J. Japanese Int.

Economies 26 (2012) 308325

Contents lists available at SciVerse ScienceDirect

Journal of The Japanese and

International Economies
journal homepage: www.elsevier.com/locate/jjie

Trade liberalization and welfare: Differentiated-good

versus homogeneous-good markets
Hajime Takatsuka a,, Dao-Zhi Zeng b,c
a
b
c

Graduate School of Management, Kagawa University, Saiwai-cho 2-1, Takamatsu, Kagawa 760-8523, Japan
Graduate School of Information Sciences, Tohoku University, Aoba 6-3-09, Aramaki, Aoba-ku, Sendai, Miyagi 980-8579, Japan
Center for Research of Private Economy, Zhejiang University, Zheda Road 38, Hangzhou, Zhejiang 310027, China

a r t i c l e

i n f o

Article history:
Received 11 August 2011
Revised 14 April 2012
Available online 23 May 2012
JEL classication:
F12
Q17
R1
Keywords:
Trade costs
Firm location
Home market effect
Trade liberalization
Welfare

a b s t r a c t
Takatsuka, Hajime, and Zeng, Dao-ZhiTrade liberalization and
welfare: Differentiated-good versus homogeneous-good markets
In this paper, we examine the effects of liberalization on industrial
location and national welfare in a framework of new economic
geography. Specically, we explicitly incorporate arbitrary trade
costs in both differentiated-good and homogeneous-good sectors
into a two-country model, and clarify the effects of trade-barrier
reduction in each sector. We show that their impacts on welfare
levels in the two countries are different, and, if an industry is liberalized while the other is protected, a conict between the countries
might occur. Therefore, appropriate liberalization in both sectors is
effective to alleviate such a conict. J. Japanese Int. Economies 26 (3)
(2012) 308325. Graduate School of Management, Kagawa University, Saiwai-cho 2-1, Takamatsu, Kagawa 760-8523, Japan; Graduate School of Information Sciences, Tohoku University, Aoba 6-309, Aramaki, Aoba-ku, Sendai, Miyagi 980-8579, Japan; Center for
Research of Private Economy, Zhejiang University, Zheda Road
38, Hangzhou, Zhejiang 310027, China.
2012 Elsevier Inc. All rights reserved.

1. Introduction
In this paper, we examine the impacts of trade liberalization on industrial location and national
welfare in a framework of new economic geography (NEG) (Fujita et al., 1999; Baldwin et al., 2003).

Corresponding author. Fax: +81 87 832 1907.

E-mail addresses: takatsuka@gsm.kagawa-u.ac.jp (H. Takatsuka), zeng@se.is.tohoku.ac.jp (D.-Z. Zeng).
0889-1583/$ - see front matter 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jjie.2012.05.003

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

309

Specically, we explicitly incorporate arbitrary trade costs in both differentiated-good and homogeneous-good sectors into a two-country model, and clarify the effects of trade-barrier reduction in each
sector. We show that reducing the trade barrier in the homogeneous-good sector has a different impact
on the welfare levels in two countries from reducing the trade barrier in the differentiated-good sector.
According to Article XXIV of the General Agreement on Tariffs and Trade (GATT), all free trade
agreements (FTAs) should aim to eliminate duties and other restrictive regulations of commerce on
substantially all the trade among participating countries. However, in real FTAs, some industries are
often protected. For example, the United States excluded about 100 items (e.g., sugar and dairy products) from the liberalized list of a USAustralia FTA in 2005. In addition, Japan carefully protects the
agricultural market by imposing high tariffs and is very reluctant to join the Trans-Pacic Partnership
(TPP). These facts indicate the necessity to examine how a countrys welfare level changes if some industries are liberalized while others are protected.
Trade costs include various barriers to trade. In addition to formal regulations such as tariffs, the
technical barriers to trade (TBT) are also signicantly restrictive. In particular, country-specic product standards have the potential to keep foreign producers out of domestic markets by imposing adaptation costs (Fischer and Serra, 2000; Gandal and Shy, 2001), while internationally shared standards
(e.g., International Organization for Standardization (ISO)) are expected to promote trade (Swann
et al., 1996). Interestingly, an empirical result of Moenius (2004) shows country-specic standards
having different impacts on the trade pattern. Although they tend to inhibit trade in non-manufactured goods such as agriculture, they promote trade in manufactured goods.
Turning to theoretical trade studies, since Krugman (1980), the literature of NEG successfully claries that industrial location is strongly related to trade costs.1 However, most NEG papers focus on the
positive aspects, particularly the home market effect (HME), rather than the welfare issues (Davis,
1998; Head and Ries, 2001; Head et al., 2002; Ottaviano and Thisse, 2004; Yu, 2005; Crozet and Trionfetti,
2008; Zeng and Kikuchi, 2009). Some studies have focused on the gains and losses from trade; however,
the comparisons are limited to completely free-trade economies with entirely autarky economies
(Krugman, 1981; Venables, 1987). Furthermore, since Helpman and Krugman (1985), most NEG papers
(e.g., Krugman and Venables, 1990, 1995; Baldwin and Venables, 1995) impose an assumption of free trade
on the homogeneous good. Although this convenient assumption makes the analysis of the differentiatedgood sector much easier, it has two theoretical defects. First, if the two countries have identical technology
in the homogeneous-good sector, the wages in the two countries are equalized under this assumption,
failing to capture the wage gap between countries. Baldwin and Robert-Nicoud (2000) showed the reduction of frictional barriers between asymmetric-sized nations improving the welfare of both nations when
wages in the two countries are equalized by the free-trade assumption in the homogeneous-good sector. It
is uncertain if their result remains true when a wage gap is possible. Second, it becomes impossible to
examine the integration of the homogeneous-good markets under this assumption.
The importance of transport costs in the homogeneous-good sector was rst recognized by Davis
(1998), who showed that the HME of Helpman and Krugman (1985) disappears if the homogeneous
good is transported with the same positive cost as the differentiated goods. Fortunately, their model
can be used in our research. We maintained the structure of two asymmetric-sized countries, one production factor (labor), and two sectors, in which arbitrary trade costs were allowed in both sectors.
This made it possible to compare the integration of the differentiated-good and homogeneous-good
markets. Furthermore, we were able to analyze the effects of integration on welfare at an arbitrary
level of trade costs. Thereby, we claried when and which market integration produces (or does not
produce) a conict between the two countries.
To the best of our knowledge, the equilibrium analysis of the HelpmanKrugmanDavis model is
incomplete. While Helpman and Krugman (1985) focused on the case of free trade in the homogeneous-good sector, Davis (1998) mainly considered the case of equal trade costs in the two sectors.
The case of arbitrary trade costs in the homogeneous-good sector remains unclear. Therefore, before
the welfare analysis, we rigorously re-examined the equilibrium of industrial location and wage for
arbitrary trade costs in the two sectors.

Fujita and Thisse (1996) is a comprehensive review of theories of agglomeration economies.

310

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

We obtained the following results. First, we found a necessary and sufcient condition for observing the HME. The condition is given as a relationship between the trade cost of the differentiated goods
and the trade cost of the homogeneous good. This result provides a comprehensive understanding of
some known results scattered in the literature. Second, when the differentiated-good markets are
more deeply integrated, the number of differentiated-good rms in the larger country (resp. the smaller country) evolves as an inverted-U-shaped curve (resp. a U-shaped curve). Meanwhile, the smaller
country is denitely better off, whereas the larger one could be worse off. Third, when the homogeneous-good markets are more deeply integrated, the number of differentiated-good rms in the larger
country (resp. the smaller country) monotonically increases (resp. decreases). Meanwhile, the smaller
country is denitely worse off along the interior equilibrium path, whereas the welfare in the larger
country is denitely improved. In summary, if an industry is liberalized while the other is protected,
a conict between the countries might happen. Therefore, appropriate liberalization in both sectors is
effective to alleviate such a conict.
The remainder of this paper is organized as follows. Section 2 sets up the model of Helpman and
Krugman (1985) and Davis (1998). Section 3 is a detailed description of the relationship among trade
costs, rm location, and the HME. Section 4 is an analysis of the welfare, presenting the main results.
Section 5 is the conclusion.
2. The model
The economy consists of two countries (the north N and the south S), two sectors (modern M and
traditional A),2 and one production factor (labor). The amount of labor in country N is denoted as L, and
its counterpart in country S is denoted with an asterisk. The worldwide endowment Lw = L + L is xed.
The share of labor in country N (L/Lw) is denoted as h. We assume that country N is larger: h 2 (1/2, 1).
Sector M consists of a continuum of product varieties and is characterized by increasing returns to
scale (IRS) and monopolistic competition, whereas sector A produces a homogeneous good under constant returns to scale (CRS) and perfect competition.
Workers are assumed to hold the same preference described by a CobbDouglas utility for the two
types of goods with a constant elasticity of substitution (CES) subutility on the varieties of good M:

U M l A1l ;

where

M

"Z

nw

mi

r1
r

#rr1
di

nw is the number of varieties in sector M, and m(i) is the consumption of variety i. Parameter r > 1 is
the elasticity of substitution between any two varieties of good M, and l 2 (0, 1) is the expenditure
share on good M.
As in most related papers, we assume Samuelsons iceberg trade costs. Specically, sM (resp. sA)
units of the good M (resp. good A) must be shipped for one unit to reach the other country, where
sM 2 (1,1) and sA P 1 are assumed in the paper. The concept of trade costs here is rather general.
In addition to transport costs, it contains adaptation costs if exporters have to adjust their goods to
the foreign market because of country-specic standards. Furthermore, as in Baldwin and RobertNicoud (2000, p. 770), sM and sA also include tariffs.3
Labor is the only production factor. Each worker owns one unit of labor, which is immobile across
countries. In the production of good M, each rm needs a marginal cost of (r  1)/r units of labor and
a xed cost of f units of labor. Meanwhile, in the production of good A, one unit of labor produces one
unit of good A. Following Helpman and Krugman (1985) and Davis (1998), we assume that the con2

Many authors call them manufacturing sector and agricultural sector, respectively.
However, even if such an interpretation is applied, in the following analysis, we assume that tariff revenue plays a negligible
role in national welfare calculations, as is the case of all nations of the Organization for Economic Co-operation and Development
(OECD) (Baldwin and Robert-Nicoud, 2000, p. 775).
3

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

311

sumption share of good A is sufciently large for both countries to constantly produce good A. We normalize the wage in S as w = 1 and denote the wage in N as w. Then, the prices of good A in countries N
and S are

pA w 1;

pA w;

respectively. Note that the wage is the only income of the workers. Therefore, the total expenditures in
N and S are

E L  ;

E Lw;

respectively. Meanwhile, the total costs of producing x units of each variety of good M in N and S are
c(x) = fw + (r  1) wx/r and c(x) = f + (r  1)x/r, respectively.
Let p be the price of a variety of good M in country N made in country N, p be the price of a variety
 be the price of a variety in country N made in country S, and p
 be the
in country S made in country S, p
4
price of a variety in country S made in country N. Then, the monopolistic competition framework of
Dixit and Stiglitz (1977) suggests that

 sM ; p
 wsM :
p 1; p

p w;

From (1), the demand (including iceberg costs) of each variety produced in N is

dM l

pr
P

1r

E sM l

 r
p
P 1r

E ;

where P and P are the price indices of good M in the two countries, respectively. They are dened by
1

1r 1r ;
P np1r n p

 1r n p 1r 1r ;

P np

where n and n are the numbers of rms in the two countries, respectively. On the other hand, from (1)
to (3), the demands of good A in the two countries are

dA

1  lE
;
pA

dA 1  lE ;

respectively.
In the model, free entry and exit of rms are assumed so that rms have zero prot. Therefore, the
output and input of each rm in the two countries are

x x f r ;


l l f r;

8
9

respectively. Thus, from (3) to (6) and (8), the market-clearing conditions for each variety of good M
produced in N and S are


hLw w
1  hLw /M
f r;

nw1r n /M n n/M w1r



1  hLw
hLw w/M
f r;
l 

1
r
n n/M w
nw1r n /M

lwr

10
11

r is the trade freeness of good M.

where /M  s1
M

3. Trade costs, rm location, and the home market effect

In this section, we examine the equilibrium rm location. Following Krugman (1995, p. 1261) and
Davis (1998, p. 1271), we formally apply the following denition in this paper:
4

Because of symmetry among varieties, this price is independent of the variety name.

312

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

Denition 1. The home market effect (HME) is the phenomenon that the larger country is a net
exporter of good M.
While the above denition is based on the trade pattern, many papers on the HME employ another
denition focusing on rm location: The larger country attracts more than proportionate share of
rms in the IRS sector. The following Lemma says that they are equivalent.
Lemma 1. The HME occurs iff

n
> h:
n n

12

Proof. Denote the labor input in sector M in N and S as LM and LM . Since the labor input of each rm in
any country is the same given by (9), Eq. (12) is equivalent to

LM >

h
L :
1h M

From the assumption L = hL/(1  h), the above inequality is rewritten as

L  LM <


h  
L  LM ;
1h

which implies that the share of labor input in sector A (=the share of good A produced) in N is less than
h. On the other hand, according to Eqs. (2), (3) and (7),
1lwhLw

dA
h
w
;


dA 1  l1  hLw 1  h
which implies that the share of good A consumed in N is h. Thus, we conclude that (12) is equivalent to
the fact that country N is an importer of good A. In other words, by the trade balance, this suggests that
country N is a net exporter of good M. h
If good A is not traded between the two countries, then the trade in good M is balanced and the
HME disappears. Meanwhile, the total output of good A is (1  l)L in country N and (1  l)L in country S. Therefore, in this case, (9) gives

lL lhLw
lL l1  hLw

; n

:
fr
fr
fr
fr

13

Substituting (13) into (11), we obtain

Fw  w1r  w/M h  wr  /M 1  h 0

14

after simplication. Therefore, the equilibrium wage is determined by (14) when A is not traded.
Clearly, F(w) decreases in w, and it holds that

F1 1  /M 2h  1 > 0;


 r1 
1
 /M 1  h < 0;
F sMr 
/M
where the inequalities are from h 2 (1/2, 1) and /M < 1. Thus, (14) has a unique solution that lies in

r1 
~ On the other hand, when A is traded, Appendix A shows that it
1; sMr and will be denoted by w.
is impossible for country S to be the importer of good A. In other words, country N necessarily imports
~ is actually the highest value of trade costs for good A to be traded.
A so that w = sA. Therefore, w
~ indicating the fact that good A is nontradable if
~A to denote w,
Accordingly, we occasionally use s
~A .
and only if sA P s

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

313

~A :
We have the following results on the HME and s
Proposition 1.
~A ;
(i) Good A is tradable if and only if sA < s
~A ; otherwise, good A is not traded and trade in good M is balanced;
(ii) the HME is observed if sA < s
(iii) it holds that

~A
@s
> 0;
@ sM


2r11
~A
~A
@s
@s
h
~A
< 0; lim s
:
> 0;
sM !1
1h
@h
@r

Proof. See Appendix A.

As shown in Fig. 1, the above result is helpful to comprehensively understand some known results
scattered within the literature. Typically, good A is tradable when sA = 1. Helpman and Krugman
(1985) examined the rms locations in this case and found the HME. Proposition 1 (ii) generalizes
~A ). Since
their result and shows that the HME is observed in the whole shaded area of Fig. 1 (i.e., sA < s
s~A < sM , the HME disappears when sA = sM. This special result was originally provided in Davis (1998),
~A . Although Davis
and the above result demonstrates that the HME generally disappears for all sA P s
(1998) tried generalization in his Section 3C, he did not obtain a necessary and sufcient condition. Yu
r1

(2005, p. 261) showed that good A is not traded if sA P sMr , which is only a sufcient condition. Crozet
r1

and Trionfetti (2008, p. 313) concluded that the sufcient condition for the HME to exist is sA < sMr in
our notations. However, their result was based on a different denition of the HME: the situation in
which there exists a 
h 2 1=2; 1 such that dn/dh > 1 holds for all h > 
h, which is neither sufcient
nor necessary for the HME.
In contrast, we found a necessary and sufcient condition for observing the HME. Thus, we could
investigate how some parameters affect the region of sA where the HME appears. Proposition 1 (iii)
~A depends on relative parameters. The rst and the second
provides four results showing how s
inequalities show that a larger trade cost of A is necessary to obscure the HME when sM or h is larger.
This is because rms save more trade costs by locating in the larger country in such a situation. Meanwhile, a higher elasticity of substitution suggests smaller price cost markups and smaller economies of
scale in equilibrium and hence, it works against agglomeration of rms. This is captured by the third
~A .
inequality. Finally, the equality gives an upper bound of s

Fig. 1. The existence of the HME.

314

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

Next, we consider how rms relocate when either sM or sA falls. We have the following result:
Proposition 2. At the interior equilibrium with tradable A,
(i) the rm number in the larger country (resp. the smaller country) monotonically increases (resp.
decreases) when sA falls;
(ii) the rm number in the larger country (resp. the smaller country) evolves as an inverted-U-shaped
curve (resp. a U-shaped curve) when sM falls.

Proof. See Appendix B.

To understand Proposition 2 (i), we note that the relative wage in N increases in sA as long as A is
tradable, since it holds that w = sA. The wage differential has two effects. On the one hand, it has an
impact on the production side. As rms pay the wages as production costs, more rms are attracted
from S to N if w or sA falls. On the other hand, it also has an impact on the demand side. When w falls,
the deducted wage income in N shrinks the market size of good M so that more rms of M are likely to
move out from the market to save transport costs. Proposition 2 (i) shows that the former productioncost effect denitely dominates the latter income effect in our setup. Therefore, the rm number in N
(resp. S) monotonically increases (resp. decreases) for a falling sA. Such a change is shown by vector (I)
in Fig. 1.
Helpman and Krugman (1985) concluded that a small country is de-industrialized when the good
M markets are more integrated. Proposition 2 (ii) shows that their result is not valid when the trade
costs of good A are positive. Specically, there is a re-dispersion process whereby rms return to the
small country for a sufciently small sM. This is because the dispersion force of a higher wage in the
larger country dominates the agglomeration force due to the market size.5 Such a change is shown by
vector (II) in Fig. 1. In summary, the argument of Helpman and Krugman (1985) is true for a falling sA
rather than sM.
4. Welfare
In this section, we focus on welfare, which is the main concern of the paper. The indirect utility of
workers in N and S is expressed as

x wl Pl ; x P l ;
respectively.6 Furthermore, from (4) and (6), the above equations are rewritten as
l

x n n /M wr1 r1 ; x

n/M
n
wr1

rl1
:

15

From the above equations, we know that the welfare in each country is determined by three factors: the trade freeness of good M (/M), the number of rms (n,n), and the wage in N (w). Clearly, if other
things are equal, the rst two factors have positive effects on the welfare in both countries. Meanwhile, the third factor has a positive effect on local welfare and a negative one on foreign welfare.
For example, a higher wage w in N implies a higher price of differentiated goods produced there,
which lowers the welfare in S. Meanwhile, in country N, a higher wage also implies a higher income,
dominating the negative effect of higher prices and leading to a higher local welfare.
First, we derive the following result for the welfare comparison:
Proposition 3. The welfare in the larger country is always higher than that in the smaller country.
5
6

Zeng and Kikuchi (2009) analytically showed this fact with a model based on Ottaviano et al. (2002).
For simplicity, a constant multiplier, ll(1  l)1l, is omitted in each equation.

315

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

~A when A is nontradable.
Proof. The previous section shows that w = sA if good A is tradable and w s
r1

In either case, we have w < sMr < sM , which implies w1r > /M > /Mw1r. Therefore, it holds that

n
w1r  /M
w1r  /M
>1>
1r
;
2

1
r
1
r
n
w 1  /M w1r
w
 /M w

16

where the rst inequality is from Proposition 1 (ii). The inequalities of (16) imply

nw1r 1  /M w1r > n w1r  /M ;

which derives x > x according to (15).

Subsequently, we examine how the welfare in each country changes when either trade cost sM or

sA decreases.
4.1. Falling sM
This subsection focuses on the decreasing sM, as illustrated by vector (II) in Fig. 1. We consider the
case of tradable A (the shaded area in Fig. 1) in Section 4.1.1 and the case of nontradable A in
Section 4.1.2.
4.1.1. The case of tradable good A
In this case, the large country imports good A from the small country (Appendix A (ii)), and thus, we
have w = pA = sA. Therefore, (15) could be rewritten as

x n n

/M
/A

rl1

x n/M /A n r1 ;

17

r is the trade freeness of good A. From the fact that w = s , (10) and (11), we have
where /A  s1
A
A

sA lLw 1  hsA r /2M  1 sA  1h/M h/A

;
fr
/A  sA /M sA  /A /M
sA /A lLw hsA /2M  1 sA  1hsA r /M 1  h
n
:
/A  sA /M sA  /A /M
fr

18
19

Note that the above equations are true only if the RHSs of (18) and (19) are nonnegative. Otherwise,7

lhsA 1  hLw
; n 0:
f rsA

20

By (17)(19), we have



r1
lhLw sA /2M 1  2srA /M
@x l

;
 r
2
@/M
f r sA  /M /A




r1
l1  hLw srA /2M 1  2/M
@x l

> 0;

2
@/M
f r sr /  1
A

21
22

at the interior equilibrium. Inequality (22) implies that x increases in /M. Furthermore, it holds that

" r1 #


hsA s2Ar  1
@
xl



2 < 0
@/M x rl1
1  h srA  /M /A

from (18) and (19). Thus, we know that x/x decreases in /M at the interior equilibrium.
On the other hand, for a corner equilibrium with n = 0, we have

Appendix A shows that the corner equilibrium with n = 0 is impossible.

23

316

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

r1

@x l
0;
@/M
r1

@x l
lhsA 1  hLw

/A > 0;
@/M
f rsA
from (17) and (20). These imply that x is independent of /M, while x must increase in /M.
Finally, from (21), we have
r1

@x l
@/M

/M 0

lhLw
> 0;
f rsrA

which implies that x increases in /M for a small /M at the interior equilibrium. However, for a large
/M, x might increase or decrease in /M. For a further examination, let

TsA 




1  h r1
1  h r1
sA  1
2sA 
sA  1 2srA  1;
h
h

24

which increases in sA because

T 0 sA

21  h
h2

sA2r3 h2r  1sA  1  hr  1 > 0:

Meanwhile, it holds that


2
1
T1  2 
< 0;
h


T

h
1h

!
1
r1

Therefore, T(sA) = 0 has a unique solution

summarize the results as follows:


2

rr1
h
 1 > 0:
1h

s]A 2 1; h=1  h1=r1 , which is illustrated in Fig. 1. We

Proposition 4. If good A is costly tradable (i.e.,

sA 2 1; s~A ),

(i) x increases in /M, and the ratio x/x decreases in /M;

(ii) at the interior equilibrium, x increases in /M when sA > s]A and has an inverted-U-shaped relationship with /M when sA < s]A ;
(iii) for a corner equilibrium, x is independent of /M.

Proof. Both (i) and (iii) are already shown in the context, so we here prove (ii). The sign of (21)


depends on the term of /2M 1  2srA /M , which decreases in /M 2 (0, 1). Thus, we need to evaluate
~ M , which is the maximum value of /M making good A tradable, given by (B.1) in Appen(21) at /M /


~ 2 1  2sr /
~
dix B. Result (ii) then follows from sgn /
h
M
A M sgnTsA .
While Proposition 2 (ii) shows that the rm number in the smaller country evolves as a U-shaped
curve when sM decreases, Proposition 4 shows that the welfare in the country monotonically increases.
This is basically because falling sM lowers the prices of imported goods and contributes to the welfare
improvement. This effect dominates the negative effect of decreasing rms in the smaller country.
Interestingly, this is not true for the larger country: the converse might be true for a small sA. Intuitively, when sA < s]A , most rms (or all rms) could agglomerate in the larger country because the dispersion force based on the wage differential is small. Nevertheless, the rm share returns to the
population share when sM becomes so small that good A becomes nontradable. The negative effect
of a rm relocation is relatively large for a small sA, which dominates the positive effect of falling
sM, so the welfare in N becomes lower.

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

317

The model of Helpman and Krugman (1985) is the case of sA = 1. In a similar manner, we obtain the
following results for their case: (i) at the interior equilibrium, both x and x increase in /M, and their
ratio x/x is independent of /M8 and (ii) for a corner equilibrium, x is independent of /M, x increases
in /M, and their ratio x/x decreases in /M. Therefore, an integration of good M markets basically improves the welfare of both countries, which is similar to the result of Baldwin and Robert-Nicoud
(2000, Section 3). This is because, in both papers, the assumption of costless trade of good A, which
equalizes wages in two countries and does not capture the U-shaped evolution of rm location, is
imposed.
4.1.2. The case of nontradable good A
In this case, rms in sector M are distributed in proportion to country size. The rm numbers in N
~ the unique solution of (14). Recall that
and S are expressed by (13) and the equilibrium wage in N is w,

r1 
r
~
~
~
w sA 2 1; sM holds and w decreases in /M (see Fig. 1and Proposition 1 (iii)).
In this case, from (15), we have

lLw
~ r1 /M ;
h 1  hw
rf
r1
lLw ~ 1r
x l
hw /M 1  h:
rf
r1

xl

25
26

~ decreases in /M.
According to (26), x increases in /M since w
On the other hand,

~ r1 /M
~
@w
dw
~ r2 /M
~ r1
r  1w
w
d/M
@/M
~ r  1 /M w
~ r  2
~ 2r 1  hr wh
~ r r  11  h  wh
w

;
~ 2 F 0 w
~
w

27

where the second equality is obtained by the implicit function theorem. Since F0 (w) < 0, the denomi~ r  2 P 0, (27) is evidently positive. Otherwise, we
nator is positive. Then, if r  11  h  wh
have

~ r  2
~ r  1 /M w
~ r r  11  h  wh
wh
~ r  1 r  11  h  wh
~ r  2 wh
~ r  11  h > 0;
> wh
r1

~ < sMr . Therefore, (27) is always positive, which implies that x inwhere the rst inequality is from w
creases in /M.
Furthermore, from (25) and (26), we have

" r1 #
r1 @ w
r1 @ w
~ r1 /
~ 1r /
@
xl
lLw 1  hx l @/M M  hx l @/M M
<0

2r1
@/M x rl1
rf
x l
since h > 1/2,x > x, and

~ 1r /M @ w
~ r1 /M
@w
>
;
@/M
@/M
~ decreases in /M. Thus, we know that x/x decreases
where the last inequality is from the fact that w
in /M.
The above results are summarized as follows:
~A ), both x and x increase in /M, and their ratio x/x
Proposition 5. If good A is nontradable (i.e., sA P s
decreases in /M.

For sA = 1, (23) becomes 0.

318

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

Fig. 2. Effects of /M on rm location and national welfare.

~ decreases in /M, while the number of rms in each country does not
Note that equilibrium wage w
change when A is nontradable. Thus, the welfare in S clearly increases, since the price of imported vari w
~ sM ) decreases. With respect to country N, the decreasing w
~ reduces the
eties in country S (i.e., p
income as well as the price index of good M. Proposition 5 concludes that the positive effect dominates
the negative one and the welfare in N also increases.
Our model is adequately general to include the model of Krugman (1980, Section II) as a special
case of l = 1 (without sector A). Thus, Proposition 5 also holds for his setup.
Fig. 2 provides simulation examples of r = 5, l = 0.4, and h = 0.7. From (24), we have s]A 1:04 in


this setting. The left side of the gure corresponds to the case with a smaller sA 1:02 < s]A , and


the right side is the case with a larger sA 1:05 > s]A . The upper graphs (a1) and (b1) show changes
of rm numbers for sector M with /M and the lower ones (a2) and (b2) show welfare changes with /M
in each case. Note that, as in Fig. 1, good A is nontradable when /M is large (i.e., sM is small). The gure
conrms Propositions 4 and 5.
4.2. Falling sA
Next, we examine the effects of
r1
l

@x
@ sA

If

sA 2 1; s~A on the welfare.9 From (17) to (19), we have



lhLw srA 1 /M 1  /2M

< 0;

2
f srA  /M

sA P s~A , good A is not traded, and, thus, decreasing sA does not change the equilibrium.

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

@x
@ sA

r1
l

319

l1  hLw srA 1 /M 1  /2M

> 0;

2
f srA /M  1



r , while x decreases in / . On the
at the interior equilibrium. In other words, x increases in /A s1
A
A

other hand, for a corner equilibrium with n = 0, we have

r1

@x l
l1  hLw

< 0;
@ sA
f rs2A
@x
@ sA

r1
l

lLw /M r1  h r  1hsA 
< 0;
f rsrA 1

from (17) and (20). They imply that both x and x increase in /A. Furthermore, it holds that

" r1 #


@
xl
@ srA 1

>0
@ sA x rl1
@ sA /M

from (15) and the fact that w = sA. Thus, we know that x/x decreases in /A.
The above results are summarized as follows:
Proposition 6. If good A is tradable (i.e.,

sA 2 1; s~A ),

(i) x increases in /A;

(ii) at the interior equilibrium, x decreases in /A, and x/x increases in /A;
(iii) for a corner equilibrium, x increases in /A, and x/x decreases in /A.

Fig. 3. Effects of /A on rm location and national welfare.

320

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

When sA falls, w decreases, and n (resp. n) increases (resp. decreases), while /M does not
change. Regarding the welfare in country N, Proposition 6 shows that the positive effect of increasing n dominates the negative effect of decreasing w. In contrast, with respect to the welfare in S,
the negative effect of decreasing n dominates the positive effect of decreasing the price of imported varieties in the country. At rst glance, the above result (ii) might be counterintuitive because increasing /A improves country S export of A. However, our model captures the rm
relocation. A large sector A in S results in a small number of rms in M and nally, decreases
the welfare in S.
Fig. 3 provides simulation examples with the same parameter values as in Fig. 2 except for /M and
/A, which conrms Proposition 6. The left side of the gure corresponds to the case with a smaller
/M(=0.2), and the right side is the case with a larger /M(=0.4). The upper graphs (a1) and (b1) show
changes of rm numbers of sector M with /A and the lower ones (a2) and (b2) show welfare changes
with /A in each case. Note that, from Proposition 1, good A is nontradable when /A is small (i.e., sA is
large). Furthermore, in the case with a larger /M, rms of sector M completely agglomerate in N for a
larger /A. This is consistent with Helpman and Krugman (1985), in which rms agglomerate in the larger country for a larger /M when /A = 1.
4.3. Discussion
For the interior equilibrium case, our results in Propositions 16 are summarized in Table 1, where
a falling sM is represented by /M"and a falling sA is represented by /A".
From the viewpoint of welfare, we nd that the impact of integrating the homogeneous-good markets is contrastive to that of integrating the differentiated-good markets. In fact, the welfare in the
smaller country is improved, and the welfare differential (in terms of welfare ratio) becomes smaller
when the differentiated-good markets are more integrated. However, the welfare in the smaller country is lowered, and the welfare differential becomes larger when the homogeneous-good markets are
more integrated. In other words, while the integration of differentiated-good markets does not threaten the smaller country, the integration of homogeneous-good markets does. This implies that increasing the trade freeness of differentiated varieties is benecial to small countries even if it drives rms
from there to relocate to a larger country. Meanwhile, although increasing /A also contributes to
decreasing the price of imported varieties of the differentiated good via decreasing the wage in the
larger country, such a positive effect is not sufciently large to dominate the negative effect of
decreasing rm share. On the other hand, the welfare in the larger country is improved when the
homogeneous-good markets are more integrated, while it might be lowered when the differentiatedgood markets are more integrated. In addition, this result shows that rm relocation is crucial for national welfare.

Table 1
Effets of globalization at the interior equilibrium.

Notes: +: Increase, : decrease, 0: no change.

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

321

Table 1 reveals that x and x may change in different directions, suggesting possible conicts
(shown by two oval boxes) in free-trade policy. More specically, integrating the differentiated-good
markets generates a conict when trade costs of both sectors are small, whereas integrating the homogeneous-good markets denitely generates a conict at the interior equilibrium. In both conict cases,
rms producing differentiated varieties go out from the country where the welfare is lowered. Meanwhile, in the rst conict case, the price of imported varieties decreases in the country, which alleviates the welfare loss and the liberalization can be benecial to the country even if rms go out. On the
contrary, it is not true in the second conict case of integrating the homogeneous-good markets. This
is one of the reasons why liberalization of agricultural products is delayed in some countries, as mentioned in the Introduction.
Consequently, liberalizing one sector may not be supported by both countries in a long-run equilibrium with rm relocation due to a possible conict. To alleviate such a conict, appropriate liberalization in both sectors is effective. The larger (resp. smaller) country is denitely better off by
liberalizing the homogeneous-good (resp. differentiated-good) sector even if the country suffers a
welfare loss from the liberalization of the other sector. In other words, when the welfare in a country
is lowered by the integration of one sector, it is possible to reduce the welfare loss in the country by
slightly liberalizing the other sector, keeping the other country better off compared to its state before
liberalization of both sectors.
Finally, our ndings are consistent with some historical facts. Following the literature (e.g., Davis, 1998; Fujita et al., 1999), we consider the differentiated and the homogeneous goods as the
manufactured and the agricultural goods, respectively. According to Kindleberger (1975, p. 33),
some political economists in Britain in the 19th Century sought free trade of agricultural goods
as a means of slowing down the development of manufacturing on the Continent. They believed
that other European countries could be diverted to invest more heavily in agriculture and to retard
the growth to manufacturing by the British repeal of the Corn Laws in 1846. This story is consistent
with our results of industrial relocation and welfare change accompanied by liberalizing the homogeneous-good sector. In fact, while the immediate impact of repealing the Corn Laws was not signicant (Grigg, 1989, p. 21), it paved the way for the agricultural depression after 1870 on the one
hand, and increased industrial activities leading to the industrial agglomeration in the UK on the
other hand.
With respect to the recent FTA issue, the conict occurs when manufacturing rms move from a
developed country to a developing country in order to save labor costs. To prevent the relocation of
rms, developed economies occasionally oppose the freer trade of manufacturing goods. For example,
the United States and Korea ratied a USKorea FTA in November 2010. However, its negotiation was
long delayed due to the controversial issues regarding automobiles. By a US request, the agreement
nally allowed for the United States to retain a 2.5 percent tariff on vehicle imports until the fth year.
This might be also consistent with our results of industrial location and national welfare.

5. Concluding remarks
This study is an examination of the effects of liberalization (i.e., falling trade costs) on industrial
location and national welfare. We use the HelpmanKrugmanDavis model with two countries, one
factor, and two industries, both of which incur trade costs. The following results were obtained.
First, we found a necessary and sufcient condition for the HME to be observed. The condition is in
regard to the trade costs of differentiated and homogeneous goods, and the result is helpful for a comprehensive understanding of some known results throughout the literature. Second, when the differentiated-good markets are more integrated, the number of differentiated-good rms in the larger
country (resp. the smaller country) evolves as an inverted-U-shaped curve (resp. a U-shaped curve).
Meanwhile, the welfare in the smaller country must be better off, while the welfare in the larger country could be worse off. Third, when the homogeneous-good markets are more integrated, the number
of differentiated-good rms in the larger country (resp. the smaller country) monotonically increases
(resp. decreases). Meanwhile, the welfare in the smaller country must be worse off at the interior equilibrium while the welfare in the larger country must be improved.

322

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

In summary, if an industry is liberalized while the other is protected, a conict between the countries might occur. Accordingly it is effective to appropriate liberalize trade in both sectors to alleviate
the conict.
Acknowledgments
The authors are very grateful to an anonymous reviewer, Carlos Bacha, Taiji Furusawa, Salvador
Gil-Pareja, Tatsuo Hatta, Ryusuke Ihara, Jota Ishikawa, Edwin L.-C. Lai, Shin-Kun Peng, Kenmei Tsubota,
and Kazuhiro Yamamoto for their useful comments. The nancial support from the Japanese Ministry
of Education, Culture, Sports, Science, and Technology and Japan Society for the Promotion of Science
through Grants-in-Aid for Scientic Research 20730183, 22330073, 24530303 for the rst author, and
24243036, 24330072, 22330073, 21243021, and the Y. C. Tang disciplinary development fund of Zhejiang University for the second author are acknowledged.
Appendix A. Proof of Proposition 1
The case of nontraded A is already considered in the text, so we analyze the case of traded A here.
First, we show that it is impossible that all rms agglomerate in country S. Otherwise, country N imports good M, which implies that country N exports good A and w = 1/sA by the trade balance. Noting
n = 0 and w = 1/sA, the market-clearing condition for good M (11) gives

n

lLw
h
:
1  h
fr
sA

The number of workers of sector A in S is



h
1  hLw  n f r 1  hLw  lLw 1  h
;

sA

from (9), and, thus, the import of good A in S is





h
h
lLw 21  h
:
1  l1  hLw  1  hLw  lLw 1  h

sA

sA

A:1

On the other hand, the export of good A from N is

hLw  1  lhLw

sA

lhLw
;
sA

which is always smaller than (A.1). Therefore, it is impossible that all rms agglomerate in country S.
~A . OtherSecond, we show that it is impossible that all rms agglomerate in country N for sA P s
wise, country N imports good A, which implies w = sA. From (10), (11), n = 0 and w = sA, the market-clearing condition for each variety of good M produced in N and the condition for no rms in S
are written as



Lw
1h
f r;
h
n
sA

w
L 1  h sA h/M
< f r;
l

n /M /A
/A

A:2
A:3

respectively. From (A.2), we obtain

lLw
1
h 1  h :
fr
sA

Substituting this into (A.3), we have

r
sA /M  s1
1  h
A

1
 sA r
/M

< 0:

A:4

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

The LHS is increasing in sA. On the other hand, when

323

sA s~A , the LHS of (A.4) is







 r

1
1
r
r
r
~
~A /M  s
~1
~
~

1

h
h s

s
s

/

s

1

h

1

h
M
A
A
A
A
/M
/M





1
1
~rA r1  r
1  h srM1  s
s~A
sM



~rA srM1 s
~rA  1
1  h srM1  s
> 0;

sMr1 s~rA

r1 
~A 0 and the last inequality is from s
~A 2 1; sMr . Therefore, the
where the rst equality is from Fs
~A , which implies that the agglomeration never occurs for sA P s
~A .
LHS of (A.4) is positive for sA P s
Next, we consider an interior equilibrium. We rst assume that the larger country N is the importer
of good A, so that pA = w = sA and E = wL = sAL hold. From (10) and (11), we have


hLw /A
1 1  hLw /M /A
f r;



n/A n /M sA n n/M /A


w
w
1  hL
sA hL /M
f r:
l 

n n/M /A n/A n /M

A:5
A:6

The above equations immediately derive

hLw
f rsA  /A /M
1  hLw
f r/A  sA /M


; 

 :
2

n/A n /M lsA /A 1  /M n n/M /A
l/A 1  /2M
Since the above two terms are positive, we obtain a necessary condition for the interior equilibrium:

/A  sA /M > 0:

A:7

From (A.5) and (A.6), the numbers of rms in two countries are expressed as (18) and (19). Then,
using these equations, we have

n  hn n

lLw /M sA  1h  1FsA 
;
f r /A  sA /M srA  /M

A:8

where F() is dened in (14). The denominator of (A.8) is positive from (A.7). Since F() is a decreasing
~A 0, we have n  h(n + n) > 0 for all sA 2 1; s
~A . According to Lemma 1, this implies
function and Fs
~A .
that the HME exists for all sA 2 1; s
~A , we have n  h(n + n) < 0. By the same logic used above, country N
On the other hand, when sA > s
must be an exporter of good A, which contradicts the assumption that N is the importer of good A.
~A .
Thus, a necessary condition for N to be the importer of good A is that sA < s
Here, we show that it is impossible for the smaller country S to be the importer of good A. Otherwise, the equilibrium wage in N is w = 1/sA < 1. Similar to the previous arguments, we have

n  hn n

s1  1 h 1 F s1
lLw /M
 1 A
 r A  > 0;
fr
sA  /M
/A  /M s1
A

r
where the inequality is from the following facts: (a) s
A > /M , which can be derived similar to (A.7);
(b) F() is decreasing; and (c) F(1) > 0 and sA > 1. Therefore, country N must be an importer of good A,
which contradicts the assumption that N is the exporter of good A.
~A . OtherSumming up the above discussion, good A is not traded and the HME disappears if sA P s
wise, the HME appears.
Finally, from (14), we have

324

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325


@F 
~ r hr  1  w
~ r1 1  hr  h/M < 0;
w
@www~

 1r
@F 
r
~
~ M < 0;
~ r  w/
w
 s1
w
M
@/M ww~

@F 
~ < 0;
1  1 wh
@hww~

@F 
~ 2r 1  h wh
~ w
~ r ln w
~ < 0;
w
@ rww~


h
;
lim Fw 1  hw1r w2r1 
/M !0
1h
r1

~ < sMr < sM (see the discussion below (14)) and the
where the second inequality is from the fact that w
~ > 1 and h > 1/2. By applying the implicit functhird and fourth inequalities are from the facts that w
~ s
~A , we obtain three inequalities in Proposition 1 (iii). The equality
tion theorem to (14) and noting w
in (iii) is true because [h/(1  h)]1/(2r1)is the only solution of lim/M !0 Fw 0. h
Appendix B. Proof of Proposition 2
r , we have
(i) Noting /A  s1
A


2 


2
@n
lLw /M srA 2 1  h srA  /M r  1 srA /M rhsA 1  srA /M

< 0;






2
@ sA
fr
1 s2Ar /M  srA 1 /2M

2

2


@n lLw /M r1  hsrA 1 srA  /M h 1  srA /M r  1srA /M

>0




2
@ sA
fr
1 s2r /  sr 1 /2
A

from (18) and (19). In other words, the rm number in the larger (resp. smaller) country monotonically increases (resp. decreases) when sA decreases.
(ii) Even if good A is traded initially, it becomes nontradable when /M increases and reaches a
threshold, as can be seen in Fig. 1. The threshold is obtained by solving F(sA) = 0 for /M, where
F() is dened by (14). Specically, the threshold is expressed as
r

~ M  sA
/

If


hsA  1  hs2Ar
:
hsA  1  h

sA < s~A , it must hold that

hsA > 1  hs2Ar ;

~ M 2 0; 1. From (18) and (19), we have

and, then, /



lLw sA r1 hsA  1  hs2Ar
@n 

> 0;
@/M /M 0
fr



lLw sA r hsA  1  hs2Ar
@n 

< 0;

fr
@/M /M 0



lLw srA 2 hsA h  12 hsA  1  hs2Ar
@n 


< 0;


2
@/M /M /~ M
f rh1  h s2Ar  1



lLw srA 1 hsA h  12 hsA  1  hs2Ar
@n 

> 0;

2
@/M /M /~ M
f rh1  h s2Ar  1
where all inequalities are from (B.2).

B:1

B:2

H. Takatsuka, D.-Z. Zeng / J. Japanese Int. Economies 26 (2012) 308325

325

Furthermore, we know that both

@n
/ 0 and
@/M M

@n
/ 0
@/M M

have at most two roots. Thus, we conclude that

/M 2 [0, 1]. h

@n
@/M

/M (resp.

@n
@/M

/M ) is concave (resp. convex) for

References
Baldwin, R., Forslid, R., Martin, P., Ottaviano, G., Robert-Nicoud, F., 2003. Economic Geography and Public Policy. Princeton
University Press, Princeton.
Baldwin, R.E., Robert-Nicoud, F., 2000. Free trade agreements without delocation. Canadian Journal of Economics 33, 766786.
Baldwin, R.E., Venables, A.J., 1995. Regional economic integration. In: Grossman, G.M., Rogoff, K. (Eds.), Handbook of
International Economics, vol. 3. Elsevier, Amsterdam, pp. 15971644.
Crozet, M., Trionfetti, F., 2008. Trade costs and the home market effect. Journal of International Economics 76, 309321.
Davis, D.R., 1998. The home market effect, trade and industrial structure. American Economic Review 88, 12641276.
Dixit, A.K., Stiglitz, J.E., 1977. Monopolistic competition and optimum product diversity. American Economic Review 67, 297
308.
Fischer, R., Serra, P., 2000. Standards and protection. Journal of International Economics 52, 377400.
Fujita, M., Krugman, P.R., Venables, A.J., 1999. The Spatial Economy: Cities, Regions, and International Trade. MIT Press,
Cambridge.
Fujita, M., Thisse, J.-F., 1996. Economics of Agglomeration. Journal of the Japanese and International Economies 10, 339378.
Gandal, N., Shy, O., 2001. Standardization policy and international trade. Journal of International Economics 53, 363383.
Grigg, D., 1989. English Agriculture: An Historical Perspective. Basil Blackwell, Oxford, UK.
Head, K., Mayer, T., Ries, J., 2002. On the pervasiveness of home market effects. Economica 69, 371390.
Head, K., Ries, J., 2001. Increasing returns versus national product differentiation as an explanation for the pattern of USCanada
trade. American Economic Review 91, 858876.
Helpman, E., Krugman, P., 1985. Market Structure and Foreign Trade. MIT Press, Cambridge, MA.
Kindleberger, C.P., 1975. The rise of free trade in Western Europe, 18201875. Journal of Economic History 35, 2055.
Krugman, P., 1980. Scale economies product, differentiation and the pattern, of trade. American Economic Review 70, 950959.
Krugman, P., 1981. Intraindustry specialization and the gains from trade. Journal of Political Economy 89, 959973.
Krugman, P., 1995. Increasing returns, imperfect competition and the positive theory of international trade. In: Grossman, G.M.,
Rogoff, K. (Eds.), Handbook of International Economics, vol. 3. Elsevier, Amsterdam, pp. 12431277.
Krugman, P.R., Venables, A.J., 1990. Integration and the competitiveness of peripheral industry. In: Bliss, C., de Macedo, J. (Eds.),
Unity with Diversity in the European Economy: The Communitys Southern Frontier. Cambridge University Press, pp. 5675.
Krugman, P.R., Venables, A.J., 1995. Globalization and the inequality of nations. Quarterly Journal of Economics 110, 857880.
Moenius, J., 2004. Information versus product adaptation: the role of standards in trade. Available at SSRN. <http://ssrn.com/
abstract=608022>.
Ottaviano, G., Tabuchi, T., Thisse, J.F., 2002. Agglomeration and trade revisited. International Economic Review 43, 409436.
Ottaviano, G., Thisse, J.-F., 2004. Agglomeration and economic geography. In: Henderson, J.V., Thisse, J.-F. (Eds.), Handbook of
Regional and Urban Economics, vol. 4. Elsevier, Amsterdam, pp. 25632608.
Swann, P., Temple, P., Shurmer, M., 1996. Standards and trade performance: the UK experience. Economic Journal 106, 1297
1313.
Yu, Z., 2005. Trade, market size, and industrial structure: revisiting the home market effect. Canadian Journal of Economics 38,
255272.
Venables, A.J., 1987. Trade and trade policy with differentiated products: a ChamberlinianRicardian model. Economic Journal
97, 717780.
Zeng, D.-Z., Kikuchi, T., 2009. The home market effect and trade costs. Japanese Economic Review 60, 253270.