You are on page 1of 30

Journal of Economics (2019) 128:147–175

https://doi.org/10.1007/s00712-019-00659-7

Indirect taxes in a cross-border shopping model:


a monopolistic competition approach

Hiroshi Aiura1 · Hikaru Ogawa2

Received: 25 January 2018 / Accepted: 11 March 2019 / Published online: 18 March 2019
© Springer-Verlag GmbH Austria, part of Springer Nature 2019

Abstract
In a model of cross-border shopping with monopolistic competition, we examine the
relative merits of an ad valorem (ADV) tax and a unit (specific) tax. Our focus is
on the effects of the opening of borders that enables consumption outside the coun-
try of the residence. The result shows that the increased cross-border shopping may
strengthen or weaken the superiority of either of the two tax methods, depending on the
relative weight that the government places on tax revenue and welfare. Specifically,
while cross-border shopping always increases the welfare of all countries in ADV
tax competition, welfare increases only when the governments place a large weight
on tax revenue in unit tax competition. Cross-border shopping lowers welfare when
governments with high weight on welfare compete in unit tax.

Keywords Monopolistic competition · Spatial tax competition · Unit and ad valorem


taxes

JEL Classification H2 · H7

1 Introduction

Traditional tax analyses suggest that an ad valorem (ADV) tax is welfare superior to
a unit (specific) tax in a monopoly (Suits and Musgrave 1953). An ADV tax reduces
monopoly power in a price-setting environment and, thus, transfers more revenue from
the monopoly to the government. Since the early research, scholars have reexamined
and substantiated this argument, focusing mainly on monopoly and oligopoly market

B Hiroshi Aiura
office@aiura.info

1 Faculty of Economics, Nanzan University, 18 Yamazato-cho, Showa-ku, Nagoya 466-8673,


Japan
2 Graduate School of Economics and Graduate School of Public Policy, University of Tokyo, Tokyo,
Japan

123
148 H. Aiura, H. Ogawa

structures.1 Among them, two intriguing studies, Anderson et al. (2001) and Schröder
(2004), challenge the view on the superiority of ADV tax over unit tax from the
long-run perspective. Anderson et al. (2001) compare the two tax methods under the
standard Cournot–Nash oligopoly model with free entry of firms and find that the short-
run superiority of ADV taxes in extracting firms’ profit reduces firms’ incentives to
enter and hence reduces welfare in the long-run. This negative aspect of ADV tax was
more clearly modeled by Schröder (2004), which applies the love-of-variety model of
Dixit and Stiglitz (1977). It models the negative aspect of ADV tax as it reduces the
variety of products by decreasing the firms’ incentives to enter. Yet it shows that ADV
tax is superior to unit tax since the tax over-shift effect outweighs the effect of loss of
product variety under ADV taxation.2
Our aim in this study is to contribute to long-run analyses of tax comparison between
ADV tax and unit tax. A particular feature of our paper is to allow consumers to
cross the border for shopping. For this, we construct a two-country model in which
consumers cross a border to buy a set of differentiated products provided by monop-
olistically competitive firms. Prior works using a single country model have made an
implicit assumption that consumers can buy goods only from the country in which
they live no matter how expensive they are. Of course, purchase of consumer goods
and firms’ procurement of intermediate goods are not always completed in their own
country. In particular, along with the increased mobility of consumers and tourists
accompanied by the decrease in shipping and transportation costs, cross-border shop-
ping and procurement has become common.3 Cross-border shopping is also important
for policy in practice. For instance, Manuszak and Moul (2009) have estimated that
the area of Chicago, the territory with the highest gasoline taxes, loses approximately
40% of the tax-raising capacity that would exist if taxes were equal throughout the
region.
The presence of cross-border shopping leads directly to standard interregional tax-
cutting competition to attract consumers. In addition, because differentiated products
are supplied in each country and consumers have love-of-variety preferences, the
number of firms and thus the variety of products emerges as a source that attracts
consumers. Compared with unit taxation, ADV tax lowers after-tax prices but reduces
product variety. While the former attracts consumers from abroad, the latter could be
the cause of the customer’s outflow. When the two contradictory effects are added,

1 See Keen (1998) for a general review of these studies. Recently, the application of tax analysis to the
newly emerging market structure has also been implemented. For instance, Kind et al. (2009) analyzed a
two-sided market, while Hamilton (2009) and Lapan and Hennessy (2011) both analyzed a multiproduct
market.
2 Since Schröder (2004), several papers have used monopolistic competitive models to compare unit tax
and ADV tax. Dröge and Schröder (2009) compare the two tax methods in the presence of environmental
externalities and show that unit tax leads to more firms in the industry, less tax revenue, but higher welfare
compared with ad valorem taxes. Vetter (2013) points out the possibility that the result of Schröder (2004)
depends on the functional form.
3 Leal et al. (2010) states that the importance of cross-border shopping increases as two trends in the world
economy intensify. The first is the globalization of the economy, which substantially reduces the transaction
costs associated with trade relations between territories. The second is the decentralization of tax-raising
powers towards jurisdictional government even within the same country, which significantly increases the
number of jurisdictions with different taxation on the same goods or services.

123
Indirect taxes in a cross-border shopping model… 149

does cross-border shopping strengthen the advantage of ADV tax or work to the
advantage of unit tax? Departing from the single country model of Anderson et al.
(2001) and Schröder (2004), this paper examines the advantage of ADV tax from
a long-term perspective focusing on the effects of cross-border shopping. We use
a love-of-variety model of monopolistic competition when considering cross-border
shopping, following Schröder (2004). This is motivated by the literature on consumer
research which has revealed that after-tax price and variety of goods are the dominant
factors for cross-border shopping (Fitzgerald et al. 1988; Mittelstaedt and Stassen
1990; Timothy and Butler 1995; Wakefield and Baker 1998; Tömöri 2010; Bygvra
2011).4 These studies reveal that the attractiveness of products through lower prices
and variety of products and services together constitute the “total attraction” of a
country, and the cross-border shopping model with love-of-variety is best suited to
describe the situation.
Our model allows for cross-border shopping based on the work of Aiura and Ogawa
(2013), who analyze the effects of a unit tax and an ADV tax using the cross-border
shopping model. They analyze a situation where there is one firm in each country
and the goods supplied by each firm are homogeneous. In other words, their research
focuses on short-run equilibrium in the sense that the number of firms is fixed. In
addition, it is an analysis reflecting only a part of the observations in the sense that
cheaper price is the only factor that led consumers to choose to shop across the border.
We analyze the tax effects in models that include firms’ entry and preference for
product variety. This enables us to clarify how the superiority of ADV taxes differs
in the presence and absence of cross-border shopping when the conflicting effects of
attracting consumers at low prices, while losing them by reducing the variety of goods
are associated with ADV tax.
Our study is also related to the literature on two-country analyses of tax policies.
Lockwood and Wong (2000) and Jorgensen and Schröder (2005) conduct comparisons
between ADV tariffs and unit tariffs in the international trade model. Lockwood (2004)
and Akai et al. (2011) show that unit tax is superior to ADV tax in the tax competition
model with mobile capital. Schröder and Sørensen (2010) confirm the dominance
of ADV tax over unit tax in a model allowing firm heterogeneity and intra-industry
reallocation. Takatsuka (2014) also compares the effects of ADV and unit taxes on firm
location in the framework of new trade theory. Akai et al. (2014) show the advantage of
unit subsidy in the model in which two competing countries export to a third country.
In a recent paper, Liang et al. (2018) also use an export model with a third country
to show that superiority of one tax method depends on production technology. All
of these studies have the same purpose as our study in that they analyze the effects
4 Fitzgerald et al. (1988) have found that, in the Republic of Ireland, the main perceived advantage for
crossing the border to shop is the variety of items, following the price difference. Mittelstaedt and Stassen
(1990) show that product variety, both within an individual store and across competing stores, is likely to
attract consumers to shop for specific product purchases across stores in different locations. Timothy and
Butler (1995) and Tömöri (2010) have shown that cheaper price and variety of goods are factors that led
shoppers to choose to shop across the border in Canada and Hungary, respectively. Wakefield and Baker
(1998) found that the variety in the shopping mall has the strongest effect on the desire to stay when
competing for mobile customers. Bygvra (2011) shows that the cross-border shopping done by Germans
is not caused by price differences to the same extent, but is mainly the result of differences in the range of
products available on the other side of the border.

123
150 H. Aiura, H. Ogawa

of ADV and unit taxes. However, our concern lies not in the trade, capital mobility,
and location choice of firms, but in the factors that lead shoppers to choose to shop in
bordering countries.
The main argument of our study is that the relative merit in terms of welfare of using
either ADV tax or unit tax critically depends on the presence of cross-border shopping
and government objectives. When the government behaves as if it is a Leviathan and
puts a higher weight on revenue than on the residents’ utilities, the increased mobility
of consumers across the border improves efficiency if the governments compete in
unit tax. In this case, the consumers’ mobility is a deterrent to Leviathan behavior
and improves the welfare in all countries. However, if the governments are benevolent
and thus attach a high weight to domestic welfare, the increased consumer mobility
for shopping lowers the welfare in the unit tax competition. By contrast, when gov-
ernments compete in ADV tax, cross-border shopping always increases the welfare
in all countries, without depending on the government objectives. This presents the
superiority of ADV tax in the sense that the country using ADV tax can avoid a decline
in domestic welfare, whatever objective the government has, even if the competition
for mobile consumers becomes intense.
The remainder of this paper is organized as follows. In Sect. 2, we set up the
basic model. Section 3 presents the baseline model in which there is no cross border
shopping. Section 4 studies the model of cross border shopping and discusses how the
equilibrium is modified in the presence of consumer mobility. Section 5 summarizes
the effects of increased mobility of consumers on tax revenue, consumers’ utility, and
domestic welfare. Finally, Sect. 6 concludes the paper.

2 Model

2.1 Technology and preferences

A line economy. The basic Hotelling model with differentiated products follows the
work of Henkel et al. (2000).5 A line economy consists of two symmetric countries,
j = 1, 2. The location space of the economy is given by γ ∈ [−1, 1] divided into two
countries at γ = 0, the length of each country. Therefore, γ = 0 is the exact middle of
this economy, so that consumers to its left are residents in country 1 while consumers to
its right are residents in country 2. A continuum of consumers is uniformly distributed
along the line. We assume that consumers cross the border for shopping but do not
change their place of residence. This is for simplicity of the analysis, but it is justified by
assuming the existence of language and cultural barriers, strong attachment to home,
and other reasons. In each country, there is a retail product market of monopolistic
competition at the end of the line.

5 In a line economy in which possible market place locations are restricted to the end of the line, Henkel
et al. (2000) study the conditions under which producers of substitutes cluster in market places in order
to attract consumers dispersed along the line. Spatial externalities in their model lead the equilibrium to
be sub-optimal, and they then analyze the role of coalitions of firms to improve the sub-optimal spatial
allocation.

123
Indirect taxes in a cross-border shopping model… 151

The retail product sector (s = M) consists of a continuum of firms producing


differentiated products under an increasing-returns-to-scale (IRS) technology in a
monopolistic competition market. The firms in the retail product sector are fixed at
their locations. Consumers shop for a set of differentiated retail products at either
market and bear the transportation cost (i.e., traveling expenses) to the market.

Technology. In addition to the monopolistically competitive retail sector, each country


has a standard homogeneous goods sector (s = H ) which produces a homogeneous
product under a constant-returns-to-scale technology in a perfectly competitive market.
Firms in this sector produce one unit of a homogeneous good from one unit of labor
in every location. Therefore, in the homogeneous goods sector, goods are priced at
the wage rate, and shipment of the goods incurs no transportation costs. Because we
assume that the homogeneous goods are the numeraire, the wage rate is unity.
In the retail product sector, because of the presence of fixed costs, firms produce
products with an IRS technology under monopolistic competition. Firms that produce
any differentiated goods need f units of labor without depending on the amount of
production, which represents the fixed investment necessary to develop a new variety
of goods. In addition, they need c units of labor to produce one unit of product, which
corresponds to the variable cost. Because each firm produces one variety, the number
of firms coincides with the number of varieties. That is, entry of a new firm introduces a
new product, which increases the product variety. In addition, new entrants will lower
long-run firm profits to zero. The number of varieties produced in country j is denoted
by the continuous variable n j . The variety i produced in country 1 is homogeneous
with the variety i produced in country 2. This implies that consumers are able to
consume a wider variety of products in country j if n j ≥ n k (k = j).

Preferences and budget constraint. Preferences are assumed to be identical across


consumers and are characterized by a quasi-linear utility function over homogeneous
goods produced by sector H and differentiated goods produced by retail product sector
M. The utility function is identical for all consumers located at any point x ∈ [0, 1] and
is given as U (x) = H + μ ln M, where M is the sub-utility derived from consuming
varieties of differentiated products and H is the consumption of homogeneous goods.6
If a consumer goes to country j to consume a set of differentiated goods, the sub-utility
M is given by
 n j σ/(σ −1)
(σ −1)/σ
M = Mj = m j (i) di , (1)
0

where m j (i) is the amount of consumption of variety i in country j and σ > 1


is the elasticity of substitution between any two products. Residents supply labor
inelastically in one of the two sectors in the country of residence. Since labor is
freely mobile between two sectors, wages are the same whether they work in either
sector. The budget constraint of a consumer living at point x ∈ [−1, 1] who buys the
differentiated products from the firms in country j is given as follows:

6 As with many studies, this formulation does not include income effects, but this makes it possible to
obtain analytical solutions.

123
152 H. Aiura, H. Ogawa

 nj
I =H+ p j (i)m j (i)di + F j (x), (2)
0

where F1 (x) = τ (1 + x) and F2 (x) = τ (1 − x).7 Here, τ (> 0) denotes the transporta-
tion cost per unit distance. To ensure H ≥ 0, τ + μ ≤ 1 is assumed. In (2), I is the
wage income which is equal to 1, F j (x) is the transportation cost when traveling to the
market in country j, and p j (i) stands for the price of variety i supplied in country j.
Because variety i in one country is homogeneous with variety i supplied in the other
country and consumers bear the transportation cost of traveling to these markets, they
buy the differentiated products from either country 1 or country 2.
Government. In each country, there is a single government that raises revenue only
through taxes on differentiated products. The government uses either the unit or ADV
tax method and chooses the tax level. If the government adopts the unit tax method,
taxes are imposed on the number of units sold. If it selects the ADV tax method, taxes
are imposed on the amount of sales.

2.2 Utility maximization

Consider first that a consumer living at point x decides to buy differentiated products
from country j. Then, solving the utility-maximization problem of the consumer, we
obtain the following demand function:

 n 1/(1−σ )
p j (i)−σ j
m j (i) = μ and H = I − μ − F j (x), where P j = p j (i)(1−σ ) di .
P j1−σ 0
(3)
Using (3), the indirect utility function of an individual located at point x who goes to
country j to shop for differentiated products can be obtained as

u j (x) = I − μ + μ ln M j − F j (x).

When u 1 (x ∗ ) = u 2 (x ∗ ), the consumer who is located at x ∗ is indifferent to consuming


in one country or the other. In this case, to buy differentiated products, all consumers
located to the left (right) of x ∗ would go to country 1 (2).
Hence, the number of consumers who buy differentiated products from the market
in country j, l j , is given by

μ
l j (M j , Mk ) = 1 + (ln M j − ln Mk ), where j = 1, 2, j = k. (4)

Because M j includes the number of firms and prices, mobile consumers are attracted
by the product varieties and lower prices. This captures the facts mentioned in the first

7 For example, consider a consumer in country 1 who lives at point x = −0.5. If he or she goes to the
n
market in country 1, (2) becomes I = H + 0 1 p1 (i)m 1 (i)di + τ (1 − 0.5), while if he or she goes to the
 n2
market in country 2, it becomes I = H + 0 p2 (i)m 2 (i)di + τ (1 + 0.5).

123
Indirect taxes in a cross-border shopping model… 153

section that cross-border consumers are attracted by lower price and variety of goods
in deciding the place of consumption.

2.3 Firms

Denoting the ADV tax rate of country j as r j and the unit tax rate as t j , each firm
maximizes its profit:

π j = [(1 − r j ) p j (i) − (c + t j )]m j (i)l j (M j , Mk ) − f , (5)

where m j (i) is given by (3). Note that in (5) it is reflected that the wage rate is unity.
Under monopolistic competition, each firm takes the prices charged by its rivals as
given and ignores the effect of its price change on the market variables. Therefore,
∂ P j /∂ p j (i) = 0 and ∂ M j /∂ p j (i) = 0. Hence, profit maximization yields

σ (c + t j )
p j (i) = p j = . (6)
(σ − 1)(1 − r j )

Substituting (6) into (5), the firm profit in country j is

(1 − r j )μ
πj = lj − f . (7)
n jσ

In (7), from (1), (3), (4), and (6), l j is given by

μ[ln(1 − r j ) − ln(1 − rk )]
l j (r j , rk , t j , tk , n j , n k ) = 1 +

μ[ln(c + t j ) − ln(c + tk )] μ(ln n j − ln n k )
− + . (8)
2τ 2τ (σ − 1)

The second and third terms in (8) capture consumers’ mobility because of the inter-
regional price gap that originates from the tax gap between the two countries. The
fourth term reflects the fact that consumers are attracted to the country with a greater
variety of differentiated products. These three terms become more important as trans-
portation costs decrease. In addition, the fourth term is more important in the economy
in which σ is small, as consumers are increasingly attracted to the country with a greater
variety when the elasticity of substitution between any differentiated products is small.
Because the long-run profit for monopolistic competitors is zero owing to free
entry and exit, the long-run equilibrium condition satisfies π j = 0. Therefore, the
equilibrium number of firms (varieties) is

(1 − r j )μ
nj = lj. (9)
σf

123
154 H. Aiura, H. Ogawa

Using (3), (6), and (9), the demand for variety i in country j is given by

μ (σ − 1) f
m j (i) = m j = = . (10)
n j pj (c + t j )l j

The equilibrium values of l j and n j satisfy (8) and (9), which are implicitly
obtained as a function of tax rates, l j (r j , rk , t j , tk ) and n j (r j , rk , t j , tk ). Substitut-
ing l j (r j , rk , t j , tk ) into (10), we implicitly obtain the equilibrium consumption of
variety i in country j, m j (i), as the function of the tax rates. The explicit solu-
tions are obtained by inserting the equilibrium tax rates shown in the subsequent
analysis.

2.4 Governments

The government in each country is assumed to behave as a moderate Leviathan having


the objective function given by:

V j (r j , rk , t j , tk ) = θ R j (r j , rk , t j , tk ) + (1 − θ )W j (r j , rk , t j , tk ), (11)

where R j (r j , rk , t j , tk ) and W j (r j , rk , t j , tk ) are tax revenue and domestic welfare,


explained in detail below. θ ∈ [0, 1] is a weight parameter attached to tax revenue.
This approach enables us to characterize both the perfectly benevolent and purely
Leviathan government by assuming θ = 0 and θ = 1, respectively (Edwards and
Keen 1996; Wrede 1998; Pal and Sharma 2013; Kawachi et al. 2018).
Using (6) and (10), the tax revenue is given by
 
σ − 1 (1 − r j )t j
R j = (t j + r j p j )m i n j l j = μ · + rj lj. (12)
σ c + tj

Note that l j is a function of r j , rk , t j , and tk . Since the long-run profit for monopolistic
competitors is zero, domestic welfare consists of consumers’ utility and tax revenue,
and is given by

Wj = max{u 1 (x), u 2 (x)}d x + R j ,
x∈L j

where L j is the location space of country j. More specifically, the domestic welfare
of country 1, for instance, W1 , is described in the following two equations according
to the value of l1 :
 0
W1+ = u 1 (x)d x + R1
−1
 0
= I − μ + μ ln M1 − F1 (x)d x + R1 if l1 ≥ 1, (13)
−1

123
Indirect taxes in a cross-border shopping model… 155

 −1+l1  0
W1− = u 1 (x)d x + u 2 (x)d x + R1
−1 −1+l1
 −1+l1
= I − μ + l1 μ ln M1 − F1 (x)d x + (1 − l1 )μ ln M2
−1
 0
− F2 (x)d x + R1 if l1 < 1. (14)
−1+l1

(13) represents the domestic welfare in country 1 when all the residents consume the
products in their own country. (14) shows the welfare of country 1 when the residents
who live far away from the border consume in country 1, but residents living near the
border cross the border and go to country 2 to consume differentiated products.
Substituting (10) into (1), M j in (13) and (14) can be obtained as
1
M j = (n j m j )n σj −1
μ σ −1
1
= nj , (15)
pj

where p j and n j are given by (6) and (9), respectively. (15) implies that the sub-utility
depends on the price and the number of firms in the equilibrium.
In the next section, we analyze the equilibrium characteristics when both govern-
ments maximize their objective function using either a unit tax or an ADV tax. Since
the two countries are symmetric, l1 = l2 = 1 will hold in the equilibrium, indicating
that it is reasonable to insert (13) into (11) to obtain the objective function of the
government. That is, the government in country i maximizes, with respect to either ti
or ri ,

V j (r j , rk , t j , tk ) = θ R j + (1 − θ )W +
j

= R j + (1 − θ ) u j (x)d x
x∈L j
= R j + (1 − θ )[v̄ + v(r j , rk , t j , tk )], (16)

where

μ (σ − 1)σ −1 μσ
v̄ = I − μ − F j (x)d x + ln
x∈L j σ −1 σσ f
σ
(1 − r j ) σ −1 σ −1
1
and v(r j , rk , t j , tk ) = μ ln lj .
c + tj

v̄ represents a part of consumers’ utility independent of tax rates and v(r j , rk , t j , tk )


indicates the utility that depends on them. To simplify the notation, let
 σ
(1 − r j ) σ −1 σ −1
1
Uj ≡ u j (x)d x = v̄ + μ ln lj . (17)
x∈L j c + tj

123
156 H. Aiura, H. Ogawa

2.5 First-best optimum

Before describing the choice of government policies, we present the optimal quantity
and number of differentiated products that maximize social welfare. Remembering
that the marginal cost of supplying differentiated products is c, from (1), the socially
optimal solution satisfies ∂U (x)/∂m j (i) = c, where
  n j σ/(σ −1)
∂U (x) ∂
= H + μ ln m j (i)(σ −1)/σ di
∂m j (i) ∂m j (i) 0
  nj −1
(σ −1)/σ
= μ m j (i) 1/σ
m j (i) di .
0

The optimal quantity of differentiated product is, thus,

μ
m j (i) O P T = m Oj P T = . (18)
cn j

The superscript O P T denotes the optimal value from the perspective of welfare max-
imization. Substituting (18) into (1), we have

1
M = M j = n σj −1 (n j m Oj P T )
μ σ −1
1
= nj ,
c

which yields

dU (x)

μ
= .
dn j
m j (i)=m O P T (σ − 1)n j
j

As n j m Oj P T = μ/c, the total output level of differentiated goods is constant as long


as each firm produces its product at the optimum level. In this case, the social cost
required for allowing one firm to enter corresponds to the fixed cost, f . Therefore,
the number of firms producing differentiated products that maximize social welfare
satisfies μ/[(σ − 1)n j ] = f , suggesting that the optimal number of firms, n Oj P T , is
given by
μ
n Oj P T = . (19)
f (σ − 1)

3 Equilibrium without cross border shopping

Section 3 presents a benchmark model in which cross-border shopping is not allowed,


which is interpreted as a closed-border economy. In Sect. 4, we present an analysis of
open border economy where consumers can cross the border for their shopping. By

123
Indirect taxes in a cross-border shopping model… 157

comparing the results obtained in these two sections, we examine the effects of cross
border shopping in Sect. 5.

3.1 Unit tax competition

When two governments compete in unit taxes, the first-order condition for the maxi-
mization of (16) is given by


d V jU

1 σ −1 c

= μ −(1 − θ ) + lj
dt j
c + tj σ (c + t j )2
r j =0
 
σ − 1 tj 1−θ dl j
+ + = 0, (20)
σ c + tj (σ − 1)l j dt j

where superscript U denotes that two governments compete in unit taxes. In the
closed-border economy, every consumer buys differentiated products from the domes-
tic markets. This means that dl j /dt j = 0, dl j /dr j = 0, and l1 = l2 = 1 hold.
Therefore, the unit tax rate of each country in the symmetric equilibrium is obtained
as follows:8
θσ − 1
t U C = t1 = t2 = c, (21)
(1 − θ )σ
where superscript C denotes the equilibrium value in the closed-border economy.
Because ∂t U C /∂θ > 0, the government reduces the tax rate as it places more weight
on domestic welfare in its objective. Specifically, the government provides a subsidy
to firms if the government is sufficiently benevolent to satisfy θ < 1/σ . The reason for
this is simple. By providing subsidies to firms, the sufficiently benevolent government
induces them to produce more because the imperfectly competitive firms produce
less output and set their price above the marginal cost. Substituting (21) into (6) with
r j = 0, we obtain pU C = c/(1 − θ ). The firm sets its price above the marginal cost
based on its market power except for θ = 0. When θ = 0, in which the governments
are benevolent and aim to maximize domestic welfare, the price is equal to the marginal
cost.
Substituting (21) and l j = 1 into (12) and (17), we obtain the tax revenue and the
sum of the consumers’ utility of each country, respectively, as:9

θσ − 1
RU C = μ, (22)
σ
(1 − θ )σ
UUC = v̄ + μ ln − μ ln c. (23)
σ −1

8 Because the two reaction curves derived from the first-order condition for i = 1, 2 are continual under
0 ≤ l j ≤ 1 and symmetric with respect to a 45◦ (t1 = t2 ) line, there is a unique symmetric equilibrium.
The same applies for other cases to be analyzed.
9 The same notation applies in the following analysis: U yz ≡
 yz
x∈L j U (x)d x, where y = U , A and
z = O, C.

123
158 H. Aiura, H. Ogawa

(22) and (23) show that tax revenue increases and consumers’ utility decreases as the
government places more weight on tax revenue, ∂ R U C /∂θ > 0 and ∂U U C /∂θ < 0.

3.2 ADV tax competition

When two governments compete in ADV taxes, the first-order condition for the max-
imization of (16) is given by

 
d V jA

σ 1 1−θ dl j

= μ −(1 − θ ) + lj + rj + = 0,
dr j
σ − 1 1 − rj (σ − 1)l j dr j
t j =0
(24)

where superscript A denotes that two governments compete in ADV taxes. Substituting
l1 = l2 = 1, dl j /dt j = 0, and dl j /dr j = 0 into (24), the equilibrium ADV tax rate
of each country is obtained as follows:

θσ − 1
r AC = r1 = r2 = , (25)
σ −1

suggesting that the government reduces the tax rate as it places more weight on
domestic welfare. Just as in the case of unit taxation, the government provides sub-
sidies to firms when θ < 1/σ . Substituting (25) into (6) with t j = 0, we obtain
p AC = c/(1 − θ ). This indicates that equilibrium prices will be equal no matter
whether the government uses unit or ADV tax.
Substituting (25) and l j = 1 into (12) and (17), the tax revenue, R AC , and the
sum of the consumers’ utility, U AC , when governments use an ADV tax in a closed
economy are given by:

θσ − 1
R AC = μ, (26)
σ −1
σ (1 − θ )σ
U AC = v̄ + μ ln − μ ln c. (27)
σ −1 σ −1

As with the case of the unit tax competition, tax revenue increases and consumers’
utility decreases as the government places more weight on tax revenue, ∂ R AC /∂θ > 0
and ∂U AC /∂θ < 0.

3.3 Comparison

From (22), (23), (26), and (27), we obtain differences in tax revenue and consumers’
utility in equilibria between unit and ADV tax competitions:

θσ − 1
R AC − R U C = μ. (28)
σ (σ − 1)

123
Indirect taxes in a cross-border shopping model… 159

μ σ (1 − θ )
U AC − U U C = ln , (29)
σ −1 σ −1

which derives the following lemma.


Lemma 1 If θ > (<)1/σ , (i) R AC > (<)R U C , (ii) U AC < (>)U U C , and (iii)
W+ < (>)W + .
AC UC

Proof “Appendix A”.


From Lemma 1, we find that the question of which taxes bring higher tax revenue,
consumers’ utility, and domestic welfare depend on government objectives, repre-
sented by θ . If the government places sufficiently high weight on tax revenue, namely
θ > 1/σ , the consumers’ utility and domestic welfare are smaller and the tax revenue
is larger in ADV tax competition than in unit tax competition. By contrast, if the gov-
ernment places sufficiently high weight on domestic welfare, that is, θ < 1/σ , the
ranking is reversed. These results show that the argument made by Schröder (2004)
is relevant if the governments are sufficiently benevolent. However, our analysis sug-
gests that his argument depends on the form of government objectives; generalization
of the governments’ objective presents the welfare superiority of unit tax when the
government aims toward a Leviathan objective.
The intuition of Lemma (i)–(iii) is explained separately as follows:
(i) Suppose θ > 1/σ where t and r are positive. In this case, firms have to pay more
tax as they increase prices. This is effective in curbing the market price, and as
proved by Suits and Musgrave (1953), ADV tax is more effective in depriving the
market power of firms than unit tax is. In contrast, suppose θ < 1/σ where t and
r are negative, meaning that the government provides subsidies for firms. In this
case, firms can receive larger subsidies as they increase prices, and therefore, the
government intervention induces price increases. This strengthens firms’ market
power. In fact,




θσ − 1

θσ − 1

|t U C | =

< | p AC r AC | =

(30)
(1 − θ )σ (1 − θ )(σ − 1)

holds, suggesting that the tax revenue (subsidy) per unit output is larger for ADV
method than for unit method when t and r are positive (negative), or equivalent
θ > (<)1/σ . Because p AC = pU C , the total supply of differential products in
equilibrium is the same for both tax methods, and the tax revenue (subsidy) will
be higher for ADV method than unit method if t and r are positive (negative).
(ii) The sub-utility obtained from the differential products is determined by their
price, p, and the number of varieties produced, n (see (15)). As the equilibrium
price is the same for the ADV tax and the unit tax, p AC = pU C , the number of
product varieties determines which tax method gives higher utility. According to
(9), the number of varieties is larger (smaller) for unit tax than for ADV tax if the
tax rates are positive (negative). Therefore, when t and r are positive (negative),
or equivalently θ > (<)1/σ , the total utility of the residents becomes larger
(smaller) under unit tax competition than under ADV tax competition.

123
160 H. Aiura, H. Ogawa

(iii) Domestic welfare is equal to the total utility of the residents minus the production
cost. The latter is the product of the varieties of differentiated goods and the
demand per variety. In our model, the demand per variety depends on the price,
but as in (ii), the price is the same for ADV tax and unit tax. Therefore, the levels
of domestic welfare under the different tax methods depend only on the total
utility of the residents, meaning that the mechanism to have Lemma 1 (iii) is
simply explained by that described in (ii).
The following lemma shows the comparison of the first-best optimum with the
equilibrium in the closed-border economy. This is useful to interpret how the govern-
ment changes its tax policy when the borders are open and the consumers are allowed
to cross the border, which is analyzed in the next section.

Lemma 2 Assume the government maximizes the social welfare, θ = 0. The equi-
librium quantity of differentiated product is equal to the optimum level in both unit
tax and ADV tax cases. However, the equilibrium number of firms is smaller than
the optimal number when the government uses unit tax, while it is optimal when the
government uses ADV tax.

Proof Substituting t U C |θ=0 = −c/σ and l j = 1 into (9) and (10), we obtain
n U C |θ=0 = μ/(σ f ) < n O P T , m U C |θ=0 = (σ f )/c = μ/(cn U C ) = m O P T . Substi-
tuting r AC |θ=0 = −1/(σ − 1) and l j = 1 into (9) and (10), we obtain n U C |θ=0 =
μ/[(σ − 1) f ] = n O P T and m AC |θ=0 = [(σ − 1) f ]/c = μ/(cn AC ) = m O P T .

When the government uses unit tax, the number of firms does not depend on the unit
tax rate (see (9)). In this case, the unit tax rate is determined by the government only
to satisfy the optimal quantity of product, and thus the government cannot lead the
equilibrium number of products to the optimal level. In contrast, when the government
uses ADV tax, the equilibrium number of firms depends on the ADV tax rate, and
therefore the government can change not only the quantity of products, but also the
number of firms in the equilibrium. In our formulation, the government using the ADV
tax instrument can achieve both the optimal quantity and number of varieties in the
closed-border equilibrium.

4 Equilibrium with cross border shopping

In the open-border economy, consumers cross the borders to shop, meaning that li
changes with the tax rates. Using (8), (9), dlk /dt j = −dl j /dt j , and l j + lk = 2, the
comparative statics yield

dl j μ(σ − 1)l j lk
=− , (31)
dt j {2[τ (σ − 1)l j lk − μ]}(c + t j )
dl j μσ l j lk
=− . (32)
dr j {2[τ (σ − 1)l j lk − μ]}(1 − r j )

In the following analysis, we make the following assumption:

123
Indirect taxes in a cross-border shopping model… 161

Assumption 1 τ (σ − 1) > μ.

This ensures the standard case that a rise in price because of raising tax rates reduces
the demand, dl j /dt j < 0 and dl j /dr j < 0.10 With this assumption, the second order
condition for government optimization is satisfied.

4.1 Unit tax competition

Focusing on the symmetric equilibrium, we substitute (31) and l j = lk = 1 into (20)


to obtain the tax rate in the open-border equilibrium when two governments compete
in unit taxes:

2(σ − 1)[τ (σ − 1) − μ] − (1 − θ )σ [2τ (σ − 1) − μ]


t U O = t1 = t2 = c, (33)
(σ − 1)2 μ + (1 − θ )σ [2τ (σ − 1) − μ]

where superscript O denotes the equilibrium value in the open-border economy. As


2τ (σ − 1) − μ > τ (σ − 1) − μ > 0, a decrease in θ decreases t U O , which implies
that the government lowers the tax rate as it places more weight on domestic welfare.
Furthermore, if the government has a sufficiently high weight for utility, that is, θ < θ̂ O
where

(σ − 1)(2τ + μ) − μ
θ̂ O = < 1,
σ [2τ (σ − 1) − μ]

it provides a subsidy to firms. Substituting (33) into (6) with r j = 0, we obtain the
equilibrium price as follows:

σ (σ − 1)μ + 2σ [τ (σ − 1) − μ]
pU O = c.
(σ − 1)2 μ + (1 − θ )σ [2τ (σ − 1) − μ]

Substituting (33) and l j = 1 into (12) and (17), the tax revenue, R U O and the
consumers’ utility, U U O in the open-border equilibrium are, respectively, obtained as

μ 2(σ − 1)[τ (σ − 1) − μ] − (1 − θ )σ [2τ (σ − 1) − μ]


RU O = · , (34)
σ 2[τ (σ − 1) − μ] + (σ − 1)μ
(σ − 1)2 μ + (1 − θ )σ [2τ (σ − 1) − μ]
UUO = v̄ + μ ln − μ ln c. (35)
(σ − 1)2 μ + 2(σ − 1)[τ (σ − 1) − μ]

(34) and (35) suggest that, because 2τ (σ − 1) − μ > τ (σ − 1) − μ > 0, the tax
revenue increases and the utility decreases as the government places more weight on
revenue rather than on welfare, ∂ R U O /∂θ > 0 and ∂U U O /∂θ < 0.

10 As l + l = 2, dl /dt < 0 for t > 0 and dl /dr < 0 for 0 < r < 1 if τ (σ − 1) > μ.
j k j j j j j j

123
162 H. Aiura, H. Ogawa

4.2 ADV tax competition

As in the previous section, we substitute (32) and l j = lk = 1 for (24) to find the
symmetric equilibrium when the two governments compete in unit tax in the open-
border economy. The tax rate is obtained as:

2(σ − 1)[τ (σ − 1) − μ] − (1 − θ )σ [2τ (σ − 1) − μ]


r AO = r1 = r2 = . (36)
(σ − 1){2[τ (σ − 1) − μ] + σ μ}

Because τ (σ − 1) > μ, a decrease in θ decreases r AO , which implies that the gov-


ernment reduces the tax rate as it places more weight on tax revenue. Furthermore, we
find that the government provides a subsidy to firms when θ < θ̂ O . Substituting (36)
into (6) with t j = 0, we obtain the equilibrium price as follows:

2[τ (σ − 1) − μ] + σ μ
p AO = c
(σ − 1)μ + (1 − θ )[2τ (σ − 1) − μ]

Substituting (36) and l j = 1 into (12) and (17), the tax revenue and the consumers’
utility in the open-border equilibrium are given by

2(σ − 1)[τ (σ − 1) − μ] − (1 − θ )σ [2τ (σ − 1) − μ]


R AO = μ, (37)
(σ − 1){2[τ (σ − 1) − μ] + σ μ}
σ σ (σ − 1)μ + (1 − θ )σ [2τ (σ − 1) − μ]
U AO = v̄ + μ ln − μ ln c. (38)
σ −1 (σ − 1){2[τ (σ − 1) − μ] + σ μ}

(37) and (38) suggest that, because τ (σ − 1) − μ > 0, the tax revenue increases and
the utility decreases as the government places more weight on revenue, ∂ R AO /∂θ > 0
and ∂U AO /∂θ < 0.

4.3 Comparison

Similar to Sect. 3.3, from (34), (35), (37), and (38), we obtain the difference between
prices, tax revenue, and consumers’ utilities under different tax methods as follows:

p AO − pU O
2(σ − 1)[τ (σ − 1) − μ] − (1 − θ )σ [2τ (σ − 1) − μ]
=− μc,
{(σ − 1)μ + (1 − θ )[2τ (σ − 1) − μ]}{(σ − 1)2 μ + (1 − θ)σ [2τ (σ − 1) − μ]}
(39)

2[τ (σ − 1) − μ]{2(σ − 1)[τ (σ − 1) − μ] − (1 − θ)σ [2τ (σ − 1) − μ]}


R AO − R U O = μ,
σ (σ − 1){2[τ (σ − 1) − μ] + σ μ}{2[τ (σ − 1) − μ] + (σ − 1)μ}
(40)

123
Indirect taxes in a cross-border shopping model… 163

σ σ (σ − 1)μ + (1 − θ )σ [2τ (σ − 1) − μ]
U AO − U U O = μ ln
σ −1 (σ − 1){2[τ (σ − 1) − μ] + σ μ}
(σ − 1)2 μ + (1 − θ )σ [2τ (σ − 1) − μ]
−μ ln , (41)
(σ − 1)2 μ + 2(σ − 1)[τ (σ − 1) − μ]

which derives the following lemma.

Lemma 3 If θ > (<)θ̂ O , p AO < (>) pU O , R AO > (<)R U O and U AO < (>)U U O .
If θ < θ̂ O , W + > W + , while the sign of W + − W+
AO UO AO UO
is ambiguous if
θ > θ̂ .
O

Proof “Appendix B”.

Lemma 3 shows the similarities and differences between open and closed border
economies. As similarities, the tax revenue is larger and the consumers’ utility is
smaller in ADV tax competition than in unit tax competition when the government puts
more weight on tax revenue in its objective, θ > θ̂ O . In contrast, if the governments
put more weight on domestic welfare, θ < θ̂ O , the ranking reverses. These features
are also seen in the closed border economy, and the mechanism behind them is the
same.
In addition to these similarities, there are two equilibrium features of an open border
economy that are different from the closed border economy. The first is the price
differential between the two tax methods. While p AC = pU C in the closed-border
economy, the two prices are not necessarily equal in the open border economy. This
is because, when the consumers cross the border, the government can attract overseas
demand by lowering the (after-tax) price in the country, and the effect of attracting
demand is different between the two tax methods. Suppose that θ is sufficiently large,
and that the government taxes the firms. In the closed-border economy with unit tax
competition, there is relatively little room for the government to reduce its tax rate,
compared with that under the ADV tax competition since the tax per output is already
lower in unit tax than in the ADV tax (see (30)). This leads to the rate of decline in tax
rate caused by the cross border shopping becoming larger in the ADV tax competition
than in the unit tax competition, and as a result the equilibrium price is lower for
ADV tax. The opposite applies when θ is sufficiently small, and that the government
provides subsidies to the firms.
The second feature of equilibrium in the open border economy is that the ranking
of welfare is not clearly determined since prices differ between the two tax methods.
In the closed border economy, only the difference in the number of varieties affects
the levels of social welfare since the equilibrium price is the same between the unit
tax and ADV tax, p AC = pU C . However, in the open border economy, not only
the number of varieties but also the price difference influences domestic welfare.
Therefore, there is a possibility that the ranking of welfare tends to change, depending
on θ . To obtain a visual understanding of this, Fig. 1 depicts the gap of domestic
welfare in the equilibrium between unit and ADV taxes, where c = 1 is assumed
for the numerical calculation. It shows that when θ is relatively small, the ADV tax
is welfare dominant over unit tax, without depending on the value of μ, τ, and σ .

123
164 H. Aiura, H. Ogawa

Fig. 1 The welfare comparison between unit and ADV taxes in the open-border economy, W + −W +
AO UO

However, the welfare ranking is ambiguous when θ is sufficiently large. For instance,
domestic welfare is larger in unit tax competition than in ADV tax competition for
sufficiently large θ for the combination of (μ, τ, σ ) = (0.1, 0.15, 2). However, for
(μ, τ, σ ) = (0.4, 0.3, 5), the domestic welfare is larger in ADV tax competition than
in unit tax competition when θ is sufficiently large.
As can be seen from Fig. 1, if the government places a high weight on domestic
welfare, that is, θ is small, the ADV tax is welfare superior to the unit tax, just as it
is in the closed border economy. However, as the government increases the weight
on tax revenue, that is, θ is large, the ADV tax loses its superiority. Specifically, in a
certain range where θ is sufficiently large, the unit tax is welfare-dominant over ADV
tax when there is cross-border shopping.

5 Effects of opening borders

In this section, we compare the equilibrium values between the closed and open border
economies and clarify the effects of opening borders, which allows consumers to cross
the border for their shopping.

5.1 Tax revenues

Suppose θ is sufficiently large and that tax rates, t j and r j , are positive in the open-
border economy. In this case, tax revenue in the home country increases compared with
the closed-border economy if the country can invite overseas consumers. Therefore,
the government placing a high weight on tax revenue would have an incentive to reduce
tax rates because a decrease in the price of differentiated products through tax-cutting

123
Indirect taxes in a cross-border shopping model… 165

attracts consumers living in other countries and increases tax revenue, which increases
domestic welfare.
The following result makes this clear.
Lemma 4 Border openings always reduce the equilibrium tax rate in unit tax competi-
tion, t U O < t U C , and weakly reduce the equilibrium tax rate in ADV tax competition,
r AO ≤ r AC . The equality holds if θ = 0.
Proof “Appendix C”.


Lemma 4 repeats the standard argument in the tax competition theory: The com-
petition for mobile factors lowers the tax rate imposed on it.11
From (12), the reduction of t j and r j bring down tax revenue. Therefore, with
Lemma 4, we can immediately obtain the following proposition:
Proposition 1 Border openings reduce tax revenue in unit tax competition, R U O <
R U C , and weakly reduce tax revenue in ADV tax competition, R AO ≤ R AC . The
equality holds if θ = 0.
Border openings reduce the equilibrium tax revenue except for θ = 0 in ADV tax
competition. This means that cross border shopping will change the relative merits, in
terms of tax collection, of using either of the two tax methods. The relative merits of
both the tax methods can be represented by the difference in the tax revenue between
unit and ADV taxes, which is summarized as follows:
Proposition 2 Define θ̌ to satisfy R AC − R AO = R U C − R U O . θ̌ is unique and
0 < θ̌ < 1/σ . If 0 ≤ θ < θ̌ , the decline in tax revenue because of the border
openings is larger in unit tax competition than in the ADV tax competition, that is,
R AC − R AO < R U C − R U O . If θ̌ < θ ≤ 1, the tax revenue drops sharply because of the
border openings in the ADV tax competition compared with the unit tax competition,
that is, R AC − R AO > R U C − R U O .
Proof “Appendix D”.


Proposition 2 shows that the relative merits in terms of securing tax revenue in using
either of the two tax methods depend on θ . If governments place a large weight on their
tax revenue, θ > θ̌ , the decline in tax revenue because of the border openings is larger
in ADV tax competition than in the unit tax competition. As R AO > R U O for θ > θ̂ O ,
this means that if the governments place sufficiently large weight on their tax revenue,
θ > θ̂ O , cross-border shopping behavior associated with border openings closes the
tax revenue gap between ADV and unit tax competitions, which weakens the revenue
superiority of an ADV tax. If the governments place a large weight on their domestic

11 It shows that when θ = 0, the border opening reduces the equilibrium tax rate in unit tax competition,
but not in ADV tax competition. As shown in Lemma 2, the quantity and number of differentiated products
satisfy the optimal levels, when the government uses the ADV tax in the closed economy. Hence, the
government has no incentive to change its ADV tax rate when the consumers cross the border for their
shopping. When the government uses unit tax, however, the number of firms in the equilibrium is insufficient
compared with the optimal level. In this case, the government has incentive to reduce its unit tax rate to
make the product variety closer to the optimum level.

123
166 H. Aiura, H. Ogawa

welfare, θ < θ̌ , the decline in tax revenue because of the border openings is larger in
the ADV tax competition than in the unit tax competition. Since R AO < R U O < 0
for θ < θ̌ , this means that if the governments place a sufficiently large weight on their
domestic welfare, θ < θ̌ , cross-border shopping associated with border openings
closes the subsidy gap between ADV and unit tax competitions.
As mentioned in Sect. 4.3, in the closed-border economy with unit tax competition,
there is relatively little room for the government to reduce its tax rate, compared with
that under the ADV tax competition since the tax per output is already lower in unit
tax than in the ADV tax (see (30)). Therefore, if the governments place a large weight
on their tax revenue, border opening causes a sharper decline in the tax per output
in the ADV tax competition than in the unit tax competition, which reduces the gap
of the tax revenue between unit and ADV taxes. If the governments place a large
weight on domestic welfare, the variety of products in the equilibrium of unit tax
competition is smaller than the optimum level, whereas the variety of products in the
equilibrium of ADV tax competition is close to the optimum level, as deduced from
Lemma 2. Therefore, the government with the unit tax method has a stronger incentive
to increase its subsidy in order to improve the variety of products, but the government
with the ADV tax method does not have such incentives because the product variety
is sufficiently close to the optimum level. This leads to the border opening shrinking
the subsidy gap between unit and ADV tax competition.

5.2 Consumers’ utility

Because a reduction in tax rate (or an increase in subsidy rate) accompanied by the
border opening decreases the price of the differentiated products, it increases the
consumers’ utility. Furthermore, a reduction in the ADV tax rate (or an increase in the
ADV subsidy rate) increases the number of firms, which also increases consumers’
utility.12 Summarizing these, we obtain the following result:

Proposition 3 Border opening increases the consumers’ utility in unit tax competition,
U U O > U U C , and weakly increases the consumers’ utility in ADV tax competition,
U AO ≥ U AC . The equality holds if θ = 0.

The effects of border openings on the relative merits in terms of consumers’ utility
of the two tax methods can be obtained as:

Proposition 4 Define θ̃ to satisfy U AO − U AC = U U O − U U C . θ̃ is unique and


0 < θ̃ < 1/σ . If 0 ≤ θ < θ̃ , compared with the ADV tax competition, the sum of the
consumers’ utility increases greatly as a result of the cross border shopping in unit
tax competition, that is, U AO − U AC < U U O − U U C . In contrast, if θ̃ < θ ≤ 1,
compared with the unit tax competition, the sum of the consumers’ utility increases
greatly as a result of the cross border shopping in ADV tax competition, that is,
U AO − U AC > U U O − U U C .

12 This can be confirmed, from (17), that U increases as t and r decrease.


j j j

123
Indirect taxes in a cross-border shopping model… 167

Proof “Appendix E”.


Proposition 4 suggests that there is a case in which the border-opening reduces the
gap of consumers’ utility between unit tax and ADV tax competitions, and it happens
in two extreme cases. The first is the case in which the governments place a sufficiently
large weight on their domestic welfare, θ < θ̃ . In this case, the border openings narrow
the gap of utilities because U AC > U U C for θ < 1/σ . The second case is the reverse
of the first case. That is, when the governments place a sufficiently large weight on their
tax revenue, θ > 1/σ , the cross border shopping shrinks the gap of the consumers’
utility between two different tax methods because U AC < U U C . These results hold
because an increase in subsidy rate is larger in unit tax competition than in ADV
tax competition if the governments place a large weight on their domestic welfare,
θ < 1/σ , and a reduction in tax rate is larger in ADV tax competition than in unit tax
competition if the governments place a large weight on their tax revenue, θ > 1/σ .

5.3 Welfare

In the symmetric equilibrium, l j = 1 holds, and thus the domestic welfare at a certain
tax rate is the same between closed- and open-border economies. In this case, the
domestic welfare is maximized at t U C |θ=0 and r U C |θ=0 even in the open-border
economy, and it decreases as the tax rate moves away from t U C |θ=0 and r U C |θ=0 .
This argument gives the following result.

Proposition 5 (i) If θ ≥ μ/[2τ σ (σ −1)−σ μ], W + > W+


UO UC
. If θ is sufficiently
close to 0, W + UO
<W + UC
.
(ii) If θ > 0, W + > W + . If θ = 0, W + = W+ .
AO AC AO AC

Proof “Appendix F”.


Proposition 5(i) shows that the border openings increase the equilibrium domestic
welfare in unit tax competition if the government puts a sufficiently high weight on
tax revenue, that is, θ is large. In contrast, it reduces domestic welfare in unit tax
competition when the government puts a sufficiently high weight on welfare, that is,
θ is small. Proposition 5(ii) shows that the border openings increase the domestic
welfare in ADV tax competition, except for θ = 0. If θ = 0, the welfare is unchanged
without depending on whether the consumers cross the border.
When 0 < θ < μ/[2τ σ (σ −1)−σ μ], we cannot obtain clear answers as to whether
border openings increase or decrease the domestic welfare in unit tax competition.
To see the ranking of the welfare clearly, Fig. 2 is depicted, where c = 1 is assumed
with reasonable parameters. It shows that when governments use unit taxes, the border
openings increase (decrease) the domestic welfare if θ is larger (smaller) than a certain
threshold value in different parameter sets.
The intuition of the results is as follows. The governments have an incentive to
reduce their tax rate (increase their subsidy rate) to attract consumers from the foreign
country to the domestic market of the differentiated products. As tax cuts lead to an
increase in firms’ production, it alleviates the inefficiency because of the under supply
of product in the imperfectly competitive markets, thereby improving the domestic

123
168 H. Aiura, H. Ogawa

Fig. 2 The impact of border openings on the domestic welfare

welfare. When governments use ADV taxes, as θ → 0 in the closed border economy,
the quantity and the number of firms in the equilibrium approach optimal levels.
Hence, the smaller θ , the smaller the incentive for tax reductions by the government
when the market is open, and therefore welfare is only slightly improved since there
is little room for the government to improve welfare by reducing the tax rate. When
governments use unit taxes, however, the government has incentive to lower tax rates
so as to increase the variety of products which is less than the optimal level, but
such tax reduction generates an over-supply of differentiated products. Furthermore,
the variety of products is unchanged even when the equilibrium tax rate is reduced
because both countries take the same action in a symmetric two-country model, and
they share the consumers equally as before. Accordingly, the border openings reduce
social welfare in unit tax competition.

123
Indirect taxes in a cross-border shopping model… 169

Proposition 5 with Fig. 2 suggests that when the government aims toward a
Leviathan objective and puts a higher weight on revenue than on the residents’ utilities,
the increased mobility of consumers across the border increases domestic welfare if
the governments are competing in unit tax. In this case, the consumers’ mobility is a
deterrent to Leviathan behavior and improves welfare. However, if the governments
are more like benevolent agents who seek welfare maximization, and thus attach a
high weight to domestic welfare, the increased consumer mobility for shopping low-
ers welfare in the unit tax competition. In this case, cross border shopping is harmful
from the welfare perspective. In contrast, when governments compete in ADV tax,
cross-border shopping always increases the welfare in all countries, without depend-
ing on government objectives. The implication based on these findings suggests that if
the government commits to ADV tax competition, it can avoid the decline in domestic
welfare associated with intense competition for cross border consumers.
Finally, the effects of border openings on the relative merits in terms of domestic
welfare can be obtained as follows.

Proposition 6 If θ is sufficiently close to 1, the increase in domestic welfare in ADV


tax competition because of border openings is larger than in unit tax competition,
that is, W + − W+ > W+ − W + . If θ is sufficiently close to 0, while
AO AC UO UC

the border openings lower domestic welfare in unit tax competition, the welfare is
unchanged or even increases in ADV tax competition, that is, W + − W +
AO AC
≥0>
W+ − W + . The difference between the two is given by |W + − W+ | <
UO UC AO AC

|W + UO
−W + UC
|.

Proof “Appendix G”.

Proposition 6 reveals clear results in the two extreme cases of government objective;
the pure Leviathan and the benevolent. If the government places a high weight on rev-
enue, the government under the ADV tax competition would lower its tax rate more
aggressively than the government under the unit tax competition. Thus, compared
with unit taxation, as the ADV taxes have a greater extent to improve distortion asso-
ciated with the under-supply and the insufficient number of differentiated products,
the improvement of domestic welfare is larger in ADV than in unit tax competition.
As a result, domestic welfare in a closed-border economy is larger in unit tax com-
petition than in ADV tax competition, while the domestic welfare in an open-border
economy is not always larger in unit tax competition than in ADV tax competition, as
Lemmas 1 and 3 show. If the government places a high weight on domestic welfare, as
Proposition 5 shows, the border opening decreases domestic welfare in unit tax com-
petition while it does not change, or slightly increases, domestic welfare in ADV tax
competition. Therefore the effect of border-openings is larger in unit tax competition
than in ADV tax competition.
These extreme scenarios imply that when θ is moderate, the effect of border-
openings is ambiguous. Panel A of Fig. 2 shows that W + − W+ > W+
AO AC UO

W + UC
for any θ if μ = 0.1, τ = 0.15, σ = 2, and c = 1. On the other
hand, Panel B of Fig. 2 shows that if μ = 0.1, τ = 0.8, σ = 2, and c = 1,
W+ − W+ < W+ − W+
AO AC UO UC
when θ ranges from approximately 0.2–0.4.

123
170 H. Aiura, H. Ogawa

This suggests the possibility that an increase in domestic welfare because of cross
border shopping may be larger in unit tax competition than in ADV tax competition
in some cases.

6 Conclusion

In this study, we examined the efficiency of two different tax methods in a cross-
border shopping model with monopolistic competition. The particular feature of our
approach, compared with preceding studies, for example, Aiura and Ogawa (2013),
is that consumers are assumed to choose the place of their shopping by considering
not only the after-tax-price levels but also the variety of product. That is, the “total
attraction” does matter to consumers who can cross the borders, and thus the govern-
ment has to account for tax effects on both prices and the varieties of products in the
country.
We compared equilibrium outcomes with or without cross border shopping to clarify
the effect of consumer mobility. The main argument from our study is that the relative
merits in terms of welfare of using either ADV tax or unit tax critically depend on the
presence of cross-border shopping and the government objectives. When the govern-
ment aims toward a Leviathan objective and puts a higher weight on revenue than on
residents’ utilities, the increased mobility of consumers across the border increases
domestic welfare if the governments compete in unit tax. In this case, the consumers’
mobility is a deterrent to Leviathan behavior and improves welfare. However, if the
governments are benevolent and thus attach a high weight to domestic welfare, the
increased consumer mobility for shopping lowers the welfare in the unit tax competi-
tion. In contrast, when the governments compete in ADV tax, cross-border shopping
always increases the welfare in all countries, without depending on the government
objectives. This leads us to conclude that if the government commits to ADV taxes,
it avoids a decline in domestic welfare even if the competition for mobile consumers
becomes intense.
The results are derived within the context of a standard monopolistic competition
model, but they obviously depend on less general assumptions. First, the equilibrium
price is identical for all varieties, which seems to be unrealistic. If the price elasticity
of demand differs between the varieties of products, the price of each variety is not
identical. Second, we compared the outcomes in the closed-border and open border
economies by assuming that both countries use the same tax method when competing
for mobile consumers. It is natural to assume that all countries use the same tax method
in a model of symmetric countries. However, this approach might be inadequate when
analyzing an endogenous tax method choice. An analysis of the equilibrium outcome
when two governments use a different tax method would contribute to the literature on
the endogenous choice of tax methods. Finally, we do not allow residential choices. The
assumption that residents are fixed in the initial location though they can move just for
shopping is reasonable in the short-run equilibrium, but there might be room to account
for inter-regional residential choices in the long run, especially for two asymmetric
regions. An analysis using simulation techniques would solve these problems.

123
Indirect taxes in a cross-border shopping model… 171

Acknowledgements An earlier version of this paper was presented at ARSC2013 and NARSC2013.
We would like to thank Takanori Ago and Johan Lundberg for their helpful comments. In addition,
the comments of anonymous referees and the editor Giacomo Corneo have substantially improved this
paper. Any mistakes herein are, of course, our own. The research has been supported by JSPS KAK-
ENHI (Grant Numbers JP25245042, JP16K12374, JP17H02533, JP18H00865, JP18K01632), and the Nitto
Foundation.

Appendix

A Proof of Lemma 1

From (28) and (29), R AC − R U C = 0 and U AC − U U C = 0 when θ = 1/σ . In


addition, R AC − R U C is an increasing function of θ and U AC − U U C is a decreasing
function of θ . Therefore, R AC > (<)R U C and U AC < (>)R U C for θ > (<)1/σ .
Furthermore,

d(W + − W+
AC UC
) d(R AC − R U C ) d(U AC − U U C )
= +
dθ dθ dθ
μ μ
= − ≤ 0,
σ − 1 (1 − θ )(σ − 1)

suggesting that W + − W +
AC UC
is a decreasing function of θ . Notice that W AC −
W U C = 0 at θ = 1/σ . Then we have W + < (>)W +
AC UC
for θ > (<)1/σ .

B Proof of Lemma 3

If θ = θ̂ O , t U O = r AO = 0. Therefore, p AO = pU O , R AO = R U O , and U AO =
U U O . Thus, W + = W+
AO UO
if θ = θ̂ O . From (39) and (40), p AO − pU O and
R −R
AO U O are the increasing function of θ . Thus, p AO > (<) pU O and R AO > (<)
R U O for θ > (<)θ̂ O . Since

d(U AO − U U O )

σμ (1 − θ)[2τ (σ − 1) − μ]2
=− · ≤ 0,
σ − 1 {(σ − 1)μ + (1 − θ)[2τ (σ − 1) − μ]}{(σ − 1)2 μ + (1 − θ)σ [2τ (σ − 1) − μ]}

U AO − U U O is a decreasing function of θ . Thus, U AO < (>)U U O for θ > (<)θ̂ O .


Furthermore,

d(W + − W+
AO UO
)
d(R AO − R U O ) d(U AO − U U O )
= +
dθ dθ dθ
2τ (σ − 1) − μ
= μ
(σ − 1){2[τ (σ − 1) − μ] + σ μ}{2[τ (σ − 1) − μ] + (σ − 1)μ}
σμ (1 − θ)[2τ (σ − 1) − μ]2
− · ,
σ − 1 {(σ − 1)μ + (1 − θ)[2τ (σ − 1) − μ]}{(σ − 1)2 μ + (1 − θ)σ [2τ (σ − 1) − μ]}

123
172 H. Aiura, H. Ogawa

which is positive at θ = 1 and negative at θ = 0 and θ = θ̂ O . Furthermore, the


second-order derivative shows

d 2 (W + − W+ )
AO UO

dθ 2
σμ [2τ (σ − 1) − μ]2 {(σ − 1)3 μ2 − (1 − θ)2 σ [2τ (σ − 1) − μ]2 }
= · ,
σ − 1 {(σ − 1)μ + (1 − θ)[2τ (σ − 1) − μ]}2 {(σ − 1)2 μ + (1 − θ)σ [2τ (σ − 1) − μ]}2

which implies that W + − W+


AC UC
has at most one extremum for 0 ≤ θ ≤ 1.
Therefore, W + AC
−W + UC
is a decreasing function of θ for 0 ≤ θ < θ̂ O . Hence,
W+ > W+
AO UO
for θ < θ̂ O .

C Proof of Lemma 4

We have

(σ − 1)[θ (σ − 1)2 + (1 − θ )]
tUC − tU O = μc > 0,
(1 − θ )σ {(σ − 1)2 μ + (1 − θ )σ [2τ (σ − 1) − μ}]
θσμ
r AC − r AO = ≥ 0,
[2τ (σ − 1) − μ] + (σ − 1)μ

which derives Lemma 4.

D Proof of Proposition 2

(R AC − R AO ) − (R U C − R U O )
{2(2σ − 1)[(σ − 1)τ − μ] + σ (σ − 1)μ} − (1 − θ )σ {4[(σ − 1)τ − μ] + σ μ} 2
= μ .
σ {2[(σ − 1) − μ] + (σ − 1)μ}{2[(σ − 1) − μ] + σ μ}
(42)

Thus, (R AC − R AO ) − (R U C − R U O ) increases with an increase in θ . Furthermore,


(R AC − R AO ) − (R U C − R U O ) = R U O − R AO > 0 if θ = 1/σ and (R AC − R AO ) −
(R U C − R U O ) < 0 if θ = 0. Therefore, Proposition 2 is proved.

E Proof of Proposition 4

(U AO − U AC ) − (U U O − U U C ) = −μ ln αg(θ )h(θ ) (43)

where

[2(σ − 1)τ − μ + (σ − 1)μ]σ/(σ −1)


α= ,
2[(σ − 1)τ − μ] + (σ − 1)μ

123
Indirect taxes in a cross-border shopping model… 173

(1 − θ )σ [2(σ − 1)τ − μ] + μ(σ − 1)2


g(θ ) = ,
(1 − θ )σ [2(σ − 1)τ − μ] + μσ (σ − 1)
 
σ − 1 −1/(σ −1)
h(θ ) = 2(σ − 1)τ − μ + μ .
1−θ

As g(θ ) and h(θ ) are the decreasing function of θ , the RHS of (43), that is,
−μ ln αg(θ )h(θ ), is an increasing function of θ . Moreover, we have
−μ ln αg(1/σ )h(1/σ ) = U AO − U U O > 0 and
2σ (σ − 1)τ + (σ 2 − 3σ + 1)μ
−μ ln αg(0)h(0) = −μ ln < 0.
2σ (σ − 1)τ + (σ 2 − 3σ )μ

Therefore, (U AO − U AC ) − (U U O − U U C ) > 0 if 0 ≤ θ < θ̃ and (U AO − U AC ) −


(U U O − U U C ) < 0 if θ̃ < θ ≤ 1. Accordingly, Proposition 4 is proved.

F Proof of Proposition 5
We have


dW +
d R j

dU j


= +
dt j
dt j
l j =1 dt j
l j =1
l j =1
μ c 
=− + tj ,
(c + t j ) σ
2


dW +
d R j

dU j


= +
dr j
dr j
l j =1 dr j
l j =1
l j =1
 
μ 1
=− + rj .
1 − rj σ − 1

These show that domestic welfare is maximized at t U C |θ=0 and r U C |θ=0 . Furthermore,
they show that domestic welfare decreases as the tax rate moves away from t U C |θ=0
and r U C |θ=0 .
Domestic welfare is maximized at θ = 0 in the equilibrium of the closed-border
economy and in the equilibrium of the open-border economy with ADV tax com-
petition. In the equilibrium of the open-border economy with unit tax competition,
domestic welfare is maximized at θ = μ/[2τ σ (σ − 1) − σ μ] because of t U C = t U O .
These derive Proposition 5.

G Proof of Proposition 6

lim [(W + − W+ ) − (W + − W+
AO AC UO UC
)]
θ→1
= lim [(U AO − U AC ) − (U U O − U U C )]
θ→1
− lim [(R AC − R AO ) − (R U C − R U O )] (44)
θ→1

123
174 H. Aiura, H. Ogawa

Substituting (42) and (43) into (44), we obtain limθ→1 [(W + −W + )−(W + −
AO AC UO

W + )] = ∞, that is W + − W+ > W+ − W+
UC AO AC UO UC
if θ is sufficiently
close to 1. Further, it is obvious, by Proposition 5, that if θ is sufficiently close to 0,
W AO − W AC ≥ 0 > W U O − U U C and |W AO − W AC | < |W U O − U U C |.

References
Aiura H, Ogawa H (2013) Unit tax versus ad valorem tax: a tax competition model with cross-border
shopping. J Public Econ 105:30–38
Akai N, Ogawa H, Ogawa Y (2011) Endogenous choice on tax instruments in a tax competition model:
unit tax versus ad valorem tax. Int Tax Public Finance 18:495–506
Akai N, Ogawa H, Ogawa Y (2014) Endogenous choice of subsidy instruments in imperfectly competitive
markets: a unit subsidy versus an ad valorem subsidy. Ann Econ Stat/Ann ’Écon Stat 113(114):81–98
Anderson SP, de Palma A, Kreider B (2001) Tax incidence in differentiated product oligopoly. J Public
Econ 81:173–192
Bygvra S (2011) The road to the single European market as seen through the Danish retail trade: cross-border
shopping between Denmark and Germany. Int Rev Retail Distrib Consum Res 8:147–164
Dixit A, Stiglitz J (1977) Monopolistic competition and optimum product diversity. Am Econ Rev 67:297–
308
Dröge S, Schröder P (2009) The welfare comparison of corrective ad valorem and unit taxes under monop-
olistic competition. Int Tax Public Finance 16:164–175
Edwards J, Keen M (1996) Tax competition and Leviathan. Eur Econ Rev 40:113–134
Fitzgerald J, Quinn T, Whelan R, Williams J (1988) An analysis of cross-border shopping. Economic &
Social Research Institute, Dublin
Hamilton SF (2009) Excise taxes with multiproduct transactions. Am Econ Rev 99:458–471
Henkel J, Konrad S, Walz U (2000) Coalition building in a spatial economy. J Urban Econ 47:136–163
Jorgensen J, Schröder P (2005) Welfare-ranking ad valorem and specific tariffs under monopolistic com-
petition. Can J Econ 38:228–241
Kawachi K, Ogawa H, Susa T (2018) Endogenizing government’s objectives in tax competition with capital
ownership. Int Tax Public Finance. https://doi.org/10.1007/s10797-018-9516-1
Keen M (1998) The balance between specific and ad valorem taxation. Fisc Stud 19:1–37
Kind HJ, Koethenbuerger M, Schjelderup G (2009) On revenue and welfare dominance of ad valorem taxes
in two-sided markets. Econ Lett 104:86–88
Lapan HE, Hennessy DA (2011) Unit versus ad valorem taxes in multiproduct Cournot oligopoly. J Public
Econ Theory 13:125–138
Leal A, Lopez-Laborda J, Rodrigo F (2010) Cross-border shopping: a survey. Int Adv Econ Res 16:135–148
Liang W-J, Cheng K, Wang A (2018) The superiority among specific, demand ad valorem and cost ad
valorem subsidy regimes. J Econ 123:1–21
Lockwood B (2004) Competition in unit vs. ad valorem taxes. Int Tax Public Finance 11:763–772
Lockwood B, Wong K (2000) Specific and ad valorem tariffs are not equivalent in trade wars. J Int Econ
52:183–195
Manuszak M, Moul C (2009) How far for a buck? Tax differences and the location of retail gasoline activity
in southeast Chicagoland. Rev Econ Stat 91:744–765
Mittelstaedt R, Stassen R (1990) Shopping behavior and retail merchandising strategies. J Bus Res 21:243–
258
Pal R, Sharma A (2013) Endogenizing government’s objectives in tax competition. Region Sci Urban Econ
43:570–578
Schröder PJH (2004) The comparison between ad valorem and unit taxes under monopolistic competition.
J Econ 83:281–292
Schröder PJH, Sørensen A (2010) Ad valorem versus unit taxes: monopolistic competition, heterogeneous
firms, and intra-industry reallocations. J Econ 101:247–265
Suits D, Musgrave R (1953) Ad valorem and unit taxes compared. Q J Econ 67:598–604
Takatsuka H (2014) Tax effects in a two-region model of monopolistic competition. Pap Region Sci 93:595–
617

123
Indirect taxes in a cross-border shopping model… 175

Timothy D, Butler R (1995) Cross-border shopping: a North American perspective. Ann Tour Res 22:16–34
Tömöri M (2010) Investing shopping tourism along the borders of Hungary: a theoretical perspective. Geo
J Tour Geosites 6:202–210
Vetter H (2013) Consumption taxes in monopolistic competition: a comment. J Econ 110:287–295
Wakefield, Baker (1998) Excitement at the mall: determinants and effects on shopping response. J Retail
74:515–539
Wrede M (1998) Household mobility and the moderate leviathan: efficiency and decentralization. Region
Sci Urban Econ 28:315–328

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

123
Journal of Economics is a copyright of Springer, 2019. All Rights Reserved.

You might also like