TECHNOLOGICAL ASYMMETRY AMONG FOREIGN INVESTORS … · construction machinery electrical mach automotive all sectors Sophistication Technological Sophistication FDI in the Russian

TECHNOLOGICAL ASYMMETRY AMONG FOREIGN INVESTORSAND MODE OF ENTRY

BEATA S. JAVORCIK and KAMAL SAGGI∗

How does the preferred entry mode of foreign investors depend on their tech-nological capability relative to that of their rivals? This article develops a sim-ple model of entry mode choice and evaluates its main testable implication usingdata on foreign investors in Eastern European countries and the successor statesof the Soviet Union. The model considers competition between two asymmetricforeign investors and captures the following trade-off: while a joint venture (JV)helps a foreign investor secure a better position in the product market vis-a-visits rival, it also requires that profits be shared with the local partner. The modelpredicts that the efficient foreign investor is less likely to choose a JV and morelikely to enter directly relative to the inefficient investor. Our empirical analy-sis supports this prediction: foreign investors with more sophisticated technolo-gies and marketing skills (relative to other firms in their industry) tend to pre-fer direct entry to JVs. This empirical finding is robust to controlling for hostcountry–specific effects and other commonly cited determinants of entry mode.(JEL F13, F23, O32)

I. INTRODUCTION

During the past several decades, there hasbeen a significant change in the attitudes ofmany countries toward inflows of foreign directinvestment (FDI). From being viewed asexploiters, foreign investors are now welcomedas a source of new technologies, know-how,better management, and marketing techniques.One only needs to consider the large-scale eco-nomic liberalization that has been undertaken byEastern European transition economies and thesuccessor states of the Soviet Union to appreci-ate the reversal in attitudes toward FDI that hasoccurred in the world. The increasingly liberalattitude toward FDI has been accompanied byrenewed interest among policy makers and aca-demic researchers in the relationship between

*The authors thank an anonymous referee, an associateeditor, the editor, and Dann Millimet for helpful comments.All errors are our own.Javorcik: Reader, Department of Economics, University of

Oxford, Manor Road Building, Manor Road, OxfordOX1 3UQ, UK. Phone (44 1865) 271-065, Fax (44 1865)285-900, E-mail [email protected]

Saggi: Department of Economics, Southern MethodistUniversity, Dallas, TX 75275-0496. Phone 214-768-3274, Fax 214-768-1821, E-mail [email protected]

the technology transfer and the entry mode usedby foreign investors.1

An interesting finding in the existing empir-ical literature on international technology trans-fer is that the technologies of joint ventures(JVs) tend to be of an older vintage relative tothose employed by wholly owned subsidiariesof multinational firms (Mansfield and Romeo1980). A possible explanation for this find-ing is that firms are reluctant to share state-of-the-art technologies with local partners inforeign countries due to the fear of potential

1. There are several reasons for this interest. Forinstance, the consequences of restrictions on foreign owner-ship that used to be prevalent in many developing countriesand are still present in some of them are likely to depend onwhether mode choice is systematically related to technol-ogy transfer. Similarly, the degree of spillovers from FDImay also vary with the mode of investment (Blomstromand Sjoholm 1999; Javorcik 2004a; Javorcik and Spatareanu2008).

ABBREVIATIONS

EBRD: European Bank for Reconstructionand Development

FDI: Foreign Direct Investment GDP: Gross Domes-tic ProductIPR: Intellectual Property RightJV: Joint Venture

415

Economic Inquiry doi:10.1111/j.1465-7295.2009.00250.x(ISSN 0095-2583) Online Early publication October 21, 2009Vol. 48, No. 2, April 2010, 415–433 © 2009 Western Economic Association International

416 ECONOMIC INQUIRY

competition.2 However, this explanation ignoresthe fact that firms have an incentive to uti-lize their best technologies to compete moreeffectively with their rivals. In other words,to fully understand the relationship betweenmode choice and technology transfer, one needsto account for competitive pressures amonginvestors. In this article, we investigate the fol-lowing question both theoretically and empir-ically: do foreign firms that possess technolo-gies superior to their rivals’ prefer direct entryto JVs? We develop a simple duopoly modelof mode choice and evaluate its main findingempirically by using data from a 1995 survey offoreign investors in Eastern European transitioneconomies and the former Soviet republics con-ducted by the European Bank for Reconstructionand Development (EBRD).

A casual examination of our sample suggeststhat there might indeed be a systematic rela-tionship between a firm’s relative technologicalsophistication and its preferred mode of entry.We measure a firm’s technological sophistica-tion by the ratio of its research and development(R&D) intensity (R&D expenditures as a per-centage of total sales) to the average R&D inten-sity of major firms operating in the same indus-try in industrialized countries. Figure 1 presentsthe average technological sophistication index offoreign investors entering directly and via JVs inthe top five investment destinations in our sam-ple broken down by industry. As is clear, JVsare associated with lower values of the indexin all five of these countries except Russia. Forexample, in the case of the Czech Republic andHungary, JVs are associated with lower valuesof the index in eight of nine industries.

Similarly, as Figure 2 indicates, a cross-country comparison of the average technolog-ical sophistication of investors operating JVsand subsidiaries for several broadly definedindustries reveals a similar pattern: in food,machinery, electronics, and automobile indus-tries, investors with higher technological sophis-tication seem to prefer direct entry to JVs.

2. This fear may be particularly strong in corrupt coun-tries where disputes between JV partners are unlikely to beadjudicated fairly (Javorcik and Wei 2009). For an alterna-tive explanation, see Ramachandran (1993). In her model,the effort expended by the local agent determines the extentof technology transfer, and such effort is undersupplied ina JV due to double moral hazard created by profit sharing.Mattoo, Olarreaga, and Saggi (2004) provide a comparativeanalysis of technology transfer under two modes of FDI:direct entry and acquisition.

In addition to these pointers from our sam-ple, two prominent stylized facts regarding FDIare relevant for our analysis. First, it is wellknown that foreign investors often choose JVsto team up with local partners that possess com-plementary skills and assets. For example, in asurvey of JVs in developing countries, more than65% of the foreign respondents rated knowl-edge of local politics, government regulations,local customs, and local markets as importantconsiderations for seeking local partners (Milleret al. 1996). Of course, while local partners inJVs often bring much needed skills to a project,they also require compensation for their servicesthrough some sort of profit sharing. The secondstylized fact worth noting is that multination-als operate mostly in oligopolistic markets andare quite responsive to each other’s decisions(Caves 1996). Thus, it is important to capture thestrategic decision making involved in the choiceof entry mode.

Motivated by these considerations and Figures1 and 2, we construct a highly stylized modelwith the goal of deriving an empirically testableimplication that relates the entry mode chosenby an investor to its technological position vis-a-vis its rivals. In our duopoly model, invest-ing firms are technologically asymmetric andtwo independent parameters quantify the rent-sharing aspect and the complementarity of JVpartners. The main result is that the more techno-logically advanced a foreign investor, the morelikely it is to choose direct entry over a JV. Infact, in equilibrium, it is never the case thatthe technologically advanced firm forms a JVand its rival firm chooses direct entry.3 Oursimple model adds value to the theoretical liter-ature on mode choice by considering compe-tition between asymmetric investors. By con-trast, most existing models typically considerthe case of a single investor, and when theydo consider multiple investors, they either donot allow for JVs or consider only symmet-ric investors (Asiedu and Esfahani 2001; Ethierand Markusen 1996; Horstmann and Markusen1992; Markusen 2001).

3. One has to be careful here: the model does not saythat the efficient firm never chooses a JV but rather that,if the inefficient firm does not choose a JV, the efficientfirm will never do so either. There certainly exist parametervalues for which both firms opt for a JV.

JAVORCIK & SAGGI: TECHNOLOGICAL ASYMMETRY AMONG FOREIGN INVESTORS 417

FIGURE 1Technological Sophistication Index of Investors Undertaking JVs versus Direct Entry by Host

Country

FDI in HungaryFDI in Poland

food

food

met

al

cons

truc

tion

mac

hine

ry

mac

hine

ry

pape

r

drug

s,co

smet

ics

drug

s,co

smet

ics

chem

ical

s

chem

ical

s

met

al p

rod

met

al p

rod

elec

tron

ics

elec

tron

ics

auto

mot

ive

auto

mot

ive

mis

c

all s

ecto

rs

all s

ecto

rs

Tec

hnol

ogic

alS

ophi

stic

atio

n

Direct entry JV

Direct entry JV Direct entry JV

Direct entry JV

Direct entry JV

FDI in the Czech Republic

Tec

hnol

ogic

alS

ophi

stic

atio

n

FDI in the Slovak Republic

0123456

0

1

2

3

4

5

6

food

mac

hine

ry

pape

r

chem

ical

met

al p

rod

elec

tric

alm

ach

elec

tric

al m

ach

elec

tron

ics

auto

mot

ive

all s

ecto

rsal

l sec

tors

cons

truc

tion

mac

hine

ry

elec

tric

alm

ach

auto

mot

ive

all s

ecto

rs

Tec

hnol

ogic

alS

ophi

stic

atio

nT

echn

olog

ical

Sop

hist

icat

ion

FDI in the Russian Federation

0.5

1.5

2.5

0

0.5

1

1.5

2

2.5

3

0

0.5

1

1.5

2

2.5

3

food

beve

rage

met

al

mac

hine

ry

prin

ting

drug

s,co

smet

ics

met

al p

rod

elec

tron

ics

auto

mot

ive

mis

c

Tec

hnol

ogic

alS

ophi

stic

atio

n

To evaluate the empirical validity of our mainresult, we use a unique data set based on a sur-vey conducted by the EBRD among major com-panies in the world asking about their invest-ments (if any) in transition countries of EasternEurope and the former Soviet Union. The sur-vey results are supplemented with firm-specificinformation from the Worldscope database.4 We

4. This data set was also used by Javorcik (2004b).

estimate a probit model with the dependent vari-able taking on the value of 1 if investor i

engages in a JV with a local partner in coun-try k and 0 if it chooses direct entry. The resultslend support to our theoretical model by indi-cating that firms possessing more sophisticatedtechnologies (relative to other firms operatingworldwide in their industry) are less likely toengage in JVs and more likely to enter the mar-ket directly. The same is true of firms withabove-average marketing sophistication, which


FIGURE 2Technological Sophistication Index of Investors Undertaking JVs versus Direct Entry by Industry

FDI in Food Sector

0123456

Tec

hnol

ogic

alS

ophi

stic

atio

n

Direct entry JV

Direct entry JV

Direct entry JV

Direct entry JV

FDI in Machinery Sector

FDI in Electronics Sector FDI in Automobile Sector

00.5

11.5

22.5

33.5

Cze

chR

ep.

Hun

gary

Pol

and

Rus

sia

Bul

garia

Est

onia

Latv

ia

Lith

uani

a

All

coun

trie

s

Cze

chR

ep.

Hun

gary

Pol

and

Slo

vaki

a

Rus

sia

Est

onia

All

coun

trie

s

Tec

hnol

ogic

alS

ophi

stic

atio

n0

0.51

1.52

2.53

Tec

hnol

ogic

alS

ophi

stic

atio

n

00.5

11.5

22.5

Cze

chR

ep.

Hun

gary

Pol

and

Rus

sia

Rom

ania

Bul

garia

Est

onia

Latv

ia

Ukr

aine

All

coun

trie

s

Cze

chR

ep.

Hun

gary

Pol

and

Slo

vaki

a

Rus

sia

All

coun

trie

s

Tec

hnol

ogic

alS

ophi

stic

atio

n

is also related to the firm’s endowment ofintangible assets. Coefficients on both variables(relative technological and marketing sophisti-cation) are statistically significant and remainso even when entered into the same equation.These results are robust to the inclusion ofother investor-specific characteristics as well ashost country and industry fixed effects. How-ever, when both host country and industry fixedeffects are included in the same equation, onlythe technological sophistication index remainssignificant.

As a further robustness check, not employedin the earlier studies of mode choice,5 weestimate a two-stage model that captures twochoices: (1) the decision to undertake FDI ina given country and (2) the choice betweena direct entry and a JV, conditional on FDItaking place. In other words, we control forselection bias that may be present when onlyactual investment projects are considered andobservations pertaining to firm-country pairswith no investment are discarded. We also notethat our results are robust to controlling forcountry characteristics such as the evolution ofthe transition process and the distance betweenthe source and the host country. Finally, wecheck whether our results could be driven by the

5. With the exception of Javorcik and Wei (2009).

possibility that the link between the entry modeand the technological sophistication reflects fearof technology leakage in countries with a weakprotection of intellectual property rights (IPRs).This, however, does not seem to be the case.While we find that foreign investors are attractedto locations with stronger IPR protection, wefind no evidence of IPR legislation affecting thechoice of entry mode.

While existing empirical studies of entrymode find a negative relationship between theimportance of firm-or industry-level intangibleassets and the probability of entering through aJV (Asiedu and Esfahani 2001; Gatignon andAnderson 1988; Gomezs-Casseres 1989), thisarticle focuses on technological and market-ing sophistication of investing firms relative toother firms operating in the industry worldwide.Thus, we concentrate on intraindustry differ-ences as a determinant of mode choice in addi-tion to controlling for interindustry effects. Theonly existing study that considers intraindus-try differences, Javorcik (2006), differs in focusas its objective is to compare high- and low-technology industries in this respect. Moreover,we employ a data set that is unique in the extentof its coverage. Previous studies on the choiceof entry mode use data on FDI originating inone source country (i.e., Sweden in the case ofBlomstrom and Zejan 1991, or the United States


as in the case of Asiedu and Esfahani 2001) oron FDI entering a single host country (typicallythe United States). Our data set covers invest-ment projects undertaken in multiple economiesby investors from all over the world.

This article is structured as follows. The nextsection presents our theoretical model of entrymode choice. Section III discusses our empiricalstrategy, the data used, and the results obtained.The last section concludes. Details of theoreticalderivations and the data used are collected inseparate appendices.

II. MODEL

In this section, we develop a partial-equilib-rium duopoly model of mode choice. Twoforeign firms are considering entry into a marketwhere the inverse demand function is given byp(q) and q denotes total output. Each firm canenter the market directly and produce the goodon its own or form a JV with a local partner wholacks the ability to produce the good alone. Lete denote direct entry and j a JV. The technologyof production and distribution depends on modechoice in the following way. If firm i decidesto enter the market directly (i.e., by establishinga wholly owned subsidiary), it requires αi unitsof labor for producing each unit and βi units fordistributing it. Thus, under direct entry, firm i’smarginal cost equals

Cei = αi + βi .(1)

By definition, under a JV, firm i must sharesome rents with its local partner. Let firm i’sshare of the total profit of the JV be given by θ,where θ ∈ [0, 1].6 The advantage of forming aJV is that the local partner brings knowledge andexpertise about the host country market, whichlowers the unit labor requirement in distributionto δβi , where δ ∈ [0, 1].7 The smaller is δ, the

6. Since in our data set we do not have informationregarding the equity structure of JVs, we leave θ as anexogenous parameter. Al-Saadon and Das (1996) constructa model of an international JV in which ownership sharesare endogenously determined as the outcome of bargainingbetween a multinational firm and a single host firm. Foranalyses of policy restrictions on the degree of foreignownership, see Mattoo, Olarreaga, and Saggi (2004) andKlimenko and Saggi (2007).

7. Note that δ is intended to represent more generally thecontribution of the local partner to the JV. Such contributionmay take the form not only of access to distributionnetworks but also of knowledge of local tastes, suppliers,and legislation as well as an improved ability to navigatethrough the bureaucratic maze in the host country.

lower a JV’s unit cost of distribution. Thus,under a JV, firm i’s unit cost is given by

Cj

i = αi + δβi .(2)

To generate technological asymmetry betweenforeign investors, assume that α1 ≤ α2 and β1 ≤β2. Note that, holding constant the mode ofentry, Firm 1 has a lower marginal cost thanFirm 2.

Now consider the following market entrygame. In the first stage, each firm choosesbetween the two modes of entry (JV vs. directentry). Next, both firms compete in quantities(Cournot-Nash competition). Firm i’s profit fun-ction at the output stage is given by πi (qi , q−i ) =(p(q) − ci)qi , and the associated first-order con-dition for profit maximization can be written asfollows:

∂πi (qi, q−i )

∂qi

= p + p′qi − ci = 0.(3)

Solving the above first-order conditions yieldsthe equilibrium in the product market. Let thepair (x, y) denote the regime where Firm 1chooses entry mode x and Firm 2 choosesentry mode y, where x, y = e, j . Further-more, let π

ey

1 denote Firm 1’s equilibrium profitunder regime (e, y) and θπ

jy

1 under regime (j ,y). Similarly, interpret Firm 2’s payoffs πxe

2 andθπ

xj

2 .Since Firm 1 is more efficient than Firm 2,

it is always the case that πey

1 ≥ πxe2 and θπ

jy

1 ≥θπ

xj

2 : under the same entry mode, Firm 1 hashigher total profit than Firm 2. Of course, itneed not be the case that Firm 1 has higherprofit when it chooses direct entry and whenFirm 2 chooses a JV. This ranking depends onthe parameters of the model.

To describe the subgame perfect equilibriumof this model, we need two definitions. Denotethe change in firm i’s profit that results from itsswitching from a direct entry to a JV, given thatits rival adopts direct entry by ui :

u1 ≡ θπje

1 − πee1 and u2 ≡ θπ

ej

2 − πee2 .(4)

Each firm can gain market share at theexpense of its rival by forming a JV andlowering its marginal cost. Of course, to do so, afirm must forsake some of the total profit of theJV to the local partner. Firm i prefers a JV toa direct entry given that its rival chooses directentry iff ui > 0. We will say that firm i has aunilateral incentive for a JV iff ui > 0.


Let �i denote the change in firm i’s profitthat results from its switching from a direct entryto a JV, given that its rival forms a JV:

�1 ≡ θπjj

1 − πej

1 and �2 ≡ θπjj

2 − πje

2 .(5)

A firm has a motive for forming a JV inresponse to a JV by its rival because it toocan lower its cost and regain some of its lostmarket share. The function �i measures thestrength of this motive. We will say that firm ihas a competitive incentive for a JV iff �i > 0.Using these two sets of incentive functions, thesubgame perfect equilibrium of the model canbe described in a succinct way.

DEFINITION 1. The equilibrium mode choiceof the two firms is as follows: (1) both firmschoose direct entry (e, e) iff ui ≤ 0, (2) Firm 1chooses direct entry, while Firm 2 a JV (e, j)iff �1 ≤ 0 and u2 > 0, (iii) both firms choose aJV (j, j) iff �i ≥ 0, and (iv) Firm 1 chooses aJV and Firm 2 direct entry (j, e) iff u1 > 0 and�2 < 0.

Our main interest is in relating a firm’s pre-ferred mode of entry to its technological capa-bility relative to its rival. To this end, we exam-ine how the likelihood of a particular regimebeing an equilibrium changes with a change inthe underlying technology of the two firms. The“likelihood” of a regime is measured by the areaof the parameter space over which that regimeemerges as an equilibrium. For example, if wesay that a change in some underlying parametermakes it more likely that a firm has a unilateralincentive for a JV, we mean that the parame-ter space over which the function ui is positiveincreases.

To facilitate analytical derivations and com-parisons of the incentive functions, assume thatthe demand function is linear: p = a − bq.

PROPOSITION 1. An increase in a firm’smarginal cost (caused due to an increase ineither αi or βi) makes it more likely that afirm has a unilateral as well as a competitiveincentive for a JV. Similarly, an increase in themarginal cost of its rival makes it less likelythat the firm has a unilateral or a competitiveincentive for a JV.8

A corollary to the above result can also bestated.

8. Proofs of proposition 1 and corollary 1 are inAppendix I.

FIGURE 3Equilibrium Mode Choice

COROLLARY 1. Whenever Firm 1 has unilat-eral incentive for a JV, so does Firm 2. Further-more, the regime (j, e) where Firm 1 chooses aJV and Firm 2 direct entry does not constitutean equilibrium.

Corollary 1 implies that we cannot have anequilibrium in which only the efficient firm (i.e.,Firm 1) forms a JV. Figure 3 illustrates a typ-ical equilibrium pattern in (δ, θ) space.9 In thisfigure, we plot the zero contours for the incen-tive functions u2 and �1. The other incentivefunctions are omitted from this figure since theyare not needed to describe the equilibrium modechoice. For example, the function u1 is notplotted since it lies below all the other threefunctions and does not play a critical role indetermining the equilibrium mode choices offirms. Furthermore, the fact that u1 lies below�2 implies that if the efficient firm has a unilat-eral incentive for a JV, the inefficient firm hasa competitive incentive to do the same, therebyruling out (j, e) as an equilibrium entry regime.

Two properties of Figure 3 are worth not-ing. First, the zero contours for all incentivefunctions are upward sloping. This commonproperty of all zero contours follows from themodel’s fundamental trade-off: as the local part-ner’s contribution becomes less valuable (i.e., asδ increases), each firm requires a higher share θ

9. The parameters used for this figure are as follows:a = 10, α1 = 1, β1 = 1, α2 = 2, and β2 = 2.5. There isnothing special about these parameter values except that theygive a clean figure. The propositions and corollaries statedin the article hold for all permissible parameter values.


of the JV’s total profit if it is to remain indif-ferent between a JV and a direct entry. Second,higher profit contours lie in the southeast region:an increase in θ and a decrease in δ make a JVmore attractive relative to direct entry.10

Figure 3 can be divided into three regions.Above the zero contour for the u2 function, (e, e)is the equilibrium. In this region, δ is large andθ is small so that the local firm receives a largeshare of the total profit of the JV even thoughit does not make a valuable contribution to theJV. As a result, in this region, direct entry is thedominant mode of entry for both firms. In theregion between the zero contours for the u1 andthe �2 functions, (e, j) is the equilibrium: here,the contribution of the local partner is not largeenough for Firm 1 to opt for a JV, whereas itis sufficient to induce Firm 2 to choose a JV.Finally, in the region below the zero contour forthe �2 function, (j, j) is the equilibrium: here,the local partner’s expertise really matters, andthe profit share of foreign firms is large (makinga JV attractive to them).

The model presented above shows how theincentives of firms to choose JVs over directentry vary with their technological capabilities.In a broad sense, the main empirical predictionof the model is that the more technologicallysophisticated a firm is relative to its rivals, theless likely it is to enter the market via a JV. Wenow turn to an econometric verification of thisprediction.

III. EMPIRICAL EVIDENCE

In this section, we test the main predictionof the theoretical model. The empirical work isdescribed in three steps. We first present somesummary statistics, then discuss our econometricspecification and report our regression results.

A. Summary Statistics

As noted earlier, the data set used in thisstudy is based on the EBRD survey of for-eign investors supplemented with the informa-tion obtained from the Worldscope database.In January 1995, a brief questionnaire wassent out to all companies (about 9,500) listed

10. As should be clear, the model has many exogenousparameters, and figures corresponding to Figure 3 can bedrawn in the space of other parameters as well. Figure 3has been drawn in the (δ, θ) space because both of theseparameters lie between 0 and 1, thereby allowing a cleanrepresentation of equilibrium.

in Worldscope. Responses were obtained from1,405 firms, which reported whether they hadundertaken investments in Eastern Europeantransition economies and the successor states ofthe Soviet Union. Our sample covers 21 hostcountries. Further details about the survey andthe data are given in Appendix II.

Table 1 presents the breakdown of entrymodes chosen by foreign investors in our samplefor each of the host countries. Note that JVsoutnumber direct entries in most host countriesand constitute 59% of all projects.

Table 2 presents the percentage of foreigninvestors who chose a given entry mode ineach industry in our sample. The figures indicatethat JVs were the dominant form of investmentin a majority of industries. However, it isstriking that in the drugs, cosmetics, and healthcare products sector, only 12% of all projectswere JVs, while direct entries accounted for88% of investments. Similarly, wholly ownedprojects constituted 84% of all investments inthe beverage sector. It is worth noting thatthe drugs, cosmetics, and health care productssector is the most R&D-intensive industry in oursample, while the beverage sector relies heavilyon advertising and investments in marketing.

TABLE 1Entry Modes Chosen by Investors in the

Sample

Host Country Direct Entry JV Total

Russia 29 72 101Poland 45 60 105Czech Republic 47 43 90Hungary 41 37 78Slovak Republic 16 22 38Ukraine 5 17 22Estonia 8 16 24Romania 10 14 24Bulgaria 11 10 21Latvia 6 10 16Slovenia 3 10 13Kazakhstan 6 8 14Lithuania 5 6 11Croatia 4 6 10Belarus 3 4 7Georgia 2 4 6Uzbekistan 1 4 5Albania 1 3 4Macedonia FYR 1 2 3Azerbaijan 1 1 2Moldova 0 1 1Total 245 350 595


TABLE 2Industry Breakdown of Entry Modes Chosen by Investors in the Sample

IndustryDirect Entry as % of AllProjects in the Industry

JVs as % of AllProjects in the

Industry

Total No. ofProjects in

the Industry

Recreational products 100.0 0.0 5Drugs, cosmetics, andhealth care products

87.9 12.1 58

Beverages 84.2 15.8 19Electrical 67.7 32.3 31Apparel 50.0 50.0 2Printing and publishing 50.0 50.0 4Metal products 42.1 57.9 19Food 40.4 59.6 57Automotive 40.0 60.0 25Textiles 40.0 60.0 5Metal 33.3 66.7 27Machinery and equipment 32.2 67.8 90Electronics 32.1 67.9 78Aerospace 22.2 77.8 9Chemicals 22.0 78.0 59Paper 19.0 81.0 21Diversified 4.8 95.2 21Tobacco 0.0 100.0 5Total 41.5 58.5 595

Table 3 compares the average R&D intensityof investors engaged in direct entry with that ofinvestors sharing ownership in each three-digitSIC (Standard Industrial Classification) sector.The sectors are grouped into high-, medium-, and low-technology category following theclassification used by Blomstrom, Lipsey, andOhlsson (1991).

As Table 3 indicates, in all but one high-technology industry, investors undertaking directentry are on average more R&D intensive thanthose sharing ownership. For instance, in thedrugs sector, the average value of R&D spend-ing is equal to 15.7% of sales in the case ofdirect entry and 10.6% in the case of JVs.For the communications equipment, the corre-sponding figures are 13.3% and 5.6%. And inthe case of electronic components and acces-sories, 5.6% and 3.4%. In medium-technologyindustries, which include industrial chemicals,motor vehicles, household appliances, and soforth, in half of the sectors in which both modesare present, investors entering a host countrydirectly are characterized by a higher level ofR&D efforts. The average R&D outlays areequal to 3.8% of sales for direct entry and 3.2 forJVs. In low-technology sectors, this is true in 10of 16 cases. In each of the three groupings, the

average R&D intensity of firms entering directlyis higher than that of firms engaged in JVs.

B. Econometric Specification

Denote a firm by i and a country by k anddefine a binary variable JVik such that

JVik ={

1 if JV∗ik > 0

0 if JV∗ik ≤ 0,

where JV∗ik is unobserved and it determines the

attractiveness of a JV relative to direct entry tofirm i while investing into country k. We furtherposit that

JV∗ik = W i� + βti + δk + εik,

where ti is an index of technological sophis-tication of firm i, W i is the vector of otherfirm-specific determinants of the mode choiceand δk captures country fixed effects. The aboveequation is estimated using a probit model withthe dependent variable taking on the value of 1if the project undertaken by firm i in country kis a JV and 0 if it is direct entry. Our modelpredicts that β < 0.

The choice of explanatory variables employedin the estimation is driven by the predictions of


TABLE 3R&D Intensity of FDI Projects in Three-Digit SIC Industries

SIC Code JVs Direct Entry All

High-technology sectorsDrugs 283 10.62 15.71 15.23Measuring and controlling devices 382 9.94 9.08 9.61Aircraft and parts 372 7.48 9.44 8.08Communications equipment 366 5.60 13.31 7.06Medical instruments and supplies 384 4.58 5.07 4.99Electronic components and accessories 367 3.39 5.63 4.14Computer and office equipment 357 4.09 4.09Search and navigation equipment 381 3.20 3.20Average 6.36 12.67 9.54

Medium-technology sectorsRefrigeration and service machinery 358 7.26 7.26Electric distribution equipment 361 7.26 7.26Hose, belting, gasket, and packing 305 6.00 6.00 6.00Plastics materials and synthetics 282 4.65 4.86 4.71Special industry machinery 355 4.22 5.68 4.70Industrial inorganic chemicals 281 4.09 6.23 4.46Motor vehicles and equipment 371 3.91 4.49 4.17Railroad equipment 374 1.49 4.60 3.05Household audio and video equipment 365 5.79 1.03 2.93Metalworking machinery 354 2.68 2.56 2.66Soap, cleaners, and toilet goods 284 2.60 2.60General industrial machinery 356 2.30 2.30Ship and boat building and repair 373 2.14 2.14Engines and turbines 351 2.11 2.11 2.11Construction and related machinery 353 1.83 2.49 2.03Industrial machinery, nec 359 1.75 1.75Miscellaneous manufactures 399 1.59 1.59 1.59Miscellaneous chemical products 289 1.31 1.31Miscellaneous plastic products, nec 308 1.22 0.11 1.11Farm and garden machinery 352 0.00 3.68 0.74Electric lighting wiring equipment 364 0.67 0.67Rubber and plastics footwear 302 0.00 0.00 0.00Average 3.21 3.76 3.35

Low-technology sectorsPrinting trade services 279 5.25 5.25Preserved fruits and vegetables 203 4.24 4.24Broadwoven fabric mills, wool 223 4.00 4.00Nonferrous rolling and drawing 335 1.54 5.11 3.16Heavy construction, excluding highway 162 2.70 2.70Electrical work 173 2.67 2.67Copper ores 102 1.75 2.84 2.29Cutlery, hand tools, and hardware 342 2.22 2.28 2.27Nonresident building construction 154 1.25 2.94 1.93Miscellaneous food and kindred products 209 1.86 1.86Sugar and confectionery products 206 1.83 1.83Miscellaneous metal ores 109 1.73 1.73Manifold business forms 276 1.43 1.43Miscellaneous textile goods 229 1.40 1.40Clay, ceramic, and refractory minerals 145 1.35 1.35

continued


TABLE 3Continued

SIC Code JVs Direct Entry All

Secondary nonferrous metals 334 1.34 1.34Primary nonferrous metals 333 1.23 1.23 1.23Iron ores 101 1.21 1.21Miscellaneous converted paper products 267 0.21 1.34 1.15Miscellaneous nonmetallic mineral products 329 0.76 2.43 1.13Metal cans and shipping containers 341 1.20 0.79 0.99Blast furnace and basic steel products 331 0.93 0.93Meat products 201 0.79 0.91 0.85Grain mill products 204 0.68 1.10 0.72Glass and glassware pressed or blown 322 0.65 0.65Miscellaneous wood products 249 0.63 0.63 0.63Paper mills 262 0.60 0.67 0.61Dairy products 202 0.57 0.57Highway and street construction 161 0.55 0.55Fabricated structural metal products 344 0.00 0.82 0.55Paperboard containers and boxes 265 0.44 0.33 0.40Carpets and rugs 227 0.36 0.36Cement, hydraulic 324 0.28 0.28Fats and oils 207 0.15 0.15 0.15Beverages 208 0.35 0.13 0.15Gold and silver ores 104 0.00 0.00Commercial printing 275 0.00 0.00Average 0.87 1.76 1.28

Note: nec = not elsewhere classified.

our model as well as by the earlier empiricalliterature. All variables, with the exception ofregional experience that comes from the survey,are taken from Worldscope and are for 1993(or the closest year for which the informationis available). Further details about each of thevariables are given in Appendix II.

Technological Sophistication. To capture thesophistication of an investor’s technology, weuse the ratio of its R&D intensity relative to theaverage value in its industry.

One caveat of using relative R&D expendi-ture as a proxy for technological sophisticationis that R&D intensity is not a perfect mea-sure of a firm’s success in innovative activities.Furthermore, in low-technology sectors, differ-ences between (small in general) R&D activi-ties may not have strong effects. Sophisticationin terms of marketing skills and ownership ofbrand names may be far more important in someindustries. To allow for this possibility and toexplicitly account for the model’s predictionsregarding marketing/distribution costs, we alsocontrol for the investor’s advertising intensityrelative to the industry average.

To capture how important these intangibleassets are for a particular industry, we includethe average values of R&D and advertisingintensity at the industry level. This allows usto take into account both intraindustry andinterindustry effects.11

Firm Size. Blomstrom and Zejan (1991) suggestthat smaller firms are less willing to take higherrisks and are, therefore, more likely to enter ahost country through a JV. Thus, we control forfirm size and expect to find that it is negativelycorrelated with the probability of a JV.

Production Diversification. As Asiedu andEsfahani (2001) note, although a multinationalmay be well endowed in intangible assets, itsrole in its investment projects may be limitedif these assets are spread over a wide range of

11. Note that the earlier literature usually employedeither firm- or industry-level proxies for intangible assets.Asiedu and Esfahani (2001) included a firm-specific measureof all intangible assets (proxied by the ratio of sales totangible assets) as well as industry-level R&D intensity.None of the earlier studies controlled for intraindustry effectsexplicitly.


industries. Following their suggestion, we con-trol for production diversification and expect tofind a positive sign on its coefficient, implyingthat diversification is positively correlated withthe probability of a JV.

Regional Experience. Our model assumes that aJV partner contributes skills complementary tothose of a foreign investor. The more familiar aforeign investor is with the region, the less theinvestor’s need for a local partner. On the otherhand, greater familiarity with a particular regionmay lower the cost of finding a suitable JVpartner. Thus, the impact of regional experienceon the propensity to seek a JV is unclear. Tocontrol for regional experience, we include adummy variable taking on the value of 1 if afirm had a trading relationship with the regionbefore 1990 and 0 otherwise.

International Experience. As Gatignon andAnderson (1988) and Blomstrom and Zejan(1991) show, firms with greater experience inforeign operations in general may be more adeptin monitoring and dealing with local employeesand thus may be less likely to share owner-ship. Since what matters is not just the country-or region-specific knowledge but overall inter-national experience, we measure internationalexperience by the share of foreign sales in afirm’s total sales.

Host Country Characteristics. The choicebetween full and shared ownership is also likelyto be influenced by a variety of host coun-try characteristics (Asiedu and Esfahani 2001).Since the investigation of these issues is not ofimmediate interest to this study, we control forhost country–specific factors with dummy vari-ables for destination countries.

C. Results

Next, we turn to the regression results. Recallthat in our probit model, the dependent variableequals 1 if investor i has engaged in a JV with alocal partner in country k and 0 if the project isa direct entry. Thus, the number of observationsis equal to the number of projects undertakenin the region by all firms in the sample. Theestimated results are presented in Table 4 interms of marginal effects.

The standard errors, listed in parentheses,are clustered for observations pertaining forthe same company. As predicted, the results

indicate that firms possessing more sophisticatedtechnologies relative to the industry averageare less likely to engage in JVs and preferto retain full ownership of their investmentprojects (column 1). The same is true of firmswith above-average investment in marketing andbrand names (column 2). Both coefficients arestatistically significant and remain so even whenboth proxies are entered into the same equation(column 3). As a robustness check, in column4, we include dummies for three-digit SICsectors and drop sector-specific variables. Thecoefficient on technological sophistication bearsthe same sign and remains significant, while thecoefficient on relative advertising intensity losesits significance.

As for other explanatory variables, as antici-pated, we find that JVs are more likely to takeplace in industries where intangible assets play aless prominent role (i.e., industries characterizedby lower spending on R&D and advertising).Furthermore, they are more likely to be under-taken by smaller and more diversified firms.Regional and international experiences do notappear to have a statistically significant impacton the decision regarding the mode of entry.

One could argue that our empirical analysissuffers from a selection bias since we only con-sider projects that took place and ignore firmsthat decided against investment in a particularcountry or in the whole region. Thus, as a fur-ther robustness check, we estimate a two-stagemodel where the first stage (investment deci-sion) describes the decision to invest and thesecond stage (ownership decision) examines thechoice of mode of entry. The dependent vari-able in the first stage is equal to 1 if firm i hasundertaken FDI in country k and 0 otherwise. Inaddition to all the determinants of the mode ofentry described in the previous section, the firststage includes controls for host country char-acteristics commonly found in studies of FDIdeterminants.12 These are market size (proxiedby population size), purchasing power of localconsumers (captured by GDP per capita), qual-ity of business environment (measured using theEBRD ratings of progress in transition process),corporate tax rate, openness to trade (definedas the sum of exports and imports divided bythe GDP), and distance between source and hostcountry.

12. See Wheeler and Mody (1992) and a survey of theliterature on the determinants of FDI by Markusen (1995).


TABLE 4Results of a Probit Model—JV versus Direct Entry

Relative R&D −0.055∗ (0.033) −0.069∗ (0.041) −0.187∗ (0.100)Industry R&D −0.039∗∗∗ (0.012) −0.058∗∗∗ (0.019)Relative advertising −0.166∗∗ (0.085) −0.244∗∗ (0.120) −0.317 (0.232)Industry advertising −0.007∗∗ (0.003) <.001 (0.004)Diversification 0.037∗ (0.021) 0.022 (0.022) 0.043∗∗ (0.022) −0.057 (0.043)Regional experience 0.096 (0.101) 0.098 (0.102) 0.141 (0.107) 0.192 (0.155)International experience −0.002 (0.002) −0.002 (0.002) −0.001 (0.002) −0.005 (0.003)Firm size −0.051∗ (0.031) −0.062∗∗ (0.028) −0.086∗∗ (0.035) −0.063 (0.059)Host dummies Yes Yes Yes YesIndustry dummies No No No YesObserved p .60 .56 .57 .51Predicted p .61 .57 .57 .52No. of observations 439 424 345 243Pseudo R2 .14 .13 .21 .41Log likelihood −255.07 −254.04 −185.61 −98.75

Notes: Dependent variable is equal to 1 for JVs and 0 for direct entry. The results are presented in terms of marginaleffects evaluated at the sample mean. All models include a constant term that is not reported. Standard errors (clustered onfirm) are listed in parentheses. <.001 denotes coefficients with absolute value below .001.*Significant at 10% level; **significant at 5% level; ***significant at 1% level.

The second stage includes all the variablesused in the simple probit model plus two con-trols pertaining to host countries: transitionprogress and distance between the source andthe host country. We expect to find a negativecoefficient on the former variable, as the moreadvanced the host country in the reform pro-cess, the less need for help from a local partnerto navigate through the bureaucracy to obtainthe necessary permits and deal with tax authori-ties. Similarly, the smaller the distance betweenthe home and the host country, the more famil-iar are foreign investors with the ways of doingbusiness in their investment destination and thusagain less need for a JV partner.

We estimate the two equations describedabove simultaneously by maximum likelihood(probit with sample selection), correcting stan-dard errors for correlation between observationsfor the same firm. The number of observations inthe first equation (investment decision) is equalto the number of firms in the sample, multipliedby the number of destination countries coveredby the data set less observations with missingvalues. In the ownership decision equation, thenumber of observations is equal to the total num-ber of FDI projects in the sample.

The results, presented in the first threecolumns of Table 5, lend support to ourhypothesis.

In Table 5, the top panel contains the find-ings from the second stage (ownership decision),

and it confirms that firms with more sophis-ticated technologies and marketing techniquesare averse to sharing ownership and prefer toenter a host country directly. As before, thedata indicate that JVs are less common in highR&D- and advertising-intensive industries andamong larger investors. Furthermore, there issome, albeit not very strong, indication thatmore diversified firms as well as those withless international experience tend to undertakeJVs rather than enter directly. As expected, thedata suggest that JVs are a less attractive optionin economies more advanced in the transitionprocess, where doing business is likely to beeasier. Finally, regional experience and distancebetween the source and the host country do notappear to have a statistically significant impacton the ownership choice.

The investment decision equation, presentedin the lower panel of Table 5, also producesthe expected results. The findings indicate thatlarger firms and those operating in advertising-intensive industries are more likely to undertakeFDI. The same is true of firms familiar withthe region, possessing international experienceand less diversified companies.13 In terms ofhost country characteristics, economies that arelarger and more advanced in the transition pro-cess are more attractive investment destinations.

13. More diversified firms may be under less pressure tosearch for new markets.


TA

BL

E5

Res

ults

ofa

Two-

Stag

eM

odel

All

Fir

ms

Inve

stor

sO

nly

Ow

ners

hip

deci

sion

Rel

ativ

eR

&D

−0.1

60∗∗

(0.0

79)

−0.1

76∗

(0.1

03)

−0.1

57∗∗

(0.0

79)

−0.1

73∗

(0.1

03)

Indu

stry

R&

D−0

.102

∗∗∗

(0.0

32)

−0.1

29∗∗

∗(0

.047

)−0

.103

∗∗∗

(0.0

32)

−0.1

30∗∗

∗(0

.047

)R

elat

ive

adve

rtis

ing

−0.3

57∗

(0.2

19)

−0.5

65∗

(0.3

07)

−0.3

57∗

(0.2

19)

−0.5

49∗

(0.3

07)

Indu

stry

adve

rtis

ing

−0.0

17∗∗

(0.0

07)

−0.0

03(0

.009

)−0

.017

∗∗(0

.007

)−0

.002

(0.0

09)

Div

ersi

ficat

ion

0.07

5(0

.052

)0.

050

(0.0

58)

0.10

2∗(0

.054

)0.

075

(0.0

52)

0.05

1(0

.058

)0.

104∗

(0.0

54)

Reg

iona

lex

peri

ence

0.22

3(0

.257

)0.

243

(0.2

68)

0.29

6(0

.284

)0.

265

(0.2

51)

0.25

7(0

.261

)0.

356

(0.2

77)

Inte

rnat

iona

lex

peri

ence

−0.0

08∗∗

(0.0

04)

−0.0

07(0

.005

)−0

.004

(0.0

06)

−0.0

08∗

(0.0

04)

−0.0

07(0

.005

)−0

.004

(0.0

06)

Log

(firm

size

)−0

.145

∗(0

.079

)−0

.154

∗∗(0

.076

)−0

.230

∗∗∗

(0.0

88)

−0.1

32∗

(0.0

78)

−0.1

50∗∗

(0.0

74)

−0.2

19∗∗

(0.0

87)

Log

(dis

tanc

e)−0

.145

(0.1

07)

−0.0

63(0

.103

)−0

.031

(0.1

16)

−0.1

76∗

(0.1

06)

−0.0

73(0

.099

)−0

.064

(0.1

13)

Tra

nsiti

onin

dex

−0.4

90∗∗

∗(0

.143

)−0

.404

∗∗∗

(0.1

35)

−0.4

21∗∗

∗(0

.149

)−0

.495

∗∗∗

(0.1

48)

−0.4

02∗∗

∗(0

.137

)−0

.428

∗∗∗

(0.1

53)

Inve

stm

ent

deci

sion

Rel

ativ

eR

&D

−0.0

03(0

.031

)−0

.008

(0.0

40)

−0.0

43(0

.031

)−0

.043

(0.0

43)

Indu

stry

R&

D0.

026

(0.0

20)

−0.0

10(0

.024

)0.

039∗

(0.0

23)

−0.0

06(0

.031

)R

elat

ive

adve

rtis

ing

0.06

5(0

.077

)0.

212∗

(0.1

26)

0.10

0(0

.090

)0.

128

(0.1

19)

Indu

stry

adve

rtis

ing

0.01

4∗∗∗

(0.0

03)

0.01

6∗∗∗

(0.0

04)

0.01

5∗∗∗

(0.0

03)

0.01

6∗∗∗

(0.0

04)

Div

ersi

ficat

ion

−0.0

59∗

(0.0

33)

−0.0

41(0

.027

)−0

.056

(0.0

37)

−0.0

79∗∗

∗(0

.028

)−0

.072

∗∗∗

(0.0

26)

−0.0

89∗∗

∗(0

.033

)R

egio

nal

expe

rien

ce0.

265∗∗

(0.1

27)

0.32

4∗∗∗

(0.1

19)

0.40

3∗∗∗

(0.1

57)

0.01

6(0

.135

)0.

064

(0.1

22)

0.10

7(0

.174

)In

tern

atio

nal

expe

rien

ce0.

003

(0.0

02)

0.00

3∗(0

.002

)0.

002

(0.0

03)

0.00

2(0

.002

)0.

002

(0.0

02)

0.00

3(0

.003

)L

og(fi

rmsi

ze)

0.20

7∗∗∗

(0.0

31)

0.19

4∗∗∗

(0.0

29)

0.20

2∗∗∗

(0.0

33)

0.16

3∗∗∗

(0.0

36)

0.17

4∗∗∗

(0.0

30)

0.17

1∗∗∗

(0.0

37)

Log

(dis

tanc

e)−0

.427

∗∗∗

(0.0

68)

−0.3

83∗∗

∗(0

.066

)−0

.407

∗∗∗

(0.0

76)

−0.3

18∗∗

∗(0

.060

)−0

.263

∗∗∗

(0.0

56)

−0.2

84∗∗

∗(0

.067

)T

rans

ition

inde

x0.

518∗∗

∗(0

.108

)0.

514∗∗

∗(0

.106

)0.

603∗∗

∗(0

.119

)0.

777∗∗

∗(0

.122

)0.

771∗∗

∗(0

.120

)0.

910∗∗

∗(0

.136

)L

og(p

opul

atio

n)0.

512∗

∗∗(0

.035

)0.

479∗

∗∗(0

.035

)0.

538∗

∗∗(0

.039

)0.

691∗

∗∗(0

.051

)0.

648∗

∗∗(0

.050

)0.

754∗

∗∗(0

.059

)L

og(G

DP

per

capi

ta)

0.07

8(0

.066

)0.

076

(0.0

68)

0.01

7(0

.073

)0.

049

(0.0

78)

0.04

4(0

.080

)−0

.039

(0.0

89)

Cor

pora

teta

xra

te−0

.014

∗∗∗

(0.0

04)

−0.0

17∗∗

∗(0

.004

)−0

.020

∗∗∗

(0.0

04)

−0.0

22∗∗

∗(0

.005

)−0

.026

∗∗∗

(0.0

05)

−0.0

31∗∗

∗(0

.006

)L

og(o

penn

ess

totr

ade)

0.41

1∗∗∗

(0.1

00)

0.21

9∗∗(0

.104

)0.

430∗∗

∗(0

.113

)0.

630∗∗

∗(0

.140

)0.

384∗∗

∗(0

.146

)0.

702∗∗

∗(0

.159

)R

ho−0

.200

∗∗∗

(0.1

65)

−0.0

62∗∗

∗(0

.185

)−0

.218

∗∗∗

(0.1

89)

−0.1

55∗∗

∗(0

.146

)−0

.039

∗∗∗

(0.1

58)

−0.1

71∗∗

∗(0

.165

)N

o.of

obse

rvat

ions

7,70

78,

589

6,25

82,

982

3,17

12,

352

Cen

sore

d7,

267

8,16

45,

912

2,54

22,

746

2,00

6U

ncen

sore

d44

042

534

644

042

534

6W

ald

stat

istic

s37

.025

.542

.835

.524

.641

.4Pr

obab

ility

Wal

d>

0.0

0.0

0.0

0.0

0.0

0.0

0L

oglik

elih

ood

−1,4

45.6

−1,4

46.4

−1,1

17.9

−1,1

38.4

−1,1

39.2

−863

.9

Not

es:

Inth

ein

vest

men

tde

cisi

oneq

uatio

n,th

ede

pend

ent

vari

able

iseq

ual

to1

iffir

mi

has

unde

rtak

enin

vest

men

tin

coun

try

kan

d0

othe

rwis

e.In

the

owne

rshi

pde

cisi

oneq

uatio

n,th

ede

pend

ent

vari

able

take

son

the

valu

eof

1fo

rJV

san

d0

for

dire

cten

try.

All

mod

els

incl

ude

aco

nsta

ntte

rmth

atis

not

repo

rted

.St

anda

rder

rors

(clu

ster

edon

firm

)ar

elis

ted

inpa

rent

hese

s.*S

igni

fican

tat

10%

leve

l;**

sign

ifica

ntat

5%le

vel;

***s

igni

fican

tat

1%le

vel.


Similarly, less distant countries and those moreopen to trade and offering lower corporate taxrates are more successful at attracting FDI. Onthe other hand, GDP per capita, which may bea proxy for labor costs, does not appear to haveimpact on the investment decision.

As noted in Appendix II, firms that engagedin FDI in the region are oversampled in our dataset. Therefore, as an additional robustness check,we reestimate the two-stage model, restrictingour sample to investors, that is, firms with atleast one investment in the countries consideredin the study. A benefit of this restriction is thatwe reduce the number of 0s on the left-hand sideof the equation, as the original data set containsmany firms that have not undertaken any invest-ment projects in the region. These results, shownin the last three columns of Table 5, do not differsignificantly from those obtained from the fullsample. The variables of interest, technologicaland marketing sophistication, retain their signs,magnitudes, and significance levels, thus againlending support to our hypothesis.

Finally, we check whether the negative rela-tionship between the technological sophistica-tion and the choice of JVs could be driven by thefear of technology leakage due to less than per-fect protection of IPRs in transition economies.To do so, we control in both the investmentand the ownership decision equations for thestrength of IPR protection in the host coun-try. The extent of IPR protection is measuredusing an index of patent protection developedby Ginarte and Park (1997) and extended byJavorcik (2004b). The index refers to 1995 or theclosest year for which the information was avail-able. Its values range from 0 to 5 with the highervalues indicating a stronger level of protection.The highest score in the group of countries underconsideration is found in Hungary (3.75), whilethe lowest score of 2.52 belongs to Uzbekistanand former Yugoslav Republic of Macedonia.For a more detailed description of the index,see Appendix IIB.

The results of the augmented model are pre-sented in Table 6. While we find that foreigninvestors are attracted to location with strongerIPR protection, we find no evidence of IPRlegislation affecting the choice of entry mode.The coefficient on the IPR index is statisticallysignificant in four of six specifications in theinvestment decision equation and in none of thesix ownership decision equations. Other resultsremain pretty much unchanged. As before, we

find that investors with more sophisticated tech-nologies are less likely to engage in JVs, withthis effect being statistically significant in allspecifications. A similar pattern is found formarketing sophistication, though now it reachesconventional significance levels only in half ofthe cases.

IV. CONCLUSIONS

The choice of entry mode by foreign investorshas been of interest to both policy makers andresearchers. Developing country governmentsare especially interested in the technology andknow-how transfer that results from FDI. To beable to assess the potential magnitude of suchbenefits, it is important to understand prefer-ences of different types of investors with respectto the entry mode. This study sheds some lighton this issue by analyzing the intraindustrydeterminants of entry modes chosen by foreignfirms entering transition economies of EasternEurope and the successor states of the SovietUnion in the early 1990s.

Our empirical work is motivated by a sim-ple theoretical model that allows for competi-tion between asymmetric foreign investors. Themodel predicts that relatively efficient foreigninvestors are less likely to choose JVs and morelikely to enter directly. The empirical resultssupport this prediction. Thus, policies influenc-ing FDI entry mode may affect technologicalcontent of the investment projects and generatedifferent implications for the extent of potentialspillovers to the host economy.

APPENDIX I : DERIVATIONS AND PROOFS

Here, we report all the analytical derivations and provideproofs for our results. Using the first-order conditions forCournot competition,

∂πi (qi , q−i )

∂qi

= p(q) + p′(q)qi − ci

= a − bq−i − 2bqi − ci = 0.

We can easily calculate the equilibrium output levels:

qxy

i = a − 2cxi + c

y

−i

3b,(6)

where i = 1, 2 and x, y = e, j .Furthermore, the equilibrium profit of a firm under

regime (x, y) is proportional to the square of its quantity.For example,

πej

1 = b[qej

1 ]2 and πej

2 = θb[qej

2 ]2.(7)


TA

BL

E6

Res

ults

ofa

Two-

Stag

eM

odel

with

Con

trol

sfo

rIP

RPr

otec

tion

All

Fir

ms

Inve

stor

sO

nly

Ow

ners

hip

deci

sion

Rel

ativ

eR

&D

−0.1

63∗∗

(0.0

78)

−0.1

88∗

(0.0

99)

−0.1

60∗∗

(0.0

79)

−0.1

86∗

(0.0

99)

Indu

stry

R&

D−0

.102

∗∗∗

(0.0

32)

−0.1

33∗∗

∗(0

.046

)−0

.103

∗∗∗

(0.0

31)

−0.1

35∗∗

∗(0

.046

)R

elat

ive

adve

rtis

ing

−0.3

50(0

.218

)−0

.556

∗(0

.300

)−0

.351

(0.2

18)

−0.5

39∗

(0.3

01)

Indu

stry

adve

rtis

ing

−0.0

17∗∗

(0.0

07)

−0.0

02(0

.009

)−0

.017

∗∗(0

.007

)−0

.001

(0.0

09)

Div

ersi

ficat

ion

0.07

4(0

.051

)0.

045

(0.0

58)

0.10

1∗(0

.053

)0.

074

(0.0

52)

0.04

6(0

.058

)0.

103∗

(0.0

54)

Reg

iona

lex

peri

ence

0.21

3(0

.253

)0.

226

(0.2

64)

0.29

7(0

.275

)0.

256

(0.2

49)

0.25

0(0

.258

)0.

359

(0.2

71)

Inte

rnat

iona

lex

peri

ence

−0.0

08∗∗

(0.0

04)

−0.0

07(0

.005

)−0

.004

(0.0

05)

−0.0

08∗

(0.0

04)

−0.0

07(0

.005

)−0

.004

(0.0

05)

Log

(firm

size

)−0

.146

∗(0

.078

)−0

.161

∗∗(0

.076

)−0

.235

∗∗∗

(0.0

89)

−0.1

32∗

(0.0

78)

−0.1

55∗∗

(0.0

74)

−0.2

23∗∗

(0.0

87)

Log

(dis

tanc

e)−0

.134

(0.1

08)

−0.0

53(0

.103

)−0

.029

(0.1

14)

−0.1

65(0

.106

)−0

.069

(0.0

98)

−0.0

63(0

.111

)T

rans

ition

inde

x−0

.472

∗∗∗

(0.1

55)

−0.3

96∗∗

∗(0

.137

)−0

.468

∗∗∗

(0.1

65)

−0.4

79∗∗

∗(0

.160

)−0

.397

∗∗∗

(0.1

39)

−0.4

78∗∗

∗(0

.170

)IP

Rin

dex

−0.0

36(0

.206

)−0

.037

(0.2

00)

0.27

1(0

.229

)−0

.035

(0.2

08)

−0.0

34(0

.200

)0.

275

(0.2

31)

Inve

stm

ent

deci

sion

Rel

ativ

eR

&D

−0.0

04(0

.031

)−0

.008

(0.0

39)

−0.0

44(0

.031

)−0

.043

(0.0

42)

Indu

stry

R&

D.0

25(0

.020

)−0

.010

(0.0

24)

0.03

9∗(0

.023

)−0

.005

(0.0

30)

Rel

ativ

ead

vert

isin

g0.

063

(0.0

78)

0.21

0∗(0

.126

)0.

093

(0.0

91)

0.12

1(0

.119

)In

dust

ryad

vert

isin

g0.

014∗∗

∗(0

.003

)0.

016∗∗

∗(0

.004

)0.

015∗∗

∗(0

.003

)0.

016∗∗

∗(0

.004

)D

iver

sific

atio

n−0

.057

∗(0

.033

)−0

.043

(0.0

28)

−0.0

54(0

.038

)−0

.078

∗∗∗

(0.0

29)

−0.0

75∗∗

∗(0

.026

)−0

.087

∗∗∗

(0.0

33)

Reg

iona

lex

peri

ence

0.26

3∗∗(0

.126

)0.

332∗∗

∗(0

.119

)0.

402∗∗

∗(0

.155

)0.

013

(0.1

33)

0.07

2(0

.121

)0.

106

(0.1

70)

Inte

rnat

iona

lex

peri

ence

0.00

3(0

.002

)0.

003∗

(0.0

02)

0.00

2(0

.003

)0.

002

(0.0

02)

0.00

2(0

.002

)0.

003

(0.0

03)

Log

(firm

size

)0.

205∗

∗∗(0

.031

)0.

192∗

∗∗(0

.029

)0.

199∗

∗∗(0

.033

)0.

158∗

∗∗(0

.035

)0.

169∗

∗∗(0

.030

)0.

165∗

∗∗(0

.036

)L

og(d

ista

nce)

−0.4

27∗∗

∗(0

.068

)−0

.388

∗∗∗

(0.0

66)

−0.4

05∗∗

∗(0

.076

)−0

.319

∗∗∗

(0.0

60)

−0.2

70∗∗

∗(0

.057

)−0

.283

∗∗∗

(0.0

68)

Tra

nsiti

onin

dex

0.54

6∗∗∗

(0.1

15)

0.55

7∗∗∗

(0.1

17)

0.66

3∗∗∗

(0.1

33)

0.81

2∗∗∗

(0.1

30)

0.83

0∗∗∗

(0.1

33)

0.98

2∗∗∗

(0.1

50)

Log

(pop

ulat

ion)

0.51

0∗∗∗

(0.0

36)

0.48

2∗∗∗

(0.0

35)

0.53

9∗∗∗

(0.0

40)

0.69

4∗∗∗

(0.0

52)

0.65

4∗∗∗

(0.0

51)

0.76

1∗∗∗

(0.0

61)

Log

(GD

Ppe

rca

pita

)0.

037

(0.0

82)

0.00

8(0

.089

)−0

.064

(0.0

97)

0.00

1(0

.096

)−0

.042

(0.1

03)

−0.1

30(0

.114

)C

orpo

rate

tax

rate

−0.0

14∗∗

∗(0

.004

)−0

.016

∗∗∗

(0.0

04)

−0.0

19∗∗

∗(0

.004

)−0

.022

∗∗∗

(0.0

05)

−0.0

25∗∗

∗(0

.005

)−0

.031

∗∗∗

(0.0

06)

Log

(ope

nnes

sto

trad

e)0.

405∗

∗∗(0

.102

)0.

228∗

∗(0

.104

)0.

434∗

∗∗(0

.114

)0.

635∗

∗∗(0

.144

)0.

400∗

∗∗(0

.147

)0.

721∗

∗∗(0

.162

)IP

Rin

dex

0.09

9(0

.092

)0.

176∗

(0.0

94)

0.18

5∗(0

.098

)0.

137

(0.1

09)

0.23

7∗∗

(0.1

10)

0.23

5∗∗

(0.1

16)

Rho

−0.2

08∗∗

∗(0

.161

)−0

.106

∗∗∗

(0.1

86)

−0.2

27∗∗

∗(0

.190

)−0

.164

∗∗∗

(0.1

42)

−0.0

79∗∗

∗(0

.159

)−0

.183

∗∗∗

(0.1

66)

No.

ofob

serv

atio

ns69

7377

7156

6226

9826

9821

28C

enso

red

6538

7352

5321

2263

2263

1787

cont

inue

d


TA

BL

E6

Con

tinue

d

All

Fir

ms

Inve

stor

sO

nly

Unc

enso

red

435

419

341

435

435

341

Wal

dst

atis

tics

37.9

26.2

143

.79

36.2

736

.27

42.4

7Pr

obab

ility

Wal

d>

0.0

0.0

0.0

0.0

0.0

0.0

0L

oglik

elih

ood

−1,4

19.3

−1,4

13.0

−1,0

92.4

−1,1

13.4

−1,1

07.8

−839

.6

Not

es:

Inth

ein

vest

men

tde

cisi

oneq

uatio

n,th

ede

pend

ent

vari

able

iseq

ual

to1

iffir

mi

has

unde

rtak

enin

vest

men

tin

coun

try

kan

d0

othe

rwis

e.In

the

owne

rshi

pde

cisi

oneq

uatio

n,th

ede

pend

ent

vari

able

take

son

the

valu

eof

1fo

rJV

san

d0

for

dire

cten

try.

All

mod

els

incl

ude

aco

nsta

ntte

rmth

atis

not

repo

rted

.St

anda

rder

rors

(clu

ster

edon

firm

)ar

elis

ted

inpa

rent

hese

s.*S

igni

fican

tat

10%

leve

l;**

sign

ifica

ntat

5%le

vel;

***s

igni

fican

tat

1%le

vel.

Thus, we have

u1 = [√

bθqje

1 ]2 − [√

bqee1 ]2 <

= (√

θbqje

1 +√

bqee1 )(

√θbq

je

1 −√

bqee1 )

so that

u1 > 0 iff e1 ≡√

θqje

1 − qee1 > 0.(8)

Using the equilibrium quantity levels given in Equation(6), we can describe the two incentive functions in termsof exogenous parameters. For example, u1(.) > 0 iff e1 > 0where

e1 =√

θ(a − 2(α1 + δβ1) + α2 + β2)

− (a − 2(α1 + β1) + α2 + β2)

= (1 −√

θ)(−a + 2α1 − α2 − β2) + 2β1(1 −√

θδ).

We can similarly show that

u2 > 0 iff e2 ≡√

θqej

2 − qee2 > 0,(9)

and

�1 > 0 iff d1 ≡√

θqjj

1 − qej

1 > 0.

Finally,

�2 > 0 iff d2 ≡√

θqjj

2 − qje

1 > 0.

As for the case of u1, we can describe the aboveincentive functions in terms of exogenous parameters usingEquations (1), (2), and (6).

Proof of Proposition 1

Note that

∂ei

∂αi

= ∂di

∂αi

= 2(1 − √θ)

3> 0

and∂ei

∂βi

= ∂di

∂βi

= 2(1 − δ√

θ)

3> 0

and

∂ei

∂α−i

= ∂di

∂α−i

= − 1(1 − √θ)

3< 0

and∂ei

∂β−i

= ∂di

∂β−i

= 1(1 − √θ)

3< 0.

Proof of Corollary 1

From Equations (8) and (9), we know that corollary 1holds iff e2 − e1 > 0. Substituting for the equilibrium outputlevels, we can show that

∂(e2 − e1)

∂θ= − 3(α2 − α1) + (β2 − β1)(1 + 2δ)

6√

θ< 0.

That is, e2 − e1 is decreasing in θ. Furthermore, at θ = 1,we have

e2 − e1|θ=1 = 2(β2 − β1)(1 − δ)

3> 0.

Thus, it must be that e2 − e1 > 0 for all θ. Since e2

is strictly greater than e1, there exists parameter values forwhich only Firm 2 has a unilateral incentive for a JV.


To prove the second statement of the corollary, it isenough to show that if u1 > 0 then �2 > 0. We know that

∂(d2 − e1)

∂θ= 3(α1 − α2) − (1 + 2δ)β2 + 3δβ1)

6√

θ< 0.

That is, d2 − e1 is decreasing in θ. Furthermore, at θ = 1,we have

d2 − e1|θ=1 = 2(β2 − β1)(1 − δ)

3> 0.

Thus, it must be that d2 − e1 > 0 for all θ. In otherwords, if Firm 1 has a unilateral incentive for a JV, Firm 2will have a competitive incentive for a JV. As a result, theregime (j, e) cannot be an equilibrium.

APPENDIX II: DATA

A. Survey

The respondents of the 1995 EBRD survey were askedto classify each of their existing or planned projects as a JVwith a local partner, acquisition, or greenfield entry. Forthe purpose of this study, we classify all greenfield andacquisition projects not associated with JVs as direct entry.In other words, if a respondent listed more than one form ofentry mode, the observation was classified as a JV if one ofthese forms was “JV with a local partner” and direct entryotherwise.

As a robustness check, we also used an alternativeclassification in which we created a separate observationfor each entry mode reported by a respondent. Then, weestimated a probit model with the dependent variable takingon the value of unity for JVs and 0 for greenfield projects.The results on the variables of interest (i.e., R&D andmarketing intensities) were very similar to those presentedin Table 4. Furthermore, we also estimated a multinomiallogit model with the dependent variable representing thethree entry modes and a multinomial logit model with threeentry modes plus the option of not investing at all. In bothcases, the results on the impact of intangible assets on thechoice between greenfield projects and JVs lent support toour hypothesis.

It is likely that firms that perceived the survey as morerelevant (for instance, firms that had invested or consid-ered investing in transition economies) were more likely torespond. To check this hypothesis, the list of major foreigninvestors in Poland compiled by the Polish State InvestmentAgency (1995) was examined. Poland was chosen for thisexercise since it was the most popular destination country inthe sample. Of 329 firms on the list, 118 received the EBRDsurvey and 50% of them responded, as opposed to the overallresponse rate for the survey equal to about 15%. Statisticaltests indicated that the means of firm-specific variables in therespondent and nonrespondent groups were not significantlydifferent from each other. Thus, among the investing firms,the decision to respond to the survey was not systematicallyrelated to firm characteristics. Unfortunately, it was not pos-sible to identify which among the firms that did not respondto the survey were not interested in undertaking investmentin Eastern Europe and the former Soviet Union. There is no

reason, however, to suspect that in the case of these firms,the decision to answer the survey was systematically relatedto their characteristics. Therefore, the data set can be treatedas if the investing firms had been oversampled.

The survey did not ask about the date when each invest-ment was undertaken. Since the magnitude of FDI inflowsto transition countries was marginal before 1989 and thesurvey was conducted in January 1995, the information col-lected pertains mostly to the period 1989–94. Furthermore,to the best of our knowledge, none of the countries in thesample had legislation specifically forbidding full ownershipby foreign investors. For instance, in the USSR, a presiden-tial decree issued as early as October 1990 allowed foreignwholly owned companies to be established in the form ofbranches or subsidiaries. The decree also created the legalbasis for foreign investors to buy out existing Soviet enter-prises as these were privatized (McMillan 1996, 50). InHungary, Act XXIV of 1988 on the Investment of Foreign-ers in Hungary allowed non-Hungarian companies to ownequity up to 100% (WTO 1998). In Poland, the 1988 Law onEconomic Activity with the Participation of Foreign Partiespermitted 100% foreign equity participation (GATT 1992). Itis possible, however, that in practice, permissions for fullyowned projects may had been denied in some economiesduring the period covered by our sample. To control for thispossibility, we included host country dummies in our model.

Since restrictions on the extent of foreign ownershipmay have been present in extractive sector and services,we excluded firms in the coal, gas, and oil industry fromour sample. We also dropped projects in service industries,such as banking, insurance, telecommunications, accounting,public relations services, and so forth. In addition to possiblerestrictions on FDI, including these sectors would alsopose some difficulties with measuring the endowment ofintangible assets.

Note that our analysis assumes that all foreign investorshave the option of engaging in a JV with a local partner,should they want to do so. In other words, the supply of localJV partners is not constrained, and the observed entry pat-terns are determined entirely by foreign investors’ demand.Considering that the aggregate FDI inflows into transitioneconomies were quite small during the period covered byour sample, this assumption is quite realistic.

B. Other Data Sources

All firm-specific explanatory variables used in the anal-ysis, with the exception of regional experience that comesfrom the survey, were taken from the commercial databaseWorldscope and are for 1993 (or the closest year for whichthe information is available). Note that the variables per-tain to the characteristics of the parent companies, not theirparticular subsidiaries in the regions. Details of variable def-initions are listed below.

Firm size: Log of firm sales in millions of U.S. dollars.Relative technological sophistication: Firm R&D inten-

sity/average R&D intensity in the industry.R&D intensity is measured by R&D expenditure expres-

sed as a percentage of total sales. To calculate industryaverages (at the three-digit SIC industry classification), weuse figures for all firms listed in Worldscope in a given


industry, not just firms included in our sample. Thus, thesevalues correspond to the average R&D intensity of majorfirms operating in developed countries in a given industry.

Relative marketing sophistication: Firm advertising inten-sity/average advertising intensity in the industry.

Advertising intensity is defined as the ratio of sales,general, and administrative expenditure to total sales, whichis a standard proxy used in the literature. The industryaverage is again calculated at the three-digit SIC level.

Product diversification: The number of four-digit SICcodes describing a firm’s activities.

International experience: The share of foreign sales in afirm’s total sales. Ideally, we would like to use the share offoreign assets in a firm’s total assets. However, using thismeasure would severely reduce the size of our sample. Theshare of foreign sales is highly correlated with the shareof foreign assets (correlation of .82). Thus, our proxy forinternational experience seems reasonable.

Population size, GDP per capita: Both variables enter inlog form, pertain to 1993, and come from EBRD (variousissues).

Transition indicators: The transition indicators rate theprogress of a country’s reforms in the following areas: priceliberalization and competition, trade and exchange system,large-scale privatization, small-scale privatization, enterpriserestructuring, and banking reform. See EBRD (1994, 11) fora detailed description. In the empirical analysis, a simpleaverage of the EBRD indicators is used.

Openness to trade: log (exports + imports)/GDP is cal-culated using figures from the World Bank World Develop-ment Indicators.

Corporate tax rate: Expressed in percentages, corre-sponds to the highest rate applicable in the host country.Source: PriceWaterhouseCoopers.

Distance: Log distance between the capital cities expres-sed in kilometers. The following source countries areincluded in the sample: United Kingdom, United States,Germany, France, Finland, Switzerland, Denmark, Norway,the Netherlands, Austria, Sweden, Belgium, Canada, Japan,Australia, Italy, Greece, Ireland, Portugal, Singapore, Spain,Brazil, Malaysia, South Africa, and South Korea.

Index of IPR protection: Originally developed by Ginarteand Park (1997) and later extended by Javorcik (2004b).The index takes into account five aspects of patent laws: (1)extent of coverage, (2) membership in international patentagreements, (3) provisions for loss of protection, (4) enforce-ment mechanisms, and (5) duration of protection. Each ofthe categories is assigned a value between 0 and 1, and theunweighted sum of these values constitutes the patent rightsindex. The index ranges from 0 to 5 with the higher valuesindicating a stronger level of protection. The index refersto 1995 or the closest year for which the information wasavailable.

REFERENCES

Al-Saadon, Y., and S. P. Das. “Host-Country Policy, Trans-fer Pricing and Ownership Distribution in InternationalJoint Ventures: A Theoretical Analysis.” InternationalJournal of Industrial Organization, 14, 1996, 345–64.

Asiedu, E., and H. S. Esfahani. “Ownership Structurein Foreign Direct Investment Projects.” Review ofEconomics and Statistics, 83, 2001, 647–62.

Blomstrom, M., and F. Sjoholm. “Technology Transfer andSpillovers: Does Local Participation with Multination-als Matter?” European Economic Review, 43, 1999,915–23.

Blomstrom, M., and M. Zejan. “Why Do MultinationalFirms Seek Out Joint Ventures?” Journal of Interna-tional Development, 3, 1991, 53–63.

Blomstrom, M., R. E. Lipsey, and L. Ohlsson. “WhatDo Rich Countries Trade with Each Other? R&Dand the Composition of U.S. and Swedish Trade.”NBER Reprint No. 1551. Banca Nazionale del LavoroQuarterly Review, 173, 1991, 215–35, June 1990.

Caves, R. E. Multinational Enterprise and Economic Anal-ysis. Cambridge: Cambridge University Press, 1996.

EBRD (European Bank for Reconstruction and Develop-ment). Transition Report. London: EBRD, 1994.

Ethier, W., and J. Markusen. “Multinational Firms, Techno-logical Diffusion and Trade.” Journal of InternationalEconomics, 41, 1996, 1–28.

Gatignon, H., and E. Anderson. “The Multinational Corpo-ration’s Degree of Control over Foreign Subsidiaries:An Empirical Test of a Transaction Cost Explanation.”Journal of Law, Economics, and Organization, 4, 1988,305–36.

GATT (General Agreement on Tariffs and Trade). TradePolicy Review: Poland, Vol. 1. Geneva, Switzerland:GATT, 1992.

Ginarte, J. C., and W. G. Park. “Determinants of patentrights: A cross-national study.” Research Policy 26,1997, 283-301.

Gomes-Casseres, B. “Ownership Structures of Foreign Sub-sidiaries; Theory and Evidence.” Journal of EconomicBehavior and Organization, 11, 1989, 1–25.

Horstmann, I. J., and J. R. Markusen. “Endogenous MarketStructures in International Trade (Natura Facit Sel-tum).” Journal of International Economics, 32, 1992,109–29.

Javorcik, B. S. “Does Foreign Direct Investment Increasethe Productivity of Domestic Firms? In Search ofSpillovers through Backward Linkages.” AmericanEconomic Review, 94, 2004a, 605–27.

. “The Composition of Foreign Direct Investmentand Protection of Intellectual Property Rights: Evi-dence from Transition Economies.” European Eco-nomic Review, 48, 2004b, 39–62.

. “Technological Leadership and Foreign Investors’Choice of Entry Mode,” in Global Integration andTechnology Transfer, edited by B. Hoekman and B.Javorcik. New York: Palgrave Macmillan, 2006, 179-206.

Javorcik, B. S., and M. Spatareanu. “To Share or Not ToShare: Does Local Participation Matter for Spilloversfrom FDI?” Journal of Development Economics, 85,2008, 194–217.

Javorcik, B. S., and S.-J. Wei. “Corruption and Cross-BorderInvestment in Emerging Markets: Firm-Level Evi-dence.” Journal of International Money and Finance,28, 2009, 605-24.

Klimenko, M., and K. Saggi. “Technical Compatibility andthe Mode of Foreign Entry under Network Exter-nalities.” Canadian Journal of Economics, 40, 2007,176–206.

Mansfield, E., and A. Romeo. “Technology Transfer toOverseas Subsidiaries by U.S. Based Firms.” QuarterlyJournal of Economics, 95, 1980, 737–49.

Markusen, J. R. “The Boundaries of Multinational Enter-prises and the Theory of International Trade.” Journalof Economic Perspectives, 9, 1995, 169–89.


. “Contracts, Intellectual Property Rights, and Multi-national Investment in Developing Countries.” Journalof International Economics, 53, 2001, 189–204.

Mattoo, A., M. Olarreaga, and K. Saggi. “Mode of ForeignEntry, Technology Transfer, and FDI Policy.” Journalof Development Economics, 75, 2004, 95–111.

McMillan, C. H. “Foreign Investment in Russia: SovietLegacies and Post-Soviet Prospects,” in ForeignInvestment in Russia and Other Soviet Successor States,edited by P. Artisien-Maksimenko and Y. Adjubei.New York: St. Martin’s Press, 1996, 41–72.

Miller, R., J. Glen, F. Jaspersen, and Y. Karmokolies.“International Joint Ventures in Developing Countries:

Happy Marriages?” IFC Discussion Paper No. 29,1996.

Polish State Investment Agency. Major Investor List. War-saw: PSIA, 1995.

Ramachandran, V. “Technology Transfer, Firm Ownership,and Investment in Human Capital.” Review of Eco-nomics and Statistics, 75, 1993, 664–70.

Wheeler, D., and A. Mody. “International InvestmentLocation Decisions: The Case of US Firms.”Journal of International Economics, 33, 1992,57–76.

WTO (World Trade Organization). Trade Policy Review:Hungary. Geneva, Switzerland: WTO, 1998.

Documents

TECHNOLOGICAL ASYMMETRY AMONG FOREIGN INVESTORS … · construction machinery electrical mach automotive all sectors Sophistication Technological Sophistication FDI in the Russian