Optimal Entry Mode and Technology Transfer of FDI with
Differentiated Products
Ho-Chyuan Chen
Department of Risk Management and Insurance
National Kaohsiung First University of Science & Technology, Taiwan
Mei-Fang Chung
Department of International Economics, Kun Shan University of Technology
Abstract
This paper employs a three-stage game to analyze the effect of product differentiation
on the foreign firm’s technology transfer and entry mode and on the host welfare. It
finds that lower differentiation induces higher (lower) technology transfer given a
sufficiently low (high) transfer cost. Secondly, given sufficiently low transfer costs,
acquisition induces less technology transfer than direct entry and the foreign firm
would prefer acquisition, and the probability of those decrease in differentiation.
Probability for the equilibrium mode of acquisition to induce higher technology (than
direct entry) decreases in differentiation. Thirdly, the equilibrium entry mode more
likely results in less technology transfer and host welfare (than the off-equilibrium
mode). Fourthly, the government and the foreign firm prefer identical entry mode
only when transfer cost is within some certain intermediate levels. But probability of
that is always less than half and generally decreases in differentiation.
Keywords: Entry mode; technology transfer; differentiation; FDI; acquisition

Corresponding author: Ho-Chyuan Chen, Professor of Economics at Department of Risk
Management and Insurance, National Kaohsiung First University of Science & Technology, No.2,
Jhuoyue Rd., Nanzih District, Kaohsiung City 811, Taiwan. Tel: 886-7-6011000 ext.3024; Fax:
886-7-6011020; E-mail: [email protected].
1
1. Introduction
The issue of technology transfer and entry mode of foreign firms has gained
increasing attention, because the enormous growth in FDI has significantly increased
the number of multinational enterprises (MNEs).1 Entry modes may be different
among MNEs. During the period from 1994 to 1999, on average about 93% of the
total FDI inflows to OECD countries took the form of mergers and acquisitions, while
direct entry has predominated over other types of FDI into China from the late
1980s.2 Bjorvatn (2001) claimed that the knowledge gaps between MNEs and the
developing countries usually result in larger costs of technology transfer for mergers
and acquisitions, which makes direct entry more preferable to MNEs. Obviously, the
entry mode of MNEs affects the efficiency of technology transfer and becomes an
important strategic decision for the foreign entrants (Mudambi, 1995; Culem, 1988;
Devereux and Griffith, 1998; Kasuga, 2003; and Eicher and Kang, 2005). But then
what factors determine MNEs’ optimal entry mode?
As Culem, (1988), Davidson (1980), Devereux and Griffith (1998), and Mudambi
(1995) claimed, one important decision to make for international diversification is to
enter the foreign market through acquisition of a currently operating firm in the
foreign market, or through the direct entry. Mudambi (2002) found that a strong
relationship exists between the entry mode and diversification. On the other hand,
Eicher and Kang (2005) showed that large countries are more likely to attract
acquisitions, intermediate-sized countries may be served predominantly through trade,
1
During the period from 1994 to 2003, according to the International Financial Statistics Year Book
(2004), the increase in FDI inflows are estimated at about 144%, 478%, and 122% in Germany, France
and the U.K., respectively, which have an average 16.3%, 10.3%, and 13.98% GDP growth,
respectively. In China, firms with FDI have contributed to over 40% of its total trade since the
mid-1990s, according to Lv (2005). It has received about one third of FDI in all emerging markets and
60% of FDI in Asian emerging markets in the last decade (Prasad and Wei, 2005), and has become the
largest overall recipient of FDI in the world in 2003.
2
See Lv (2005). Razin and Sadka (2003) showed that a host country can have the ex ante gains from
opening up its economy to the flow of capital and intrinsic gains associated with the superior
micro-management by FDI investors.
2
and small countries are most likely to experience either or no entry. When the capital
market of local countries is inefficient, Kasuga (2003) found that acquisition mode
becomes more difficult. When the costs of technology transfer are high (low), Mattoo
et al (2004) showed that domestic welfare is generally higher under acquisition (direct
entry), whereas the foreign firm prefers direct entry (acquisition) to acquisition (direct
entry). Moreover, Muller (2007) showed that direct entry becomes optimal if the
technology transfer cost is sufficiently high and the local market is with low
competition.
However, product differentiation of MNEs, which can produce a significant effect
on their own profitability and hence their decision behaviors (including decisions on
technology transfer and mode of entry), has not yet been studied in the literature.
Porter (1990) claimed that product differentiation and lower costs are the two
competitive advantages, which are the main reasons for the rise of Japan as a major
exporter in the 1970s (Liu and Song, 1997). In the patent licensing contract,
Mukherjee and Balasubramanian (2001) found that the best technology is transferred
if products are sufficiently differentiated. As Lin (2004) observed in the Server and
the PC industry, the entry modes of MNEs into a target market are generally different,
because their different degrees of product differentiation matter.
In order to analyze the effect of product differentiation to a foreign firm’s decision
of entry mode and technology transfer, we employ a three-stage game model. The
foreign firm chooses an entry mode (acquisition or direct entry) in the first stage,
determines the level of technology transfer (which lowers its marginal production
costs at the production stage) in the second stage, and finally produces heterogeneous
products and Cournot-competes with the domestic firms in the host country.
The existing literature relevant to our paper mainly focused on the decision of
3
entry mode, either licensing or direct entry. The MNEs are seeking optimal entry to
prevent the dissipation of their technological advantages (e.g., Ethier and Markusen,
1996; Markusen, 2001; and Saggi, 1996 and 1999). On the other hand, Lee and Shy
(1992) showed that the quality of technology transferred by the foreign firm may be
lower if there are restrictions on foreign ownership. Roy et al (1999) identified market
structure as a crucial determinant of optimal domestic policy by assuming costless
technology transfer in a duopoly model.
Note that our paper analyzes only the horizontal FDI (as in Mattoo et al, 2004)
and the horizontal differentiated product markets, in order to handle the analysis of
host welfare.3 There is substantial evidence that most FDI is horizontal, in which all
firms (including the domestic and entrant firms) compete for sales of their final
products in the host country (Markusen and Maskus, 1999). Instead, for vertical FDIs,
the MNEs usually use the intermediate products that the domestic firms produce as
the inputs of their final products, which are mostly exported to other countries (Busse
and Hefeker, 2007).4 Furthermore, a large share of world trade happens as horizontal
differentiated products, especially among the OECD countries (Greenaway and
Torstensson, 2000). Storper, Chen, and De Paolis (2002) manifested that globalization
brings into a given market many more versions of very similar products and benefits
consumers’ variety preference, i.e., greater horizontal product differentiation, for EU
and OECD countries. What is more, the horizontal differentiated products introduced
to developing countries by MNEs have been increasing recently. MNEs, for lower
labor cost in early days, introduced vertically differentiated products into the markets
of China, who now instead welcomes MNEs to lead horizontal differentiated products
3
For vertical differentiation, the host welfare analysis will become more complicated and will remain
until further study is done.
4
According to Aizenman and Marion (2004), the vertical FDI often arises when the multinationals
fragment the production process internationally, locating each stage of production in the country where
it can be done at the least cost.
4
into its markets for enhancing competition (Fang and Zhao, 2003).5 Brambilla (2006)
empirically linked technological differences to product differentiation expansions of
MNEs in China. 6 As commonly happens in OECD countries and recent Asian
countries, our paper focuses on the horizontal FDI and horizontal differentiated
products, and that helps keep the welfare analysis tractable.
Our paper assumes that all domestic products are undifferentiated, but are
differentiated from the entry products with a degree (denoted as  ) to extend Mattoo
et al (2004). The results then contain what they observed: the foreign firm is more
likely to choose direct entry when the technology transfer cost is relatively high.
Moreover, most of the results depart from their findings. We show that: (1) lower
differentiation induces higher (lower) technology transfer given a sufficiently low
(high) transfer cost; (2) given a sufficiently low transfer cost, acquisition induces less
technology transfer than direct entry, and probability of that decreases in
differentiation; (3) when transfer cost is sufficiently low, the foreign firm would
prefer acquisition, and probability of that decreases in differentiation; (4) the
equilibrium entry mode more likely results in less technology transfer (than the
off-equilibrium mode), and probability for the equilibrium mode of acquisition to
induce higher technology (than direct entry) decreases in differentiation; (5) when
transfer cost is sufficiently low, the government would prefer acquisition, and
probability of that decreases in differentiation; (6) only when transfer cost is within
some certain intermediate levels, do both the government and the foreign firm prefer
5
Recently, Suntech Power, the top company of the solar energy industry in China, acquired MSK, a
solar energy company in Japan, in order to smoothly enter the Japanese market. On the other hand,
Sharp (from Japan), Siemens (from Germany), and Shell (from England) all set up their subsidiary to
compete with Suntech Power in China (Lin, 2008). They provide horizontal differentiated products for
consumers in China.
6
In the context of Brambilla (2006), product variety expansions (or new varieties) are not innovation
but rather a horizontal expansion or renovation of the product portfolio of firms. That is, new varieties
are those horizontal differentiated products of firms.
5
identical entry mode; (7) the foreign firm’s equilibrium entry mode more likely leads
to smaller host welfare than the off-equilibrium mode. Under equilibrium mode of
acquisition, the probability of the government preferring acquisition increases in  but
is always less than half.7
This paper is organized as follows. Section 2 discusses the theoretical setup and
describes the two entry modes (acquisition and direct entry) for the foreign firm to
enter the host country. Section 3 provides a three-stage game model to analyze the
extent of technology transfer and the optimal entry mode when a foreign firm decides
to enter the market of the host country. Section 4 focuses on the impact of the foreign
firm’s entry decision on the host country’s welfare. Section 5 offers our conclusions.
2. The Model
Extending the basic model structure of Mattoo et al (2004), our model includes
the product differentiation issue. There are n-1 domestic firms and one foreign firm.
The domestic firms have identical marginal production costs and homogeneous
products. The foreign firm enters the domestic market via acquisition or direct entry,
introduces a differentiated product, and has lower marginal production cost through
technology introduction. 8 To investigate the relation of product differentiation,
technology transfer, and mode of entry, we employ a three-stage game.
In the first stage of the model, the foreign firm has two options for entering the
domestic market: acquiring a domestic firm or setting up a wholly owned subsidiary
7
Mattoo et al (2004) obtained similar results without taking into account the issue of product
differentiation. They found that when  increases, the foreign firm is more likely to choose direct entry,
but under this circumstance the host country’s welfare is generally higher under acquisition relative to
direct entry.
8
Without loss of generality, the assumption that all domestic firms produce homogeneous products can
simplify our analysis.
6
that directly competes with domestic firms.9 If it chooses to acquire a domestic firm,
it makes a take-it-or-leave-it offer, specifying a fixed transaction price (v), to the
target firm. If the target firm accepts the offer, they form a new firm owned by the
foreign firm. If the target firm refuses the offer, the foreign firm can enter the market
by establishing its own subsidiary or by acquiring some other domestic firm. N
denotes the total number of firms in the domestic market after the foreign firm enters.
Thus, N = n-1 if the foreign firm chooses an acquisition mode, and N = n if it chooses
the direct entry mode.10
After selecting the mode of entry, the foreign firm in the second stage chooses the
amount of technology transfer (x) to its subsidiary, which costs the foreign firm C(x)
with C(x) =x2/2. Assume that the technology transfer will lower the foreign firm’s
marginal production cost at the third stage by x. Therefore, a higher parameter 
implies lower transfer efficiency and higher marginal transfer costs. In the last stage,
firms compete in a Cournot-Nash fashion.
Let a linear form of p(q) be the inverse demand function and notation qh, qi, and qf
represent the total outputs of the N-1 domestic firms, the individual domestic firm,
and the foreign firm, respectively.11 Therefore, the inverse market demand functions
for each firm can be expressed as follows:
N 1
pi  a  qh  q f and p f  a  qh  q f , where q h   q i a > 0, and 0    1. (1)
i 1
9
Mattoo et al (2004) shows that, in equilibrium, the foreign firm does not choose partial acquisition.
Throughout the paper in order to proceed to a concise mathematical analysis, we consider only the case
of full acquisition of the domestic firm.
10
Of course, there is a fixed cost of setting up plant for direct-entry MNEs. We do not contain the
fixed cost in our model because it does not affect the decisions of optimal choice.
11
Hereafter, subscript h, i, and f denote the domestic firms, individual domestic firm, and the foreign
firm, respectively. When the foreign chooses direct entry, the number of domestic firms N-1 = n-1;
otherwise N-1 = n-2.
7
Parameter  measures the degree of horizontal product differentiation. When  = 1,
the goods are perfect substitutes and the goods are unrelated when  = 0. As 
increases, the products become more standardized (i.e., less differentiated).
3. Model analysis
3.1. Product market
By backward induction, the analysis starts from the last stage (i.e., the product
market), in which the domestic firms and the foreign firm simultaneously choose their
output levels. The firms’ profit functions are given by:
 i (qi , q f )  (a  q h  q f  c)qi .
 f (qi , q f )  (a  qh  q f  c  x)q f  C ( x).
(2)
(3)
Solving equation (2) and (3) for the optimal output levels under the Cournot
competition is straightforward, which gives the first order conditions and then
individual equilibrium output as follows:
a  Nqi  q f  c, for any firm i.
a   ( N  1) qi  2q f  c  x, for firm f.
qi 
[ N   ( N  1)]( a  c)  Nx
(2   )( a  c)  x
and q f 
.
2
2 N   ( N  1)
2 N   2 ( N  1)
To guarantee positive individual output, Assumption 1 is made:
8
(4)
(5)
(6)
Assumption 1. x 
2 

(a  c)  x .
The comparative statics of individual equilibrium output with respect to technology
transfer, product differentiation, and the market competition are in Lemma 1:
Lemma 1. The comparative statics are as follows:
dq f
dq
dq h
(1) i < 0;
< 0;
 0.
dx
dx
dx
dq f
dq
(2) i > (<) 0 if x < (>) x and
< (>) 0 if x < (>) x, where x = [(4-)(N-1)-2N]
d
d
(a-c)/ [2N+2(N-1)] and x = [2(N-1)+ 2(1-)N] (a-c)/2N.
dq f
dq
dq h
(3) i < 0;
> 0;
 0.
dN
dN
dN
d ( f / qi )
d 2q f
d 2q f
d 2 qi
d 2 qi
(4)
 0;
 0;
0;
 0;
0;
dxdN
ddN
dN
dxdN
ddN
d ( f / qi )
d 2q f
d 2 qi
 0 if x >x;
 0;
 0.
d
dxd
dxd
Proof. See Appendix (A1) and (A2).
Given a higher level of technology transfer, as shown in part (1), the foreign firm
has a higher marginal production cost reduction, leading itself to a higher output and
the domestic firms to a lower output under the Cournot competition. An increase in
 (i.e., less differentiated) increases market competition, decreases the foreign firm’s
marginal revenue (as shown in (5)), and decreases qf if its cost reduction of
technology transfer is not too high. It then causes qi to increase under Cournot
competition and is stated in part (2). When the number of firms increases, each
individual equilibrium output decreases, while the total domestic output increases due
to higher competition, as part (3) claims. Moreover, higher competition due to the
9
increase of N (which lowers market price) increases the impact of x and  to each
individual’s output, as part (4) shows. According to part (4), the partial effect of x and
 to each individual output under the direct entry mode is higher than the acquisition
mode because NE > NA.12
3.2. Technology transfer
In the second stage, given an entry mode and expecting qi and qf in (6), the
foreign firm maximizes its profit by choosing the level of technology transfer x,
which reduces its marginal production cost by x and incurs transfer costs by x2/2.
Taking derivative of equation (3) with respect to x, we obtain the first order
conditions for technology transfer as follows, by the given (6) and the first order
condition in (5):
d f
dx

 f dq f
q f dx
 ( N  1)
 f dqi
dC x 
dC ( x)  f
 qf 
 0, where q f 

. (7)
qi dx
dx
dx
x
The first term in RHS of the first equality equals zero due to the first-order
condition of the foreign firm’s profit maximization. Following the terminologies used
in Mattoo et al (2004), the second term is the strategic effect (SE) while the third
term represents the scale effect (EE) of technology transfer. SE implies that an
increase in technology transfer lowers the domestic firms’ output, thereby increasing
the foreign firm’s profit. EE measures the direct effect of marginal production cost
advantage due to technology transfer. SE plus EE represents the marginal revenue for
technology transfer (hereafter, termed as MRTT), and the last term in (7) represents
Throughout the paper, superscript A and E will be used to denote the “Acquisition mode” and the
“Direct-entry mode”, respectively, when there is a need.
12
10
the marginal cost of technology transfer (termed as MCTT, hereafter). Therefore we
obtain equation (8) by rewriting equation (7):
MRTT  SE  EE  {( N  1)[ q f ][

dC x 
]}  {q f }  MCTT 
 x. (8)
2
dx
2 N   ( N  1)
By Lemma 1, we can show that the SE is concave in N while the EE is
decreasing in N. We then build Lemma 2:
Lemma 2. The strategic effect of technology transfer is concave in the number of
firms, while the scale effect always decreases in the number of firms.
Proof. See Appendix (B1).
Lemma 2 indicates that MRTT will decrease in N after some certain large
number of firms (given  and ). This implies that xE < xA, if it exists, will occur only
when N is sufficiently large.
Similarly, by taking derivative of SE and EE with respect to  and applying
Lemma 1, we obtain the impact of product differentiation to SE and EE in Lemma 3.
Lemma 3. If x > x, then dEE/d > 0 and dSE/d > 0. If x < x, dEE/d < 0 but the
sign of dSE/d is indeterminable.
Proof. See Appendix (B1).
11
Therefore, MRTT is higher when product differentiation becomes smaller under
a sufficiently large technology transfer. However, Even though MRTT decreases due
to the N-impact in Lemma 2 and/or -impact in Lemma 3, the equilibrium technology
transfer may still increase if MCTT is sufficiently low due to a sufficiently low . To
see the equilibrium technology transfer, we solve (8) for x:
x( N , , ) 
2 N [ N   ( N  1)]( a  c)
[2 N   2 ( N  1)]2  2 N 2
(9)
It is obvious that 2n 2 [2n  2(n  1)] 2 is the lower bound of  to guarantee a
non-negative technology transfer. 13 Moreover, to guarantee both non-negative
technology transfer and non-negative individual outputs, we need the following
assumption for technology transfer costs:14
Assumption 2.  
2n
 LL(n,  ) .
(2   )[ 2n   2 (n  1)]
In Appendix (B2), we show that dx dN < 0 if  > 1 and dx d < 0 if  > 2,
where 1 ( N , ) 
2N 2
2N 2
,  2 ( N , ) 
, and T2N
T [T  2 ( N  N   )]
T [T  4 ( N  N   )]
-2(N-1). Since LL < 1 < 2, we then build Proposition 1:
Proposition 1. When  > 2, the technology transfer is decreasing in both N and .
For x 0, it requires   2N2/[2N-2(N-1)]2, which is increasing in N. Therefore, n (for the direct
entry mode) instead of n-1 (for the acquisition mode) is used in the lower bound for x 0.
14
The necessary condition for non-negative x and individual outputs is   2N/{[(2-)(2N-2(N-1)],
which is increasing in N. Similar to footnote 9, n instead of n-1 is used in the lower bound LL.
13
12
When  < 1, the technology transfer is increasing in both N and . If it is
intermediate such that 1 <  < 2, then the technology transfer is decreasing in N but
increasing in .
Proof. See Appendix (B2).
There are several implications in Proposition 1. First, when  is sufficiently high,
a market with more competing firms induces less technology transfer of the foreign
firm if it enters. Therefore, under a discrete number of firms, x A  x E if  is
sufficiently high so that   TT (n, ) , where15
TT (n, ) 
2n(n  1)
.
4n(n  1)(1   )   (2n 2  4n  1)   4 (n  1)( n  2)
3
(10)
Secondly, given a ,   1 ( N , ) is satisfied only if N is sufficiently large. That
is, when N is sufficiently large under a given , dx/dN < 0 and x A  x E . The reason is
associated to what Lemma 2 describes, owing to the concavity of MRTT with respect
to N. Thirdly, given N, an increase in  results in MRTT increase (when x > x
according to Lemma 3) and a higher equilibrium x. However, the MRTT increase is
dominated by the MCTT increase if    2 ( N , ) , leading to a decrease of equilibrium
x.
Fourthly, less differentiation (i.e., higher) leads to a higher substitution effect
and market competition, which makes technology transfer more valuable for the
15
See Appendix (B3).
13
foreign firm. Therefore, the increase in  leads to higher dx/dN.16 This implies that
some result with x A  x E given fixed  and n under some lower level of  may turn
into “ x A  x E ” when  increases, which implies that TT(n,) becomes larger and
shifts up as in Figure 1(a).17 Thus, we are ready for building Proposition 2:
Proposition 2. If  < TT, xA < xE; otherwise xA > xE. If  is larger, then the possibility
of xA < xE is greater.
Proof. See Appendix (B3) and (B5) and simulations in Figure 1(a).
In an economic sense, when the technology transfer becomes much more
expensive (and therefore technology transfer decreases), the profit of the foreign firm
becomes smaller when the market is with more competing firms. Therefore, it could
be profitable for the foreign firm to adopt an acquisition mode to avoid fierce
competition, in which the increase of technology transfer follows and results in xA >
xE. On the contrary, the technology transfer becomes much cheaper when  is
sufficiently low. Therefore, facing more competing firms under the direct entry mode
(which results in smaller profit), the foreign firm can largely increase technology
transfer for taking advantage of the cheap cost-reduction and result in xE > xA. The
threshold between these two cases is represented as a TT curve.
[Figure 1 about here]
3.3. Optimal entry mode of the foreign firm
16
17
See Appendix (B4).
Also see Appendix (B5) for the mathematic proof.
14
In the first stage, the foreign firm chooses the mode for entering the host country.
Assume that the foreign firm has all the bargaining power under acquisition and it
offers the take-it-or-leave-it price v = iA(xA) to buy out a domestic firm.18 The net
profit (denoted as  fA ) under the acquisition mode for the foreign firm is thus  fA =
 fA ( x A )  v =  fA ( x A ) -  iA ( x A ).
Plugging (6) and (9) into (2) and (3), we obtain the following reduced-form
profits:
(a  c){( 2   )[ 2(n  1)   2 (n  2)]  2(n  1)} 2
 ( x )  (q ) = {
}.
[2(n  1)   2 (n  2)] 2   2(n  1) 2
A
i
A
A 2
i
 iE ( x E )  (q iE ) 2 = {
(a  c){( 2   )[ 2n   2 (n  1)]  2n} 2
}.
[2n   2 (n  1)]2  2n2
 fA   fA ( x A )  v   fA ( x A )   iA ( x A )  (q Af ) 2  (q iA ) 2
[( n  1)   (n  2)][ 2(n  1)   2 (n  2)]( a  c) 2
{
}  (q iA ) 2 .
2
2
2
[2(n  1)   (n  2)]   2(n  1)
[n   (n  1)][ 2n   2 (n  1)]( a  c) 2
} .
[2n   2 (n  1)] 2   2n 2
 Ef   Ef ( x E )  (q Ef ) 2  {
(11)
(12)
(13)
(14)
Let    fA   Ef and then solve for equation  /( a  c) 2 = 0, which gives equal
profit contours. By way of simulations, we find two possible solutions defined as
FF(n,) and FF1(n,), which give rise to  fA   Ef when FF(n,) >  > FF1(,).19
Since the simulations in FF1(n,) are almost equal to or less than LL(n,), we use LL
to substitute for FF1. Therefore, we conclude that  fA   Ef if FF(n,) >  > LL(n,)
18
Assuming that the foreign firm has all the bargaining power simplifies our analysis, which is also
used in Mattoo et al (2004).
19
Throughout the paper, all simulations are performed within the space of (0, 0, 2)  (, , n)  (3, 1,
200) with interval 0.015, 0.1, and 1, respectively.
15
and  fA   Ef if  > FF(n,).
FF curves are drawn in Figure 1(b), which shows up-shifting when product
differentiation is lower (i.e., when  is higher). Lower differentiation (i.e., higher )
results in a larger substitution effect and reduces more profit of each firm under direct
entry than under acquisition (for the number of firms is greater under direct entry).
Accordingly, a higher  (therefore a higher buy-out fee) is required in order to achieve
 fA =  Ef . Thus, the area below FF curve increases, which represents a probability
of  fA >  Ef to increase.
Proposition 3. If  < FF,  fA   Ef ; otherwise  fA   Ef . If  is larger, the probability
of the acquisition entry being the equilibrium mode increases.
Therefore, when the product introduced by the foreign firm is relatively highly
differentiated from the domestic products, direct entry is more likely to become the
equilibrium mode. While product differentiation is relatively low, acquisition entry is
more likely to become the equilibrium mode.
This finding accords with some practical facts. For example, in the Server
industry, the products are highly differentiated because the server product line is
highly technology-oriented. As we observe in the industry, the major players- Sun
Microsystems and IBM - enter each target market mostly by adopting the direct entry
strategy, which includes three steps: set up a branch, use the technology and engineers
from headquarters to expand into this target market, and find a cooperative distributor
to perform such a support role.20 In contrast, in the PC industry, the products become
20
See National Applied Research Laboratories in Taiwan (2004).
16
one kind of commodity and more homogenous. For the purpose of expanding the
market, firms in this industry mostly adopt the acquisition strategy. For example,
Lenovo, the number one PC market share leader in China, made an acquisition of
IBM’s PC division in 2004 to enter the US and European markets. Acer, one of the
top three PC makers in the world, made an acquisition of Gateway in 2007 to enter
the US market. In the telecommunications industry, mobile phones are also becoming
more homogeneous. In order to quickly expand its mobile phone market, BenQ
merged with the Siemens mobile phone division in 2006.
Combining the FF and the TT curves, there are four separated regions, which are
shown in Figure 2 and summarized in Table 1. Notations in Figure 2 and Table 1 are
defined as:21
R12  R1 + R2 = {LL < (n,) < min{TT, FF}}
R36  R3 + R6 = {FF < (n,) < TT}
R4  {TT < (n,) < FF}
R5  {(n,) > max{TT, FF}}
Rp  R124 {LL < (n,) < FF}={fA > fE}
Table 1. Classification of profits and technology transfer
fA > fE
fA < fE
xA>xE
R4
R5
xA<xE
R12
R36
The reason for these results is as follows. When  is sufficiently small: (1) xA <
xE according to Proposition 1 because the foreign firm can largely increase the
cheaper technology transfer to take advantage of production cost reduction; (2) The
buy-out fee of acquisition entry is driven substantially down; (3) the x-increase
Throughout the paper, whenever the notation “Rabc” is used, it means Ra + Rb +Rc, where a, b, c
represent any numbering.
21
17
brings about higher qf increase for direct entry mode because N increases, according
to Part (1) and (4) in Lemma 1. When  is sufficiently small, the benefit of (2) may
dominate the benefit of (3), making the foreign firm more likely to adopt acquisition
entry. This is represented in Figure 2 as the region R1 plus R2 (i.e., R12), indicating
 fA   Ef but xA < xE. 22 Nonetheless, if n is very small, the buy-out fee of
acquisition entry becomes very much higher and may outweigh the benefit of (3),
resulting in  fA   Ef with xA < xE, which is indicated as R3 plus R6 (i.e., R36) in
Figure 2.
Results turn out to be different when  is sufficiently high: (1) According to
Proposition 2, under a sufficiently highly expensive cost of technology transfer, xA >
xE; (2) the buy-out price of acquisition entry is higher; (3) the x-decrease brings about
lower qf decrease for direct entry mode because N increases, according to Part (1) and
(4) in Lemma 1. For a sufficiently high , the costs of (2) may outweigh the costs of
(3), making direct entry more likely to become the equilibrium entry. It is represented
in Figure 2 as the region R5, indicating  fA   Ef but xA > xE.23 If  is intermediately
high, then xA > xE still holds but the costs of (2) may not outweigh the costs of (3),
resulting in an equilibrium mode of acquisition. This is represented in Figure 2 as the
region R4 for  fA   Ef and xA > xE. We then build the following corollary.
Corollary 1. When  < FF, acquisition becomes the equilibrium entry mode, where xA
> xE if TT <  < FF and xA < xE if  < TT. When  > FF, direct entry becomes the
22
Eicher and Kang (2005) found similar results that larger markets are covered by acquisition because
the foreign firm can use FDI as threats to reduce purchasing price when the cost of technology transfer
is small.
23
The result is consistent with Salant et al (1983) and Kamien and Zang (1990), which stated that
higher technology transfer costs are more likely to induce direct entry because the profit (and hence the
buy-off fee of acquisition) of each domestic firm increases.
18
equilibrium mode and xA > xE.
[Figure 2 about here]
The size of each region can be a good measure of possibility for its occurrence.
In region R4 (in which acquisition is the equilibrium mode) and region R36 (in which
direct entry is the equilibrium mode), more technology transfer of the foreign firm is
induced (than under the off-equilibrium mode). However, simulations show that the
area ratio R4/Rp (i.e., R4/R124) and the area ratio R36/R356 are relatively small.
Figure 3(a) shows that R4/Rp is increasing in  and always less than 0.5, while Figure
3(b) shows that R36/R356 is always less than 0.07 but exhibits no regular relation
with . That implies that the equilibrium entry mode of the foreign firm more likely
results in less technology transfer than the off-equilibrium mode; and probability of
“xA > xE” under equilibrium mode of acquisition decreases in differentiation. However,
even when products are undifferentiated, probability of “A > B and xA > xE” is less
than probability of “A > B and xA < xE”.
Proposition 4. The equilibrium entry mode more likely results in less technology
transfer (than the off-equilibrium mode). The probability for the equilibrium mode of
acquisition to induce higher technology (than direct entry) decreases in
differentiation.
[Figure 3 about here]
4. Host country welfare
19
To analyze the change of host welfare, we first investigate consumers’ surplus (CS)
and firms’ profits.24 The consumer surplus can be expressed as:
1
1
1
1
(a  pi )q h  (a  p f )q f  (q h  q f )q h  (q h  q f )q f
2
2
2
2
1
1
 [( q h ) 2  (q f ) 2  2q h q f ]  [Q 2  2(1   )q h q f ], where Q  q h  q f .
2
2
CS 
(15)
Therefore, the host welfare can be approximately expressed as the consumer surplus
plus the domestic profits (denoted as D):
W  CS  D ,
(16)
where D  i 1 i for direct entry and D  v  i 1  i for acquisition, in which
n 1
n2
the buy-out price v   i and  i  (qi ) 2 . We then derive the host welfare under the
acquisition mode and the direct entry mode as follows, respectively:
1
W A  [(qhA ) 2  (q Af ) 2  2qhA q Af ]  (n  2)( qiA ) 2  v
2
1
 [(n 2  2n  2)( qiA ) 2  (q Af ) 2  2 (n  2)qiA q Af ].
2
(17)
1
W E  [( qhE ) 2  (q Ef ) 2  2qhE q Ef ]  (n  1)( qiE ) 2
2
1
 [( n 2  1)( qiE ) 2  (q Ef ) 2  2 (n  1)qiE q Ef ].
2
(18)
The consumer surplus, as in equation (15), decreases in  when the cost of
24
Our paper starts with a given demand function, rather than the consumers’ utility function. Therefore,
we approximate welfare by summing consumers’ surplus and firms’ profits, as in Shy (1995), p.68.
20
technology transfer is sufficiently high, as shown in Appendix (C1). Lower product
differentiation (i.e., higher ) reduces each domestic firm’s marginal revenue,
inducing a decrease of output and profit. A lower cost of technology transfer gives the
foreign firm lower production costs, driving down the domestic market price.
Therefore, when  is sufficiently high, a higher product differentiation can induce
higher consumer surplus and higher domestic firms’ profits, which is consistent with
the findings of Symeonidis (2003) that quality heterogeneity is beneficial for both
consumers and firms.
Lemma 4. When the cost of technology transfer is sufficiently high, larger
heterogeneity is beneficial for both consumers and firms, and therefore increases the
host welfare.
Proof. See Appendix (C1).
Proceeding with welfare comparison between different entry modes, we solve
for (WA-WE)/(a-c)2 = 0, with the solution defined as WW(n,). By equation (17) and
(18) and through simulations, we find that WA < WE if  < WW(n,); otherwise, WA >
WE. When  increases, the profit of an individual firm decreases, resulting in a lower
buy-out fee for acquisition. Accordingly, the welfare difference WE-WA increases,
which then requires a relatively higher  to increase the buy-out fee (and therefore
increase WA) for achieving WE = WA. This shifts the WW curve up and enlarges the
area below WW when  increases, indicating that the probability of WA < WE increases
for lower product differentiation. Figure 1(c) shows the result, in which the WW curve
shifts up when  increases.
21
Proposition 5. If  < WW, then WA < WE; otherwise WA > WE. If  becomes larger,
the probability of WA < WE increases.
Combining together the curves of FF, TT, and WW, there are six separate regions
as depicted in Figure 2 and summarized in Table 2. Notations are defined as:
Rw  R1 + R6 {LL < (n,) < WW}
R1  {LL < (n,) < min{WW, FF}}
R2  {WW < (n,) < min{FF, TT}}
R3  {max{WW, FF} < (n,) < TT}
R6  {FF < (n,) < WW}.
Table 2. Classification of welfare, profits, and technology transfer
WA < WE
WA > WE
xA > xE
xA < xE
xA > xE
xA < xE
A > E
--
R1
R4
R2
A < E
--
R6
R5
R3
The reasons which result in the classification in Table 2 are as follows. When  >
LL, the increase of one more firm in the market increases CS but decreases the
domestic firms’ profits. 25 Therefore, consumers always prefer direct entry mode
while the domestic firms always prefer acquisition mode. For the acquisition mode to
create higher welfare, the buy-out fee paid to some domestic firm has to outweigh the
CS increase of direct entry mode, which requires a sufficiently high . Accordingly,
we observe in Figure 2 that WA > WE only when  is sufficiently high so that  > WW,
which contains regions R2, R3, R4, and R5. In contrast, when  is too small, the
25
See Appendix (C2).
22
buy-out fee paid to some certain domestic firm is too small to outweigh the CS
increase of direct entry mode, resulting in WA < WE, shown as the region below WW
(denoted as Rw). The key factor to make WA < WE is that the technology transfer
under acquisition is relatively smaller than under direct entry (i.e., xA < xE).
When WW <  < FF (i.e., in R2 and R4), both the government and the foreign
firm prefer the acquisition mode in which a sufficiently high  induces higher
technology transfer than under the direct entry mode. When FF <  < WW (i.e., in R6),
they both prefer direct entry, which induces higher technology transfer than
acquisition. When  > max(WW, FF) (i.e., in R3 and R5), the government prefers
acquisition but the foreign firm prefers direct entry, in which xA is most likely to be
greater than xE. Lastly, when  < min(WW, FF) (i.e., in R1), the government prefers
direct entry but the foreign firm prefers acquisition, in which xA < xE.
Corollary 2. When  > max(WW, FF) (i.e., in R3 and R5), the government prefers the
acquisition but the foreign firm prefers direct entry. When  < min(WW, FF) (i.e., in
R1), the government prefers the direct entry but the foreign firm prefers acquisition.
When  is an intermediate value, both the government and the foreign firm prefer
acquisition if WW <  < FF (i.e, in R2 and R4) but both prefer direct entry if FF < 
< WW (i.,e in R6).
Proof. See simulations in Figure 2.
We use area size of each region as a measure of probability for its occurrence.
When the foreign firm chooses acquisition entry, the probability of WA > WE is
measured by the ratio of area R24 to area R124, which is increasing in  but never
23
goes beyond 0.5, as shown in Figure 4(a). Contrarily, when the foreign firm chooses
direct entry, the probability of WA < WE is measured by the ratio of area R6 to area
R536, which is less than 0.0075 and exhibits no regular relation with , as shown in
Figure 4(c). In sum, the foreign firm’s equilibrium entry mode more likely leads to
smaller host welfare than the off-equilibrium mode. That is, the possibility for both
the foreign firm and the government to prefer an identical entry mode is shown to be
relatively small.
Moreover, the probability of “WA > WE and xA > xE” under equilibrium mode of
acquisition, which is measured by R4/R124, is also increasing in  and is less than 0.5,
as shown in Figure 4(b). Contrarily, the probability of “WA < WE and xA < xE” under
equilibrium mode of direct entry, which is measured by R6/R536, is less than 0.0075
as in Figure 4(c). Therefore, there is relatively small possibility that the equilibrium
entry mode leads to higher host welfare as well as higher technology transfer (than
the off-equilibrium mode).
Interestingly in Table 2, as long as host welfare under direct entry is larger than
under acquisition entry (i.e., WA < WE), the foreign firm’s technology transfer will
always be xA < xE (i.e., R1 and R6), indicating that technology transfer to reduce
market price is a main source for welfare increase.
[Figure 4 about here]
Proposition 6. The foreign firm’s equilibrium entry mode more likely leads to smaller
host welfare than the off-equilibrium mode. Under equilibrium mode of acquisition,
the probability of the government preferring acquisition increases in  but is always
less than half.
24
5. Conclusions
Multinational enterprises and their products have been and will continue to be
an important source for improving competitiveness of domestic firms in developing
countries. Accordingly, multinationals’ entry mode, technology transfer, and product
differentiation will have distinct impact on the host countries’ welfare. This paper
shows that an entrant’s product differentiation does affect the choice of technology
transfer and entry mode and the host country’s welfare.
Our analysis shows that lower differentiation induces higher (lower) technology
transfer given a sufficiently low (high) transfer cost. Given sufficiently low transfer
costs, acquisition results in lower technology transfer but higher foreign firm’s profit,
which occurs with a decreasing probability in differentiation. In other words, when
transfer cost is sufficiently low, acquisition more likely becomes the equilibrium entry
mode if the foreign firm introduces less differentiated products. This finding is
supported by stylized facts. For example, firms in the PC industry (in which products
are commonly acknowledged as being highly undifferentiated) mostly adopt the
acquisition strategy to expand the market. Lenovo acquiring IBM’s PC division in
China in 2004 and Acer acquiring Gateway for entering the US market in 2007 are
the recent events.
However, acquisition as the equilibrium entry mode more likely results in less
technology transfer and lower host welfare (than direct entry). Probability of its
occurrence decreases in differentiation. We further find that either entry mode as an
equilibrium will most likely induce less technology transfer and less host welfare
(than the off-equilibrium entry mode) regardless of product differentiation level.
In fact, the foreign firm and the host government generally prefer different entry
25
modes. When the cost of technology transfer is very high, it is most likely that the
foreign firm chooses direct entry, while it most likely chooses acquisition entry under
a sufficiently low cost of technology transfer. Both of them result in smaller host
welfare (than the other entry mode). Only when transfer cost is within some certain
intermediate levels, do the government and the foreign firm prefer identical entry
mode. But probability of that is always less than half, which decreases in
differentiation.
Lastly, the extension issues of these lines can include the spillover effect of
technology transfer, the joint ventures as another entry mode, and the vertical FDI.
Appendix
Appendix A. Calculations for Section 3.1
(A1)
dqi
2
2
dqh
dq
 ( N  1) i  0 .



 0.
2
2
2
dx
dx
dx
2 N   ( N  1)
N (2   )  
dq f
N
=
> 0.
dx
2 N   2 ( N  1)
dqi
[( 2 N   (4   )( N  1)]( a  c)  [ 2 ( N  1)  2 N ]x
> 0 if x < x 

[2 N   2 ( N  1)] 2
d
[(4-)(N-1)-2N] (a-c)/[2N+2(N-1)].
dq f
( N  1) 2Nx  [ 2 ( N  1)  2(1   ) N ]( a  c)
=
< 0 if x < x  [2(N-1) +
2
2
d
[2 N   ( N  1)]


2(1-)N] (a-c)/2N.
Comparing with x in Assumption 1, we get x < x < x .
dq f   [( 2   )( a  c)  x]
dqi  (2   2 )[( 2   )( a  c)  x]

 0.
=
<
0.
dN
dN
[2 N   2 ( N  1)] 2
[2 N   2 ( N  1)] 2
dqh
dq 2[( 2   )( a  c)  x]
 qi  ( N  1) i 
 0.
dN
dN
[2 N   2 ( N  1)]2
(A2)
26
d (dqi dx)
 2 (2   2 )
d (dq f dx)
2


0
.

 0.
dN
dN
[2 N   2 ( N  1)]2
[2 N   2 ( N  1)] 2
d (dqi d ) [ N (4   4 )   2 (6   2 )]( a  c  x)   (a  c)[2 N (4  2 2 )  4(4   2 )]

 0.
dN
[2 N   2 ( N  1)]3
d (dq f d ) 4 N[(2  2   2 ) N   2 ](a  c)  2 [(2   2 ) N   2 ]x

 0.
dN
[2 N   2 ( N  1)]3
q
d ( f qi )
q f
d ( dqi dx)
 4N
d ( f qi )
 q f   f .

 0.
 
 0.
d

d
[2 N   2 ( N  1)]2
dN
N
d (dq f dx)
2N ( N  1)

 0.
d
[2 N   2 ( N  1)] 2
Appendix B. Calculations for Section 3.2
(B1)
( f qi ) dqi  f (dqi dx)
dSE  f dqi
2

 ( N  1)[

] =

2
dN
qi dx
N
dx
qi N
[2 N   ( N  1)]3
(-) (-)
() (-)
(-) ()
N[(2  4   2 )(a  c)  (2   2 ) x]   [( 4   )(a  c)  x] < 0 if N is sufficiently
 [( 4   )( a  c)  x]
dEE dq f
large so that N 
.

 0.
[( 4   2  2)( a  c)  (2   2 ]) x
dN
dN
d 2 SE
 2 2

 {[ 4 (3   )   3 ( N  1)( 4   )  4 N (1  2 )]( a  c)
2
2
4
dN
[2 N   ( N  1)]
 [(4   2 )( N   2 )   2 N (1   2 )]x}  0.
 ( f qi ) dqi  f  ( dqi dx)
dSE
K
 ( N  1)[

] and dEE  dq f , where
d

dx
qi 
d
d
(-) (-)
(-)
 q f 
and
dq f
d
. According to (A1), if x > x, then
dq f
d
 ( d f dqi )
>0, which results in

=
dEE
>0
d
dSE
> 0.
d
(B2)
dx
 2 (a  c)
 2
 {T [2 N (1   )   2 ( N  1)]  2 N 2 }  0, if   1 ( N , ) , where
2 2
dN [T   2 N ]
1 ( N , ) =
2N 2
and T = 2N - 2(N-1).
T [T  2 ( N  N   )]
dx [2 N ( N  1)( a  c)]{[ 3 2 ( N  1)  2 N (1  2 )]T  2 N 2 }
< 0 if  >  2 ( N , ) ,

d
[T 2  2 N 2 ]2
27
where  2 ( N , ) 
2N 2
and T = 2N - 2(N-1).
T [T  4 ( N  N   )]
dx
[ N   ( N  1)][ 2 N   2 ( N  1)]
 4 N (a  c) 
 0.
d
{[ 2 N   2 ( N  1)] 2   2 N 2 }2
(B3)
2(a  c)
 {( n  1)[( n  1)   (n  2)][( 2n   2 (n  1)) 2   2n 2 ]
RS
 n[n   (n  1)][( 2(n  1)   2 (n  2)) 2   2(n  1) 2 ]}
x  x A  x E 
2(a  c)
 {[ 4n(n  1)(1   )   3 (2n 2  4n  1)   4 (n  1)( n  2)]  2n(n  1)},
RS
where R  [2(n  1)   2 (n  2)] 2   2(n  1) 2 and S  [2n   2 (n  1)] 2   2n 2 .

We then obtain that x  0 (i.e., xA > xE) if   TT (n,  ) , where TT ( n,  ) =
2n(n  1)
.
4n(n  1)(1   )   (2n 2  4n  1)   4 (n  1)( n  2)
3
(B4)
d (dx
)
dN  2(a  c) {T 2 [4 N 2  16N ( N  1)  22 2 N ( N  1)  4 3 ( N  1) 2 ]
d
[T 2  2 N 2 ]3
2
2
 2 N [4 N  2 N ( N  1)(8  3 )  12 3 ( N  1) 2 ]},
which is always positive, proven by way of simulations.
(B5)
dTT (n, )
 2 3 [2n 2  (2   )n  (1   )]

 0.
dn
[4n(n  1)(1   )   3 (2n 2  4n  1)   4 (n  1)( n  2)] 2
d 2TT (n,  )
2 3


dn 2
[4n(n  1)(1   )   3 (2n 2  4n  1)   4 (n  1)( n  2)] 3
{4n 3 [4(1   )   3 (2   )]  3n 2 [8  4 (1   )   3 (4   2 )] 
n[24  4 (1  5 )  2 3 (6  5 )  7 5 ]  [8  2 2 (4  3 )   4 (5  4 )]}  0.
dTT (n, ) 2n(n  1)[ 4n(n  1)  3 2 (2n 2  4n  1)  4 3 (n  1)( n  2)]

 0.
d
[4n(n  1)(1   )   3 (2n 2  4n  1)   4 (n  1)( n  2)] 2
Appendix C. Calculations for Section 4
(C1)
With the equilibrium x in equation (9), we derive the equilibrium outputs in second
period: qh  ( N  1)qi 
( N  1)(a  c){( 2   )[ 2 N   2 ( N  1)]  2 N}
, and
[2 N   2 ( N  1)]2  2 N 2
28
qf 
(a  c)[ N   ( N  1)][ 2 N   2 ( N  1)]
. By simulations under Assumption 2
[2 N   2 ( N  1)]2  2 N 2
(i.e.,  > LL), we obtain that:
dq h
 ( N  1)( a  c)

 {[ 2 N   2 ( N  1)] 2 [( 2  4   2 ) N
2
2
2 2
d {[ 2 N   ( N  1)]   2 N }
  (4   )]  2 N [( 2  4  3 2  4 3 ) N 2  N (4  3  8 2 )  4 3 ]}  0;
 ( N  1)( a  c)
 {[ 2 N   2 ( N  1)][( 4  4  2 3 ) N 2
2
2
2 2
d {[ 2 N   ( N  1)]   2 N }
 2 2 N (1     2 )   4 ]  2 N 2 [( 2  2  3 2 ) N  3 2 ]}  0; and
dq f

dq
dCS
dq
 (qh  q f ) h + (q f  qh ) f  qh q f < 0 when  is sufficiently large.
d
d
d
d
dW dCS
d i d (qi ) 2
dq
Since
=
+ ( N  1) i < 0

 2qi i  0 according to (A1),
d
d
d
d
d
d
only when  is sufficiently large.
(C2)
With the second-period equilibrium x in (9) and qh and qf in (C1), simulations show
dq f
dq h
  (a  c)

that
> 0 holds under Assumption 2. We also derive that
= 2
dN [T   2 N 2 ] 2
dN
{( 2   )T 2  2 N[ N (2     2 )  2 2 ]}  0 if  >
2 N (2 N   2 ( N  2)  N )
 . It
(2   )[ 2 N   2( N  1)] 2
dQ dqh dq f
> 0 under


dN dN dN
dq f
dq h
Assumption 2 (i.e.,  > LL), indicating that
dominates
. Although
dN
dN
dq
dQ
dCS
dqh
= Q
q f  f qh < 0 under a sufficiently large , we obtain that
dN
dN
dN
dN
dq f
dq
(1   )[ h q f 
q h ] > 0 always hold under assumption 2 because the first term
dN
dN
of its RHS dominates the second term when  is not that high. On the other hand,
is shown that  > LL. But simulations show that
Simulations show that
d h d ( N  1) i
d
dq

  i  ( N  1) i  qi2  2( N  1)qi i <
dN
dN
dN
dN
0, indicating the first term in the RHS of the third equality is dominated by the
negative second term.
References
Aizenman, J., Marion, N., 2004. The merits of horizontal versus vertical FDI in the
presence of uncertainty. Journal of International Economics 62, 125– 148.
29
Brambilla, I., 2006. Multinations, technology, and the introduction of varieties of
goods. NBER Working Paper No. 12217.
Busse, M., Hefeker, C., 2007. Political risk, institutions, and foreign direct investment.
European Journal of Political Economy 23(2), 397-415.
Bjorvatn, K., 2001. Foreign ownership and market entry. Nordic Journal of Political
Economy 27, 13-32.
Culem, C., 1988. The locational determinants of direct investments among
industrialized countries. European Economic Review 32, 885-904.
Davidson, W.H., 1980. The location of foreign direct investment activity: Country
characteristics and experience effects. Journal of International Business Studies
11(2), 9-22.
Devereux, M.P., Griffith, R., 1998. Taxes and the location of production: Evidence
from a panel of U.S. multinationals. The Journal of Public Economics 68,
335-367.
Eicher, T., Kang, J.W., 2005. Trade, foreign direct investment, or acquisition: Optimal
entry modes for multinationals. Journal of Development Economics 77,
207-228.
Ethier, W.J., Markusen JR., C.D., 1996. Multinational firms, technology diffusion and
trade. Journal of International Economics 41, 1-28.
Fang, J., Zhao, I.S., 2003. The institution of economic integration between China and
Taiwan. The Academic Journal of Geography 58.
Glass, A.J., Saggi, K., 1999. FDI policies under shared factor markets. Journal of
Economics 49, 309–332.
International Monetary Fund, 2004. International financial statistics year book.
Bernan Press.
Kaimien, M.I., Israel, Z., 1990. The limits of monopolization through acquisition. The
Quarterly Journal of Economics 105, 465-499.
Kasuga, H., 2003. Capital market imperfections and forms of foreign operations.
International Journal of Industrial Organization 21, 1043-1064.
Lee, F.C., Shy, O., 1992. A welfare evaluation of technology transfer to joint ventures
in the developing countries. The International Trade Journal 7, 205-220.
Lin, Z.X., 2004. The influence of Lenovo acquires IBM PC division. Industry &
Technology Intelligence Services in Taiwan.
Lin, W.C., 2008. The solar energy market in China. Micro-Electronics 268, cover
story.
Liu, X., Song, H., 1997. China and the multinationals- A winning combination. Long
Range Planning 30, 74-83.
Lv, Y.J., 2005. The chance and challenge in China for Taiwan’s enterprises. Learning
and Development 598, 62-68.
Markusen JR, C.D., Maskus, K.E., 1999. Estimating the knowledge-capital model of
the multinational enterprise. American Economic Review 89, 693-708.
Markusen JR. C.D., 2001. Contracts, intellectual property rights, and multinational
investment in developing countries. Journal of International Economics 53,
189-204.
30
Mattoo, A., Olarrega, M., Saggi, K., 2004. Mode of foreign entry, technology transfer,
and FDI policy. Journal of Development Economics 75, 95-111.
Mudambi, R., Mudambi, S.M., 1995. From transaction cost economics to relationship
marketing: A model of buyer-supplier relations. International Business Review
4, 419-433.
Mudambi, R., Mudambi, S.M., 2002. Diversification and market entry choices in the
context of foreign direct investment. International Business Review 11, 35-55.
Mukherjee, A., Balasubramanian, N., 2001. Technology transfer in a horizontally
differentiated product market. Research in Economics 55, 257-274.
Muller, T., 2007. Analyzing modes of foreign entry: Greenfield investment versus
acquisition. Review of International Economics 15(1), 93-111.
National Applied Research Laboratories in Taiwan, 2004. Science &Technology
Policy Research and Information Center e-paper.
Porter, M., 1990. The Competitive Advantage of Nations. Macmillan, London.
Prasad, E., Wei, S-J., 2005. The Chinese approach to capital inflows: patterns and
possible explanations. NBER Working Paper No. 11306.
Razin, A., Sadka, E., 2003. Gains from FDI inflows with incomplete information.
Economics Letters 78, 71-77.
Roy, Prithvijit, Kabiraj, T., Mukherjee, A., 1999. Technology transfer, merger, and
joint venture- A comparative welfare analysis. Journal of Economic Integration
14(3), 422-466.
Saggi, K., 1996. Entry into a foreign market: foreign direct investment versus
licensing. Review of International Economics 4, 99-104.
Salant, S.W., Switzer, S., Reynolds, R.J., 1983. Losses from horizontal merger: The
effects of an exogenous change in industry structure on Cournot-Nash
equilibrium. Quarterly Journal of Economics 98, 185-199.
Shy, O. 1995. Industrial organization: theory and applications. Massachusetts Institute
of Technology.
Symeonidis, G., 2003. Quality heterogeneity and welfare. Economics Letters 78, 1-7.
Symeonidis, G., 2003. Comparing Cournot and Bertrand equilibrium in a
differentiated duopoly with product R&D. International Journal of Industrial
Organization 21, 39-55.
31
Figures
Figure 1. Contours
 = 0.2
 = 0.4
(a) TT : x
E
=x
 = 0.6
 = 0.8
 = 0.9

(b) FF :  = 
A
E
 = 1.0

A
E
(c) WW : W = W
2.5
2.5
2.5
2.25
2.25
2.25
2
2
2
1.75
1.75
1.75
1.5
1.5
1.5
1.25
1.25
1.25
1
1
1
0.75
0.75
0.75
0.5
3 6 9 12 15 18 21 24 27 30 33 36 39 42 45
n
0.5
3 6 9 12 15 18 21 24 27 30 33 36 39 42 45
32
n
0.5
A
n
3 6 9 12 15 18 21 24 27 30 33 36 39 42 45
Figure 2 Technology Transfer, profit, and welfare
Simulations are performed within the space of (0, 0, 2)  (, , n)  (3, 1, 200). Rx  {LL < (n,) < TT},
Rp  {LL < (n,) <FF}, and Rw  {LL < (n,) < WW}, R1  {LL < (n,) < min{WW, FF}}, R2 
{WW < (n,) < min{FF,TT}}, R3  {max{WW, FF} < (n,) < TT}, R4  {TT < (n,) < FF}, R5  {(n,)
> max{TT, FF}}, R6  {FF < (n,) < WW}.
LL
 TT
FF
 WW
0.65


 = 0.2

 = 0.4
0.85
R5
R5
R3
R3
R2
R6
 = 0.6
1.15
R5
R2
1.05
R6
R4
R3
R2
R1
0.75
R1
1.25
R1
0.95
0.85
n
0.55
3
6
9

12
15
18
3
 = 0.8
1.75
n
0.65
21
1.2
1.95
6
9

R5
R3
1.85
1.1
1.65
R5
1.55
R3
TT
A
E
R1 I : x  x

0.8
1.45
n
1.15
3
6
9
12
15
nn
0.7
1.35
23
7
6
12
3
9 17
2.55
2.9
2.8
2.7
2.45
2.6
2.5
2.35
2.4 R2
2.3
2.25
2.2
2.1
2.15
2
1.9
2.05
1.8
1.7
1.95
1.6
1.5
1.85
1.4
1.3
1.75
23
2212
Figure 3. Area ratio for foreign firm’s profit
 = 0.2
 = 0.4
 = 0.6
 = 0.8
(a) R4/Rp
 = 0.9

 = 1.0

(b) R36/R536
0.07
0.5
0.06
0.4
0.05
0.3
0.04
0.03
0.2
0.02
0.1
0.01
0
3
6
9
12
15
18
n
21
n
0
3
6
33
6
9

R1
1.25
n
0.75
21
R2
0.9
1.55
1.35
18
0.9
 ==0.6
II : x A  x E
1.65
R2
15
R4
1.75
1
R4
1.45
12
12
15
18
21
 == 1.0
1.0
R5
II : x  x E
A
R4
I: x  x
A
TT
E

R1
nn
7
126
17
229
Figure 4. Area ratio for welfare
 = 0.2
 = 0.4
 = 0.6
 = 0.8
(a) R24/R124
 = 0.9
 = 1.0


(b) R4/R124
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
(c) R6/R536
0.0075
0.005
0.0025
0.1
0.1
n
0
3
6
9
12
15
18
21
n
0
3
6
9
12
34
15
18
21
n
0
3
6