case 14,

if constraints so allow it. Therefore, it won't choose the first case.

Another issue that goes to the heart of the benefits of control to the MNE is the

one related to transfer prices and fees. Clearly, the more economically significant is the

internal transactions between a parent and its subsidiary, the greater the total size of

such fees. These cash movements can be seen as riskless rents that the MNE derive

from the JV. If the MNE exerts no control over the critical decisions of the venture, the

transfer of these riskless cash flows is not likely. Another scenario could be that the JV

does not come with the convenience of frequent internal transactions, leaving the MNE

with the need to obtain maximum results out of the risky part of the cash flow, plus any

riskless part that has to be shared with the subsidiary. Assuming that the MNE will

prefer riskless gains to risky ones, it is useful to recognize the importance of any

riskless cash flow that results from internal pricing and fees, as opposed to any riskless

cash flow that is shared by the partners. In other words, instead of writing γL , the shared

riskless cash flow, we should write lc + γL as the complete expression for the riskless

14
This is so because the second term eventually has a higher slope than the first, for I large enough.

component, where l represents the amount stemming from internal pricing and fees.

Observe that if the MNE has no actual control, then the additional profits arising from

pricing and fees vanishes.

Realizing that the description of bargaining power may be a more complex

matter than it is usually taken for in much of the present literature; it is worth to

consider a more direct representation of this concept. It is possible to picture the

bargaining process between the MNE and the local firm/government with a dynamics

where the final outcome in terms of control and equity participation is uncertain to some

degree. At an early stage, when the MNE is considering whether or not to enter a JV, it

will have to decide on the convenience of the proposed JV pretty much based on

expectations of some kind. The setup of the problem for the MNE must consider this

ex-ante expectation. There is then the matter of how to make operational this

expectation. Clearly what is meant by bargaining power is a relative notion. It is not

reasonable in general to suppose that the MNE can exert it equally well with different

companies spread across different countries. In the case of Latin-American countries, as

is remarked in Fagre and Wells (1982), there are obvious differences in negotiating

with, say the Dominican Republic, than doing so with Mexico. The lesson that we can

extract from these considerations is that we can talk only of relative bargaining power,

and that if this power is to be measured in terms of the more or less uncertain outcome

of some negotiation, then a likely candidate to represent it in a formal way is a

conditional probability. This probability should meet at least three criteria: First, we

assume that the most relevant element in the negotiation is the extent to which the MNE

can gain control of critical parts of the project. Second, the extent to which it is able to

keep such control at a given level of equity participation. Third, once the first two issues

have been settled, there are no other significant elements left that can characterize the

outcome of the bargaining process. This is not to say that in order to reach a final state

in terms of control and equity participation there might not be other factors that weigh

in as chips used in the bargaining game. This could be the case, for instance, of the size

of the investment.

Given the above criteria, we propose ρ (β ) = Pr (c = 1 | β ) as a direct measure of

the true relative bargaining power of the MNE. We argue that this conditional

probability summarizes the conditions that best reflect the findings in the empirical

literature, in particular those in LeCraw (1984) and Fagre and Wells (1982). It is

necessary to carefully state the interpretation given to this choice for the bargaining

power of the MNE. First, it means that if the MNE is not able to retain control of critical

parts of the project, then this outcome is a strong indicator of a relatively low bargaining

power with respect the local firm/government. Second, the ''given  part of the
''

formula must be read as ''given that it is able to get at least  in the bargaining process''.

Much too often reaching agreements for carrying out a JV between a MNE and a local

firm, with the local government as a third interested party, brings the spotlight on the

issue of ownership, as represented by equity participation. Arrangements that may lead

to successful outcomes - or possible equilibrium solutions if we see it as a bargaining

game - include either having a majority participation, up to 100%, by the MNE, as well

as having full control, or alternatively they may lead to the MNE settling for a minority

participation, so that it can have full control. A different sort of outcome results if equity

considerations are closely tied to the issue of control of critical operations. This may be

the case if the local government is, for example, a strong advocate that the JV must also

be a learning process that contributes to increase local managerial skills. Assume that

the informational advantage possessed by the MNE truly represents a superior

organizational and technical know-how. If the MNE gets into a situation in which both

sides of the table see more equity as more control, then this may result in a mixed

governing body, quite possibly in the neighborhood of a fifty-fifty composition. Then

its knowledge advantage may be seriously diluted (i.e. c = 0 ) by a hampered and

somewhat unpredictable decision making process. The resulting inefficiencies are likely

to have bad consequences for the project. From these considerations we can gather that

the particular shape of ρ (β ) is not necessarily simple; this function is somehow

summarizing all the essentials of what the MNE and the locals can achieve, or are

willing to concede, in the negotiation. Further discussion of this model is deferred to

section 1.7.

Structuring international cooperative ventures

Noe, Rebello and Shrikhande (2002) explore the relationship between

bargaining power, regulations, information asymmetry and financing policies in

international joint ventures. They consider a cooperative arrangement between a

multinational and a local firm; the partners determine the scale of the venture and its

financial structure. This scenario assumes that the local firm is capital constrained, and

it will not undertake direct investment unless circumstances, in terms of its own private

information, would indicate otherwise. The multinational may find legal barriers

established by the host government for it to realize its desired equity ownership

participation in the venture. Either one or both of the parties may face competition and

may hold private information about the venture's prospects. The approach in NRS

consists of studying four different allocations of information and bargaining power:

First, the multinational has the bargaining power and also has the information

advantage; this could be the case, for instance, if the local firm faces competition and

the main product from the venture is marketed outside the host country. Second, the

multinational has a bargaining advantage, but the local firm has more information about

the venture's prospects; this may happen, for instance, when the multinational has a

recognized brand name, but the product of the venture is directed toward the host

country's market, where the local firm has better knowledge of market's conditions.

Third, the local firm finds several multinationals competing to enter the local market; in

this case the local firm has both the bargaining power and the information advantage.

Fourth, bargaining power rests with the local firm, but the multinational has an

informational advantage. In the four cases mentioned, the authors stress the importance

of contract structures that favor firm-value contingent payments over simple transfer

payments for the firm with the information advantage. This is in line with the literature

on takeover bids, when the target firm has private positive information (Eckbo,

Giammarino and Heinkel (1990), Fishman (1989), and Hansen (1987)). This intuition

about the optimal contract is not fully in line though with actual JV experience 15, where

an important part of the motivations underlying the interest of a multinational enterprise

in entering a JV is related to riskless transfer payments.

In the first case, when the multinational has both the bargaining power and an

informational advantage, the multinational has to make sure that the local firm values

highly its participation in the venture. The multinational may achieve this goal by

signaling the goodness of the project through taking as much equity as possible.

Government regulations on capital participation restrictions can prevent this form of

signaling and then the multinational will be forced to use a costlier form of signaling via

15
See, for instance, the account in Killing (1982).

overinvestment. Overinvestment caused by capital regulations will have a positive

impact on the level of local employment and on the possible rents that the local firm

could derive from this situation. Therefore, in the present case of bargaining power and

information allocation, the local firm and government stand to benefit from capital

restrictions 16.

In the second case, when the MNE holds the bargaining power, but the local

firm has an informational advantage, as it might be the case when the end product of the

JV is directed to the local market. In this case, the local firm has an incentive to declare

the state to be B. With this news the local firm is interested in convincing the MNE that

it needs a larger share of the profits (i.e., a larger β ) in order to meet its opportunity

costs. The right response of the MNE, should the local firm declare B, is to offer a

contract that ''severely restricts project investment, reduces to a minimum domestic firm

investment participation, and deprives the domestic firm of upside cash flows'' 17. The

way in which the MNE limits upside the upside cash flows to the local firm is by

maximizing its own equity participation, and therefore capital restriction rules are

relevant, as in the first case.

The third case is when the local firm has both the bargaining power and an

informational advantage; this might be exemplified by an arrangement where the MNE

acts as a simple financier of the venture. In this case the local firm is the one interested

in taking an all equity position, as a signal of the project's good prospects. This measure

alone might not be enough, and a degree of overinvestment might still be a necessary

signal. The local firm will indicate in this way that it can bear the costs of
16
As noted earlier, this sort of reasoning makes strict sense only in the total absence of the possibility of
internal transfer payments.
17
Noe, Rebello and Shrikahnde (2002), page 3.

overinvestment. In this case any rule restricting the capital participation of the local firm

would benefit the MNE.

The fourth case considered in NRS occurs when bargaining power rests with the

local firm, but the MNE has the informational advantage, situation that may arise if the

local is in a monopoly position, but the end product is directed to the MNE's home

market. The MNE may have an incentive to report bad news, in which case the response

from the local firm must be to offer a contract were investment is restricted to a

minimum and it will take a100% equity participation. Again, any rule restricting the

capital participation of the local will go in favor of the MNE.

In conclusion, capital participation restrictions on FDI have an effect on the

MNE mainly when it has the bargaining advantage, while any rule restricting

investment participation by the local firm affects it mainly if it has the bargaining

power. Investment distortions result from asymmetry of information between the

partners, with overinvestment occurring when the partner having the bargaining power

also possesses an informational advantage, and underinvestment occurring when the

partner that has the bargaining power is at an informational disadvantage. In what

follows, we shall explore the combined effects of capital restrictions and information

asymmetry in the first case, when the MNE has both the bargaining power and an

informational advantage. The main question to be addressed is whether capital

restrictions can eliminate the effects due to asymmetric information, restoring

investment to its Pareto efficient level, or even inverting the original effect and creating

underinvestment.

The model

The project requires of a total investment I that is bounded by exogenous

conditions, so that I ∈ [ I min , I max ] . Throughout this work we assume that the

multinational (from now on the MNE) makes all the necessary investment and its

partner - from now on the domestic firm - doesn't have to put any cash up front. This

assumption, as explained earlier, is consistent with the case when the multinational has

the bargaining power, which is the main focus of the subsequent analysis. The project's

future cash flow can only be one out of two possible outcomes, either H or L , where

H > L > 0 . The probability for realizing the high cash flow H depends on the value of

the information signal. The information types are either G , for good news, or B for bad

prospects. These states determine probabilities Pt ( I t ) for cash flow H , with t ∈ {G, B} .

These probabilities are assumed to be common knowledge, and are such that

DI Pt ( I ) > 0 and PG ( I ) > PB ( I ) , for I ∈ [ I min , I max ] , and they are strictly concave over the

same interval. In order to ensure that there is a solution to the optimal investment

problem we assume

lim I → I min DI Pt (I ) → ∞

Since information in the G state is more valuable than in the B state, we also

require (Reily, 1979)

DI PG (I ) DI PB (I )
>
PG (I ) PB (I )

This will ensure that there is no more than a single crossing in the iso-utility

diagrams in the plane of investment-equity participation, condition that then guarantees

the existence of a separating equilibrium for the adverse selection problem.

The domestic firm incurs in an irreversible cost Vmin if it undertakes the project;

this will constitute its reservation value. The MNE's reservation value will be taken to

be zero, and all its initial sunk costs are included in I min ; this simplification helps to

produce more transparent expressions, without any loss of generality18. The project is

not so good as to be riskless; the riskiness of the project is summarized by the

conditions

L < I min and L < Vmin .

Presumably, both the MNE and the local firm borrow the necessary funds in

order to carry out the project; therefore, L plays the role of that (riskless) part of the

final cash flow of the project that help repay the initial debt, and the rest of the

obligation will have to come from the risky part ( H − L ) Pt ( I ) of the cash flow. The

expected net present value for the overall project in state B would be enough to cover all

costs, so that there is a positive incentive to enter the joint venture in the first place. This

condition is written as

N B (I min ) := L + (H − L )PB (I min ) − I .

Where N t (I ) − Vmin (I ), with t ∈ {G, B}, is the net present value for state t . Notice that we

are working with a zero rate of return on capital.

The contract between the two parts can be specified by stating the partition of

both equity and debt that each will take. β is the equity participation for the

multinational, then the risky part of its profits will be β t ( H − L ) Pt ( I ) . There is also the

proportion γ t that the multinational takes from the risk-free part of the project,

18
The choice of a reservation value U 0 > 0 just defines the lowest utility acceptable for the
MNE to be willing to undertake the project in association with the domestic firm. It would act as
the equivalent of an opportunity cost.

represented by γ t L . The full specification of the contract is contained in the triad

St := ( I t , β t , γ t ) ; the set of all feasible contracts in state t will be denoted Σt . From these

definitions we can construct the expected utility of the multinational, which will be

U t (S t ) := γ t L + β t (H − L )Pt (I t ) − I t

From this expression is not difficult to see that the MNE will have an incentive

to seek an equity participation as long as the risk-free component of the project does not

become important compared with the risky part, i.e., as long as L < β t (H − L )Pt (I t ) for

all St ∈ Σt . If this were not the case, the riskless cash flows would generate a tension

with the signaling purpose for the MNE to take a large portion of equity in the venture.

Such tension will indeed be present for any amount L , and there is an inverse

relationship between the maximum portion of equity that the MNE is interested in

taking and the size of L . From now on we shall assume that the above mentioned

inequality applies.

The expected utility for the domestic firm is given by

Vt (S t ) := N t (I t ) − U t (S t ) = (1 − γ t )L + (1 − β t )(H − L )Pt (I t )

As will be stated in the next section, this expected value needs to be at least the

reservation value Vmin . This condition would turn equality, if it weren’t for the effect of

the capital restriction rules. The constraint on capital participation is taken to be a state

independent rule that puts an upper bound to the equity portion of the multinational.

This rule is formally expressed as

βt ≤ α .

Where α (≤1) is the capital participation constraint. As discussed below, this constraint

will have the effect of substituting investment distortion for equity participation as a

signaling mechanism for the multinational.

The optimization problem

We address the case when the multinational has the bargaining power and also

the informational advantage. In the G state of the world the optimization problem is

(MB/MI-G):

max S U G (S G ) = γ G L + β G ∆PG (I G ) − I G

s.t. U B (S G ) − U B (S B ) ≤ 0

Vmin − VG (S G ) ≤ 0

With

I G ∈ [I min , I max ] , β G ∈ [0,α ], γ G ∈ [0,1] .

The notation ∆ = H − L will be used throughout the rest of the paper. The first
:

constraint is the incentive compatibility, or ''no-mimicry'' constraint, which prevents the

multinational with B B information to mimic the strategy of the one with G

information. U B ( S B ) is the optimal solution to the maximization problem if the signal is

B . The second constraint is the reservation value constraint for the domestic firm. The

problem when the MNE has information B is stated as (MB/MI-B):

max S U B (S B ) := γ B L + β B ∆PB (I B ) − I B

s.t. Vmin − VB (S B ) ≤ 0

With

I B ∈ [I min , I max ], β B ∈ [0,α ] , γ B ∈ [0,1] .

The optimal solution to MB/MI-B is then used as the constraint for no mimicry

in the G problem. How this constraint acts on the problem with good information

depends in general on the level of the capital participation restriction α and on the

domestic firm's reservation value Vmin . In Appendix A we show the precise form of this

dependence. The main facts about the solution can be intuited from the situation faced

by the multinational when its information is B . Under this condition the multinational

will not use costly overinvestment to signal its type, instead, it will keep all equity

participation compatible with constraints. If the value of α is so high as to be too

lenient, then we may expect that the MNE will proceed to take an all equity position 19,

PO
( )
up to a maximum β ≤ α such that L + (1 − β B )∆PB I B = Vmin , where I B is the Pareto
PO

optimal investment level. The multinational, having the bargaining power, will structure

a contract that keep the domestic's rationality constraint binding, if this is possible.

Under informational disadvantage, the domestic firm will be more interested in the

riskless cash flow component L , in order to cover as much as possible of its reservation

costs. A second regime comes if the capital restriction rule is such that α < β B . If the

( )
reservation value is high, and by this we mean L + (1 − α )∆PB I B ≤ Vmin , the
PO

multinational will invest at the Pareto optimal level, but no more than that. This requires

that the multinational take some share γ B > 0 of the riskless cash flow L . If the

( ) ( )
reservation cost is not so high, such as when (1 − α )∆PB I B > Vmin > (1 − α )∆PB I B ,
PO F

F
where I B is the full information investment, the multinational will take all the share of

( )
debt ( γ B = 1 ) and will invest at a lower level I * ∈ I B , I B in order to just meet the
F PO

( )
reservation constraint. If the reservation value is too low, that is if (1 − α )∆PB I B > Vmin ,
F

19
See the observation made about the value of L in the previous section.

then the reservation constraint will not bind and the multinational will choose the full

F
information level of investment I B . Since this level of investment corresponds to the

global maximum of the multinational's utility, we can not expect it to lower it even

further, no matter how small Vmin may become, therefore the domestic will derive

( ( ))
forced rents in the B state for any α < 1 − Vmin / ∆PB I B .
F

If the multinational type is G then it will need to convince the domestic firm of

its goodness through signaling. This means taking as much equity as allowed by the

imposed capital constraint and the reservation value of the domestic firm. NRS show

that as long as the reservation value constraint remains active, signaling will take the

form of overinvestment, and the contract structure will be such that either β G = α or
*

γ G = 0 . 20
*

The way investment strategy changes with α can be depicted as follows: if α is

too high, then the capital restriction constraint will not be binding, and overinvestment

will come solely as a consequence of the no-mimicry constraint. As α grows smaller,

the effect is to suppress the signaling value of equity participation, which will further

distort the investment policy, and from that point on we get β G = α and γ G > 0 . The
* *

reservation value constraint remains active provided α is not too low. In other words, it

depends on the truth value of the inequality

(1 − α )∆PG (I G ) < Vmin
*
.

Once  is so low that this inequality fails, the reservation constraint stops


being binding and the multinational's signaling through overinvestment no longer

matters, since once the domestic firm starts deriving forced rents, it will not care for any

20
We use ''*'' to indicate the values of the quantities that correspond to the optimal solution.

other form of signaling. As α diminishes further, the no mimicry constraint wi l stop
l

being binding, since paying rents for the G MNE is costlier than for the B type. Since at

this point both constraints have become non-binding, the MNE's investment level will

F
reach the full information, unconstrained value I B . Therefore, for a range of values of

α the investment policy will have gone from overinvestment to underinvestment.

Consequently we can state the following proposition

Proposition 1: When the multinational has both bargaining power and the

information, there is a degree of capital restriction for which the investment level will

be the Pareto efficient level.

The proof is left to Appendix B.

Continuity arguments indicate that there must be some value α c for which the

PO
multinational will invest at the Pareto optimal level I B . In Appendix B we show that

this value is given by the equation

α c ∆PB (I G ) = I G + ∆PB (I B ) − I B − Vmin .
PO PO PO PO

Notice that this formula only involves quantities related to general structural and

technological aspects of the project, meaning that α c can serve as an objective reference

for any negotiation process.

Proposition 2: When the multinational has both bargaining power and the

information, capital participation restrictions can be Pareto improving even when the

information is G. The level of the Pareto optimizing restriction depends only on the

structural characteristics of the project and not on any particular strategic variable.

Observe that as α grows smaller, the value of expected utility for the domestic

firm can only either to increase or stay at the reservation value. This proves the

following

Proposition 3: When the MNE's information is G and it has the bargaining

power and the informational advantage, capital restrictions are beneficial for the

domestic firm.

As is argued above, and more formally demonstrated in Appendix C, capital

restrictions increase the degree of overinvestment to a certain point, namely up to the

point where the MNE with G news starts paying rents to the local firm. After that

threshold is surpassed, then investment levels quickly descend until they reach the

unconstrained level for the optimization problem. Therefore, we can state the following

Proposition 4: Capital restrictions imposed on the MNE when it has the

bargaining power and an informational advantage are good for employment level up to

a certain point, beyond which become detrimental.

A numerical example

At this point it is useful to introduce a case example in order to illustrate all the

features of the optimal solution described in the previous paragraphs. This is the same

case example used in NRS, but we will make a fuller use of its possibilities, in order to

illustrate in detail the evolution of some of its main variables with changes in  .

The specific form for the probabilities is

Pt (I t ) = Pt 0
+
(I t − I min )c
t

.
100

With I t ∈ [25,33] , and PG = 0.75 , PB0 = 0.5 , cG = 0.4 , c B = 0.001 .
0

For the cash flows we choose L = 20 and H = 300 .

The specific value of Vmin that is used to produce the output graphs is Vmin = 29 .

This particular value has no special significance, it was chosen only for illustration

purposes.

The best way to see all the features described in the previous section is through a

graphic representation of some of the more significant variables as a function of α .

Figure 1.1 illustrates the changes undergone by the optimal investment as α descends

from 100% down to 70% of maximum participation. This graph is a plot of the

parameter κ versus α , where κ is a more easily readable measure of I G , defined by the
*

relation I G = I min + κ (I G − I min ) . In terms of this measure κ = 1 is the Pareto optimal
* PO

investment level, κ > 1 means overinvestment and κ < 1 corresponds to

underinvestment. Figure 2 shows the comparative evolution of the utilities for the

multinational and the domestic firm in the G state across the same range of values of α

as in Figure 1.1.

Reading the graph in Figure 1.1 from right to left we can see that very loose

levels of capital restriction have no effect on the multinational's level of investment,

which corresponds to a degree of overinvestment ( κ = 2.5 ) that is just the necessary to

meet the reservation value constraint, and avoids imitation by a B copycat. How this

reflects in both partners' utilities is shown in Figure 2. As the restriction tightens,

additional overinvestment, up to κ = 8.2 , substitutes for lost equity participation as a

signal. At this point the domestic firm starts getting more utility than its reservation

value and the multinational's starts to decline in the same amount. As the maximum

participation goes on reducing, it reaches the point where the multinational starts paying

forced rents to the domestic firm in the G state. At this point signaling through

overinvestment is no longer meaningful, because the B copycat would start having

even a more difficult time sustaining any attempts at mimicry. The investment level

plummets quickly across a very narrow range of values of α , until it reaches the value

I G , the unconstrained level of investment. At this value of α the mimicry constraint
F

ceases to be binding, the multinational with B information no longer capable of

reaching the necessary contract structure.

Further considerations on the relation between effective control and bargaining
power

In several respects, the treatment of the problem as presented in NRS presumes

circumstances and attitudes toward equity ownership in JVs that are at variance to those

known to the author, at least in the Latin American context. It is not frequent the case

that, even in the presence of adverse selection problems, the locals will simply seat

waiting for the MNE to choose its most convenient level of equity as a signal. It is not

in general the case either that governmental restrictions on equity participation act as

sort of ''stone wall'', against which the MNE has no negotiating power. The statements

in LeCraw (1984) are a more realistic guide. There, equity ownership is the result of a

negotiation process in which there are no firmly defined stop points, and this result is

seen as a relatively complex conjunction of four basic conditions 21:

1) The desired ownership level of the MNE.

21
LeCraw (1984), p. 28. See also Vernon (1971), Stopford and Wells (1972) and Franko (1971).

2) The bargaining power of the MNE.

3) The desired level of local equity participation of the host country.

4) The bargaining power of the host government (including the bargaining

power of locally-owned firms in the host country).

The combination of these four factors brings about the immediate consequence

that for MNEs operating within the same industry actual ownership participation varies

within the same country. No such thing as a constant, repetitive top ownership share is

observed. The observed distribution of actual ownership is more the ''equilibrium'' result

of the negotiation process, where the above mentioned factors act as forces. Being this

the case, a MNE seeking to establish a JV in any given country has to consider at least

three variables: its desired equity level, the equity level that is most likely to achieve

after negotiations, and the degree of effective control attained after negotiations. Other

considerations, such as the degree of competition that is likely to find may be seen as

part of the bargaining power that MNE has.

This part of the essay is mostly concerned with the explicit introduction of the

concept of effective control in the description of the bargaining process. One of the

simplifications made in NRS is the implicit assumption that control of critical

operations within the project is not relevant. Whichever of the partners that carries

through the project, or whatever the composition of governance, will have no effects on

either total welfare or the efficacy to get results. However, it can easily be argued on the

basis of the available empirical evidence that this assumption is not justified. In

particular, the ''J'' shaped form in FIGURE 3 of LeCraw (1984) would become hard to

explain within the NRS model, not to mention that it would not be able to say anything

at all about the results presented in FIGURE 2 of the same reference. The present

treatment is an attempt at including these components missing in NRS, so that with a

minimum of additional complications, the expanded model becomes rich enough to deal

with the above mentioned facts.

It is important to clarify from the outset that the maximization problem faced by

the MNE is ex-ante. Therefore, it does not know before hand the results of negotiations

with the local firm/government. In what follows we will voluntarily limit our analysis to

the case of no adverse selection problems, when there is no need for the MNE to be

constrained by an incentive compatibility condition. This is done only for the sake of

transparency of the arguments that follow. Full treatment in the case with adverse

selection will be done elsewhere.

If the MNE has the bargaining power, then when considering the bargaining

process with the local firm/government there are at least two possibilities for the role

that the size of the investment can play: either it is part of the bargaining or it is not. In

case the size of the investment does matter, the ex-ante problem that the MNE must

solve is given by 22

max U (β , γ , I ) = ρ (β )U 1 (β , γ , I ) + (1 − ρ (β ))U 0 (β , γ , I ) .
β ,γ , I

22
Note that this treatment is simplified. The fact that the MNE has no control does not mean
automatically that the local firm has the control. It may be the case that none of the parties has
effective control, situation that seems to happen often enough. In this case the analysis would
entail a third term. The only change with respect to the formulas given in the text is that we
would have
P(β , I ) := ρ MNE (β ) p MNE (I ) + ρ LOC (β ) p LOC (I ) + ρ NONE (β ) p NONE (I ).
Where we must have
ρ MNE (β ) + ρ LOC (β ) + ρ NONE (β ) = 1 ,
with all terms being non-negative. But at this stage we choose to keep matters as simple as
possible.

Where U 1 (β , γ , I ) and U 0 (β , γ , I ) are the utilities for the MNE in case of full control or

no control, respectively. Expressions for each are

U 0 (β , γ , I ) = γL + β∆p0 (I ) − I

and

U 1 (β , γ , I ) := l + γL + β∆p1 (I ) − I

In these expressions all symbols used are as they are explained in other parts of

this essay. If we substitute the above equations back into the maximization problem

we get

max U (β , γ , I ) = ρ (β )l + γL + β∆P(β , I ) − I .
β ,γ , I

Where

P(β , I ) = ρ (β ) p1 (I ) + (1 − ρ (β )) p0 (I ) [1.1]

Notice that if the site of control is irrelevant, i.e., if p1 (I ) = p0 (I ) and l = 0 , then

the above expressions are indistinguishable from those found in Section 1.5 of this

essay, where we discuss the model in NRS. Under the present formulation, the

effective conditional probability for project success P(β , I ) incorporates all the risk

structure associated with the possible outcomes from negotiations. This function

represents the best information that the MNE can count on in order to make its own

estimates for the prospects of success of the JV.

The absence of adverse selection in this problem is not tantamount to an absence

of information asymmetry. The information asymmetry of the problem resides in the

fact that the MNE has no credible way to pass on the knowledge contained in the

effective conditional probability function to the local firm/government. The knowledge

condensed in P(β , I ) is the product of a long process of organizational learning, and

there is simply no way in which this can be meaningfully revealed to an external party

over a short period of time. It is telling how is it likely that project's success prospects

are going to be influenced by ownership structure, since ownership structure will most

likely have a bearing on how decisions are made and implemented. This knowledge-trap

brings us to the matter of the rationality constraint imposed by the local firm. First,

consider the not unlikely case where the MNE is seeking an already established firm to

serve as a service provider for a line of products for the MNE's home market. The local

firm/government does not have information about the prospects for success of the

project, beyond that contained in some ''average'' conditional probability p(I ) . This

probability function does not contemplate the effects of the structure of governance,

because the locals may not believe the MNE's claims to the contrary. From such claims

they may construe that the ''true'' conditional probability p(I ) is somewhere between

p1 (I ) and p0 (I ) , but no more and no less. For the sake of this discussion we shall

assume that p(I ) = (1 2 )( p1 (I ) + p0 (I )) . This understanding of the problem by the locals

is fully known to the MNE. Under these circumstances, if the local firm/government is

assumed to be risk neutral, the rationality constraint takes the form

(1 − γ )L + (1 − β )∆ p(I ) ≥ 1 I + Vmin .
2

Where Vmin corresponds to the reservation value of the local firm. It is important to

realize at this point that this value may not, in general, be invariant with respect to the

condition of which of the parties - if any - does exert effective control of the JV.

Continuing with our simplified scenario of either/or for the possession of effective

0
control, let's call Vmin to the reservation value if the local firm exerts effective control,

1
and Vmin when the MNE has the control. Then, on general grounds related to the

structure of costs, one might expect these two values to differ. If we can attribute these

reservation values as due mostly to operating costs, then the main cost factors are those

related to managerial compensation. Certainly, the costs associated with a local

managerial team, assuming control rests with the local firm, are going to differ

significantly from those caused by managerial fees coming from the MNE, if it has

control. Call C 0 the managerial costs in the first case, and let them be C1 in the second

case. Then in general we must assume C0 ≠ C1 (and in the Latin American context, we

can assume that most likely it is C0 < C1 .) With this notation, if the described factors

make the only important difference between the two circumstances, then we could write

Vmin = Vmin − C 0 as the connection between the two values, while C1 (or at least an
1 0

important part of it) would become a part of I . In all likelihood, there are other reasons

why the two reservation values will be different; for example, the reservation value may

be altered by the efficiencies that one style of management may bring over the other,

independent from the changes in the risk structure represented by the pi (I ) . In the

reservation value constraint given above, the right hand side could be seen only as some

expected reservation value.

Another feature of the rationality constraint is the presence of the internal

transfer value l . The concept that it represents may not be assimilated to that of L . One

has to keep in mind that, contrary to L , which acts as value added by the project to total

welfare, l is an internal - forced - type of payment, and therefore it adds nothing to total

welfare. It would not be adequate to absorb l into a redefinition of Vmin either, because

the presence of l is conditioned by the access to control by the MNE, and includes rents

due to internal pricing, while Vmin is mainly related to operating costs. Finally, and

perhaps most important, the constrained quantity is not the result of the remaining

welfare, once the MNE has taken its share; rather, it is the result of what the locals are

expected to believe about the value of such residual claim. Therefore, it will have a

distortionary effect over the investment policy. Obviously, the extend of the distortion

will depend on how misaligned are the beliefs of the locals with respect to the

knowledge contained in P(β , I ) .

The particular role played by l becomes even more significant if in order to

solve its optimization problem the MNE must assume that the locals are risk averse.

Then, instead of the single constraint above we must write

(1 − γ )L + (1 − β )∆p1 (I ) ≥ l + Vmin
1
[1.2a]

and

(1 − γ )L + (1 − β )∆p0 (I ) ≥ Vmin
0
[1.2b]

It must be remarked that, depending on the value of l , these two constraints do

not necessarily produce feasibility sets such that, for instance, the one for the first

constraint would be a subset of the one for the second constraint. This could be the case

perhaps if l = 0 , but in general for l > 0 this will not be the case.

Second, consider the case when the MNE wants to set up a subsidiary from

scratch in the foreign country, because of labor and/or tax advantages, or perhaps

because of the existence there of natural resources that represent interesting potential

profits. Ideally the MNE would wish to fully own its subsidiary, if it weren't because of

governmental restrictions that may be negotiated within limits dictated in part by the

degree of relative bargaining strength of the MNE. In this case, the proposed firm would

share the full know-how of its parent, and it would be willing to share as well the

associated risks resulting from negotiation. Therefore we would expect the following

rationality constraint to apply:

(1 − γ )L + (1 − β )∆P(β , I ) ≥ ρ (β )l + Vmin [1.3]

We may see the two cases presented above as opposite corners in the space of

possible JV scenarios. The latter case will in general preserve the Pareto optimal level

of investment, while the former will in general distort investment and produce

inefficiencies. These scenarios are depicted by the specifics of the constraints and by the

shape of ρ (β ) . The solution set to the optimization problem composed of (D1) plus any

{ }
of the appropriate constraints will be characterized by a triplet S = γ * , β * , l * .

The conception of relative bargaining power introduced here presents a

challenge when the time comes to be more specific about the properties of ρ (β ) . If its

interpretation is the conditional probability for the MNE to keep effective control, given

that it is able to get at least β in the bargaining process, then we can imagine at least

one particular situation: any MNE should be able to ''bargain'' for a β = 0 equity

participation with a 100% chance to succeed. This could even be the case in practice, if

l is sufficiently large. Therefore we believe that the boundary condition ρ (0 ) = 1 is a

reasonable proposition. Then we could model the bargaining process itself as one in

which, as the MNE advances in the negotiation to reach and achieve a position in β , it

does so by ''spending'' probability for a further gain in equity participation. In other

words, the marginal probability for advancing from β to β + δβ grows smaller with

β ; therefore we make the speculation that ρ ' (β ) ≤ 0 , for all β ∈ [0,1] . Without making

any claims for the universality of the properties proposed above, we believe them

supported by the dynamics expected in many negotiations. Finally, we must remember

the necessary condition 0 ≤ ρ (β ) ≤ 1 for all β ∈ [0,1] . The simplest case that serves as

an example of ρ (β ) is when the MNE has absolute bargaining power, in which case we

get ρ (β ) = 1 , for β ∈ [0,1] . As relative bargaining power is less complete, we may

expect in general to have ρ (β ) smaller, for any given value of β . Except for these

properties, it seems hard to be more specific about the shape of ρ (β ) . There is the

additional problem that this construct is not, in principle, an observable. This is not to

discourage us from representing ρ (β ) with a ''gauge function'', or perhaps we should

call it an ''envelop function'', of a rather simple structure, that meets the general

requirements already mentioned. Only to leave on the record an example that is easy to

interpret, consider the function

ρ (β ) :=
1
,
(1 + β )ν
with ν ≥ 0 .

This expression meets all the requisites made earlier. The case of absolute

bargaining power would correspond to ν = 0 . In this case the MNE keeps 100%

probability of achieving any desired level of ownership. As ν grows larger, the relative

bargaining power of the MNE vis-à-vis the local firm/government is weaker, since the

marginal probability for achieving β + δβ starting from β is smaller the higher ν . This

provides an easy way to rank relative bargaining power for the MNE; its only drawback

is that ν relates monotonically decreasing to the bargaining power of the MNE. ν has

the convenient feature that its range covers the entire non-negative real axis, we could

then interpret the value 1 /ν as a positively related index for the relative bargaining

power of the MNE, since this has the same qualities as ν , but its relationship to

bargaining power is now monotonically increasing.

It is worth to summarize here some of the main results that follow from this

approach. In order to prove them we shall need to write down the Lagrangian for a

particular problem. We choose the case where internal transfers are negligible from the

start, therefore l = 0 . We also focus on the case of a subsidiary created from scratch,

represented by the choice of constraint (1.3). In this particular case the Lagrangian

function is

Λ = [γ + (1 − γ )λ ]L + [β + (1 − β )λ ]∆P(β , I ) − I − λVmin

Where λ is the Lagrange multiplier, positive if and only if the constraint is binding,

and zero otherwise. The first order conditions that follow from this Lagrangian are

given by

∂Λ
= (1 − λ )L [1.4g]
∂γ

∂Λ
= (1 − λ )∆[βP(β , I )]β + ∆λPβ (β , I ) [1.4b]
∂β

and

∂Λ
= [β + (1 − β )λ ]∆PI (β , I ) − 1 = 0 [1.4i]
∂I

In the above set of equations a subindex represents partial differentiation with

respect to the corresponding variable. The last equation is a first order condition that

requires the investment level to be an ''interior'' solution to the problem. This should be

always possible, since both p1 (I ) and p0 (I ) will be required to satisfy the Inada

condition

dpi (I )
lim →∞
I → I min dI

for i =0,1 and I ∈ [I min , I max ].

By assumption p1' (I ) > p0 (I ) for all I ∈ [I min , I max ]. In addition we want the
'

second Lagrangian derivative to be negative at the optimal solution. To make sure this

is the case we impose the sufficient condition p1'' (I ) < p0' (I ) < 0 for all I ∈ [I min , I max ].
'

Then P(β , I ) will allow interior solutions.

Investment policy is determined in a very similar way as in the case for NRS

already shown in Appendix C. From (1.4i) we can get an expression for λ :

1 − β∆PI (β , I )
λ= .
(1 − β )∆P(β , I )
The Pareto optimal investment policy for a given β obtains from the condition

( )
∆PI β , I PO = 1 .

Notice that, different from the case in NRS, Pareto optimality, and therefore the

expected total social welfare, is conditional on how the cake is cut. This is not

surprising, because in this case there is an association between ownership participation

and the decision about who runs the project, which clearly affects total welfare; we shall

refer to this form of Pareto optimality as conditional Pareto optimality (CPO). If we

choose the CPO investment level, then from the expression for λ we get λ = 1 . Since

the general shape of P(β , I ) as a function of I is the same as for the pi (I ) , then we can

conclude that underinvestment will correspond to λ < 1 and overinvestment to λ > 1 .

Underinvestment may go down to the point where the MNE reaches its unconstrained

maximum profit, given β . This investment level I U can be calculated from the

equation:

β∆PI (β , I U ) = 1

The MNE would not have any incentive to invest below this point for the

obvious reason that it would start paying rents to the subsidiary; observe that the above

equation implies λ = 0 .

We shall derive next some of the most immediate results that come out of (1.4g),

(1.4b) and (1.4i). In the absence of transfer payments, the MNE needs to get something

out of the project, therefore it must be the case that

 ∂Λ ∂Λ 
max  ,  ≥ 0 [1.5]
 ∂γ ∂β 

In order to be able to discuss the consequences of this statement, it is necessary

first to recall some of the basic arguments brought to bear in this part of the essay. The

correlation between control and equity ownership makes it important for the MNE to

bargain for as much share of ownership as it is possible in order to undertake a

successful JV. The product ∆βP(β , I ) represents the stakes of the MNE, the reason why

it will want to bargain for more equity, at least up to some value, is because it has the

incentive to do so due to reasons of control. This incentive can be expressed by

imposing the condition that

∂
[βP(β , I )] ≥ 0 , β ∈ [0, β ]. [1.6]
∂γ

Where 0 < β ≤ 1 .

Otherwise the MNE would have no reason why to enter the venture in the first

place. The implication of the above condition is that [βP(β , I )]β ≤ 0 . Therefore, we

should not expect this latter range of values of β to be included in any feasible solution

to the MNE's optimization problem 23. This condition leads to some concrete

consequences.

First, from the general properties of ρ (β ) it follows that Pβ (β , I ) ≤ 0 .

Combining this result with the condition above, it is immediate to see that the MNE

would never overinvest. Because in such case, using condition [1.6], we have λ > 1 ;

both derivatives of the Lagrangian would be negative, and therefore overinvestment is

inconsistent with the given equity share of the project for the MNE. This result can be

argued as well from the very setup of the problem, since there is no asymmetry of

information associated to adverse selection, the MNE does not need to incur in

overinvestment as a signal. Second, in general, investment at the Pareto optimal level

for any given β will only be possible if the MNE has absolute bargaining power, i.e. if

ρ (β ) = 1 for β ∈ [0, β ]. This is so because, from [1.4g] and [1.4b], if the MNE does not

have absolute bargaining power then it would attempt to take all the riskless part of the

cash flows (i.e., γ = 1 ) and, due to the local firm's rationality constraint, it would also

take as low an equity participation as possible, consistent with the holding of effective

( )
control. The rationality constraint is binding, calling β * , I * any pair of values that

would solve the constraint, we have

(1 − β )∆P(β
* *
)
, I * = Vmin .

Differentiating this expression with respect to β * we get

( ) ( ) ( ) ( ) ( ) dβ
*
dI
− P β * , I * + 1 − β * Pβ * β * , I * + 1 − β * PI * β * , I * *
=0.

23
At the extreme, should we have [βP(β , I )]β ≤ 0 for all β , a rational MNE should find
preferable to offer a licensing agreement instead of internalizing through a JV.

dI *
From the stated properties for P(β , I ) , this equation leads to the result >0.
dβ *

Therefore, as β * gets smaller, it takes down the value of I * , below the CPO optimum.

The only exception may happen only if we have Pβ (β , I ) = 0 , in other words, when the

MNE's bargaining power is absolute, in which case the associated Lagrangian derivative

would be nil and it would not require the optimal value of β to drift downward. This

tendency to underinvest is the result of uncertainties in the outcome of negotiations

about the true degree of exertion of effective control.

Discussion

Results in the last section show that investment policy can be significantly

affected by the particular level of capital participation restriction imposed on the MNE,

when it has the bargaining power and also an informational advantage. The MNE will

have an increasing trend of overinvestment as a response to stricter capital participation

restriction, as long as this does not lead to a the MNE paying rents to the local firm.

Once this is the case, and with further reductions in its participation, the MNE will

quickly abandon such trend, and will reduce its investment level to the minimum,

unconstrained value. The utility going to the MNE is a monotonically increasing

function of α , while the utility to the local firm shows the opposite behavior. This is

consistent with NRS appraisals. Overall welfare also shows a dependence on the level

of capital participation restrictions, through its dependence on the level of investment,

with its minimum at the point of maximum overinvestment.

When we inspect the particular numerical example presented in the previous

section, it is interesting to see that the relationship between investment policy and

capital restriction predicted is not one-to-one. For a wide range of values of α , there are

two possible values associated to a single value of the investment level. This suggests

the interesting possibility that any regulator seeking to increase the level of employment

only needs to consider the lower end, positive-slope, side of the investment-restriction

graph, since on that side the local firm also gets to derive rents, beyond its opportunity

cost. Additionally, the extreme variations shown by the optimal investment over the

analyzed range of α , seems to confirm the statements in NRS, that countries with strict

capital restriction policies must show wide variations in the size of investments coming

from multinationals. These features though deserve some further comments, more

related to their empirical relevance. This is admittedly anecdotal evidence; nevertheless

we believe it of some use in the context of the present essay. When actual entrepreneurs

from Colombia who have some international experience were told of these results, they

found it somewhat at odds with their working knowledge of the subject. They argued

that what really happens is that increased capital restriction policies induce the

prospective MNE to underinvest systematically. In terms of the results of this paper

they thought Figure 1.1 should be a simple one-to-one, positive slope curve. This

perception could be due to more than one reason: First, they are most likely right, and

the present model simplifies too much the problems related to asymmetric and/or

imperfect information. Perhaps those MNEs seeking a potential partner for a JV read in

a too strict equity participation restriction signs of uncertainty about the future actions

of the local government, and this form of information asymmetry induces

underinvestment. Second, it might be that a too strict capital restriction policy actually

''scares away'' MNEs with big investment projects, and therefore such policies act as a

sort of filter that is likely to result in that most of the FDI made in the country is de

facto small scale, and therefore what the entrepreneurs see is just a statistical effect.

References

Anonymous, “Alliances: growing popularity over mergers”, Economist, May 15, 1999.

Contractor, F. J. 1990, “Ownership Patterns of U.S. Joint Ventures Abroad and the

Liberalization of Foreign Governments Regulations in the 1980s: Evidence from the

Benchmark Survey”, Journal of International Business Studies , vol. 21, issue 1, 55-73.

Darrough, M. and N. Stoughton, 1989, “A bargaining approach to profit sharing in joint

ventures”, The Journal of Business, vol. 62, 237-270.

De la Torre, J. 1981, “Foreign Investment and Economic Development: Conflict and

Negotiation”, Journal of International Business Studies, vol. 12, Autumn issue, 9-32.

Eckbo, E., R. Giammarino and R. Heinkel, 1990, “Asymmetric Information and the

Medium of Exchange in Takeovers: Theory and Tests”, Review of Financial Studies,

vol. 3, 651-675.

Fagre, N. and L. Wells, 1982, “Bargaining power of multinational and host

government”, Journal of International Business Studies, vol. 13, 9-23.

Fishman, M., 1989, “Preemptive Bidding and the Role of the Medium of Exchange in

Acquisitions”, Journal of Finance, vol. 44, 41-57.

Franko, L. G., 1971, Joint Venture Survival in Multinational Corporations, Praeger,

New York.

Froot, K., 1993, Introduction in Foreign Direct Investment, ed. K. Froot, University of

Chicago.

Gomes-Casseres, B., 1990, “Firm Ownership Preferences and Host Government

Restrictions: An Integrated Approach”, Journal of International Business Studies, vol.

21, 1-22.

Green, R., 1984, “Investment incentives, debt, and warrants”, Journal of Financial

Economics, vol. 13, pp. 115 - 136.

Hansen, R., 1987, “A Theory for the Choice of Exchange Medium in Mergers and

Acquisitions”, Journal of Business , vol. 60, 75-95.

Harrigan, K. R., 1988, “Joint Ventures and Competitive Strategies”, Strategic

Management Journal , vol. 9, issue 2, 141-158.

Hennart, J-F., 1988, “A Transaction Costs Theory of Equity Joint Ventures”, Strategic

Management Journal, vol. 9, 361-374.

Hladik, K., 1985, International Joint Ventures: An Economic Analysis of U.S.-Foreign

Business Partnerships, Lexington Books, Lexington, MA.

Killing, P., 1982, “How to make a Global Joint Venture Work”, Harvard Business

Review, May/June issue, 120-127.

Kogut, B., 1988, “Joint ventures: Theoretical and empirical perspectives”, Strategic

Management Journal, vol. 9, 319-332.

LeCraw, D. J., 1984, “Bargaining Power, Ownership, and Profitability of Transnational

Corporations in Developing Countries”, Journal of International Business Studies , vol.

15, issue 1, 27-43.

Markusen, J. R., 2002, Multinational Firms and the Theory of International Trade, MIT

Press, Cambridge, Mass.

Noe, T. H., M. J. Rebello and M. M. Shrikhande, 2002, “Structuring International

Cooperative Ventures”, Forthcoming in the Review of Financial Studies.

Reily, J., 1979, “Informational Equilibrium”, Econometrica, vol. 47, 331-360.

Sercu, P. and R. Uppal, 1995, International Financial Markets and the Firm, South-

Western College Publishing, Cincinnati, OH.

Stopford, J. and L. T. Wells, 1972, Managing the Multinational Enterprise, Basic

Books, New York.

Stoughton, N. and E. Talmor, 1994, “A mechanism Design Approach to Transfer

Pricing by the Multinational Firm”, European Economic Review, vol. 38, 143-170.

Stulz, R. and W. Wasserfallen, 1995, “Foreign equity investment restrictions, capital

flight, and shareholder wealth maximization: Theory and evidence”, Review of

Financial Studies, vol. 8, 1019-1057.

Vernon, R., 1971, Sovereignty at Bay, Basic Books, New York.

Williamson, O. E., 1975, Markets and Hierarchies: Analysis and Antitrust Implications,

Basic Books, New York.

Williamson, O. E., 1985, The Economic Institutions of Capitalism, Free Press, New

York.

Yoffie, D., 1993, Review of Rival States, Rival Firms: Competing for World Market

Shares, by J. Stopford and S. Strange with J. Henley, New York, NY: Cambridge

University Press, in Journal of Economic Literature, vol. 31, 1462-1463.

FIGURES

9

8

7
G investment policy

6

5

4

3

2

1

0
60% 65% 70% 75% 80% 85% 90% 95% 100%
capital constraint level

Figure 1.1: The vertical axis shows the investment policy, as represented by the scaled
investment variable κ. The Pareto optimal investment level corresponds to κ =1. Notice
the tendency of the G MNE to overinvest, as long as the B MNE could mimic its
strategy. If capital participation constraints are too tight, excess investment by the G
MNE quickly drops.

180
160
profits (arbitrary scale)

140

120
100
80

60
40

20
0,4 0,5 0,6 0,7 0,8 0,9 1
capital constraint level

Figure 1.2: The vertical axis shows profits for the G MNE (upper, continuous) and for
the local partner (lower, dashed) as a function of capital participation constraint level.

Ownership and control in multinational joint ventures

More Related Content

What's hot (18)

Viewers also liked (7)

Similar to Ownership and control in multinational joint ventures (20)

Recently uploaded (20)

Ownership and control in multinational joint ventures