HETEROGENEOUS FIRMS IN OPEN ECONOMY WITH CORRUPTION
Vanessa Pham
April 20th, 2012
Third-Year Paper
This paper studies the e¤ect of trade liberalization on corruption using Melitz’s model.
We …nd that corruption is reduced because of an increase in total revenues when exporting
cost is lower. Moreover, when home country trades with a less corrupt foreign country, fewer
domestic …rms will stay active than in the symmetric case but many of them will become
exporters. In the case where foreign …rms have to pay higher exporting cost, home country
will have more …rms operating domestically but only a smaller number of them will export.
E¤ect on corruption level is determined by changes in total revenues due to these changes
in home country’s market environment.
1
1
Introduction
Corruption and its e¤ect on economic growth have been a popular topic of debates among
economists. The general consensus is that corruption a¤ects growth negatively. Mauro
(1995) is one of the …rst papers using systematic cross-country empirical analysis to study corruption, measured by indicators of bureaucratic honesty, and e¢ ciency to economic growth.
He found that corruption lowers private investment and reduces economic growth. A popular
explanation for this negative e¤ect is that corruption allows ine¢ cient usage of resources and
distorts economic activities.
Some have found that other variables play an important role in determining the relationship between corruption and economic growth. Ehrlich and Lui (1999) study the investments
in accumulating human capital and accumulating political capital and their consequences on
long-term growth under two political regimes. Neeman, Paserman, and Simhon (2008) have
found evidence that the relationship between corruption and output per capita is strongly
related to a country’s degree of openness. That is in open economies, corruption and GNP
per capita are negatively correlated, but closed economies show no relationship. Other empirical studies, including Ades and Di Tella (1999), Fisman and Gatti (2002), Goldsmith
(1999), Treisman (2000), and Wei (2000), have found that higher trade intensity and/or
small populations are associated with lower corruption levels. On the other hand, Dutt and
Traca (2010) show that in high tari¤ environments, corruption can actually enhance trade
‡ows between countries by allowing …rms to evade these tari¤ barriers.
This paper contributes to the literature on corruption by establishing a theoretical model
based on Melitz (2003) to study the e¤ect of trade liberalization on corruption. Melitz models
how international trade reallocates resources among heterogeneous …rms in a monopolistically competitive industry. In this paper, corruption is modelled as the fraction of total
revenues from …rms across the economy spent on bribery to government o¢ cials to reduce
their variable costs in production. Examples of this type of bribery include bribing a public
2
o¢ cial to gain signi…cant advantage over its competitors or to skim on required production
protocols. In 2009, William Je¤erson, a Louisiana representative, was found guilty of corruption. He took at least $100,000 in bribes and recommended a tech company named iGate,
Inc. to the U.S. Army and o¢ cials in Nigeria, Ghana, and Cameroon. The article "For
Bribing O¢ cials, Chinese Give the Best" in the New York Times in March 2009 describes
the situation in China where businessmen lavish government o¢ cials with luxury goods year
round. Stores carrying these luxury products even admitted that they have special accounts
for government o¢ cials, their relatives, and sometimes their mistresses. The expenses on
these bribes are considered an increase in …xed costs that may help reduce variable costs
in production for businesses involved in bribery. For example, …rms can bribe government
o¢ cials to evade taxation. Sequeira and Djankov (2010) call this type of corruption "collusive corruption" since public o¢ cials and private agents collude to share rents generated
by involving in illegal transactions. Public o¢ cials, such as health inspectors, safety inspectors, and …re inspectors, can also create unnecessary troubles to force …rms to pay bribes.
Sequeira and Djankov call this "coercive corruption". However, in both types of corruption,
paying bribes help …rms reduce their variable costs. The approach used in this paper is
similar to the innovation literature which includes several studies showing how …rms choose
to spend more on …xed costs to reduce variable costs in Melitz’s international trade model,
such as Bustos (2011) and Yeaple (2005). Bustos shows that more productive …rms will
choose higher technology and trade openness encourages more …rms to upgrade their technology. Falvey et al (2007) studying the case where home country has a superior technology
compared to foreign country concludes that the home survival cuto¤ is lower and the home
export cuto¤ is higher than the foreign ones.
The main research question here is whether a country will experience more or less corruption with trade openness. We …nd that trade liberalization does reduce corruption and
increase social welfare. Departing from the symmetric case, this paper also studies the situation where foreign country is less corrupt than home country and shows that home country
3
will have fewer …rms but many of them will be exporters. When foreign …rms have to pay
higher exporting cost, home country has more …rms but only a small number of them will
export. The e¤ect on corruption level depends on how total revenues change due to these
changes in the home country’s market. The paper is organized as follows. Section 2 provides
the basic setup of Melitz’model with CES preferences demand and Cobb-Douglas production
function. Section 3 introduces an option for …rms to produce and export goods to foreign
markets. Section 4 analyzes …rms’decisions on whether to export or not and whether to use
a low or high level of bribery. Section 5 analyzes the case where countries are asymmetric.
Lastly, section 6 summarizes key results from the model.
2
The Setup
2.1
Demand
We have an economy with homogeneous consumers and heterogeneous …rms. A representative consumer has income I and CES preferences over a set of di¤erentiated goods x 2 X;
where X is a set of all potentially available goods. Income consists of wages normalized
to 1 paid for inelastically supplied amount of labor L and …rm pro…ts
which are equally
distributed among all consumers.
Consumer maximizes his utility as follows:
max U =
s.t.
Z
q(x)
Z
1
;0 <
q(x) dx
<1
x2X
p(x)q(x)dx = I = L +
x2X
where q(x) is the demand for good x, p(x) is the price of good x;
substitution between any too goods with
index P as:
P =
Z
1
> 1 and
1
1
p(x)
x2X
4
1
dx
:
is the elasticity of
: We de…ne the aggregate price
The demand for good x is derived from the consumer maximization problem:
q(x) =
I
P
p(x)
P
;
and the price elasticity of demand is:
"p =
2.2
dq p
= :
dp q
Production
We consider a monopolistically competitive market with N …rms where each …rm produces
a di¤erentiated good x. They can choose to involve in a low or high level of bribery to
government o¢ cials to reduce their variable costs. The increasing returns technologies for
low and high level of bribery are de…ned as follows
1
q;
'
1
T C hb (') = f +
q:
'r
T C lb (') = f +
where f > 0 is the …xed cost of production which is the same for all …rms, '
1 is a
…rm’s productivity level, r is the amount of variable cost that can be reduced by involving
a high level of bribery to government o¢ cials, and
with a high level of bribery. r and
is the additional …xed cost associated
are greater than 1:
Each …rm draws its productivity from a Pareto distribution with the cumulative distribution function:
F (') = 1
The assumption
> M axf1;
' ;
> M axf1;
1g:
(1)
1g assures that in equilibrium the size distribution of
…rms has a …nite mean. Firms also face an exogenous probability of exit :
5
Firm maximizes its pro…ts:
max
lb
(') = pq
or max
hb
(') = pq
p
p
lb
s.t.
0,
1
q f;
'
1
q f ;
'r
hb
0:
Solving the …rm’s problem, we have:
plb (') =
1
1
; phb (') =
'
'r
q lb (') = IP
lb
3
(') =
1
( ') ; q hb (') = IP
I(P ')
1
hb
f;
(') =
1
( 'r) ; and
I(P 'r)
1
f :
Model with International Trade
Let the economy as described above trade with another economy that has the same parameters as the domestic economy, except that it does not have corruption.
3.1
Production in the Domestic Market
The equilibrium price pd , quantity qd , and pro…ts
d
in the domestic market given the
aggregate price index in the open economy PT and consumer income IT are as follows:
plb
d (') =
1
1
; phb
d (') =
'
'r
qdlb (') = IT PT
lb
d (')
=
1
( ') ; qdhb (') = IT PT
IT (PT ')
1
hb
d (')
f;
6
=
1
( 'r) ; and
IT (PT 'r)
1
f :
3.2
Exports
Following Melitz (2003), we model international trade with two new cost parameters. Let
> 1 be a per-unit costs for exporting such as transportation costs and tari¤s. Also, let fx
be the …xed cost of exporting and fx > f: Then, the total cost of an exporting …rm is:
T Cxlb (') = fx +
or T Cxhb (') = fx +
qx ;
'
qx :
'rG
An exporting …rm maximizes its pro…ts in the foreign market as follows:
max
lb
x (')
= p x qx
or max
hb
x (')
= p x qx
px
px
s.t.
'
qx
'r
fx ;
qx
fx
0:
x
Then, the equilibrium price, quantity, and pro…ts for an exporting …rm are:
plb
x (') =
qxlb (')
lb
x (')
'
; phb
x (') =
1
= IT PT
=
IT
(
'
'r
1
) ; qxhb (') = IT PT
PT '
1
fx ;
hb
x (')
=
(
'r
IT
) ; and
PT 'r
1
fx :
Since exporting …rms operate both in the domestic market and the foreign market, their
pro…ts are
X (')
=
d (')
+
x ('):
Note that exporting …rms do not have an additional
bribery for their exporting production if they choose the high level of bribery since it is
already paid in the domestic market. We also have that IT = I =
L: Here, all costs are
paid to domestic labor so it does not make a di¤erence whether …rms pay to a public o¢ cial
in the form of bribes or to labor in the form of wages.
7
4
Corruption in An Open Economy
4.1
Firms’Exporting and Bribery Decisions
From the above sections, we can summarize the pro…t functions for …rms depending on their
choices of whether to export or not and whether to use a low or high level of bribery.
IT
lb
d
=
hb
d
=
lb
X
= (1 +
1
)
hb
X
= (1 +
1
)
IT
1
( 'PT )
( 'PT r)
IT
IT
f;
1
f ;
1
( 'PT )
(f + fx );
1
( 'PT r)
(f + fx ):
Since only …rms that earn non-negative pro…ts will produce, the cut-o¤ level of producing
…rms 'd satis…es ('lb
d ) = 0; so
1
'd =
PT
f
IT
1
1
(2)
:
Producing …rms face two decisions: bribery level and exporting choices. Let 'b and
'x denote the cut-o¤ level of …rms using high level of bribery and exporting …rms. As
represented in Figure 1, we assume that 'b < 'x , that means that …rms with productivity
level between 'b and 'x …nd it more pro…table to only operate in the domestic market with
a high level of bribery than operate in both domestic and foreign markets with a low level
of bribery.
The …rm with productivity level 'b is indi¤erent between a low level of bribery and a
higher level:
IT
( ' b PT )
1
f=
IT
8
( 'b PT r)
1
f :
'b =
f(
IT (r
= 'd
r
1
1
(3)
:
(4)
1)
1
1
1
r
1
PT
1
1
1
1
1
If
1
1)
1
1
> 1, then 'b > 'd . That means only the more productive …rms choose
a high level of bribery. It is a reasonable assumption since …rms have to make the decision
whether to stay in business …rst before deciding whether to export or not. We also observe
from empirical studies that not all …rms in the market are exporters.
Similarly, the …rm with productivity level 'x is indi¤erent between only producing in the
domestic market and producing in both markets using a high level of bribery
IT
( 'x PT r)
1
f = (1 +
'x = 'd
If
h
fx
fr
1
i
1
fx
fr
)
IT
1
1
( 'x PT r)
1
(f + fx ):
1
1
:
(5)
1
1
> 1, then only more productive …rms choose to export. We also assume
i 11
h
1 (r
1 1)
'x
fx
that 'b < 'x . That means ' =
> 1.
f r 1 ( 1)
1
b
9
Level of Bribery and Exporting Choices
4.2
Market Equilibrium
The market equilibrium price, the distributions of productivity level for active …rms, and the
fractions of …rms exporting and …rms choosing a high level of bribery are determined by the
free entry condition. This condition requires that …xed entry cost equals the present value
of expected pro…ts:
fe = [1
where 1
F ('d ) = ('d )
F ('d )]
1
(6)
;
is the probability of …rms remaining active and
is the
expected pro…ts of active …rms.
=
where
d
d
+ px
x;
is the expected pro…t from domestic market, px =
10
1 F ('x )
1 F ('d )
is the probability of
exporting conditional on survival, and
x
is the expected pro…t from foreign market. Using
the cut-o¤ productivity levels obtained in the previous section, we can derive :
1
f ;
+1
=
where
= 1+(
+
By substituting the solution for
fr
fx
(7)
1
1)
1
(r
1
1
fx
:
f
1
1
1)
1
in equation (6) and using equations (3) (5), we can
obtain all the cut-o¤ productivity levels:
Note that
'd =
1 f
+ 1 fe
'b =
1 f
+ 1 fe
'x =
1 f
+ 1 fe
> M axf1;
1
(8)
;
1
1
1
r
1
fx
fr
1
1
;
1
1
(9)
1
1
(10)
:
1
1g from expression (1), so
+ 1 > 0. Finally, the price
index PT can be derived from equation (2):
PT =
1 f
+ 1 fe
1
1
f
L
!
1
1
:
(11)
When the variable exporting cost, , decreases, the cut-o¤ level of existing …rms, 'd ,
increases and the cut-o¤ level of exporting …rms, 'x , decreases. That means, there are fewer
…rms producing in the market due to more competition from foreign …rms but the fraction
of …rms exporting is larger because exporting …rms earn additional pro…ts from exporting
goods to foreign market.
Proposition 1 When a country opens up its economy, the cut-o¤ productivity level for …rms
choosing a high level of bribery, 'b , increases.
11
When
decreases, …rms face two changes: a reduction in their domestic pro…ts and an
increase in their foreign pro…ts. If their total pro…ts are lowered, they can choose to bribe
government o¢ cials more aggressively. However, the result in this speci…cation shows that
although the fraction of …rms choosing a high level of bribery out of the total number of
active …rms remains unchanged, higher productivity …rms choose to bribe more aggressively
when an economy opens up. That is because …rms have to choose whether to use a high
or low level of bribery before they have to choose whether to export or not. The bribery
decision in this case does not depend on business in foreign markets. Therefore, when the
competition is more aggressive in open economy and only more productive …rms can survive,
the cut-o¤ productivity level of …rms involved in a high level of bribery also increases.
Proposition 2 The bribery expense to …rms as a fraction of their total revenues decreases
when the exporting costs, , decrease.
Total revenues across all …rms increase since …rms face lower exporting costs. Because
the total bribery expense conditional on probability of survival remains unchanged, it takes
a smaller proportion of the total revenues …rms receive from their production. Therefore,
we can say that bribery level or corruption in the economy as a whole decreases when the
country liberalizes trades with foreign markets.
Proposition 3 If we assume that 'x < 'b , the cut-o¤ productivity level of …rms choosing a
high level of bribery decreases when
decreases. Since 'd increases due to more competition
in domestic market, the fraction of surviving …rms choosing to bribe aggressively increases.
However, the total bribery expense as a fraction of their total revenues still decreases since
total revenues increase more than total bribery expense does.
Based on the results derived in Bustos (2011), we have
=
where
= 1+
1
f ,
+1
1 fx
f
1
fx
+
f
(1 +
12
1
1
)(r
1
1
1)
(
1):
'd =
1 f
+ 1 fe
'x =
1 f
+ 1 fe
'b =
1 f
+ 1 fe
1
;
1
1
fx
f
1
;
1
(1 +
1
1
)(r
1
1
1
1)
We see that the assumption on the relationship between 'x and 'b changes some results
on the e¤ect of
on 'b but the …nal conclusion on the fraction of bribery expense out of total
revenues, or the corruption level in the economy as a whole, is the same. When a country
opens up its economy for trading with other countries, corruption tends to decrease.
5
Analysis on Asymmetric Countries
The discussion so far has assumed that the foreign country is identical to the home country.
Therefore, the cut-o¤ levels of surviving …rms, exporting …rms, and …rms with a high level
of bribery from the home country can be applied to the foreign country. However, when we
assume di¤erent characteristics of the home country, these results do not hold anymore.
5.1
Foreign Country Is Less Corrupt Than Home Country
First, we can look at the case where foreign country is less corrupt than home country. Then,
in the case of "coercive corruption", that means foreign …rms receive the bene…t of high level
of bribery in the home country without having to pay its costs. The total cost for foreign
…rms is
1
Let M =
1
IT
1
PT
1
g
qe;
T
C=f+
'
er
f denote
denote the market environment in home country and M
13
the market environment in foreign country. If 'x < 'b ; we can write the cut-o¤ productivity
levels as
Then,
'd
'
ex
=
'
ed
'x
=
h
'd = M f
f
M
'
ed =
f
r
r
1 fx
f
i
1
1
=
1
1
1
f (fx )
; 'x = M
1
1
;'
ex =
M
(fx )
r
1
1
1
1
:
> 1 as long as r < :
:
'
ex =
'x =
'd ;
'
ed:
We also have 'd < 'x and '
ed < '
e x . From Figure 2, the ordering of these cut-o¤
productivity levels is as follows
'
e d < 's < 'd < 'x < 'xs < '
ex;
where 's and 'xs are cut-o¤ level of survival and of exporting in the symmetric case.
Proposition 4 When foreign country is less corrupt than home country, home country has
fewer active …rms than in the symmetric case but a larger number of them are exporters.
Since …rms in foreign country do not have to pay a high level of bribery to have the cost
reduction, this market is less competitive. So, less productive …rms can survive in foreign
country. However, when these …rms want to export to home country, they face competition
with more productive …rms in home country. Therefore, it requires very high productivity
level of …rms in foreign country to make a pro…t from exporting to home country. On
the other hand, …rms in home country have to pay more bribes and face competition from
foreign …rms; so, the number of active …rms is less than that in foreign country. But,
14
those surviving …rms in home country can easily make a pro…t from exports due to a less
competitive environment in foreign country. The result also holds for the case where 'b < 'x
as long as r < :
Then,
'd
'
ex
= r ''e d =
x
h
'
ed
r
1 fx
f
1
f (fx )
; 'x = M
r
f 1
M
M
f 1; '
(fx )
=
ex =
r
r
'd = M f
i
1
1
=
1
1
1
1
1
1
:
:
'
ex =
'd ;
'x = r '
ed:
In summary, when trading with a less corrupt country, home country has fewer active
…rms but these …rms have a higher chance of being exporters. Similarly, when trading with a
more corrupt country, home country will have more active …rms but only the very productive
…rms can export.
15
Foreign Country Is Less Corrupt Than Home Country.
5.2
Firms From Foreign Country Pay Higher Exporting Cost Than
Firms From Home Country
In this section, …rms in foreign country have to pay more exporting cost when they want to
export to home country than when …rms in home country export to foreign country.
'd = M f
where
ff
'
ed = M
0
>
1
1
1
1
> 1:
f (fx )
; 'x = M
; '
e x = M 0 (fx )
16
1
1
1
1
;
We have
'd
'
ex
=
0
'
ed
'x
h
0
= ( )
i
1 fx
f
1
1
'
ex =
1
=
'd ;
'x =
0
where
<
:
0
:
'
ed =
0
'
ed;
From Figure 3, the ordering of these cut-o¤ productivity levels is:
0
'd < 's < '
ed;
'
e x < 'xs < 'x :
Proposition 5 When foreign …rms have to pay higher exporting cost, home country will have
more active …rms than in the symmetric case but a smaller number of them are exporters.
Market in home country is less competitive due to higher importing restrictions; so the
cut-o¤ level of survival in home country is lower than in foreign country. Market in foreign
country, on the other hand, is more competitive because home country’s …rms have to pay
less exporting cost to have their products in foreign country. Therefore, exporters in home
country are those with very high productivity level. On the other hand, exporters in foreign
country face less competition in home country but have to pay higher exporting cost; so it
is ambiguous whether '
e x is lower or higher than in the symmetric case.
17
Foreign Firms Face Higher Exporting Cost
5.3
Corruption in Asymmetric Countries
1
1
We have from Section (4.1) that 'b = 'd r 11 1
in the case of 'b < 'x and 'b =
h
i 11
in the case of 'b > 'x : So, when home country trades with a less
'd (1+ 1 )(r1 1 1)
corrupt foreign country, the cut-o¤ productivity level of …rms choosing a high level of bribery
in home country increases due to an increase in the cut-o¤ productivity level of active …rms.
However, given the probability of survival, the total bribery expense stays the same. On the
other hand, depending on the magnitude of the increase in 'd and the decrease in 'x , if total
revenues across home country’s …rms decrease, then corruption, measured by the fraction
of total revenues spent on bribery, in home country is higher than when two countries are
identical, and vice versa.
When foreign country’s …rms have to pay higher exporting cost than home country’s …rm
do when exporting to foreign country, the cut-o¤ productivity level of …rms choosing a high
level of bribery in home country decreases due to a decrease in the cut-o¤ productivity level
18
of active …rms. The total bribery expense increases, given the probability of survival. If total
revenues across home country’s …rms decrease, then corruption, measured by the fraction
of total revenues spent on bribery, in home country is higher than when two countries are
identical, and vice versa.
6
Conclusion
Economists have established that corruption, in general, is detrimental to economic growth.
It pushes economies even further away from e¢ ciency than they normally are. This paper
has shown the bene…ts of trade liberalization. When a country opens up its economies,
…rms’revenues from domestic market are reduced but they have an opportunity to export
to foreign countries and gain additional revenues. With the assumption 'b < 'x , only the
most productive …rms choose to bribe more aggressively. When we assume 'x < 'b , more
…rms choose a high level of bribery. However, since total revenues increases when bribery
expense remains unchanged (in the …rst case) or decreases by less than the increase in total
revenues (in the second case), …rms tend to spend a smaller fraction of their revenues on
bribery. Therefore, trade liberalization can help reduce corruption.
Moreover, the paper extends to the study of asymmetric countries. In particular, when
foreign country is less corrupt than home country, home country has fewer …rms but a larger
number of them are exporters. In the case where foreign …rms have to pay higher exporting
cost, home country have more …rms since it can protect its market but a smaller number of
those …rms are exporters. The conclusion on whether corruption is lower or higher than in
the symmetric case depends on how total revenues change due to the changes in the market
environment.
19
7
Appendix
A. The expected pro…ts include both the expected pro…ts in domestic market and foreign
market given that …rms are exporters:
=
d
x,
+ px
where px =
1 F ('x )
1 F ('d )
=
'x
'd
:
The expected pro…ts in domestic market is calculated as follows
d
=
IT
where 'd =
=
Z
(
1
(PT 'd )
'b
'
1
'd
+1
f
1 F ('b )
;
1 F ('d )
Z 1
r 1'
d' +
f(
1)
1
'
1 F ('d )
"
1
('d )
'b
1
1 + (r
1
1)
1
1
1
r
1
1
'
d'
1 F ('d )
1
+1 #)
1
1
1
The expected pro…ts from exporting is calculated as follows
x
=
Z
where 'x =
1
PT ' x
IT
1
1
r
fx ;
1
'
'x
1
1
1
=
So, we can obtain
1
'
d'
1 F ('x )
+1
r
1
('x )
1
1
:
as in equation (7).
B. In the case where 'b < 'x , total bribery expense is
B =
Z
1
1
'
d'
1)
1 F ('d )
f(
'b
= f(
1)
1
1
r
1
:
1
Total revenues across the economy are
TR =
+1
r
1
( "
f 1 + (r
1
1
1)
20
1
r
1
1
1
#
+ fx
fr
fx
1
1
1
)
:
Total bribery expense is independent of
when we take into account the probability of
survival. However, total revenues decrease in . Therefore, when
decreases, the fraction
of total revenues spent by …rms in the form of bribery decreases.
C. In the case where 'x < 'b , we have
Z
B =
1
f(
1)
'b
= f(
+ f + f(
1 fx
+fx
1
(1 +
1)
1
f
+1
TR =
1
'
d'
1 F ('d )
f
1
)(r
1
1)
1)
1
(1 +
1
;
1
)(r
1
1
1)
1
:
We see that B is a term in T R. Therefore, when
decreases, both B and T R decrease
but T R decreases by a larger amount. The end result is a decrease in the fraction of total
revenues spent for bribery.
D. From Section 4.2, we can write the expected pro…ts for …rms in foreign country when
it is identical to home country:
e0 =
IT
+
(PT )
IT
1
(
PT
@e
'd
=
@'d
1
(e
' )
+1 d
1
(
h
1
"
1
( 'd )
+1
1 + (r
1
1
1 + (r
1)
fr
fx
1
1
1
1
1
fr
fx
1
1
r
)
1
+1
1
1
r
1)
1
1
i
+1
1
1
1
1
#)
1
1
1
1
:
1
1
1
1
We see that the slope of the zero-pro…t curve for foreign …rms as a function of '
e d and
21
'd is negative and the curve is downward sloping. If more …rms can stay in business, then
it takes fewer …rms exporting to have expected pro…ts equal zero. Similarly, the expected
pro…ts for …rms in foreign country when it is less corrupt than home country are:
e =
IeT e
(PT r)
+
'd
'
ed
1
2
e
4 IT
+1
(e
'd )
PeT r
!
1
f
1
+1
( 'd )
1
3
fx 5 :
When foreign …rms do not have to involve in a high level of bribery, their expect pro…ts
are higher than before. The zero-pro…t curve shifts to the left, resulting in lower '
e d and
higher 'd compared to the original cut-o¤ productivity level of survival in the symmetric
case, 's . Similarly, when foreign …rms have to pay higher exporting cost, the zero-pro…t
curve for foreign country has a non-parallel shift to the right, as in Figure 3.
22
References
[1] Alberto Ades and Rafael Di Tella. "Rents, Competition, and Corruption". The American Economic Review , Vol. 89, No. 4 (Sep., 1999), pp. 982-993.
[2] Bustos, Paula. "Trade Liberalization, Exports, and Technology Upgrading: Evidence
on the Impact of MERCOSUR on Argentinian Firms". American Economic Review 101,
February 2011, 304-340.
[3] Djankov, Simeon and Sandra Sequeira. "An Empirical Study of Corruption in Ports".
April 2010.
[4] Dutt, Pushan and Daniel Traca. "Corruption and Bilateral Trade Flows: Extortion or
Evasion?" The Review of Economics and Statistics, November 2010, Vol. 92, No. 4, pp.
843-860.
[5] Ehrlich, Isaac and Lui, Francis T., Bureaucratic Corruption and Endogenous Economic
Growth. Journal of Political Economy, Vol. 107, No. S6, pp. S270-29, December 1999.
[6] Falvey, Rod, Greenaway, David and Yu, Zhihong, Extending the Melitz Model to Asymmetric Countries (May 2006). University of Nottingham Research Paper No. 2006/07.
[7] Fisman, Raymond and Roberta Gatti. "Decentralization and Corruption: Evidence
Across Countries." Journal of Public Economics, Vol. 83, No. 3, March 2002, pp. 325345.
[8] Goldsmith, Arthur A. "Slapping the Grasping Hand. Correlates of Political Corruption
in Emerging Markets." The American Journal of Economics and Sociology, Vol. 58, No.
4, October 1999, pp. 865-883.
[9] Mauro, Paolo. "Corruption and Growth". The Quarterly Journal of Economics, Vol.
110, No. 3 (Aug., 1995), pp. 681-712.
23
[10] Melitz, Marc J. "The Impact of Trade on Intra-Industry Reallocations and Aggregate
Industry Productivity". Econometrica, Econometric Society, vol. 71(6), p. 1695-1725,
Nov. 2003.
[11] Neeman, Zvika, Paserman, M. Daniele and Simhon, Avi, Corruption and Openness
(November 2004).
[12] Treisman, Daniel. "The Causes of Corruption: A Cross-National Study." Journal of
Public Economics, Vol. 76, 2000, pp. 399-457.
[13] Wei, Shang-Jin. "How Taxing Is Corruption on International Investors?" The Review
of Economics and Statistics, Vol. 82, No. 1, February 2000, pp. 1-11.
24