ms

gob? ' LIBRARY
Michigan State
University

 

 

 

 

This is to certify that the
dissertation entitled

FINACIAL CONSTRAINTS AND INTERNATIONAL TRADE FOR

HETEROGENEOUS FIRMS

presented by

Kwang-Myoung Hwang

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Economics

 

 

C h u n Z k u & #61 4 if:

 

Major Professor’s Signature"

(jam 2;, L009

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

A. -.-A-.-._._.-._.—.—.-.----—.-.—-.-.-_-—

-.-—.~A- --.—.--.—-— ---—o-.—.— —.--—.—---—...—.—--.-.-.—n...

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 Kthroj/Acc8PrelelRC/DateDue‘indd

 

FINANCIAL CONSTRAINTS AND INTERNATIONAL TRADE FOR
HETEROGENEOUS FIRMS

By

Kwang-Myoung Hwang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
ECONOMICS
2009

Professor Susan Chun Zhu

ABSTRACT

F INACIAL CONSTRAINTS AND INTENATIONAL TRADE FOR
HETEROGENEOUS FIRMS

By
Kwang-Myoung Hwang

This dissertation consists of three essays exploring international trade models for
heterogeneous firms regarding productivity and financial constraints. It also investigates
the effects of financial liberalization on trade patterns using this model.

In the fist chapter, I present a general equilibrium model of heterogeneous firms
in which firms differ in productivity and face financial constraints to pay fixed costs for
production. I consider how the relaxation of financial constraints for domestic sellers and
exporters affects firms' entry and export decisions. I have the following findings. (1) The
relaxation of financial constraints for domestic sellers helps less productive firms survive
in the domestic market, while driving less productive exporters out of the export market.
(2) The relaxation of financial constraints for exporters helps less productive firms
survive in the export market, while pushing less productive domestic sellers out of the
market. (3) Financial liberalization may help less productive exporters survive in the
export market, while driving less productive domestic sellers out of the market.

In the second chapter, I examine empirically the impacts of financial liberalization
and the relaxation of financial constraints on exporters and domestic sellers. Unlike the
existing literature, I use a ratio of exports to domestic production instead of export
volumes as the dependent variable. I find that financial liberalization and real interest
rates (a measure of costs of external funds) are important determinants of international

trade patterns. In particular, financial liberalization and a reduction in the real interest rate

increase the ratio of export to domestic production disproportionately more in industries
with a higher degree of external finance dependence. In addition, I find that the effects
are greater in financially under-developed countries. I also find the role of changes in
costs of external financing in making export decisions, which implies that changes in
costs of external credits have different effects on exporters and domestic sellers. My
finding supports that exporters have lower contractibility than domestic sellers because of
their higher risks in exporting.

In the last chapter, I present a heterogeneous firm model in which firms differ not
only in productivity but also in financial constraints. My model is general in the sense
that it incorporates many factors affecting firms' variable costs into the heterogeneity in
productivity levels, and those affecting firms' fixed costs into the heterogeneity in
financial constraints. I show that firms with low productivity levels and severe financial
constraints will immediately exit the market while firms with high productivity levels and
few financial constraints can stay in the market. I also show that even if the fixed costs
for exports are bigger than those for domestic sales, it does not imply that only high
productive firms can export. Moreover, I show that firms make different decisions on
exports and domestic sales even when they have the same productivity levels. My model
has the strength in explain the stylized fact in international trade that some firms with low
productivity levels are exporters while some firms with high productivity levels sell only
domestically. It can also explain the extreme case that some exporters do not sell

domestically.

ACKNOWLEDGEMENTS

First, I would like to thank Professor Susan Chun Zhu for serving as my advisor
since my second year at Michigan State University. I will never forget her words of
advice about my research and the first time we met. She said to me that “You have to
prepare for your dissertation during the class. If you have any idea, please don’t hesitate
to contact me”. Whenever I came to see her with a vague idea, her comments were direct,
which encourage me to study the literature in the frontier deeply. Kindly, she accepted to
be my adviser, and discussed on my idea several hours whenever we met. Without
doubts, most of my idea in the dissertation came from the discussion. Her deep
knowledge of the literature in international trade is impressive and inspiring, and her
suggestions for my research have always greatly improved my dissertation.

Professor Raoul Minetti is so kind to participate the discussion even if he was
very busy with many works of his own. His perception always pointed out weak links of
the structure of my model. I met him in my first year as a student of his macro-economic
class, but I am more impressed with his perception on finance. My dissertation is quite
related to international finance, which I had little chance to deal with. Without his advice,
it would have taken me a long time to complete the dissertation.

I would like to thank Professor Steven J. Matusz for his kind help. In addition to
his valuable advice on my papers, he allows me to have experience on teaching students.
He teaches me many skills about online class. Furthermore, his advice on how to deal
with undergraduate students is of great help for me who was a beginner in this field. His

guidance leads me to the position as an instructor, which is the greatest experience in my

life.

I'also thank Professor Charles J. Hadlock in the finance department. He
generously accepts to be the committee member, and suggests much idea regarding the
models. Especially, his kind words always encourage me to go on further with
confidence.

I would also like to thank Professor Carl Davidson and Paul L. Menchik for their
kind help in managing my class. They also allow me to have an experience as an
instructor, which improves me a lot not only academically but also financially. I also
thank Margaret Lynch and Jennifer Carducci for their continuing support during my
school years.

A number of other people have been extremely helpful along the way. Especially,
I owe a great debt to my classmates; Jaesoo, Young-Gui, Nicolas Sly, Wei-Chih, Meng-
Chi, Terry-Ann, Lenisa, Nicole, Jingiing, Jongwoo, Seek-Woo, Youngjoo, Sung-Guan,
Dong-Ryeol, and Kyungmin. The many hours that we spent studying and enjoying
together in MSU certainly paid off.

Finally, none of this would have been possible without the support of my wife
Un-Joo and my children, Jun-Ho and Seon-Ho. I also thank my parents, Keum-Chun
Hwang and Jung-Ia Kim, and my parent-in-laws, Nae-Moon Lee and Song-He Lee, for

their unconditional love even if they are far away.

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... vii
LIST OF FIGURES ......................................................................................................... viii
CHAPTER ONE: THE HETEROGENEOUS FIRM TRADE MODEL WITH
FINANCIAL CONSTRAINTS ...................................................................................... l
1. Introduction ..................................................................................................... 2
2. A Heterogeneous Firm Model with Financial Constraints .............................. 6
3. Comparative Statics ......................................................................................... 12
4. An application: the Effects of Financial Liberalization ................................... 16
5. Conclusions ....................................................................................................... 22

CHAPTER TWO: ESTIMATING THE EFFECTS OF FINANCIAL

LIBERALIZATION ON EXPORTERS AND DOMESTIC SELLERS ........................ 24
I. Introduction ..................................................................................................... 25
2. A Glance of Data ............................................................................................. 28
3. Empirical Specifications .................................................................................. 31
4. Empirical Results ............................................................................................. 34
5. Conclusions ....................................................................................................... 40

CHAPTER THREE: A TWO-DIMENSION HETEROGENEOUS FIRM MODEL... 48

1. Introduction ..................................................................................................... 49
2. The Model for the Closed Economy ................................................................ 51
3. The Model for the Open Economy .................................................................. 59
4. Conclusions ....................................................................................................... 66
APPENDICES ................................................................................................................ 73
1. Nash Bargaining Solution ............................................................................... 74
2. Comparative Statics ........................................................................................ 76
3. Partial External Financing ............................................................................... 82
4. Collaterals ....................................................................................................... 83
5. Existence and Uniqueness of the Equilibrium Solution .................................. 84
REFERENCES ............................................................................................................... 87

vi

LIST OF TABLES

2-1 The Effects on Trade patterns: Basic Results ............................................................ 42
2-2 The Effects on Trade patterns: Switchers Only ......................................................... 43
2-3 Long-term Analyses ................................................................................................... 44
2-4 Controlling Trade Openness ...................................................................................... 45
2-5 Comparison with Cases that use Trade Volumes ...................................................... 46
2-6 Robustness Checks .................................................................................................... 47

vii

 

LIST OF FIGURES

3-1 Ratio of Plant Labor Productivity to Overall Mean ................................................... 67
3-2 Ratio of Plant Labor Productivity to 4-Digit Industry Mean .................................... 68
3-3 The ZCP condition in the Closed Economy ............................................................. 69
3-4 The ZCP condition in the Open Economy ................................................................ 70
3-5 The Distribution of Productivity and Financial Constraints ..................................... 71
3-6 The Distribution of Financial Constraints at the Given Productivity ....................... 72

viii

Chapter 1: The Heterogeneous Firm Trade Model with

Financial Constraints

1. Introduction

There is an emerging literature on financial constraints and international trade.
Mostly, they augment the Melitz (2003) model which emphasizes the role of produc-
tivity in firms’ export decisions. Chaney (2005), for instance, focuses on liquidity
constrained exporting firms, assuming that firms inherit an exogenous amount of liq-
uidity. Manova (2006) assumes that firms can only partially finance the fixed costs
of trade internally with fractions differing across industries exogenously. Garcia-Vega
and Guariglia (2007) analyze the effects of firms’ exogenous income shocks on their
probability of survival and their decisions to enter export markets.

I follow the basic idea of Manova (2006) in constructing the financially constrained
heterogeneous firm model. She explains the effects of financial constraints in a micro—
based way. She argues that credit constraints interact with firms’ heterogeneity and
reinforce the selection of the most productive firms into exporting. Her intuition
is that more productive firms can offer creditors a greater return in the case of re—
payments and are more likely to secure the outside capital necessary for exporting
because they raise higher revenues.

In her model, credit constraints have different effects on firms in different countries
and sectors. Credit constraints vary across countries because contracts between firms
and investors are more likely to be enforced in more financially developed countries.
In addition, sectors differ in their endowments of tangible assets that can serve as
collateral. If a financial contract is enforced, a firm makes payments to investors;
otherwise the firm defaults and creditors claim collaterals. Firms, therefore, find
it easier to obtain external finance in countries with a high level of financial con-
tractibility and in sectors with large endowments of tangible assets. Therefore, the
productivity cut-off value for exporting varies systematically across countries and
sectors. It is higher in financially vulnerable industries which require more outside
finance or have less collateralizable assets, and is lower in countries with high levels of

financial contractibility or with many collateralizable assets. The effects of financial

2

 

development are more pronounced in financially vulnerable sectors.

However, not all firms or industries face the same amount of financial constraints.
Chaney (2005) argues that more productive firms are less likely to be credit con-
strained because they generate large liquidity from their domestic sales. Therefore,
these productive or wealthy firms that inherit a large amount of liquidity are more
likely to export. However, he does not examine the implication of financial constraints
in domestic sectors.

In reality, every firm or sector confronts different degrees of financial constraints.
Thus, questions arise in which firms or sectors and how much their financial con-
straints change. Will the relaxation of financial constraints in some firms or sectors
make favorable effects on the other firms and sectors? To answer these questions, it
is necessary to analyze the effects of financial constraints with a general equilibrium
framework. 1

This chapter mainly focuses on how the relaxation of financial constraints in some
sectors affect the decisions of firms in other sectors on entry and export. I divide
sectors into two categories; one for domestic sellers and the other for exporters.1 In a
general equilibrium framework with heterogeneity in firms’ productivity, I show that
changes in financial constraints in each sector have different effects on cut-off produc-
tivity levels of entry and export. Specifically, the relaxation of financial constraints
for domestic sellers helps less productive firms survive in the domestic market, while
driving less productive exporters out of the export market. The relaxation of financial
constraints for exporters, however, helps less productive firms survive in the export
market, while pushing less productive domestic sellers out of the market.

My model can be applied to analyze the effects of financial liberalization. Fi-

nancial liberalization is one of the most interesting topics these days. It serves as a

 

lThe degrees of financial constraints differ between exporters and domestic sellers because of
different loan contracts such as export subsidy. Different amount of information also leads to differ-
ent degrees of financial constraints between exporters and domestic sellers. For example, if foreign
investors have more information on exporting firms than non-trade firms, contractibility or collat-
eralizability of exporters would become higher after financial liberalization.

3

 

source of exogenous changes in financial conditions. Levine (1997) classifies channels
that financial liberalization improves firms’ financial conditions. The improvement of
firms’ financial conditions would affect their decisions on entry and export. He argues
that financial liberalization would lead to financial development in the home country
through 1) facilitating the risk amelioration by trading, hedging, diversifying, and
pooling risks, 2) acquiring information about investments and allocating resources,
3) monitoring managers and exerting corporate control, 4) mobilizing savings, and
5) facilitating the exchange of goods and services. Due to these channels, finan—
cial development can be accomplished, which further reduces firms’ costs of external
financing.

Demirguc—Kunt and Levine (1996), however, introduce the other channels in stock
markets through which financial liberalization makes contrary effects. 1) Greater liq-
uidity in stock markets after financial liberalization may reduce savings rates by
increasing the returns to investments and by making negative impacts on uncertainty
because less uncertainty may decrease the demand for precautionary savings. 2) Due
to the euphoria and myopia that may be encouraged by highly liquid stock markets,
dissatisfied participants find it easy to sell quickly, which can lead to disincentives
to exert corporate control, affecting adversely corporate governance and hurting eco-
nomic growth.

The above literature is all based on the assumption that all sectors and firms are
affected at the same time, in the same direction, and to the same extent because of the
changes in financial constraints due to financial liberalization. The effects of financial
liberalization on firms’ financial constraints, however, can differ across sectors and
times. Thus, previous models cannot analyze the precise effects of various changes in
financial constraints on firms’ decisions on entry and export. Therefore, it is useful
to consider the changes in financial conditions in specific sectors, and their effects on
other sectors after financial liberalization.

Financial liberalization can be interpreted as free flows of credits among countries.

 

Free flows of credits induce two different effects on domestic financial markets. One
is to change firms’ financial constraints. If foreign investors have more information
on exporting firms than domestic investors, contractibility and collateralizability of
foreign investors regarding exporters will be higher than those of domestic investors.
This will lead to lower repayments of external financing to foreign investors, and they
can start to invest in exporting sectors by making contracts with exporters.

The other effects of financial liberalization are changes in the return to capital.
Due to changes in the amount of credits, the return to capital changes after the lib-
eralization. There is consensus in the literature that the return to capital decreases
after financial liberalization, especially in developing countries. [See Bacchetta (1992)
and Henry (2000a)] Bacchetta (1992) shows that a likely outcome of financial liber-
alization is an initial net inflow, followed by an outflow. His intuition is that the
liberalization of financial markets leads to better allocation of resources and makes
the country more attractive to both domestic and foreign investors. Thus, there are.
few incentives for outflows and strong incentives for inflows at the beginning. There-
fore, the return to capital will decline after financial liberalization.

I show that in both cases, financial liberalization will help less productive ex-
porters survive in the export market, but keep less productive domestic sellers out
of the domestic market because financial liberalization induces higher contractibility
and collateralizability in financial markets for exporters due to the participation of
foreign investors. I also Show that the number of exporters relative to domestic sellers
decreases with the contractibility of exporters but increases with their collateralizabil-
ity when the return to capital decreases. It is generally accepted that contractibility
of exporters is lower than that of domestic sellers. Therefore, financial liberalization
will increase the exporting sector while decreasing the domestic sector.

This chapter makes two contributions to the literature on international trade.
First, the existing literature largely assumes that changes in financial conditions hap-

pen in all sectors, at the same time, and to the same extent. In contrast, in this

0"!

chapter, I consider the case where financial constraints are relaxed only in some sec-
tors, and analyze the impacts on the other sectors in which financial constraints are
not relaxed. Thus, my analysis provides a general equilibrium view of the impacts of
financial constraints on production and exports. Second, my model sheds new light
on the effects of financial liberalization on different sectors in the economy. Specifi-
cally, financial liberalization reduces interest rates, through which raises the number
of firms in both domestic and export sectors. In addition, the number of exporters
relative to domestic sellers decreases with contractibility of exporters but increases
with their collateralizability when overall interest rates decrease.

The structure of this chapter is as follows. In section 2, I present a model of het-
erogeneous firms which incorporates financial constraints to exporters and domestic
sellers, and do comparative statics in section 3. In section 4, I apply the model to

analyze effects of financial liberalization, and conclusions follow in section 5.
2. A Heterogeneous Firm Model with Financial Constraints

I incorporate financial constraints to the Melitz (2003) model. Consider the home
country and 72. foreign countries. A continuum of heterogeneous firms produces differ-
entiated goods in each country, and varieties produced by country j are distinct from
those produced by country 2'. Consumers exhibit love of variety and can consume all
available differentiated products. The utility function for the representative consumer
in each country is given by the CES preference function with a constant elasticity of
substitution (0), which is greater than one.

The problem of the representative consumer in the home country can be written

maxU 2 ”cold? q(w)(‘7‘1)/wa]0/(U‘1), subject to R. = LEQMMqWMw, (1)
where U and R represent her total utility and revenue. p(w) and q(w) are the price

of and the demand for each good.

The utility maximization implies that the demand for variety w is

6

 

q(w) = Ri%t¥a—l, and P = lfwenP(w)1""dWI1/(""’, (2)

where P is the price index.

As in Melitz (2003), firms incur a sunk cost to enter a market. Then, they learn
their productivity levels, and make entry and export decisions. Production requires
only one input, labor (I), where wage is normalized to one. Each firm uses 1rd unit
of labor for investment, which is necessary prior to production. Production for the
domestic market involves constant variable costs, which are lower in more productive
firms. Thus, q/cp units of labor are necessary to produce q units of goods, where «,9
represents a firm’s productivity level.

According to the optimal pricing rule, the individual price and profit for the firm

with productivity level 90 can be derived as

pea) = ,1; (3)
«(a = iRlear-l—kd.

where % E #:771— is the mark—up.

An equilibrium is characterized by a mass of firms M and the distribution of
productivity level u(<p) over a subset of (0,00). The aggregate price level can be

written in terms of A! and the average productivity level.

P = [me P(99)”"Mu(e)del’/(1””) = 111’“”"’I)(39), (4)

where 95 = [[000 990—1u(<p)d99]1/("‘1) is the average productivity level.

To enter, firms must make an initial investment as a fixed entry cost kg > 0,
which is thereafter sunk. They also face a constant probability 6 in each period of a
bad event that will force them to exit. The 6 is independent of firms’ productivity.
Therefore, an entering firm with a productivity level «p will immediately exit if the
expectation of profits is negative, or will produce and earn non-negative profits in
every period until it is hit with a bad event, and is forced to exit. Assuming that

there is no time discounting, each firm’s expected value function is given by

7

 

vac) = maxio,t§_°30<1-6>tn(a} ——— maxi‘o. are». (5)

If the exit procedure does not affect the equilibrium productivity distribution,
this distribution must be determined by the initial productivity draw, conditional on

successful entry. Hence, the conditional distribution of 9(99) can be derived as

W) = ,—j’—§,%r, w _<. «.2. (6)

= 0 otherwise,
where (p' is the cut-off productivity level of entry and pm E 1 — C(tp’) is the ex—ante

probability of successful entry.

The average productivity level {,3 is a function of the cut-off productivity level 99’ .

~

9009’) = [p—G’m f3? e””’g(e)del’/(”‘1). (7)

The economy has 71 identical foreign countries. Ex ante, firms in each country have
an identical productivity distribution. Consumers share the same utility function.
These symmetry assumptions ensure the same wage rate, which is still normalized
to one. There is an ice-berg transportation cost 7‘, which represents trade resistance
costs, where 7' > 1. For simplicity, I do not allow transportation costs to differ
across countries. I assume, however, that an additional fixed cost kg; is required for
exporting. It can be shown easily that if ra'lkx > kd, low productive firms exit the
markets immediately, intermediate productive firms only sell domestically, and high
productive firms sell both in the domestic market and the export market. When firms
produce for the export market, there are no changes in their productivity levels.

Once again, the ex-ante probability of successful entry is Pin E 1 — G (cp' ). More-
over, px .=_ [1 - G (9933)] / [1 — G(<p')] now represents the ex-ante probability that one of
these successful firms will export. <pr is the critical productivity level for exporting.
The mass of exporting firms in the country will be M5,; 2 pg; M, and the total number

of varieties available to consumers in any country will be Mt = [W +nM 1.. The average

 

productivity level of domestic sellers and exporters (from foreign countries) can be

written as

a = {filMeo—l+nMx(T%)"‘ll}1/<°—1>, (8)

where $3: = [1—517]; f2: 990—19(99)d99l’/("“1)-

The way of incorporating financial sectors is similar to that of Manova (2006). I
assume that variable costs are covered by internal finance, which implies that firms
can compensate their production costs by selling their goods. Fixed costs (led and km)
however, cannot be covered by internal finance, and firms should make loans from
financial sectors before making investments.2 In the following, subscript d and :1:
represent the domestic sector and the export sector respectively. In obtaining outside
finance, firms pledge tangible assets as collateral. I also assume that a fraction (rd
and tx) of fixed costs goes towards collateralizable assets, (e.g. plants, properties,
and equipment), which depend on properties of sectors and goods. For example, the
fraction of collateral in the manufacturing sector may be greater than that in the
service sector because investors can obtain inventories in the manufacturing sector
when firms default.

Financial institutions cannot make perfect enforcement in terms of contractibility
because of imperfect information. In particular, investors can expect to be repaid with
probability Ad and A15, which are exogenous and less than one when they make loans
to domestic sellers and exporters. With probability (1 — A), financial contracts are
not enforced, firms default, and creditors claim the collaterals (tdkd and txkx). Even
though investors cannot make perfect enforcement because of imperfect information,
they know in which sector, the domestic one or the export one, the investments have
been made.

Financial contracts proceed as follows. At the beginning of each period, every

 

2The underlying assumption is that firms cannot use profits from one period to finance future
operations. This assumption can be justified if firms cannot retain earnings but have to distribute
all profits to shareholders at the end of each period.

9

firm makes a take—it-or-leave-it offer to potential investors.3 This contract specifies
the amount of credits that a firm needs to borrow, its repayments [Fd(<p) and 152502)]
when the contract is enforced, and the collaterals [tdkd and txkgg] in the case of
default. Revenues are then realized and investors receive the repayments at the end
of the period.

The problem of profit-maximizing firms for domestic sales and exports is

max YQHW)=10d(<i9)q(1(90) - WWW? - MFA?) - (1 - )‘dltdkd
+ni1):(99z)qv(99)" (117(91)/Y_)‘ITFJ:(99)— (1 —/\;l‘)t.'lfk£l‘j7 (9)
subject to 1) ([4 ($9): 41731?
(99): —T—; ’0’?

palqum - qdw )/e 2 Fave)

( ) ( )-Tr11()/992 P‘s-(99)
46+ MFA )+ (1 * Adm/Yd 2 0
E— k, + A F ( )+(1— A1.)r,,.k;,, 2 0.

PI?

2(1)

3) Adv?)
4) x09)
5) Bdfifi)
619) 2:09)

All firms maximize their profits subject to the demand function in equation (2).

With external financing, two additional constraints bind firms’ decisions. When the

financial contract is enforced, an entrepreneur can offer at most her net revenues,

which are represented by Adl'kP) and Adm), to creditors. In addition, investors only

extend finance to the firm if they expect a non-negative return. Bd(go) and 817(99)

represent net returns to creditors. Equations (5) and (6) express investors’ participa-

tion constraints with their outside options normalized as zero. These constraints will

be generalized in section 4.

With competitive credit markets, all investors break even and make zero expected

profits.4 A firm adjusts its repayments so as to bring investors to their participation

 

3The assumption of a take-it-or-leave-it offer does not affect the results. See appendix A for the
Nash bargaining solutions.

lln the case of uncompetitive markets, the outside options will not. be zero. However, the results
do not change qualitatively.

10

 

constraints. Thus, Bd(<,9) = B$(<p) = 0 in equilibrium, and the maximization problem
reduces to the firm’s problem in the absence of financial frictions, except for credit
constraints that repayments are not greater than the firm’s net revenues. Each firm
optimally chooses the same quantities and prices, raises the same revenues and earns
the same profits before repayments as in Melitz (2003).

Therefore, the solution for the above problem is the same as one without finan-
cial constraints except substituting repayments [Fd(cp) and Fm(<,9)] for fixed costs [kid
and km] in firms’ ordinary maximization problems. Constraints (5) and (6) in the
maximization problem are binding because of competitive financial markets. Thus,
repayments of domestic sellers and exporters do not depend on firms’ productivity

levels, and can be written as

1— 1—A t k _' _,\..‘_,,/,..
Fd<99):[ ( A("""d,and Fz(e)=[’ ” Milfl‘l. (10)

 

 

The equilibrium solution for critical productivity levels for entrance and export
can be derived from two conditions. One is the zero cut-off profit condition (ZCP),
and the other is the free entry condition (FE)5

The ZCP condition for incumbent firms implies that all firms should have. non—
negative profits in order to stay in the market. The ZCP condition implies 7rd(t,9') = 0
in the domestic market and 7r(99§,) = 0 in the export market. These conditions can be
written in terms of the average productivity level because of the relationship between
the average productivity level and cut-off productivity levels for entrance and expert.

The ZCP condition can be written as

”(450 = ”(1(§)+Prmix(<5x) = FdiAdvtdvk'(1)f(99,)+P.1:7lFx()‘ratrak.77)f(99.’r)i (11)
where f (99’) = [CD/993"" - 1-

The free entry condition can be found from expected average profit flows of in-

cumbents. For a potential entrant to make investment in order to enter the market,

\

5see Melitz (2003) for details.

11

 

the expected value for the entrant should be equal to zero. If the expected value for
the potential entrant is less than zero, no firm will enter the market. On the other
hand, if the expected value for the potential entrant is greater than zero, all potential
firms will enter the market. Both cases cannot be the equilibrium. Let a = (1 /6)7r((79)
represent the present value of expected average profit flows. Further define we as the

net value of entry. The FE condition can be written as

198 .—_ Wit—k... = 0 4:» was) = F6557? (12)

Finally, the equilibrium cut-off productivity level can be derived from the two

conditions. From equations (11) and (12),

if? = Fdw) Matures.» where f(e’) = [WW—1 —1 (13)

There is a unique solution for the equilibrium cut-off value because the ZCP curve
is decreasing in (0 while the FE curve is increasing in 99.6 In a stationary equilibrium,
all aggregate variables must remain constant over time. This requires a mass Me of
new entrants in every period, such that they exactly replace the mass of incumbents
who are hit with bad events and exit the market; pin Me = (MI. The equilibrium (,9’
determines the export productivity cut-off level (,9’, as well as (,9, C93,, (5}, and ex-ante

probabilities of entry and exports (pm and 193;) because (,9’, is defined as a function of

I

90-
7" ‘”, 99’, 0— F F); a—
FIT;— )) =l;1r,l 1 — JP}, 4: 993; = ‘7 w/[+Q]1/( 1’ (14)

Market clearing conditions for labor and goods lead to the aggregate labor supply
L 2 LP + Le, where Lp and Le represent the aggregate labor used for production
and the amount of investments by new entrants respectively. According to the ZCP

condition, aggregate profits (II) equal to L9, and R = L.

3. Comparative Statics

 

“See Figure 1 in Melitz (2003) for details.
12

_I

-——-
.‘

0 Proposition 1: The relaxation of financial constraints for domestic sellers helps
less productive firms survive in the domestic market. However, it keeps less productive

exporters out of the export market.

_ . dF [1—1: )1: dF _ —[1—.\ lie
Proof. From equatlon (10), jg: __/\2d_d < 0. id —— ——j;d—d < 0. From

d
I I I) ,~(r—1 70—1 1—0 .I . ~a—1
* : 'f [1“ ”W )Iilfl “#0 l+T I1_G(.¥9:r)l”ii91:l 0
' {‘7‘ } [1- C(so’)lel¢l"”+T1_Ul1-G(so.’1:)lenls'oEl”” > '

I 0— 1 I I
_ so <93;- 11- G(so)l $1..
{7%}70‘1l1—Glso’lF mo 1+l1— Ge: ’)lFa:nl«;~§l" 1<0'Th“s’d\d “

’dF d ’ d dF da’. dF (1’. dc’ dF
dailid)’ td Zaficitd< O’did 317: d>0’and77i: Fcit771d>0

equation (13),

fig

Eli?

 

(Sip;

The relaxation of financial constraints for domestic sellers can be represented by
higher contractibility and more collateralizability in financial markets for domestic
sellers, which implies high Ad and td. If Ad and td increase, repayments of domestic
sellers to investors (F d) will decrease. Due to competitive financial markets, the break
even point for investors goes down as contractibility or collateralizability increases.
Therefore, contracts are made with lower repayments for domestic sellers. If Ad and tar
increase, (,0’ decreases while (,9; increases. This implies that the relaxation of financial
constraints for domestic sellers helps less productive firms survive in the domestic
market. However, it keeps less productive exporters out of the export market, and
makes them focus only on the domestic market.

It is not counter—intuitive that the relaxation of financial constraints for domestic
sellers helps less productive firms survive in the domestic market. Higher contractibil-
ity and more collateralizability in financial markets for domestic sellers lower their
repayments for external financing. Thus, less productive firms, which earned negative
profits in the domestic market before, can repay to investors, and finally survive in
the market.

The cut-off productivity level for exporters, however, increases, which implies that
lower productive exporters give up the export market and sell only to the domestic
market. They still have high productivity levels, compared to domestic sellers. Hence,

it will not happen for less productive exporters to exit both markets at the same time

13

 

due to the relaxation of financial constraints for domestic sellers.

Two forces drive these results. One is the increase of profits in the domestic market
for incumbents. Due to the entrance of less productive firms into the domestic market,
firms can earn more profits in the domestic market. The aggregate price level in the
domestic market increases due to the entrance of less productive firms. The demand
for each good increases because of the inverse relationship between the demand and
the aggregate price level. Therefore, profits from domestic sales increase, and the
cut-off productivity level for domestic sellers decreases.

The other force is due to the increase of labor costs. The demand for labor
increases because of new entrants in the domestic market. Wages increase as a con-
sequence. The increase of the number of firms reduces the amount of labor that
goes to an individual firm because the labor supply in the country is fixed. Thus,
less productive exporters, who are not profitable enough to pay high wages, give up
exporting. In other words, less productive firms, which barely earned profits in the
export market before, exit the export market, and focus only on the domestic market
because of increased profits in the domestic market and the inability to obtain inputs
for exporting.

In brief, lower repayments for domestic sellers encourage potential entrants to
enter the domestic market. Domestic sellers are competing with exporters in the
labor market. The increase of the number of domestic sellers increases the degree
of competition in the labor market, and the exporters with low productivity levels

become losers.

0 Corollary 1: The effects of the relaxation of financial constraints for domestic
sellers on cut-off productivity levels are greater under lower contractibility, lesser

collateralizability, and a smaller elasticity of substitution between varieties.

This implies that changes in financial constraints have strong effects when finan-

cial constraints are severe. They also produce strong effects when the elasticity of

14

substitution between varieties is smaller. A smaller elasticity of substitution means
a higher degree of product differentiation. If products are highly differentiated, it
makes it easier for less productive domestic sellers to survive when their financial
constraints are relaxed. Consider new entrants with low productivity levels. Their
prices are high, compared to the aggregate price level. If their products are highly
differentiated, however, their high prices will not reduce the demand for their goods
much. This makes it easier for new entrants with low productivity levels to survive

in the market.

0 Proposition 2: The relaxation of financial constraints for exporters helps less
productive exporters survive in the export market. However, it keeps less productive

domestic sellers out of the market.

Proof: From equation (10), Q = :kiyL—lk—‘E < 0. (#54: M— < 0.

H
All?

I0- 1
99 I1 Gii’rlnl
Fromequation(13), 349—: —{J0—I} [1—G(,o~’)]Fd[,E]U*I+TI— U[1_ C(so ,7, 1<0.
l1——G(soé,)lnln{ls7cl” 1— ’0 lie” ’ll— 0( oils." 1
={—1’}i [1 Gm: ]F_1;n[(,9$]0 1+70 1l1 Cf“ )leL'liWr 1

3:de d ([99. _dv:r dFir . dfir _ .r dF
021%: ale—fl? 37ft >0’d.\""f 3?;de <0>andUG—ZIT;?IQ£<O-

 

'C‘I \
s?
T
L
Q
l

 

32:31.?

I
{1-9 _
> 0. Thus, (Th—Ag; --
(1,9,

\

The relaxation of financial constraints for exporters is represented by higher con-
tractibility and more collateralizability in financial markets for exporters, which im-
plies high Am and tr. If Am and tag increase, repayments of exporters to investors (Fm)
decrease. Due to competitive financial markets, the break even point for investors
goes down in the case of higher contractibility and more collateralizability. Therefore,
contracts can be made with lower repayments for exporters. If A3; and t3; increase, (,9,
increases while (p; decreases. This implies that the relaxation of financial constraints
for exporters helps less productive firms survive in the export market. However, it
keeps less productive domestic sellers out of the market.

It is not counter-intuitive that the relaxation of financial constraints for exporters

helps less productive firms survive in the export market. Higher contractibility and

15

more collateralizability in financial markets for exporters lower their repayments for
external financing. Hence, potential exporters, which would earn negative profits in
the export market before, can repay to investors and finally survive in the export
market.

The cut-off productivity level for domestic sellers, however, increases, which im-
plies that lower productive domestic sellers give up production. When financial con-
straints of exporters are relaxed, lower productive firms enter the export market due
to lower repayments in export financing. The production for exporting increases.
However, the labor supply for production is fixed in the country. The increase of
exporting goods increases the demand for labor, which increases the wage rate. As
a result, lower productive domestic sellers exit the market because they cannot be
profitable enough to pay high labor costs for production. In other words, the wage
rate increases because new exporters try to obtain more units of labor, and lower pro—
ductive domestic sellers, which cannot afford the increase of labor costs, earn negative
profits and exit the market.

'In conclusion, lower repayments for exporters encourage domestic sellers with high
productivity levels to enter the export market. Exporters are competing with domes-
tic sellers in the labor market. Therefore, the increase of the number of exporters
increases the degree of competition in the labor market, and the domestic sellers with

low productivity levels become losers, give up production, and exit the market.

0 Corollary 2: The effects of the relaxation of financial constraints for exporters
on cut-off productivity levels are greater under lower contractibility, lesser collateral-

izability, and a smaller elasticity of substitution between varieties.

This implies that changes in financial constraints have strong effects when finan-
cial constraints are severe. They also produce strong effects when the elasticity of

substitution is smaller. The intuition is similar to that of corollary 1.

4. An application: the Effects of Financial Liberalization
16

The above results can be applied to analyze the effects of financial liberalization
on international trade. Financial liberalization has broad aspects. For example,
Kaminsky and Schmukler (2003) consider three dimensions of financial liberalization:
(1) in a liberalized capital account economy, banks and corporations are allowed
to borrow from foreign countries freely; (2) a liberalized domestic financial system is
characterized as the lack of controls on lending and borrowing interest rates as well as
the lack of credit controls; (3) a liberalized stock market implies that foreign investors
are allowed to hold domestic equity without restrictions and capital, dividends and
interest can be repatriated freely within two years of the initial investment.

In this chapter, my analysis focuses on the first aspect of financial liberalization-
the deregulation of the foreign sector capital account. Due to the liberalization,
foreign investors can invest in loan markets in the home country without any govern-
mental intervention. Domestic investors also can invest in foreign financial markets
without any intervention. These free movements of credits among countries induce
two different effects on domestic financial markets.

One is to change firms’ financial constraints because of the severe competition
between domestic investors and foreign investors. Kletzer and Bardhan (1987), Beck
(2002), Matsuyama(2004), Ju and Wei(2005), and Becker and Greenberg(2005) show
that different financial conditions become a source of the comparative advantage in
the presence of credit constraints. For example, Kletzer and Bardhan (1987) Show
that countries with relatively well-developed financial sectors have a comparative
advantage in industries and sectors that rely more on external finance. Beck (2002)
sets up an open economy model with two production technologies, one with constant
returns to scale and the other with increasing returns to scale. He shows that the
sector with scale economies makes more profits than the other sector in countries
with a higher level of financial development. Therefore, more financially developed
countries have a comparative advantage in the sector with scale economies, and are

net exporters of goods that are produced in the sector with scale economies.

l7

Different financial constraints can be interpreted as differences in contractibility
and collateralizability in financial markets. For simplicity, I assume that contractibil-
ity and collateralizability of domestic investors do not change after financial liberal-
ization.7 As financial liberalization proceeds, many foreign investors with different
contractibility and collateralizability will enter domestic financial markets. Because
foreign investors have more information on exporting firms in the home country than
on non-exporting firms, at the beginning of financial liberalization, foreign investors
will enter the financial market for exporters first.

If contractibility and collateralizability of foreign investors for exporters are lower
than those of domestic investors, it will not have any effect. There is no incentive
for exporters in the home country to sign contracts with foreign investors because
foreign investors require higher repayments due to their lower contractibility and
collateralizability. Therefore, there will be no inflows of credits or changes in firms’
financial constraints regardless of financial liberalization, which implies no effects on
cut-off productivity levels for entry and export.

If foreign investors have more information on exporting firms than domestic in—
vestors, however, they will begin to invest in the export sector. If contractibility
and collateralizability of foreign investors are higher than those of domestic investors,
foreign investors can sign contracts with exporters because of lower repayments of
exporters in external financing.8 According to proposition 2, the cut-off productivity
level for exporters decreases while that for domestic sellers increases. Potential ex—
porters, which would earn negative profits in the export market before, can repay to
foreign investors and finally survive in the export market. The cut-off productivity
level for domestic sellers, however, increases. This implies that low productive domes-

tic sellers make negative profits and exit the market because of high wages, resulting

 

7I analyze the effects of changes in financial conditions for both domestic. sellers and exporters in
proposition 4.

1’There is also the possibility that contractibility and collateralizability of domestic investors for
exporters increase because of the severe competition with foreign investors. Either case will reduce
repayments of exporters in external financing.

18

from the high demand for labor due to the increased number of exporters.

The other effects of financial liberalization are due to changes in the return to
capital. Due to increased flows of credits, the return to capital changes after the
liberalization. There is consensus in the literature that the return to capital decreases
after financial liberalization, especially in developing countries. [See Bacchetta (1992)
and Henry (2000a)]

Bacchetta (1992) analyzes the dynamic impacts of financial liberalization on the
return to capital. He considers two experiments, a simultaneous liberalization of
the domestic financial sector and the capital account, and a sequential liberalization
in which the liberalization of the domestic financial sector is followed by the capital
account liberalization. He shows that a likely outcome of liberalization is an initial net
inflow followed by an outflow. His intuition is that the liberalization of the domestic
financial sector leads to better allocation of resources and makes the home country
more attractive to both domestic and foreign investors. As a result, initially there
are few incentives for outflows and strong incentives for inflows. Thus, the return to
capital will decline after financial liberalization.

If financial liberalization attracts foreign credits into domestic financial markets,
the return to capital will go down. The effects can be analyzed as a variant of
the maximization problem (9). Consider inflows of credits in financial markets for
exporters. The return to capital decreases because of an increase of capital supply.
This case can be represented by the generalized constraint (6) in the maximization
problem (9); Bx((,9) E —k;,; + AxF$((p) + (1 — Admkr 2 "barking, where 7‘1: is the risk
free return to capital in financial markets for exporters.

Repayments of exporters can be calculated as
Fri?) = [I‘m—Axltx‘l'rxlkm/Ax- (15)

A decrease of the risk free return to capital in financial markets for exporters

reduces their repayments because (if? 2: k1. / A,- > 0. The direction of the effects on

19

cut-off productivity levels is identical to the case of higher contractibility and more

collateralizability in financial markets for exporters.

0 Proposition 3: Financial liberalization helps less productive exporters survive

in the export market, but keeps less productive domestic sellers out of the market.

H
a
V
V
O
H
in.
at-
at):
/\
0
DJ
5
Cl.

Proof: From equation (13) and (1.5), 5575—}
do’ d
V (P

dF
7‘1: =fifi>0

Financial liberalization leads to higher contractibility and collateralizability in fi—
nancial markets for exporters due to the participation of foreign investors, and reduces
the risk free return to capital in financial markets for exporters because of inflows of
foreign capital. In both cases, the export sector expands, while the domestic sector
shrinks. Due to financial liberalization, foreign investors invest more in the export
sector because they have more information on it. Inflows of credits also reduce out—
side options of domestic investors in financial markets for exporters due to the severe
competition with foreign investors. This has favorable effects on exporters, decreasing
their repayments. However, domestic sellers with lower productivity levels exit the

market because they cannot obtain labor due to high wages.

0 Corollary 3: The effects of a decrease of the return to capital for exporters on cut-
off productivity levels are greater under lower contractibility, lesser collateralizability,

and a smaller elasticity of substitution between varieties.

This implies that changes in the return to capital in financial markets for exporters
have stronger effects on cut-off productivity levels when their financial constraints are
more severe. The changes also produce stronger effects under a smaller elasticity of
substitution between varieties because of the same reasoning described in corollary 1.

Increased flows of credits due to financial liberalization can affect financial markets

for domestic sellers as well. Inflows of credits reduce the overall interest rate in the

20

home country. In general, outside Options in financial markets for both domestic

sellers and exporters will decrease after financial liberalization.

0 Proposition 4: The decrease of interest rates due to inflows of credits increases
the number of both domestic sellers and exporters. The number of exporters relative
to domestic sellers decreases with contractibility of exporters but increases with their

collateralizability when overall interest rates decrease.

Proof: Let r' be the overall interest rate. Fd((,-9) = [1 — (1 — Ad)td + rlkd/Ad
,,,I
and Fx((p) = [1 — (1 — Afltm + rlkm/Ax. From equation (13), (if is positive when

,I
Ag; and tag are not much different from Ad and tag, and {—15} is positive when AI and

1/
tm are not much different from Ad and I‘d. Moreover, let ratio = 74.9 Eff—"ratio =

F’F —F F’ . . . . . . .
Ta‘lLdEQx—d- Thls lS pOSItive when A3; is smaller than Ad or ta; lS blgger than td.
d

See appendix B-3.

If inflows of credits decrease overall interest rates, repayments of both domestic
sellers and exporters decrease. The number of firms in both sectors increase, when A1.
and tag are not much different from Ad and td. Furthermore, the effects on exporters
are greater than those on domestic sellers when A3; is smaller than Ad or tag is bigger
than td. The intuition is that the effects of a decrease of overall interest rates on
exporters are greater than those on domestic sellers when exporters are more finan-
cially constrained than domestic sellers. If exporters can offer more tangible assets,
the effects on exporters are also greater.

There is little evidence for differences in the fraction of collaterals between domes-
tic sellers and exporters in the same industry because the fraction depends mostly on
properties of goods. In light of the substantial turnover rate in the product compo-

sition in exporting, however, the probability of default will be higher in the export

 

9The increase of the ratio implies the decrease of the number of exporters relative to that of
domestic sellers.

21

sector than in the domestic sector.10 Exporting has more risks than domestic sales.
It needs higher fixed costs and higher transportation costs. Exporters also face more
credential risks than domestic sellers because of higher probability that they may
not have payments from buyers. More risks lead to lower contractibility, which im-
plies that exporters are more financially constrained than domestic sellers. Thus,
a decrease of overall interest rates has bigger effects on exporters than on domestic
sellers, and the relaxation of financial constraints due to financial liberalization leads

to more favorable outcomes for exporters than for domestic sellers.
5. Conclusions

In this chapter, I have presented a general equilibrium model of heterogeneous
firms in which firms differ in productivity and face financial constraints to pay fixed
costs for production. I consider how the relaxation of financial constraints for domes-
tic sellers and exporters affects firms’ entry and export decisions. I have the following
findings. (1) The relaxation of financial constraints for domestic sellers helps less pro-
ductive firms survive in the domestic market, while driving less productive exporters
out of the export market. (2) The relaxation of financial constraints for exporters
helps less productive firms survive in the export market, while pushing less productive
domestic sellers out of the market. (3) Financial liberalization may help less produc-
tive exporters survive in the export market, while driving less productive domestic
sellers out of the market.

This chapter makes two contributions to the literature on international trade.
First, the existing literature largely assumes that changes in financial conditions hap-
pen in all sectors, at the same time, and to the same extent. In contrast, in this
chapter, I consider the case where financial constraints are relaxed only in some sec-

tors, and analyze the impacts on the other sectors in which financial constraints are

 

1"According to Manova (2006), more than a quarter of exported products are discontinued from
One year to the next and replaced by new ones, resulting in a reallocation of 16% of bilateral trade
by value.

22

 

not relaxed. Thus, my analysis provides a general equilibrium view of the impacts of
financial constraints on production and exports. Second, my model sheds new light
on the effects of financial liberalization on different sectors in the economy. Specifi-
cally, financial liberalization reduces interest rates, through which raises the number
of firms in both domestic and export sectors. In addition, the number of exporters
relative to domestic sellers decreases with contractibility of exporters but increases

with their collateralizability when overall interest rates decrease.

23

 

Chapter 2: Estimating the Effects of Financial
Liberalization on Exporters and Domestic Sellers

24

1. Introduction

In recent years, international economists start investigating the effects of financial
liberalization on trade patterns. If there are no financial constraints, the Heckscher—
Ohlin model predicts that a country will export the goods which use abundant input
factors. Financial liberalization does not affect trade patterns in this case. In the
presence of financial frictions, however, borrowing constraints, which vary across in-
dustries and firms, affect the composition of a country’s export by limiting investment
opportunities open to producers with lower financial credits. This serves a ground for
financial liberalization, which changes firms’ financial constraints, to affect the pro-
duction of an individual firm differently, and thus the export composition of goods.

Beck (2003), Becker and Greenberg (2005), Svaleryd and Vlachos (2005), and Hur
et al. (2006) find that financially developed countries export relatively more in sectors
that require more outside finance or are intensive in fixed up-front costs. Recently,
by examining the impacts of equity market liberalizations on the export behavior
of 91 countries in 1980—1997, Manova (2008) also shows that financial liberalization
increases exports disproportionately more in financially vulnerable sectors that require
more outside finance or having fewer collateralizable assets.

These empirical studies use the volume of exports as the dependent variable. Their
focus is on the effects of financial development only on exporters. In contrast, I am
interested in the effects of financial liberalization on both exporters and domestic
sellers. Thus, unlike the existing studies, I use the ratio of exports to domestic
productions as the dependant variable. My analyses will show how much financial
liberalization affects the export sector compared to the domestic sector.

Financial liberalization has different effects on exporters and domestic sellers be-
cause the liberalization will relax financial constraints of exporters more than those of
domestic sellers, and help the export sector expand while causing the domestic sector
to shrink. As financial liberalization proceeds, many foreign investors will enter do-

mestic financial markets. They may have different amount of information on exporters

25

and domestic sellers in the country, which leads to different contractibility or collat-
eralizability for exporters and domestic sellers from those of domestic investors. It is
widely accepted that foreign investors have more information on exporting firms than
on non-trade entrepreneurs. If foreign investors have more information on exporting
firms than domestic investors, foreign investors can make contracts with exporters
because they can offer lower repayments than domestic investors to exporters due to
their higher contractibility or collateralizability.11

According to proposition 2 in chapter 1, the relaxation of financial constraints
of exporters allows low productive exporters, which would earn negative profits in
the export market before, to make repayments to investors and thus survive in the
market. Low productive domestic sellers, however, have negative profits and exit the
market because of high input costs, resulting from intense competition with exporters
in labor markets.

In addition to the effects of financial liberalization on trade, I find the role of
changes in costs of external financing in affecting export decisions. To my knowledge,
this topic has not been examined in the literature. Changes in interest rates have
been considered only as affecting import demands through changes in total incomes
in the home country. The channel that changes in costs of external credits can have
different effects on exporters and domestic sellers has been neglected. According to
preposition 4 in chapter 1, a decrease of the overall interest rate increases the number
of both domestic sellers and exporters. Furthermore, the number of exporters relative
to domestic sellers decreases with contractibility of exporters but increases with their
collateralizability when overall interest rates decrease. In other words, when the
overall interest rate decreases, benefits to exporters are greater than those to domestic
sellers if contractibility of exporters is smaller than that of domestic sellers or their

collateralizability is higher.

 

11There is also a possibility that contractibility and collateralizability of domestic investors for
exporters increase because of intense competition with foreign investors. Either case will reduce
repayments of exporters in external financing.

26

There is little direct evidence showing differences in the fraction of collaterals
between domestic sellers and exporters in the same industry. This is because the
fraction of collaterals depends mostly on properties of goods. However, the sub—
stantial turnover rate in the product composition in the export market implies that
contractibility of exporters is lower than that of domestic sellers.12 In part, it re-
sults from higher risks in exporting than those in domestic sales. Exporting requires
larger amount of fixed costs while it loses more portions of values than domestic sales
due to higher transportation costs. Exporters also have higher credential risks than
domestic sellers because of the higher probability that they cannot receive payments
from buyers. Higher risks make lower contractibility. Thus, changes in the overall
interest rate have more effects on exporters than on domestic sellers, which implies
that a decrease of the overall interest rate expands the export sector more than the
domestic sector.

Using data for 91 countries in 27 industries from 1980 to 1997, I find strong
evidence for the effects of financial liberalization and interest rates on trade patterns.
Here, interest rates are used to measure costs of external financing. Coeflicients
on the interaction between financial liberalization and external financial dependence
are positive and statistically significant in all model specifications, which implies
that financial liberalization benefits exporters more than domestic sellers, and that
the effects are greater if they depend more on external financing. Asset tangibility,
however, does not play a significant role in this process. I also find that real interest
rates perform an important role in firms’ export decisions. Due to the different degree
of sensitivity between exporters and domestic sellers to changes in interest rates, the
export sector expands relatively more than the domestic sector when real interest
rates decrease. The more an industry depends on external financing, the larger will

be the effects of changes in real interest rates on trade patterns. In addition, I find

 

l"ZAccording to Manova (2006), more than a quarter of exported products are discontinued from
one year to the next and replaced by new ones, resulting in a reallocation of 16% of bilateral trade

in value.

27

that the effects are greater in countries with under-developed financial markets.
The structure of this chapter is as follows. In section 2, I describe the data and
present some descriptive statistics, and introduce models for the estimation in section

3. In section 4, I show the results of analyses, and conclusions follow in section 5.
2. A Glance of Data
2.1 The Ratio of Exports to Domestic Sales

I obtain most data from Manova (2008). Feenstra’s World Trade Database pro—
vides export flows in dollars at the 4—digit SITC Rev.2 industry level in each country.
Using Haveman’s concordance, they are aggregated in terms of 3-digit ISIC indus-
tries, for which the industry-level data on financial vulnerability is available. The
data on industrial productions at the 3-digit ISIC Rev.2 comes from the Industrial
Statistics Yearbook and International Yearbook of Industrial Statistics published by
the United Nations. The production data is available in national currency while the
export data is provided in dollars. Thus, I convert the export data using the exchange
rate obtained from the Penn World Table.

Calculating the ratio of exports to industrial productions, I find that 2,105 out
of 18,908 observations (11.1%) are greater than one, which means that the volume
of exports is greater than domestic production. This is mainly due to the fact that
domestic productions are net outputs, but experts are gross outputs which may double
count intermediate inputs. However, because I am interest in changes in the ratio of
exports to domestic sales, this should not cause a problem to my analysis as long as
the differences in standards are stable and persistent.

Hence, I normalize industrial export ratios in each industry and country using
the formula: (ERcz-t — AERCi) / ST DC,- a: 100, where ER, AER, and STD represent.
the export ratio of industry 2' in country c in year t, the average export ratio, and
the standard deviation during the period respectively. On average, the average of

the export ratio is 86.9%, and its standard deviation is 80.1%. This normalization

28

method will alleviate the effects of extreme values on estimates by making all averages
of trade ratios in each industry and country the same as those in other industries and
countries. The most extreme values may occur in the sector where the amount of
export or production is small. They, however, will dominate the other observations
and distort analyses if they are used without standardization. The standardized trade
ratio ranges between —333.9 and 393.2 with a mean of zero and a standard deviation

of 95.8.
2.2 Measures of Financial Liberalization

The equity market liberalization is used as a proxy for financial liberalization. It
is available for 91 countries between 1980 and 1997 in Bekaert, Harvey and Lundblad
(2005) (BHL). 39 countries opened their stock markets to foreign investors during
this period, and 16 countries liberalized prior to 1980 while 36 countries did not
remove equity market restrictions until 1997. BHL dates both the official year of stock
market reforms and the "first sign" of the liberalization. This first-sign year is the
earliest of three dates; official liberalization, first American Depository Receipt (ADR)
announcement, and first country fund launch. BHL also constructs measures of the
intensity of openness, reflecting the degree of the equity market in which foreigners can
invest. The intensity measures of the liberalization are purely cross-sectional, which
varies between zero and one, where a ratio of one implies no restrictions regarding
foreign ownership. Oflicial and first-sign liberalization intensities are set at zero prior

to the liberalization, and later at the intensity level in the year of the liberalization.
2.3 Measures of External Finance Dependence and Asset Tangibility

Measures of industry—level external finance dependence and asset tangibility for 27
3—digit ISIC industries come from Braun (2003). They are constructed from the data
on all publicly traded U.S.-based companies in COMPUSTAT.13 The indicator of

 

l3Constructing industrial measures of external finance dependence and asset. tangibility from the

29

the industrial reliance on outside finance is made by the ratio of capital expenditures
minus cash flow from operations to capital expenditures for the median firm in each
industry. This ratio ranges from -0.4512 to 1.1401. I add 0.5 to the original value
in order to set the measure as positive. It will not change the results of analyses
qualitatively because only differences in the measure across industries matter in the
estimation. Another reason for setting the value as positive is to have reasonable
values when interaction terms are made. In the case of the negative real interest rate,
for example, the interaction term between the real interest rate and negative external
finance dependence will be positive, which will mislead analyses.

Similarly, the measure of asset tangibility is constructed as the share of net prop-
erty, plant, and equipment in total book-value assets for the median firm of all publicly
traded U.S.—based companies in a sector in COMPUSTAT. It ranges from 0.0745 to
0.6708. Both measures for external finance dependence and asset tangibility are used
as averages for the 1986-1995 period. They appear very stable over time.

The average and standard deviation of external finance dependence across all 27
industries is 75% and 32% and those of asset tangibility are 30% and 13%, respectively.
Industries in the greatest need for external funds such as professional and scientific
equipment and electric machinery are mostly intensive in large up-front investment.
On the other hand, non-ferrous metals, apparel, and beverages, which do not need
much up-front investment, do not depend much on external finance. Industries with
the lowest levels of tangibility are pottery, china, and earthenware; leather products;
and apparel. Industries with the highest levels of tangibility are petroleum refineries;

paper and products; iron and steel; and industrial chemicals.

 

US data can be supported by following reasons. The United States has one of the most advanced
and sophisticated financial systems, which makes it reasonable to believe that the measures reflect
firms’ true demand for outside capital and tangible assets. Using the US. as the reference country
is also convenient because of the limited data on many other countries. Moreover, it can reduce the
possibility for the measures to endogenously reSpond to a country’s availability of external credits.
If entrepreneurs use more internal financing in countries with stricter equity market restrictions, for
instance, estimates of the effects of financial liberalization will be biased downwards.

30

2.4 Other Data14

I obtain interest rates and CPI for 91 countries from International Financial Sta-
tistics (IFS) provided by the IMF. Both are measured in percentage. The lending rate
is used as the measure of the overall interest rate because my analysis focuses on the
changes in firms’ decisions when facing borrowing constraints. The real interest rate
is obtained by subtracting CPI from the lending rate. I standardize the real interest
rate using the same formula as that of the export ratio. To control the trend of the
export ratio, I use the total export ratio to GDP in each country. It is calculated
from the value of exporting goods and services divided by GDP in National Account,

available in the United Nations Statistics Division.
3. Empirical Specifications

Based on the financial constraint model in chapter 1, I use a generalized difference—
in—difference approach to test for the differential effects of financial liberalization on
the export ratio across sectors. I include interaction terms of the country-level mea-
sure of financial liberalization (Liberalct) with industry-level measures of external
finance dependence (F linDepi) and asset tangibility (Tang). I also add the real

interest rate (Interestct) and its interaction with external finance dependence.

Tc“ = (30*Liberalct+,81*LiberalcfiFinDepi+,32Liberalct*Tang,-,+7-OInlcrcstCt-l-
71 Interestct * FinDepz- -+- 6C0'ntrolscit + 00 -+- a 1 Tet + 176 + r), + I” + cm, (1)
where Tat is the standardized export ratio of industry 2' in country c in year t. Liberalct
is a binary variable which is equal to one in the year of and all years after financial
liberalization, and zero otherwise. For robustness checks, I also use liberalization
intensity measures which reflect the fraction of equity market openness in which
foreigners are permitted to invest. F inDepi and Tang, correspond to the level of

external finance dependence and asset tangibility in sector 23 respectively. Coefficients

 

14See the Manova’s paper (2008), "Credit Constraints, Equity Market Liberalizations and Inter-
national Trade", for the detailed description of other variables used in the estimation.

31

of my interests are 331, 52, and 71. If .31 > 0 and [32 < 0, financial liberalization
increases export ratios relatively more in sectors intensive in external financing or
having more tangible assets. If 71 < 0, a decrease of the real interest rate increases
export ratios relatively more in sectors intensive in outside capital.

I allow for country, sector, and year-fixed effects, and cluster errors by countries.
I also control for the ratio of total export to GDP in each country. I do not esti-
mate main effects of FinDep, and Tang,- in themselves because they are absorbed in
industry-fixed effects which also capture other industry-specific omitted characteris-
tics. Time—fixed effects account for changes in global economic conditions that affect
all countries and industries equally, such as technological improvements, demand
shifts, or world price movements. Country-fixed effects control for country-specific
characteristics that would affect export ratios of all industries in a country such as
remoteness or institution systems that do not change during the sample period. Thus,
the main effect of financial liberalization (:30) is identified from within-country over
time variations.

The coefficient on the interaction between the real interest rate and external fi-
nance dependence (71) expects to be negative, which implies that exporters are more
sensitive to changes in the real interest rate and benefit more from a decrease of the
interest rate. Moreover, considering industrial differences in the degree of external
finance dependence, the effects of changes in the real interest rate on the export. ratio
will be greater in the sector that relies more on external financing.

In panel analyses, the estimates, [31 and fig, come from the combination of cross-
sectional and time-series variations due to financial liberalization across countries
regarding external finance dependence and asset tangibility. These coefficients, thus,
reveal the comparative advantage in financially vulnerable sectors that a country with
open financial markets has relative to the financially closed economy. In other words,
if financial liberalization relaxes financial constraints of exporters more than those of

domestic sellers, exporters in the industry with a higher degree of external finance

32

dependence benefit more from the liberalization than those in other industries with a
lower degree of external finance dependence. Hence, the export ratio would increase
more in industries with higher external finance dependence after the liberalization.
Similarly, the interaction term between the real interest rate and external finance
dependence shows the changes in the pattern of the comparative advantage due to
changes in costs of outside capital.

Recall that the industrial characteristics, F inDepi and Tangi, are obtained only
from the US. data. While the measures do not require that all firms in each industry
have exactly the same tangibility and external finance dependence levels across all
countries, the effectiveness of the results relies on the fact that the ranking of sectors
remain relatively stable across countries. Rajan and Zingales (1998) and Braun (2003)
argue that the measures, which they constructed, capture technological components
that are innate to a sector, and thus the measures are good proxies for external
finance dependence and asset tangibility in all countries. They also point out that
the measures have significant variations more across sectors than across firms within
an industry.

Controlsm-t includes factor endowments and trade liberalization. I include factor
endowments to control for traditional explanations for trade patterns. For robustness
checks, I analyze only switching countries which open their financial markets during
1981-1997 in order to avoid distortions arising from including countries which do
not make financial reforms during the period. Moreover, I divide the sample into
two groups in terms of income levels in order to examine if the effects are different
among countries. Furthermore, I analyze long-term effects of financial liberalization
by averaging out unobserved systematic differences at the time of the liberalization.

Interpreting the results of model specification (1) as a causal relationship relies
in part on the assumption that financial liberalization provides exogenous shocks to
the availability of external finance and its costs. In the absence of credit constraints,

if the liberalization is anticipated, export ratios should. not respond either ahead of

33

or after the liberalization. When financial frictions exist, however, export ratios may
increase prior to the official liberalization date if the easier external financing in the
future is expected. This suggests that specification (1) may underestimate the actual

impacts of financial market liberalization on trade patterns.
4. Empirical Results
4.1 The Effects on Trade patterns: Basic Results

The results of the basic estimation are presented in Table 2-1. I estimate spec-
ification (1) using the full panel of standardized export ratios for 91 countries and
27 industries in the 1980-1997 period. I trim the data by dropping outliers of ex-
port ratios and real interest rates whose values are outside of the range of 3—standard
deviations. I control for country, industry, year-fixed effects, and the standardized
total export ratio to GDP. I also include factor endowments to control for traditional
explanations of trade patterns. Those variables are per capita physical capital, hu-
man capital, and natural resources, and their interaction terms with intensities by
industries. The data on physical capital and human capital is obtained from Caselli
(2005), and that of natural resources is available in the World Bank (1997). Their
intensities by industries are available in Braun (2003). I use both official liberalization
and first-sign liberalization dummies in the estimation.

According to Table 2-1, the coefficients on the interaction between liberalization
and external finance dependence are positive and significant at the 5 or 1 % level.
This implies that financial liberalization benefits exporters more than domestic sellers.
The more the exporters depend on outside funding, the bigger are the effects.

After financial liberalization, the industry in the upper 75% percentile in external
finance dependence exports more by 3% points on average than the industry in the

lower 25% percentile in external finance dependence.15 The average of the standard

 

1"Due to the standardization of variables, estimates implies different amount of the effects on
actual export ratios, which depends on the specific industry and country.

34

deviation of export ratios across all industries and countries is 67%. Hence, the esti-
mate can be converted to the amount of changes in export ratios using the formula:
estimates*the difference in the explanatory variable*the average of the standard devi-
ation of export ratios/ 100. Unlike Manova (2008), the coefficients on the interaction
between liberalization and asset tangibility are negative but statistically insignificant.

I also find that the coefficients on the interaction between real interest rates and
external finance dependence are negative and significant at the 1% level. A reduction
of the real interest rate can also be used as a proxy for the relaxation of financial
constraints. The negative estimates show that exporters are more sensitive to changes
in the real interest rate. The more the exporters depend on outside credits, the
bigger are the effects. This may suggest that exporters have less contractibility than
domestic sellers due to higher risks and costs in exporting (e.g. higher credential risks
and transportation costs).

A decrease of the overall real interest rate by 1% point increases the export ratio
by 0.09% point more on average for the industry in the upper 75% percentile in
external finance dependence compared to the industry in the lower 25% percentile
in external finance dependence. Similar to the previous calculation method, the
estimate can be converted to the amount of changes in the export ratio using the
formula: estimates*the difference in external finance dependence*the average of the
standard deviation of export ratios/ the average of the standard deviation of the real
interest rate.

In order to examine whether the effects of financial liberalization differ across
countries at different development levels, I divide the sample into two groups based on
financial market capacity. I use stock market capitalization (the value of all publicly
listed companies) relative to GDP averaged between 1980 and 1984 as the proxy for
financial market capitalization. The more-developed group contains countries whose
stock market capitalization is above the sample average while the less-developed group

includes countries whose stock market capitalization is below the sample average.

According to Table 2-1, the effects of financial liberalization are bigger for finan-
cially under-developed countries. The coefficients on the interaction between liberal-
ization and external finance dependence are positive and statistically significant at the
1% level for under-developed countries. The coefficients on the interaction term of the
real interest rate are negative and significant at the 1 % level for those countries. On
the other hand, the estimates for more-developed countries are statistically insignifi-
cant. These results support Manova’s argument that countries with under-developed
stock markets gain more benefits from financial liberalization. Therefore, exporters
in countries with under-developed stock markets and those in industries which need

more external financing can get bigger benefits from financial liberalization.
4.2 The Effects on Trade patterns: Switchers Only

In Table 2—2, I analyze only the countries which opened their financial markets
during the sample period. The analysis in Table 2-2 includes 16 countries that lib-
eralized equity flows before 1980 and 36 countries that remained financially closed
until 1997, which would distort the results of the analysis. Thus, I analyze only 39
countries that underwent financial reforms during the sample period. Precisely, 30
countries are included in the analysis because of the limitation of the data on in-
dustrial productions and real interest rates. I also control for factor endowments,
country, industry, year-fixed effects, and the standardized total export ratio to GDP.

The results do not change qualitatively when I focus only on 39 switchers that
removed capital flow restrictions between 1981 and 1997. The coefficients on the
interaction between liberalization and external finance dependence are positive and
significant, and those between real interest rates and external finance dependence are
negative and significant. Overall, the estimates become stronger while their t-values
do not increase much because of a sample size. The results show that financial liber-
alization provides a large direct boost to the export sector and results in a substantial

reallocation towards exporters in the sector with greater reliance on external finance.

36

Moreover, exporters have higher sensitivity to changes in the real interest rate than

domestic sellers, and the effects are bigger in financially under-developed countries.
4.3 Long-term Analysis

The basic specification (1) can be rewritten in terms of the time of financial
liberalization. t = 0 represents the time before the liberalization and t = 1 represents
the time after the liberalization. I take a difference between them, and have the

following specification;

ATC, = 80 * Liberalc + [7’1 * ALiberalc =1: FinDepi + SQALiberalC >1: Tang,- +

yoAlnterestc + ’71A1nterestc * FinDepi + 01 ATC + Arm; (2)

Note that the constant term 00 is dropped out of the regression. F irst-differencing
also removes all country and sector-fixed effects, (0C, 7),), and thus provides cleaner
estimates for the impacts of financial liberalization on trade patterns. As in the previ-
ous panel analysis, this may still incur downward biased estimates since export ratios
incorporate any response of exports to an anticipated liberalization. Since countries
liberalize their financial markets in different years, I control for liberalization-year
fixed effects.

I average out all the data in the three-year period. Technically, I calculate averages
of Ta-t_1, Tdt_2, Tait-3, and Tat“, Tat-Hg, Tat”, where t is the time of financial
liberalization. I then take a difference between them, and use OLS regression. Ac—
cording to Table 2-3, I have the similar results as those of the basic specification. The
coefficients on the interaction between liberalization and external finance dependence
are positive and significant at the 10 or 5 % level and those between real interest rates
and external finance dependence are negative and significant at the 10% level, which
means that financial liberalization benefits exporters more than domestic sellers and
a decrease of the real interest rate increases export ratios. The more the industry

depends on outside finance, the bigger are the effects.

37

4.4 Controlling Trade Openness

Financial liberalization is sometimes part of a broader program of deregulations
that may include reforms of trade policies. In Table 2-4, I confirm that my findings
are not driven by simultaneous changes in trade policies. The data on trade openness
is obtained from Wacziarg and Welch (2003), who updated the binary indicator orig-
inally developed by Sachs and Warner (1995). According to their criteria, a country
is labeled closed to international trade if at least one of the following conditions is
met: average tariff rates are at least 40%; non-tariff barriers cover at least 40% of
trade; a black market exchange rate exists and is on average depreciated at least 20%
relative to the ofl‘lcial exchange rate; the country holds a monopoly on major exports;
or it has a socialist economic system. According to the classification, a country can
be "closed" to international trade due to high trade costs, but may still participate
in the trade.

The data on trade openness is available for 70 of the 91 countries in my sample. I
focus on countries that remove credit flow restrictions during the 1981-1997 period. I
also allow trade openness to affect sectors differently by interacting it with industry
measures of financial vulnerability. I control for country, industry, year-fixed effects,
factor endowments, and the total export ratio to GDP. As in Table 2-4, the inclu-
sion of trade openness changes little the estimated effects of financial liberalization
and real interest rates on export ratios. The coefficients on the interaction between
liberalization and external finance dependence are positive and significant, and those
between real interest rates and external finance dependence are negative and signif—
icant, while those between liberalization and asset tangibility are still negative but

insignificant. The effects are bigger in financially under—developed countries.
4.5 Comparison with Cases that use Trade Volumes

In Table 2—5, I follow the model specification of Manova (2008) in order to compare

the two different specifications. She uses export volumes as the dependent variable,

38

and I use export ratios. I use the same samples with those in Table 2-1. I control
for country, industry, year-fixed effects, and factor endowments. I also control for the
log of GDP instead of the total export ratio to GDP when export volumes are used
as the dependent variable. The last two columns come from Table 2—1.

In the first two columns, I do not include the real interest rate and its interaction
term. The model specification is the same as that in her paper. As in Table 2-5,
the coefficients on the interaction of the liberalization with external financial depen-
dence are positive and significant, and those with asset tangibility are negative and
significant. These results show that financial liberalization increases exports dispro—
portionately more in financially vulnerable sectors that require more outside finance
or employ fewer collateralizable assets.

In the third and fourth columns, I include the real interest rate and its interaction
with external finance dependence into her model. I do not standardize the real inter-
est rate because the dependent variable is not standardized. Her arguments still hold
with the inclusion of the real interest rate. Estimates of the interaction of the liberal-
ization with external financial dependence are positive and significant, and those with
asset tangibility are negative and significant. Furthermore, I find weak evidence that
exporters have higher sensitivity to changes in the real interest rate than domestic
sellers. Estimates of the interaction of the real interest rate with external financial
dependence are negative, and they are significant at the 10% level when the first-sign

liberalization dummy is used.
4.6 Robustness Checks

In Table 2-6, I do robustness checks. The first two columns come from Table 2-1.
In the third and fourth columns, I use liberalization intensity measures. The results
are qualitatively the same as those when liberalization dummies are used. In the last
two columns, I use different liberalization measures and include trade openness. My

results are changed little for all these alternative specifications.

39

Finally, there is concern that financial liberalization may be endogenous. How-
ever, this does not appear problematic for my results because of the following reasons.
First, the exact timing of financial liberalization is the procedure of complex political
decisions and thus plausibly exogenous from the view point of an individual firm and
investor. Second, prior evidence suggests that equity market liberalizations do not
follow surges in investment (Henry, 2000b) and that to control for growth opportuni-
ties or world business cycle effects does not eliminate the impacts of the liberalization
on growth (Bekaert et al., 2005; Gupta and Yuan, 2004). Finally, if domestic credit
markets are frictionless and a country expects higher export demand for sectors inten-
sive in external finance, it can liberalize financial markets to increase the availability
of funding. Capitals will then flow freely and allow firms to meet the demand. Hence,
we can observe that financial liberalization follows higher exports in financially de—
pendent sectors even in the absence of credit constraints, which will lead to downward

biased estimates.
5. Conclusions

I have examined empirically the impacts of financial liberalization and the relax—
ation of financial constraints on exporters and domestic sellers. Unlike the existing
literature, I use a ratio of exports to domestic production instead of export volumes
as the dependent variable. I find that financial liberalization and real interest rates (a
measure of costs of external funds) are important determinants of international trade
patterns. In particular, financial liberalization and a reduction in the real interest
rate increase the ratio of export to domestic production disproportionately more in
industries with a higher degree of external finance dependence. In addition, I find
that the effects are greater in financially under-developed countries.

I also find the role of changes in costs of external financing in making export
decisions. Previous literature focuses only on the role of interest rates in affecting

import demands through changes in total incomes. However, I find that changes

40

in costs of external credits have different effects on exporters and domestic sellers.
My finding supports that exporters have lower contractibility than domestic sellers
because of their higher risks in exporting.

This chapter contributes to the growing literature on determinants of trade pat-
terns. My analysis focuses on how much financial liberalization affects the export
sector compared to the domestic sector. Therefore, my results provide stronger evi-

dence on the effects of financial liberalization on international trade.

41

Table 2-l: The Effects on Trade patterns: Basic Results

 

Official Liberalization Dummy

First Sign Liberalization Dummy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

More Less More Less
Trimmed financially financially Trimmed financially financially
developed developed developed developed
Liberalization - 12.739 -28.544 -1 l . I44 -3.750 - I 8.252 -7.485
(-I .13) (- I .37) (-0.85) (-0.34) (-0.77) (-0.63)
Lib*External 12.893 4.497 l9.207 l5.442 6.735 22.429
finance (2.51)” (0.78) (2.70)*** (3.06)*** (1.34) (3.]4)***
de endence
Lib*Asset -l3.702 -l4.039 - 12.364 -l9.450 -l2.877 -22.525
tangibility (-0.85) (- l .40) (-0.46) (- l .39) (- l .33) (- l .00)
lnterest*Extemal -0.077 0.020 -0. I42 -0.082 0.018 -0.l 5]
finance (-2.59)*** (0.73) (-3.42)*** (-2.73)*** (0.66) (-3.6l)"’**
dependence
Interest 0.018 -0.0 l 4 0.062 0.01 7 -0.0] 7 0.068
(0.60) (-0.29) (l.58) (0.57) (-0.39) (l.72)*
R-square 0. l 88 0.190 0.217 0.188 0.188 0.2 l 8
# of observations [2,797 5,078 7,7l9 I2,797 5,078 7.7 l 9
# of exporters 49 22 27 49 22 27

 

Note: I control for country. industry, year-fixed effects, the total export ratio to GDP. and factor
endowments. t-statistics are reported in parentheses. ***. **, * indicate significance at the l%, 5%, and

l0% level.

42

 

Table 2-2: The Effects on Trade patterns: Switchers Only

 

Official Liberalization Dummy

First Sign Liberalization Dummy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

More Less More Less
Trimmed financially financially Trimmed financially financially
developed developed developed developed
Liberalization 3.395 H.001 -l4.730 14.348 I99 I 8 -7.809
(0.22) (0.44) (-0.69) (0.89) (0.74) (-0.40)
Lib*External 20.659 2.024 29.008 23.793 7.991 31.232
finance (1.92)‘ (0.13) (2.40)" (2.61)*** (0.65) (2.97)***
dependence
Lib*Asset -21.688 ~37.836 - 14.895 -30.902 -36.186 -30.404
tangibility (-0.68) (-I.98) ** (-0.30) (-l .23) (-2.08) ** (-0.83)
lnterest‘Extemal -0.086 0.035 -O. l 38 -0.098 0.022 -0.157
finance (-2.1 l)“ (0.69) (-2.97)** (-2.4 l )** (0.45) (-3.42)***
dependence
Interest 0.029 0.061 0.072 0.024 0.062 0.082
(0.98) (0.91) (1.90)* (0.79) (0.93) (2.14)“
R-square 0.158 0.135 0.221 0.161 0.136 0.222
# of observations 7,240 2,421 4,819 7.240 2.421 4.819
# of exporters 26 8 18 26 8 I8

 

Note: I control for country, industry, year-fixed effects, the total export ratio to GDP, and factor
endowments. t-statistics are reported in parentheses. ***. **, * indicate significance at the l%, 5%. and

10% level.

43

 

Table 2-3: Long-term Analyses

 

3 year average

Official Liberalization Dummy

First Sign Liberalization Dummy

 

 

 

 

 

 

 

 

 

More Less More Less
All financially financially All financially financially
developed developed develggcd developed
Liberalization 5.614 -73.079 83.384 48.874 -100.229 61.612
(0.20) (-l .49) (2.82)*** (1.20) (- l .86)* (1.94)*
Lib*External 42.566 58.952 23.004 51.838 63.291 52.686
finance (1.81)* (1.10) (0.93) (2.07)" (1.26) (1.79)*
dependence
Lib*Asset -20.4 1 7 1.843 -43 .870 -9.994 92.055 -3 7.980
tangibility (-0.39) (0.02) (-0.76) (-0.20) (0.75) (~0.73)
Interest*External -0.363 -0. 191 -0.587 -0.3 72 -0.280 -0.499
finance (-1.74)* (~0.51) (-2.25)** (- l .69)* (-0.80) (- l .76)*
dependence
Interest -0. 132 0.348 -0.044 -0. 154 0.421 -0. 107
(-0.65) (1.15) (-0.19) (-0.7l) (1.37) (-0.44)
R-square 0.201 0.091 0.363 0.167 0.130 0.254
# of observations 207 72 135 216 54 162

 

 

 

 

 

 

 

Note: I calculate an average of all variables before and after the liberalization, take their differences,
and do OLS estimation. 1 use the difference of the total export ratio as a control variable. 1 control for
liberalization-year fixed effects. t-statistics are reported in parentheses. ***, **, * indicate significance
at the 1%, 5%, and 10% level.

44

 

Table 2-4: Controlling Trade Openness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Official Liberalization Dummy Official Liberalization Dummy
More Less More Less
Trimmed financially financially Trimmed financially financially
developed developed developed developed
Liberalization -13.162 -26.708 -8.7 l 4 -15.894 -28. 127 -11.707
(- 1.13) (-l .36) (-0.69) (-l.30) (-1.50) (-0.93)
Lib*External 10.578 3.851 16.099 14.467 12.018 18.1 14
finance (2.09)" (0.62) (2.30)“ (2.04)" (1.24) (2.01 )**
dependence
Lib'Asset - 1 9.900 - 17.71 5 -20.505 -20.480 -33.142 -15.583
flibility (-1.29) (-l.84)* (-0.78) (-l.l4) (~2.58)"** (-0.55)
Interest*External -0.069 0.018 -0.133 —0.067 0.024 -0.132
finance (~2.26)** (0.66) (-2.93)*** (-2.20)"‘* (0.98) (-2.92)***
dependence
Interest 0.018 -0.016 0.045 0.016 -0.021 0.045
(0.53) (-0.34) (0.95) (0.49) (-0.46) (0.94)
Trade openness -'I .662 0.758 -8.7 I 4 4.293 2.300 0.486
(-0.12) (0.06) (-0.47) (0.25) (0.14) (0.02)
Trade openness* -8.454 -13.930 -5.523
External (- l .23) (-1.43) (-0.65)
finance
dependence
Trade openness 1.336 29.848 -l6.660
*Asset (0.08) (1.34) (-0.72)
tangibility
R-square 0.181 0.196 0.204 0.181 0.196 0.204
# of observations 11,872 4,899 6,973 11,872 4.899 6.973
# of exporters 44 19 25 44 19 25

 

 

 

 

 

Note: 1 control for country, industry, year-fixed effects, factor endowments, and the total export ratio to
GDP. t-statistics are reported in parentheses. ***, **, * indicate significance at the l%, 5%, and 10%
level.

45

 

Table 2-5: Comparison with Cases that use Trade Volumes’

)

 

 

 

 

 

 

 

 

 

 

 

Official First Sign Official First Sign Official First Sign
Liberal Liberal Liberal Liberal Liberal Liberal
Dummy Dummy Dummyz’ Dummyz’ Dummy Dummy
Dependent Log(export) Log(export) Log(export) Log(export) Export Export
Variable ratio ratio
Liberal 0.027 0.103 0.028 0.102 - I 2.739 -3.750
(0.14) (0.52) (0.14) (0.51) (-l.l3) (-0.34)
Lib*Ext 0.850 0.947 0.852 0.951 12.893 15.442
Fin Dep (4.74)"” (4.54)*" (4.72)"‘ (4.53)"* (2.51 )** (3.06)“u
Lib*Asset -1.770 -2.l73 -l.77l -2.|74 -13.702 -l9.450
TEL (-3.03)"“ (-4.00)"* (-3.03)’” (-4.00)"‘ (-0.85) (-l .39)
Interest*Ext -0.01 1 -0.017 -0.077 -0.082
Fin Dep (-l.l4) (-1.77)* (-2.59)"" (-2.73)*”
Interest -0.004 0.000 0.018 0.017
(-0.55) (0.01) (0.60) (0.57)
R-square 0.795 0.796 0.795 0.797 0.188 0.188
# of Obs 12,797 12,797 12,797 12,797 12,797 12,797
# of 49 49 49 49 49 49
Exporters

 

 

 

 

 

 

 

Note: 1) I control for country, industry, year-fixed effects, and factor endowments. I also control for the
log of GDP when the dependent variable is log(export), and the total export ratio to GDP when the
dependent variable is the export ratio. t-statistics are reported in parentheses. ***, **, * indicate
significance at the l%, 5%, and 10% level.

2) I use real interest rates without standardization.

46

 

Table 2-6: Robustness Checks

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Official First Sign Official liirst Sign First Sign First Sign
Liberal Liberal Liberal Liberal liberal liberal
Dummy Dummy Intensity Intensity dummy dummy
Liberalization -12.739 -3.750 -34. 146 - 10.933 -2.995 -6.547
(-1.l3) (-0.34) (-l.44) (-0.38) (-0.27) (-0.55)
Lib*External 12.893 15.442 4.024 4.753 13.044 19.034
finance (2.51)" (3.06)‘" (1.53) (1.88)* (2.65)‘” (2.48)"
dependence
Lib*Asset -13.702 - 19.450 -10.420 -11.872 -24.740 -27.855
flgibility (-0.85) (-1.39) (-1.76)“' (-2.06)"‘* (-1.81)* (-l.69)*
lnterest*Extemal -0.077 -0.082 -0.066 -0.067 -0.073 -0.071
finance (-2.59)"‘ (-2.73)"‘ (-2.28)** (-2.29)** (-2.40)** (-2.35)**
dependence
Interest 0.018 0.017 0.011 0.009 0.017 0.015
(0.60) (0.57) (0.35) (0.29) (0.51 ) (0.47)
Trade openness -3.259 3.755
(-().22) (0.21)
Trade Open -12.080
* External (-1.55)
finance dcp
Trade Open 6.892
*Asset (0.39)
tangibility
R-square 0.188 0.188 0.188 0.187 0.180 0.180
# of Obs 12,797 12,797 12,797 12.797 11.872 11,872
# of Exporters 49 49 49 49 44 44

 

Note: I use the trimmed data. I control for country, industry, year-fixed effects. the total export ratio to
GDP, and factor endowments. t-statistics are reported in parentheses. **"'. **. "‘ indicate significance at
the l%, 5%, and 10% level.

47

 

Chapter 3: A Two-Dimension Heterogeneous Firm Model

48

1. Introduction

Melitz (2003) explains export decisions based on firm—level productivity differences
and fixed costs of exporting. He assumes higher fixed costs for exporting than those
for domestic sales, and shows that only the most productive firms which can make
enough profits to cover higher fixed costs for exports will export. The intermediate
productive firms which can make enough profits to cover the fixed costs for domestic
sales choose to sell to the domestic market while the least productive firms exit the
market.

However, the Melitz model cannot explain the interesting fact in international
trade that some less productive firms export while some high productive firms do not
export. Bernard, Eaton, Jensen, and Kortum (2003) show the productivity dispersion
among firms, and find that exporters, on average, have higher productivity levels than
domestic sellers. According to the 1992 US. Census of Manufactures, exporters had a
33—percent advantage in labor productivity overall and a 15-percent advantage relative
to non-exporters within the same 4-digit industry. Interestingly, their Figure 2A and
2B show that some firms with low productivity levels are exporters while some firms
with high productivity levels sell only domestically.16

The Melitz model cannot explain this fact because he only allows productivity
to differ across firms while assuming fixed costs to be the same among exporters or
domestic sellers. The same amount of fixed costs interacting with firm-level hetero-
geneity in productivity leads firms to sort into serving the domestic or foreign markets
based on their productivity levels.

Many researches find that fixed costs for export are important components in af-
fecting firms’ export decisions. Roberts and Tybout (1997) report firm-level evidence
for sunk costs in exporting. Using data on Colombian firms, they find that a firm’s

current exporting status is largely determined by its previous export experience.17

 

1"Figure 3-1 and 32 come from their Figure 2A and 28.
1"'Similar findings are reported for other countries For example, see Bernard and Wagner (2001)

49

Becker and Greenberg (2007) also show that the effects of financial development on
exports are stronger when fixed costs are larger.

An interesting survey carried out by the World Bank, which is summarized in
the World Bank Standards and Trade Database, investigates the costs incurred by
exporting firms.(See Otsuki and Wilson (2004)) The survey reveals that exporting
firms commonly make additional investment in both compliance costs and new plant
or equipment, and that firms perceive the access to credits to be a major obstacle
to exporting. The survey also shows that fixed up-front costs have long gestation
periods, and that they are firm- or even employee-specific due to the different degree
of accessibility to credits. Thus, this survey points out that financial constraints can
lead to firm-level heterogeneity in fixed costs.

Therefore, in this chapter, I am going to present a heterogeneous firm model
in which firms differ in both variable and fixed costs. My model is general in the
sense that it can easily incorporate the factors affecting firms’ variable costs such as
productivity and labor quality, and the factors affecting firm’s fixed costs such as
financial constraints and searching costs.

My way of incorporating financial constraints into a trade model with heteroge-
neous firms follows Manova (2006). She extends the Melitz model and finds that
credit constraints interact with firm-level heterogeneity in productivity and reinforce
the selection of only the most productive firms into exporting. She incorporates fixed
costs in a way with repayments. If a firm cannot cover fixed costs before production,
the firm should borrow loans from financial sectors before making investments. In
obtaining outside finance, firms pledge tangible assets as collateral. She also assumes
that financial institutions cannot make perfect enforcement in terms of contractibility
because of imperfect information.

In her models, credit constraints affect firms in different countries and sectors dif-

ferently. Credit constraints vary across countries because contracts between firms and

 

for Germany and Bernard and Jensen (2004) for the United States.

50

investors are more likely to be enforced in countries with a higher level of financial
development. Sectors differ in their endowments of tangible assets that can serve as
collateral. If a financial contract is enforced, a firm makes payments to investors;
otherwise, the firm defaults and creditors claim collaterals. Firms, therefore, find it
easier to obtain external finance in countries with a high level of financial contractibil-
ity and in industries with more tangible assets. However, Manova focuses only on
partial equilibrium outcomes and does not allow firm-level differences in financial
constraints. Thus, repayments for exporters and domestic sellers are identical in the
same industry regardless of their productivity levels.

However, firms in the same industry can have different degrees of financial con-
straints. One reason for this is because investors may have different information on
individual firms. Some firms, for example, may have long credit histories or good rep-
utations. Investors will take this into account when they make investment contracts,
which specify repayments when the contracts mature.

Using my model with firm-level heterogeneities in both fixed and variable costs,
I can explain several interesting stylized facts when firms make decisions on entry
and export that cannot explained by the Melitz model. For example, my model can
explain why some less productive firms export while some high productive firms do
not export. My model can also explain the fact that some exporters do not sell in
the domestic market. According to the Korea Investors Service Incorporation (KIS),
46 out of 1,576 (2.9%) Korean firms in the survey reported exports without domestic
sales in 2005.

The structure of this chapter is as follows. In section 2, I set up the closed
economy model of two-dimension heterogeneous firms. I extend the model for the

open economy in section 3. Conclusions follow in section 4.
2. The Model for the Closed Economy

I incorporate financial constraints into the Melitz (2003) model of international

51

trade in order to allow firm-level heterogeneities in both variable costs and fixed costs.
Consider a country with a continuum of heterogeneous firms producing differentiated
goods. Consumers exhibit love of variety and can consume all available differentiated
products. The utility function for the representative consumer in each country is
given by the CES preference function with a constant elasticity of substitution (0),
which is greater than one. The maximization problem of the representative consumer

can be written as

maxU = [waQ q(w)(0"1)/"dw]”/(0‘1), subject to R = fwd) p(w)q(w)dw. (1)
where U and R represent her total utility and revenue. p(w) and q(w) are the price

of and the demand for each good.

The optimal consumption and expenditure decisions for each variety can be char—

acterized as

4(4) = 017931144): R1941”. (2)

where aggregate expenditure and aggregate price index P is denoted by
R = PQ = waQ r(w)dw, P = [fw69p(w)1—0(fw]1/(1—U). (3)

Production requires only one factor, labor (I), where wage is normalized to one. If a
firm does not have any financial constraint, it uses kd units of labor for the investment
that is necessary before production. Marginal production of labor is constant for all
firms with a given productivity level. Hence, to produce q units of goods, q / (,9 units
of labor are necessary, where (,0 represents the firm’s productivity level.

Variable costs are covered by internal finance, which implies that firms can com-
pensate their production costs by selling their goods. Fixed costs (kid), however,
should be covered either by internal financing or external funding. Firms that do
not have enough internal financing should borrow loans from financial sectors before
making investments. In obtaining outside finance, financial institutions cannot make

perfect enforcement in terms of contractibility because of imperfect information. In

52

particular, investors can expect to be repaid with probability Mm), which is less than
one. /\(w) also represents financial conditions for an individual firm. With probability
[1 — A(w)], financial contracts cannot be enforced, firms default, and creditors cannot
have repayments.

The timing is the following. At the beginning, a firm invests an entry cost (Ate).
After the investment, the firm learns its productivity level ((,9) and its financial con-
ditions (5 E %).18 Productivity levels and financial conditions may depend on each
other. Let g((p,l3) be ajoint distribution ofgo and (3, where (0 E [0, 00), and ,8 E [1, 00).

Given ((p, 13), a financial contract proceeds as follows. At the beginning of each
period, every firm makes a take-it-or—leave-it offer to potential investors.19 This con-
tract specifies the amount that a firm need to borrow, and its repayments [Fd((,9,13)[
when the contract is enforced. Revenues are then realized and investors receive the
repayments at the end of the period. When the firm defaults, investors receive noth-
ing.20

The problem of a profit-maximizing firm with ((9,3) can be written as

gainer) = poo..9)q<w..3>—q((a..9>/4—54%4—l. (4)

subject to 1) q((,0,l3) = W

2) 4409.13) E p(so.fi)q(so,.3) - (Mom/so 2 E1023)
3) B0213) E 44,, + “19”” 2 0

I
5

When there is external financing, two additional constraints bind the firm’s de-
cisions in the maximization problem. When a financial contract is enforced, an en-

trepreneur can offer at most her net revenues to creditors, which is represented by

 

1”If a firm’s internal funds are insufficient to cover all fixed costs, the firm will need to obtain
external financing to cover its fixed costs. 6 can be adjusted properly in this case. Thus, 6 represents
the firm’s financial conditions and the portion of external financing in general. See appendix C.

19The assumption of take-it-or-leave—it offers does not change my results. See appendix A for the
Nash bargaining solution in chapter 1.

20Some papers such as Manova (2006) consider collaterals in analyzing financial contracts. In my
model, the fraction of collaterals which goes to investors when firms default can be incorporated
into 3. See appendix D.

A((p,13) In addition, investors only extend finance to the firm if they expect a non-
negative net return. B((,9,13) represents the net return to creditors. Thus, constraint
(3) is the investors’ participation constraint with their outside options normalized to
be zero.21

Financial markets are assumed to be competitive.22 In competitive credit markets,
all investors earn zero expected profits. A firm, therefore, adjusts its repayments
so that the investors’ participation constraint is satisfied. Thus, B((p,fi) = 0 in
equilibrium, and the maximization problem reduces to the firm’s problem in the
absence of financial frictions, except for the credit constraint that repayments are no
greater than the firm’s net revenues. Hence, each firm optimally chooses the same
quantity and the same price, raises the same revenues, and earns the same profits
before repayments as in Melitz (2003). The amount of repayments does not depend

on firms’ productivity levels in equilibrium because Fd((,9, 6) = fikd.

The optimal pricing rule and individual profits can be derived as

p((p,)’3) = 31;, where % E a—‘i—I is the mark-up. (5)

«((9,3) = r(so) - q/so - F403) = %r(so) — filed. where 7‘09) = RIP/491"”-

Firm-level heterogeneities come from various sources such as differences in produc-
tivity, input quality, financial constraints, or searching costs. These heterogeneities
can be classified into two categories. One includes factors affecting firms’ variable
costs, and the other includes factors affecting firms’ fixed costs.

Factors that affect firms’ variable costs include productivity and input quality.
Both result in affecting prices of goods. If a firm has a high productivity level, for
instance, it can sell its good at a low price. The quality of labor may also change

variable costs. For example, assume that each firm find inputs with the quality t,

 

21This can be generalized. See chapter 1.

22The assumption of competitive financial markets is not critical. If financial markets are not
competitive, there can be the extra amount of the return to capital. In other words, outside op-
tions for creditors will not be zero in uncompetitive markets. However, the results do not change
qualitatively. See chapter 1.

where t differs across firms and follows a certain distribution. A well—known firm can
attract workers with higher quality than little-known firms because many workers are
eager to work in the well-known firm. Thus, the firm can select workers from a larger
pool of job applicants than little—known firms, and have high qualified workers.

If a firm has to pay 11) for one unit of labor with the quality t, the optimal price
of the good produced by the firm with a productivity level (p is $.23 If (p and t
have a joint distribution b((p, t), there exists a new distribution of 7,0 E (pt. The price
of the good is p((p,t) E p(t/)) = 5113' Therefore, productivity and input quality are
incorporated into one variable (1)9), which represents the heterogeneity in variable
costs.

Factors that affect firms’ fixed costs include financial constraints and searching
costs. Heterogeneity in firms’ financial constraints results in different amount of
repayments to investors, which can be interpreted as differences in fixed costs without
financial frictions. Searching costs may also be included in fixed costs. All firms
have to pay more or less for collecting information when they enter a new market.
For example, they need to pay for collecting data on the market demand. Some
entrepreneurs may have much experience in the market. Thus, they do not have to
pay much for collecting the market information, but those with little experience may
have to pay much for it. Firms\also have to pay some fixed amount of money for
searching qualified workers, which is differently distributed across firms. If there are
searching costs before production, constraint (2) in equation (4) should be changed to
A((p,B, S) E p((p,)3)q((p,/3) — q((,0, fi)/(p 2 Fd((p,,l3) + S, where S represents searching
costs that differ across firms. If 15’ and S have a joint distribution of d(l'3,S), there
exists a distribution 6(6), where e E Fd((,9,,3) + S = fikd + S. Therefore, financial
constraints and searching costs are incorporated into one variable (6) that captures

the heterogeneity in fixed costs.

To summarize, my model incorporates firm—level heterogeneities into two dimen-

 

2"’Recall that wage is normalized to one.

Ci"!
CI!

sions. Specifically, I allow for heterogeneities in both variable costs and fixed costs.
In the following, I will refer to all the heterogeneities in variable costs as that of
productivity, and all the heterogeneities in fixed costs as that of financial constraints.

Now, I examine firms’ decisions on entry and export. To enter, firms must make
an initial investment as a fixed entry cost (Ice > 0), which is thereafter sunk. After the
investment, they draw their own productivity levels and financial conditions. If their
productivity levels are low and their financial conditions are bad, they immediately
exit the market. If their productivity levels are high and their financial conditions are
good enough, they can make borrowing contracts with financial institutes specifying
repayments [Fd(/3)] in order to finance fixed costs for production. Even if they make
financial contracts successfully and stay in the market, they face a constant probability
6 in each period of a bad event that would force them to exit the market. The
6 is independent of productivity levels. Therefore, an entering firm with a specific
productivity level and financial conditions ((,9, 6) will immediately exit if its expected
profits are negative, or will produce and earn profits in every period until it is hit
with a bad event and is forced to exit. For simplicity, I assume that there is no time
discounting.

Equilibrium cutoff productivity levels and financial conditions can be obtained by
the zero cutoff profit (ZCP) condition and the free entry (FE) condition. The ZCP
condition implies that all firms in the market should have non-negative profits. The

marginal firm that has a zero profit can be represented by

7r(«19.13) = 0 4E r(so) = 131:.) <==> RIP/Nola" = .3de (6)
4095/3 = 1) = 0 <=> r(<o’) = kd <=> le’pso"l"‘1 = 01a
where (0' represents the critical productivity level of incumbent firms when they have

no financial constraints.

If a firm with (99(1), 6(1)) has a zero profit, there is another firm with (99(2),)3m)

that earns a zero profit if 13(1) 2 R[Pp(,9(1)["—1/crkd and 3(2) 2 R[Pp(,9(2)]"“1/rrk:d.

56

Therefore, the ZCP condition can be represented by a curve that satisfies 3(99, (,9’) =

[fly—1.

The FE condition implies that for potential entrants to make investments in order
to enter the market, the expected profit should be equal to zero. If the expected profit
for potential entrants is negative, no firm will enter the market, and if the expected
profit is positive, all potential entrants will enter the market. Hence, equilibrium
solutions cannot be found in both cases.

Each firm’s expected value function can be written as follows

(20.0.13) = max{0,t§(1—6)t7r((,9,3)} = max{0, -(157r((,9,)3)}. (7)

If the exit process does not affect the distribution of productivity and financial
constraints, the distribution g((,9, [3) must be determined by the initial draw of produc-
tivity levels and financial constraints conditional on successful entry. Hence, )u(',c, 3)

is the conditional distribution of g((,9, 3) on (,9 E [0, co) and )3 E [1, 00), where

)u((,9, /3)= gig—7:) if the firm IS in the market, (8)

= 0 otherwise,
where pr,” is the probability of successful entry. Let E = (1 / 0 )7r((p, 3 ) represent the
present value of expected profits, where fr((,9, {3) is the average profit of incumbent

firms in one period. Furthermore, we is defined as the net value of entrance. The FE

condition can be written as

v.=pr.-..v— k =0<=>rr<s2 9): 54’s (9)
m

Considering that decisions of potential entrants depend on the conditional distri-

bution of successful entry, average profits can be written as follows

4)=:/ ff ’0 ’0 ma) (42*..4)d.3d(2*. (10)

where 3*( ((,9* ,(,9) =l§rl04

An equilibrium can be characterized by a mass (M) of firms and the distribution
of productivity and financial conditions u((p,13) over the subset of (p 6 [0,00) and
)8 6 [1,00). Hence, the aggregate price level can be written by a mass of firms and

the average productivity level weighted by financial constraints:

[3*(30
P=l/: [1M (Tel— ”4114* 4144441141”)=Ml/<1—0>p(r2), (11)

/3*(so
where (a: [f:: /1p: (,9” 1 11(99 *,,3)d/’)’d(,9*[1/(‘7—1) is the weighted average of

incumbent firms’ productivity levels.

The ZCP condition can also be restated in terms of (,9’ and the average profit
because the amount of production depends only on productivity levels.24 Finally,

equilibrium solutions can be derived from the following two equations:

1r(<,9, l3): /: /W (,9 *,,i'3),1.t((,9*, l3)d[id(,9* (ZCP) (12)

i094 5): mm (FE)

Figure 3—3 shows the equilibrium solution for the cut-off productivity level and
financial constraints. (p* is obtained from the fact that the average profit of incum-
bents under the ZCP conditions should be equal to—4L p,” T.h,us the equilibrium ZCP
condition can be drawn at the cut-off productivity level (,9" . Firms that have drawn
((,0, [3) under the equilibrium ZCP condition can make positive profits and stay in the

market until they are hit with bad events while firms that have drawn ((,9, 3) above

the equilibrium ZCP condition will not enter the market.

0 Proposition 1: The entry decision depends not only on a firm’s productivity

level, but also on her financial constraints.

Proof: The equilibrium solution that IS obtained from equation( ()12 )%j=

 

 

2"From (2) and (5), 1442:) = [i—i]°“l.

74¢)

58

00 5*(99*,<.0’) .
f , / (3* — fi)g(<p*,5)d13d«p* depends on both (,9 and ,3. See appendix E for
Lp 1

the proof of the existence and uniqueness of the equilibrium solution.

Firms with low productivity levels and severe financial constraints will immedi-
ately exit the market while firms with high productivity levels and little financial
constraints can stay in the market. A more productive firm can charge a lower price,
sell larger amount of goods, and collect bigger revenues. Therefore, a more produc-
tive firm can earn more profits and cover the repayments to investors. A firm with
little financial constraints does not have to promise large amount of repayments to
investors when it makes contracts with them. Even if investors cannot collect large
amount of repayments, they can break even in the financial market and survive after
all because of the high contractibility of the firm. Lower repayments result in higher

net profits, and help the firm survive in the market.
3. The Model for the Open Economy

In the open economy model, 71 represents the number of foreign countries. Firms
are identical in terms of expected productivity levels and financial conditions. The ex
ante distribution of productivity and financial constraints is the same across all coun-
tries. Consumers share the same utility function. These symmetrical assumptions
ensure the same wage rate across countries, which is still normalized to one. I also
assume an ice-berg transportation cost (7'), which is greater than one. Furthermore,
I assume fixed costs for export (km), which are covered in part by external financing.
It can be shown easily that if Ta‘lkx > kd and there are no financial constraints, low
productive firms exit the markets immediately, intermediate productive firms only sell
domestically, and high productive firms sell both in the domestic and export markets.

The selling prices, revenues, and profits of a firm with the productivity level (,9

and no financial constraints are written as

p(w) = 791—»? if it is sold in the domestic market, and (13)

59

— ‘1'— o o I
_ p90 1f 1t 1s exported.

r(<p) = rd(c,o) if a firm does not export, and

= r(1(c,9)+ nrfip) 2 (1+ 717'1'”)rd(c,9) if it exports.

7r11(<p)== Mil—13d: and Track?) = T‘M—kx,

L
0 0'

 

where subscript d and :1: represent the domestic sector and the export sector, respec-

tively.

Variable costs of production for exports are covered by internal finance, which
implies that firms can compensate their production costs by selling their goods. Fixed
costs for exports (km), however, cannot be covered all by internal funding. Thus, firms
without enough internal funding should obtain loans from financial sectors before
producing goods for international markets.

In obtaining outside finance for exports, financial institutions cannot make perfect
enforcement in terms of contractibility because of imperfect information. In partic-
ular, investors can expect to be repaid with probability Agra), which is less than
one. Aisha) represents financial constraints of a firm that produces goods for foreign
markets. With probability [1 — A$(w)], financial contracts are not enforced, the firm
defaults, and creditors cannot have repayments.

With the given productivity level (90), a firm should also draw its financial con-
ditions for export (6 E X1;)° Thus, firms’ heterogeneities depend on the distribution
of productivity and two financial conditions, one for domestic sales and the other
for exports. They are jointly distributed, which can be written as h(1,o,[3,6) where
4p 6 [0,00), ,6 E [1, 00), and 6 E [1, 00). I assume no timing gap between the decision
on domestic sales and that on exports while Melitz does.25 In other words, potential
entrants confront both the domestic and export markets at the same time when they

enter the markets. If their productivity levels and financial conditions for domestic

 

2‘5 Melitz assumes that the decision on exports is made after the decision 011 domestic sales. There-
fore, there is no uncertainty regarding productivity levels when firms make decisions on exports.

60

sales are good enough, they will sell their goods in the domestic market. If their
productivity levels and financial conditions for exports are good enough, they will
export.

I make another assumption that even if investors cannot make perfect enforcement
because of imperfect information, they know in which sector, the domestic one or
the export one, the investments are made. This implies that financial constraints
in the domestic and exporting sectors are separable. However, it does not mean
that financial conditions in those sectors of a certain firm are independent. This
assumption can be described by the joint distribution h(<.0,13,6).

Financial contracts proceed as the same way as the case of the closed economy. At
the beginning of each period, every firm makes a take—it—or—leave-it offer to potential
investors. Contracts specify the amount of loans which a firm needs to borrow and its
repayments, Fd(<p, B, 6) and F3199, £3, (9), when the contracts are enforced.26 Revenues
are then realized, and investors receive repayments at the end of the period. In the
case of defaults, investors receive nothing.

The problem of a profit-maximizing firm with ((,9, .116) can be written as27

, F (90.3.6)
max Hm 30) = paw, 3,0)q(1(99./3, 9) - (mt/J, ,3, We - 41—3-— (14)

PdsPIaFdan
F J a
+n{Px(99,/J’,9)q(w,30) — 'rqa:(99,39)/99 -———’“( g )}.

‘ —0
subject to 1 qd(g0,/3,9) = w

, 'n‘UR

9:)“ pd(99,.3,9)q(99,x3.0) - qufiflVW 2: FaMfifl)
a: «a3 9)E pxtxb’ 9)q(<p,13 9) -Tqa:(¢,3,9)/<F 2 1%(9239)
)
)E

a;
a
§
Pv

__Ad+_d(_,3_6>()

_ kn: +

E7(79~‘330)
fi__ 2 0_

 

26If a firm can use internal financing, but it is not enough to cover all the fixed costs, the firm
uses both external financing and internal financing in order to cover its fixed costs. 6 and 0 can be
adjusted in this case. Therefore, 6 and 6 represent firm’s financial conditions and the fraction of
external financing in the domestic and export sectors.

27The maximization problem does not change qualitatively in the case of partial external financing.
See appendix C.

61

In competitive financial markets, repayments do not depend on firms’ productivity
levels: Fd(9o,fi,9) = filed and Fx(99,/3,9) = 91cm. Once again the ex-ante probability
of successful entry in the domestic market is denoted by prd. Moreover, pry; and 17rd,,
now represent the ex-ante successful entry in the export market and in both markets,
respectively. The survival ratio of potential entrants, which can be characterized by

the probability of entrance in either of the markets, is defined by pr,” 2 pm + W51: —

pTd$°28

co .3*<;*,»:’) >0
p'rd =/ / / h(99*,[3,9)(19d,13(199* (15)
90' 1 1
00 “($3,033) 00
pm pT’Z/ ‘/1*I)1‘/‘ h(99*,13,9)(113(19(199*
89:1:
3*93 93 393* 9:35) _ _
deOO/ /-(v*n0,)1/ M99“, ,5, 9)(19(1.3(199*.
‘Pfi:

The mass of exporting firms in the country is 111$ = prxi’lle. The total mass of
varieties available to consumers in the country is Mt = M + nMx, where M is the
total number of domestic sellers. The numbers of firms in both markets and the

aggregate price level are described as

1i_1__1l4x (16)

[3*
— P-—[/: /1 pd( 99, [3,9 9)1 WAIdpd(99,L3)(13(199*+/0:> /:*})l(99*,13,91 ‘7 nAIT
90:1:

1133(90 *,9 )d9d99*]1/( (1— —)0 , where p(1(99 *,;‘3)— — Wed/(x 11(99*,3,9)d9 ifa firm sells in the
1

00
domestic market and zero otherwise, and ,ur(99* ,9): pm.- 1/ h(99*, 3, (1)013 if a firm
1

exports and zero otherwise.

99;, represents the critical productivity level of exporters when they have no fi-
nancial constraints in exporting. 99;, can be defined as a function of 99’ because from

equation (2) and (13),

ra:( ' 99,416- I ,,I km 1/(0—1)
71171: 1. 1" =xi‘=>‘3°:r=7rlxil ' (”1

 

281n the Melitz model, prI = prdx, and pr“, 2 pm.

62

The ZCP condition in the export sector is similar to that in the domestic sector.
If an exporter with (99,1) ,9 (1)) earns a zero profit in the export market, there is
another exporter with (99(2 1,9(‘)) that also earns a zero profit in the market if 9(1) =
er_0[Pp99§,;1)]0‘1/0k$ and 9&1:er ”[Pp99x( 2)]‘7 l/akm. Therefore, the ZCP

conditions for the domestic and export markets can be written as

7rd(99, ,3, 9)— — 0 4:) rd(99 )2 /‘3Ad ¢=> R[Pp99]"' = 13015,, for the domestic sector
7rx(9p,3’3,9) = 0 4:) 7",-,(99) = 91:, 4:) er‘”[Pp99]"_1 2 901:, for the export
sector. Using
”(3(99 3:19)=0<=H‘d(99"_)=kd<=>R[PWla1=01~d
7r,(93{,,,,3,6_—1)—_ 0 4:» W 5,)— _. A2, :3 R71 "(P/3991]" 1:03,,

a = [:310-1 and 0 =['{}]0-1. (18)

we
The FE condition is the same as that in the closed economy with 77(99, ,3, 9) rep—

resenting the average profit of incumbents that are in either of the markets:

”(99 3 0)= --f- where prim: prd+13rx-IN'(1.T- (19)

Equilibrium cutoff productivity levels and financial constraints can be calculated

using the new ZCP and FE conditions.

3393.33) = gimww ns—mgtsxa (ZCP) (20)

7703.36) =

Ed. 00 3* IL 00 6* *
* ,* " - I' . * I * a * ‘

prm (p, /1 ”(1(99 73611111“? ifi)d1jd¥9 +n1fl‘i1l [993: f1 7713(90 361111159 361(16dS‘9

3393.33) = iii (FE)

Figure 3—4 shows the equilibrium cut-off productivity level 99* obtained from equa-
tion (20). The equilibrium cut-off productivity level determines the cut-off produc—
tivity level for exports 99; as well as marginal financial constraints for domestic sellers
and exporters (13 ,9). Firms that draw (90,13 ,9) under the Z C Pd earn positive prof-

its in the domestic market and stay in the market until they are hit by bad events

63

while firms that draw (99,13, 9) above the ZCPd cannot enter the domestic market.
Phrthermore, firms that draw (99, 13,9) under the Z CPJ; earn positive profits in the
export market and stay in the market until they are hit by bad events while firms that
draw (cp, ,3, 9) above the Z C Pm cannot export. Due to the assumption on higher fixed
costs for exports, the cutoff productivity level for domestic sales without financial

constraints is lower than that for exports.

0 Proposition 2: Even if fixed costs for exports are bigger than those for domestic
sales, it does not imply that only high productive firms can export. It is also possible

that some exporters do not sell in the domestic market.

Figure 3-5 shows the equilibrium ZCP conditions for exporters and domestic sell-
ers. Firms under the Z CPd can stay in the domestic market while firms under the
Z CPg; can stay in the export market. According to Figure 3-5, some low productive
firms which have little financial constraints can stay in the export market, while some
high productive firms which have severe financial constraints cannot export.

With a productivity level 991, Figure 3-6 shows cutoff values of financial constraints
in the domestic and export markets. The values of ,3, and 91 are obtained from Figure
3-5. The 91 and 91 are the values of the points on the ZCP conditions for the domestic
and export markets at the productivity level 991, respectively. 131 is bigger than 91
because of lower fixed costs for domestic sales than those for exports. Recall that
high 91 implies severe financial constraints. Due to lower fixed costs for domestic
sales, firms with more severe financial constraints can earn non-negative profits in
the domestic market, and stay in the market. Therefore, as in Bernard, Eaton,
Jensen, and Kortum (2003), exporters have higher productivity levels on average
than domestic sellers.

According to Figure 3—6, firms can stay in the domestic market if ,3 is lower than ,231.
If 9 is lower than 91, firms can also stay in the export market. The new equilibrium

ZCP conditions contain the face made by the L—shaped line from the vertical and

64

horizontal lines at 131 and 91. It can be shown using a graph with 3 dimensions. Each
axis should represent 99, 13, and 9, respectively. Even if the productivity level is high,
firms cannot export if their financial constraints for exports are not good enough,
which incur high repayments to investors.

If a firm has financial conditions (13 ,9) where 9 is lower than 91 and ,3 is higher
than 131, it will export without domestic sales. This result mainly comes from the
possibility that some exporters may have severe financial constraints in production
for domestic sales.29 According to the COMPUSTAT, 22 US. firms out of 23,425
produced only for exports, not for domestic sales in 2005.30 In small size economies,
I can find a larger number of those firms. According to the Korea Investors Service
Incorporation (KIS), 46 out of 1,576 (2.9%) Korean firms that reported the amounts
of domestic sales and exports focused only on exports without domestic sales in 2005.
The models that allow for heterogeneity only in productivity cannot explain these

phenomena.31

0 PrOposition 3: There is a large overlap zone of productivity levels between

exporters and domestic sellers.

According to Figure 3—6, the area below 91 shows the number of exporters while
the area above 91 and below [31 shows the number of domestic sellers. They all have
the same productivity level. This implies that there is a separation between exporters
and domestic sellers within firms with the same productivity levels. Thus, my model

can explain the pattern in Figure 3-1 and 3—2.

4. Conclusions

 

29If the distribution of (3,9) is highly correlated, most firms will be around the 45 degree line.
The probability of exports without domestic sales will be very low. If 90, B, and 9 are perfectly
correlated, the result is the same as that of Melitz, which implies that only high productive firms
can export and sell domestically.

30Precisely, their exports were greater than their net sales. I could not find the data on their gross
sales However, the data will Show at least that some firms focus mostly on exports, not on domestic
sales.

3‘ Melitz (2003) eliminates this case from the beginning. He assumes that. firms make decisions on
whether they will export or not, after they enter the domestic market.

65

In this chapter, I have presented a heterogeneous firm model in which firms differ
not only in productivity but also in financial constraints. My model is general in
the sense that it incorporates many factors affecting firms’ variable costs into the
heterogeneity in productivity levels, and those affecting firms’ fixed costs into the
heterogeneity in financial constraints. I show that firms with low productivity levels
and severe financial constraints will immediately exit the market while firms with
high productivity levels and few financial constraints can stay in the market. I also
show that even if the fixed costs for exports are bigger than those for domestic sales,
it does not imply that only high productive firms can export. Moreover, I show that
firms make different decisions on exports and domestic sales even when they have the
same productivity levels. This result mainly comes from the possibility that firms
with the same productivity levels may have different degrees of financial constraints.

My model has the strength in explain the stylized fact in international trade that
some firms with low productivity levels are exporters while some firms with high
productivity levels sell only domestically. It can also explain the extreme case that

some exporters do not sell domestically.

66

$2.885 .83 mo 25.
8.”. on...“ 8N

  

..9.. J. , .
mww xi. w .. . .rm .
A.“ . l odhl q. ...

Eatonxm ran. :3“ flotoaxoéoz can co..—

wm.~ 8.N we; 3..— o_._ 8; 3.6 :6 and and Nvd nmd and
A54. .92.. -mw.N -wm. .8.N .3; -3..— .a_._ .8.— .wwd .:.o .36 .owd -NVd .26 .33 .36 nNdv

.4.

 

 

 

 

   

., . m a
.x. .. . .1.
.5 ML rd . .m...
.. .5...“ .01... ..
- a. .. .. ..
.5. . p. .
.nh‘. Mm .3
”41.1. 1 u . .
_ _ . .,
.H. . \.
1.. ..
Wm... .11». . ;
.__.. «A. 7...
Li.— . {nu “M.Q
NR . .. w. .w,
.. H . \ ....... mm“
.. .4... _ ..
”n :4). _tpt... L .,,.J
I...» ...... a... . V
4 a a. .
. . .— . 1,.
.m .1... a...
hub ... ._. .. ..
.3.“ . ......n.
.c ..
i!“ _‘ h

7| ‘
-\

 

 

0&1 r '
V‘L.

L ”(wt
6
\O

 

 

.u\
awn
..
a... . w

u f...
1‘...
'1
f
,....
5..
v.
.....
...h
.r.
.m

I.»

a"?

 

 

o—

N—

 

 

 

#—

 

 

 

 

2
355 «o easy—3.6m
=82 _Eoo 9 $3632.. .23 .5: .6 sax ”3 2:5

67

£36.62; .33 We 23. eutaxm as .52 Etaxvcoz us c3
2: 92 23 an.” 2: «3 3... a: 8.. as .3 and and «I. “no and
$3. .02 $3 -me .8.~ -3. -:.._ -o... .8._ .33 a; 4.3 .85 -98 and .85 -25 «Nov

.t 2.. _ ... m. .
r... x
i . W r .
I m ~.I . H”.
'u.
a... . .. .W
W...
_ ..

WW.

.3 m.»
w» a.

.W

 

   

    

\ I "'
“'5
Ir "‘3‘
- i: ‘W:
I .V.; ‘ t$

Is};
9.“

«A .31. .-
. ”W5"
3‘?

‘h
a.
.

 

  

 

' - “mm

 

 

W
W

..
.-

.

.

I.
.

o

»

tfibfi '-
oo

 

   

:. . ..,. ‘ :1
”durum
‘R .

. ' . . ‘ ‘._ .. .. - . '-
. .-_. .'.I ' L '. . ' "' u _ ‘ ' x
u' l- .' ‘L .. .h _ I
p. .-

flfifia- '

-
in .

   

'\-

 

o—

éa.

 

 

 

thud“

 

N—

 

 

 

, ‘_.‘ '. I fis‘_',_1~.-“-s“v\'.'
hf

 

3

 

 

o _
$55 .3 owficoobm

 

:82 $32.. .35-.. 2 £2.88... .33 .55 .8 23. as 2&a

68

Figure 3—3: The ZCP condition in the Closed Economy

 

 

 

 

B _/
ZCP condition /
Exit
/ " Enter
/ ’/J/'
1 ./
O (0*

 

69

 

 

Figure 3—4: The ZCP condition in the Open Economy

 

36

Exh

/

/

ZCPd,/
, Doni
/’ Sde

I

CPxf

4/

Export

 

 

 

 

7O

 

Figure 3-5: The Distribution of Productivity and Financial Constraints

 

36

e.

ZCPd
I ZCPxf

. . Donf
Exn /
’/ Sde /
1" , ~” Export

 

 

 

0* 0X* 0'

 

71

 

Figure 3—6: The Distribution of Financial Constraints at the Given Productivity

. I
l l

 

9

Domestic l . Exit
Sale '

 

 

e. ____-____.___._- 1-

l
Donn’ Sale 8t l
Export l

 

 

 

 

72

APPENDICES

73

<Appendix A: Nash Bargaining solution>

In the case of the Nash Bargaining situation, the problem of profit-maximizing
firms in domestic sales and export will change as follows. I assume that contracts
include a portion of profits, which goes to investors at the end of the period. 9d and
(93: represent those portions in the domestic sector and the export sector respectively.

WI 193% Fflgflho) = pd(w)qcz(¢) - (MM/<19 - AdFdW) — (1 - mid/rd — 0.11.1199)

“1991:09qu (‘10) - quiipl/‘P — AxFI(99) ‘ (1 _ )‘Ifirkr _ grnri‘Pll (A-l)

subject to
(99—”R

1) (1.160) = 34—517
2) (11:0?) = W

3) A1169) 5 (1 - 6d)i1)(l( )(MW) - (MM/99- Fdlv )} > 0

4) AM?) 5 (1_6:1:)ipa:(99)qx(90) - TQmi‘rQ)/99 - F1199} _>_ 0

5) Bdi‘r?) E —kd+)\dFd(<P)+(1-Ad)tdkd+9d{Pd(99)(Id(99)—qd(<P)/<P—F§(99)} Z 0
6) 31.019) E —kx+A F’(so)+(1—Ax)t.;kx+6${px(so)qx(99)-Tq.r(99)/99-Fé(sv)} 2 0.

Due to competitive financial ma1 kets, constraints (5 ) and (6 ) are binding. Thus,

I can obtain

 

 

 

 

k 1—(1—A)t 9 v () ( )— .( )/"‘}

F9019): (fl Ad—Odd d}_ dwd “pad (1?dequ y ’ (A2)
km - - (1‘ 4r 0x (1 7’ r( )/+

F1019): {lxil— 1; )t }_{12(.:z~)1/\l(_)0L q 1

Solutions for optimal prices in (A.1), pd and 191:, are the same as those in the
maximization problem (12). Constraints (3) and (4) also are binding if the firm has

the cut-off productivity level. Then,

Ads?) E (1 - 0d)ipd(99)(1d(99) - Qd(‘P)/‘P - REM} (A 3)
211—0..){pd1wqd1w—w1>/«.o— A.11;1_-e.0:.>r.}+ "’d‘ @53qu ”’1:
Fé(90)={pd(99)q MU )-q(1(99 V90}

 

 

74

,\ -9 k {1—(1—,\ )t } k {1—(1-/\ )icl _
={ $51} d Act—Odd d z d '\d d l _FdW).

14:1:(90) E (1 - 0x){px(99)(h(99) - MAW/<19 — Iii-(99)}
E (1—9;c){pg;(cp)qx(<,0)-qu(99)/<,0- kr{1:\(1-0’\I)(1‘}+ 9:17{P;r(¢)qr(¢)-Tq.r(¢)/WJ} : 0.

:r- .r. Aer-9.1:

 

Fifi?) = {103(99)qu — 79$(99)/99}
: {A;It);6$}k37{l_(l_’\$)t$} = k‘T{.1-(l"’\I)t-1'} : 1756(59)

A1? - 1' Ar

Constraints (3) and (4) in the Nash Bargaining situation are the same as those
in the take—it-or—leave-it offer. This implies that there are no changes in critical
productivity levels for domestic sellers and exporters in the Nash Bargaining case. The
only difference is the way of repayments. In the Nash Bargaining case, repayments
are diversified into two ways, ordinary repayments and specific portions of profits.
If investors can have certain portions of profits, there is an incentive for them to
decrease ordinary repayments. As a consequence, total repayments to investors will
be the same, and cut-off productivity levels do not change due to different methods

of repayments.

75

<Appendix B: Comparative Statics>

Equilibrium cut-off productivity levels can be derived from the ZCP and the FE

conditions as in equation (13).

 

 

[17%)‘1 = Fde') + Fl‘pl’nf(9‘9;’z‘)a (B.1)
where Fd(<p) = [1_(l—i\f)td]kd Fr (90)— — “_(lffi'flrjk‘r, w; = w’iffill/(O‘I),
_ [1 0 ’>1

f(99')= (93/997071 -1,and Pa: - m

From (8.1),
61:, = [1 — G(<,9’)]Fdf(<p’) + [11— G (99;) ]F,nf( ((,9?)
“ex?” l-l1-G(19'515)]1519 l”— “[1 C(e )l E199” 1+Tl "ll—00191)] whom] I

 

_ ‘PM—lll —G( (99$)lFxn (3.2)
akeey 1:70 1[1— G( ,a’)]F,z,3[ }0- ;-0 1[1—G(')]F,+[1—G(.,9M’]Fn[,]0-
wéa‘lll - Gwmn (B-3)

1) Effects of the relaxation of financial constraints for domestic sellers
a) Effects on the 99’

From (B.2),

51,104 = (1—G(90’)l1~“.1[§0l"“1—(1—G(w’)le¢’””1+rl 011 G(e3)len[§0‘;l”‘1
—<P""1[1- G (wit) an

where (QI— — T99 ’IF:{—32 d(}l/ ‘7 1) )Fd(99) : __[1_’d::_d (silks!) and FAQ) = ll-(r 0-/\.rt1)k.p.

.l‘

 

 

e:_aéfia
dAd — dFd dAd'
dF _ Adi dkd d-——[1 —td+)‘dtfilkfi= —[1— —;1]kd

d
Differentiate (B. 2) in terms of F,.
(0 "1)5ke99'”(2% = -9(90 90’) )Fdl‘FlU 1% +[1-G(w’)ll§9l"“

+(0~1)l1-—: FWlel l” 2$§§3+ 9(s:’)Fd»0’” l%é— (1*G('~2’)l«2’””‘
—(0—1)[1——<,9’),]F '0 2;" —TI 0 gen, n[~ 104103321121
‘1‘9 (“if d 99.1 dtp’ d1"d

76

I

+01‘011— G<<21>1n1.’2:10-1+<a — 1> 1-011— 01.2r>11:1~1:,10 051104100—

d‘rr d‘r’ (“‘d
—(0 — 1).,9'0-211— G(.2£,)]F.m + .9'0 1 g.2( 5,,)F 1.401%
Apply (B2) in LHS and using 99:, = Tcp’{%}l/(U“1),
’ d ’ _ _ 0_ 0_ I MI YAI ;
gig—£3; = T 1{%} 1/( l)T{£FZ(11;}1/( l) : Land (1%,—’5}? = girl—l%{1—[rélU—I}
d~ I
The proof of 3:95, $—
.~2(.2'><0-1> = my I, so” 19(90)d<p
~ —2 d? ’ _ _
(0 — 1W" (73,3 = $.24)? f3 90" 19(90)d90 + WU? 10" 19(90)d10l’-

Apply Leibniz Intergral Rule.
_. (if: n— (.l (i U 1
[f°°<p 0 191.2>d.21'=.2;: 191.2,», —.2' lg<.2'>,i§+1:,° LEM...

where 99,, is the upper limit of productivity.

”‘lgwwcp + WFF’U‘IQW’H

d~—1~g(fi,) 1 d7"- ' (.9) 0 —1
dgp —U——_TQI:5W73{l—[%]U },and fi%_0—1T%FJ){1 [55:10 }

Then. 11— c.2><'1{1:210-1— .2'0—1>+21-011—G<.2;.>>n1.?;>0-1=

A (7— ~ _ {-0, 0——
[l—G .2>1Fd1.210 1117,1‘ ‘>+21 011— 01.2;>1F,n[.2.10 1,117,117,121.

d J?— [1‘G(‘P’)l{léla_l-som—l}+rl’”[1—G(¢§:)ln[¢}la"1
37%;: {01} l1”C(‘PlllelCPP—I+7I‘”[l—C(¢’l.)][~‘dn[$iln—l > O»

I AI (1F
EL — _Ld_d
Thus, (Md — dFddM < 0.

Examples: Pareto distribution with a > (a — 1), {,3 = [——‘—‘—]1/("‘1)<,9’.

a—0+l
’ d“ _ ’ —1 _ n-(a—l) 32
Therefore, figd—f; — 1, and [5:15]" — —.

(I

1—G(’) _ "I _ —a F —a 0— _ . F, (7—
W-(W—T {-131} /< Ike—2.211111“ 1)

 

“ ‘P 50.7:
d! _ I —l I F AI
(7% _ U—l{a-0‘+l + (fig)ana——3+1(Tfi)}/{a—g+l + (3)071”) 0+1FI } > 0
1d @951

 

32For .70 to have the finite value in Pareto distribution. a should be greater than ((7 —— 1).

77

b) Effects on the «,9;

From (3.3),
“#3:"1=70-111-G(19’)1Fx[§9]”‘1w’” ‘11- 0119’)+1Fd+[1—G(19;)1Fxn[§931°“
wéo’lil - C(sOQNFx"

Differentiate (B. 3) in terms of Fd.
d _ ~ _ n d

(0 - 1)5ke<p;" 23%;: -T" 19(10’)Fx[19]” 133713;;-

+1a—1170-111—011o'11a1151 2‘5

x .
+1933 19(10’)Fdd 1,0353%? 99;” ll‘GW'N 9%)]: "M1” 01711

211;: d , ——2 da,
+<>Ga—111— ¢'>1Fxn1;r 1’ Jia—g—w—lw [1-G(192)]Fx"3%5

Io—l

+901: (Wx)F$nd1d

 

 

d‘ra: d 4d»??? d
_‘LIEifi:_ T-l{%}_ l/(U- 1),”? d}l/(o—) 1) _1 m d (17:: = f, KIWI) {1_[fi]o—l}
103: (1:; ‘ d»: v“ 0-1 l—Gw’) I '
I 1
223;: _{_:.L "a 11 G111 <0
“”71 1" 1 30‘111-G<1~'111:1¢10 1+11— 61 ~151:1~rn1:x10- 1 ’
dF

%_%$1§1=12r10=r-01%§105<-01>,1o;=w1%;<}1/ 1

 

 

110:1:
(1 I I _ I
£73:_in:—1)/{70 la— o+l[%—]o— 1+a—a+1(£’_)an} <0'
d_.; claw

' dF -— 1—/\ 1;
5’43 => Same process with the case ofj—L (“d except mg = J’Kjfl—d < 0.

2) Effects of the relaxation of financial constraints for exporters
a) Effects on the 1,0’

From (B2),
6ke‘P'o"1 = [1‘ “993113111751“ -[1- G(¢’)]Fd19"’”1 + 71””[l — C(wiflFdnk’Zia”

-<P"”1[1 - G(19§:)]Fa:n
78

\

I

 

£13. _ d2 (1F;

d/\_’1j — (111; (1A;

dF; __ Art xkx— [1- tm+r\.r’:rlk:r _ —[1-l.TI/\7:r < 0

(”at _ A2 _ A '
J: :1:

Differentiate (B2) in terms of F x.
, I ~ ,, I
(a — 1161812’0-233; -= —g1<p’)Fd1<191” 1% +10 — 1111— G1 W1F31 1” 2 43%—

_f’
_w d", d rig, d~’
+9(99')F11<P'0 lfi—(U—llll—GWNFW'U 237LL—Tl ”9(%)Fd”l¢zla 13353—5;
_ ~17 d~ d~’ 1-’ _ d~’
+(0--1)‘r1 ”[1- C(wfcllenlwl Qfififfifi-(U-IWU 2ll-G<s93>lfln31fg

_ d~ d ’
+19” ‘9(19§;>F:rn 3:1- 31,: — 99'” ll1-- C(wé) 11-

I

Apply (B. 2) in ’LHS and using ——’r‘ = T—1{Fl} 1“” ”fl? d(}l/” 1) =1,

 

9’31 (1'?
I
99102 " —1
an nddi dp’ 1,0 — 11_ G( )_{l l%la }
de': ' sow—11141135113 1 < 0

‘fiI-lo— }11 0—13'11F31121I-Tw1-I11—<;1¢;11I«3n1ag10-

#331113) —~r-I1F1I/<I-l> sax—r191d11/“I 1)

/¢
1113; = {31%}/-(————- (_n 0+1)“ ‘)aFd+nF3} <0.

b) Effects on the 1,035

From (13.3),
511819;" 1 - —7”“lll—G(¢’)1Fxl§91"’1—19.’3"“[1—G1125)1F3+11—G(1,~;)1Fcn1q10-1

wé" ll1-- G(193)lF1-n

Differentiate (B. 3) in terms of F3.

10 — 116183;“ 3%: —TI-Ig1so'1Fx1¢1I~1§§;% + III-111 — 0112'11131I-1+

(0 - 1)T"’1l1- G<19( ’)lF.rlC9l” 43,5373; 3% — (0 -1)<193~"“2l1- 0119’)le3%

+993" 19119’ )Fdffrfifi‘Z—f- 91193)F n "ll 13%,? +l1- OWL-)lnliilfl

+(0 —1)l1—G(I:r)lF "[193]”“2—I} %3— —(0 -1)<23;°‘2l1- G(193)lF1-n3%
G(

_ d ’
w 19112111:an II — 11': 111—

Al Al . , . ,
Apply (B. 3) 1n the LHS and using iii—j = T—1{%}—1/(0—1)T{%§}11(a—1)= 1,

 

8911
d: I MI ,I
and33%=L.%£%&r,11— 11;. 1 11
ésefz={2.’.}11—G1.o§.11v11.931° 1— 10’.” 11+r" 111 G1 ’1.:11 10 1>0
.1... 0,-1 ll-G(¢&)1Flrnl~7rl” 901w 111—1.:1I1I.1.:1II-1 .

Examples: Pareto distribution with a > (0 — 1),

l-G'Ir " —a a (0— I .1: 0-
[1 0(127))']l=(£r)a“T if; d} /( 1)99x=719{p'd(}1/ 1)

‘Fa:
d! I _ 1 I ,.
fizgifo—gi{(§;)a {a—dU+l}n+a—U+l FI}/{Fd+( {7-)0F1;TI}>0
I .

\

 

(A

d“, _ _ ~k-
4.1:. => Same process with the case of —IC e111 ept ‘1’ 1" = _legfl; < 0_

3) Proof of Proposition (4)

Fd(<,9) = [1 — (1— Ad)td + filed/Ad, and F111,?) = [1 — (1 — ATV; + r]km/AI.

Then,7d dF — —kd//\d > 0, and “it - —k$//\x > 0.

. 5,. : 11—1 11—1.)t.+r11-.1d
A150 Fd [1—(1—Adfld‘l'rlkd)“ '

 

 

I I . I
80111 = £154 + Edi...
dr de dT‘d C11,]: (173

Q

U C(‘FlllehPlo 1+” ”[1——G(105.)1Fdn[.3’.10 1 Ad
‘9“,- [1 C(‘Pmlln k
[1 G(¢/)]Fd[¢]a— 1+Tl-011 G(‘P.z:)le”l¥/‘;rl0j X:
(l Arflr-i—Tj

_ {1; 1.1311141. 1111.21” 1— w’” 11+11— 01.11.12.” 11 1 11—1I1" 111:1, 1.113%“
‘ 0-1! 131r111— 1.1.0 11131.21” I+Ir1 011- 61.1111321510—11 ’
which is positive when A... and t... are not much different from Ad and t...

={_.§’_1}11—G1.o’1111.91° 1 -19"’ 11+71 ”11——G1..’.11n1.~.1°‘153
I

 

-{J.9_—}

 

 

 

 

 

 

b) 211’. _ fluid + 1.2.4.5.
dr — did drd (1%}; (171:
4193.} 1“" 111— G1. 11 g.

0 1 I” I11- G1.o11F.1.i10 7+11 G..1F.rn1.1" 11..
+191- ’}11 01.9.17111..1 -«I99 0* 1+r"‘ 11— (11.11.:111" 1..

11—G1.o$.11F.n-1.27.10*1+r0—111—G1.c1'11:11.11”- 1 1.
~ __ _ 1— —\ 1+

_ {A 1...1311_c1¢5.11..11.s.1” 1 —..’” 11w.” 111 a1 91131.115‘71" 1111111133 11
_ 0-1 131x17“*111—G1.p 11Fx1¢10-1+11—G1.oé.1 1117.74.31" 11} ’

which is positive when A; and t. are not much different from Ad and id.

80

 

 

,, 1
) Let rat2o :31} = TU—1{%§}
I w I

d . a_1F;de—P3:F
—7‘at20 = T ——2——4.
dr Fd

I __ I—ESE _Ed _ El. L1-(1—A1)tr+fl FA (1— AT)I;—r ()l—Adyd
Fde F‘BF _ Ade AdF‘T —Fd/\;r{1— [1—(1-\d)l)td+r]}= FdAa —[1—(1-—\d) t.d+r]

When Ax is smaller than Ad, which means that exporters have higher probability
of default, firatio > 0, which implies that the decrease of overall interest rates will
decrease the cut-off productivity level of exporters more.

When tx is bigger than td, which means that exporters can offer more collaterals,

agratz'o > 0, which implies that the decrease of overall interest rates will decrease the

cut-off productivity level of exporters more.

81

 

<Appendix C: Partial External Financing>

I assume that a firm finances a fraction 0: of its fixed costs by internal financing
and (l - a) by external financing, where 0 < a < 1. The firm’s profit maximization

problem is changed as

, , F .a
max IRea/.3) = p(w,fl)q(w,/5) - (Mam/99 - aka - 4552—)

P,Fd(<P)
. ,. r —0
subjeccto 1>q<to,.a>=”‘—*;;EL,—”

2) AMB) E pm 5mm?) - (1(99, 3%? — aka 2 Fdw, x3)
3) B(ap,,8) E —(1 — o)kd + w 2 0, where 0 < 0 <1.

In competitive financial markets, Fd(rp, [3) = 5(1— a)kd. The total amount of fixed
costs that the firm has to pay is akd+d(1—a)kd __-: Féw, [3). Define 13' E a+/3(1—a);
F c“to, [3) = ,Blk‘d. The firm’s problem is the same as that of no internal finance with
contractibility of ,5”. Therefore, it is enough to analyze the case that all fixed costs

are covered by external finance.

82

<Appendix D: Collaterals>

If there is a fraction td of collateral which goes to investors when a firm defaults,

the maximization problem should be changed as

max H(99,,3) = p(<,9,/3)<1(99,/3) — (Nam/«P — :lgF‘dW) - (1 — Eil’dkda
pJ’dW)

subject to 1) q(<p,,8) = w
2) Ame) E 20(99, .B)q(so, £3) - (1(99, BM? 2 Fd(99,/3)
3) B(<p,,/3) —_—; —kd + MW) + (1 — §)tdkd 2 0.

In competitive financial markets, Fdw, ,3) = d[1 — (1 -— §)td]kd = [23 — (,3 — 1)td]kd.
The firm’s problem is the same as that of no collaterals with contractibility of ,3' E
[B — ([3 — 1)td]. Hence, it is enough to analyze only the case of contractibility without

collaterals.

83

<Appendix E: Existence and Uniqueness of the Equilibrium Solution>
I prove them in the case of the closed economy.

0 Variations of the ZCP conditions

3(so“ so’) Pool"1

5”")(sosso 1— 1 1_1
THEM] / so*,II> ”MIIIsoI II>IIIIIIsol I" W

*(*so so)
=W[/: /6 W 1 W I"'I3>(13do*l HIM“

(so* so)
because P = {/00 I /1 p(99*,I3)1‘”]l[p.(99*,I3)d,3(l99*]1/(1“’).

5*(so*sso’) = afmfto’flsflo“,

I 00 53*(so*sso') 1 1
where f(so)E [ f , [I so*”- II(so*,II>IIIIIIso*1— .
‘P

By dividing both side of equation (6), I obtain 3"(99‘, 99’): [4]”‘1.
0 Considering the average profit

Ir*(so II) =§< <so*)—I'III.I= 1IIIIDIso 1‘” III
I)

3* so("' so)
= lR[/: /1( p(99*, I3)l "AI/399, 3) )dI3d99* ] 1[p99* )0 1 —3kd
R I or 0—1 I I 3713,99) Ito—1 t at —l
= mflso )[so l -I3kd, where f(so) E [ I/l so II(so .mdfi’do l
90

00 so Iso)
3333) =/.pI /,( {rIiIon'Nso’la‘l - I3/I‘d}os(so*I/3)df3d<o* =

5 (IP s0)f *.(os* 50’)
0T1]: f1 99l0_ 1/1 #(99 @333“? —A:d/ /f XII-(99‘ ,I’3)d3(l99* .

Horn the ZCP and the FE conditions,
[3* (so Io)
ZCP: 7r (99,3)—- —:/ (/)1[ 99**,3);I(99 , 3)(13d99*— —

B (9* aflf {3("K (99* $3,)
mg / o’l" 1I I(*so II) )IIIIsoII «If/'13 I3/I(993)(l13d99*.

FE: 7T(99,,3) =

przn

The equilibrium solutions can be obtained from

84

6k R 00 3 ($330,) I 1 .
p75; = mfg), f1 f(so )lso*l"‘ II(so*,B)d,3d-,o*
00 5*(sp*,so’)
‘kd / / I‘III(so*,I’I)dIIII9o*.
so’ 1

00 B (so Iso)
Using II*<so> .——I§,,f<so[so10-,%;= / , f (II*—Io)g<so*,o)IIIIIIso*.

99 1
B*( (so so)
Define1(99 )=:/ /f ([3 —3)9 (99, 3)d3d99*.

Apply Leibniz Integral Rule.
, II*(so*,<oU ) U II*<so*,so’)
1351 = /l (3* — II>g(so*,II)IIII%,I:7 — II (III' — II>g<so*.II)IIII

00 d I?*(so*sso')
+/‘p’ WAR/3* (6* —I3)g(so*,I3)d5dsI*
“3*(soso
ity/159) (so I3)d II:

(II -— o*>g<so*,o*>% - <o*—1>g<so*.1>:7}I + /

B*(so*sso') d * ,

1

I9(sosso) 1, .. 1, .(oo)
= / (1-0)so*"‘ so“’9(so*,I3)d.5 = (1-0)s9*"‘ 'o‘” /1 g(s9*sI3)d-I3-

III<so’) _
dpl _-
I3 (so*.so') * 00 fi (o‘oo')
= — [I (I3 —/3)9(so*,I3)dfi+/ ,(1-0)s9*""so"" [I 9(so*s/3)dI3d<o*
sp
5*(sp*,spl) I (U— 01) 00 l B#(¢t,¢I)
=— [I (II— —II>g*(so II)IIII— T/ Iod- /l g(so*,II)IIIIIso*
‘P
< 0

Therefore, 1(99’) is a monotonic decreasing function.

When sp’ => 99U => ooU
II (soU f 7
limp,_>‘pU90’U)=:/U/1)(,3*—,I3)g(,p* mclddw :0,

where IpU is the )highest productivity level. Thus, the lower limit 18 zero.

Whengo’ =,>O
12'me 99)=:/ U/W II -II>g(so*,II>IIIIIIs*=

85

BU _
=ZO/ /U( (30— (so I3)dI'3dso

oo fil ~
z]: /:3U [3Ug( UMP" ,8)d3d(p* _ f0 [39(90*y/I3)d.3d§7* : X3U _ [3 = 00

1

because /: fog g(99* ,[3) d[3d<p*— — 1. [3 IS the average of financial constraints.

Hence, the upper limit 18 infinity.

Therefore, the sufficient condition for the existence of equilibrium solutions is
[3U —B > 673’s. In the case of extremely high 67:15, there may be no solutions. However,
it is not plausible because of very high value of [3U when there are firms with severe
financial constraints. For instance, 5U = 00 for the firm that cannot make contracts
with any investor. Moreover, if there is a solution, the solution is unique because of

the property of monotonic decreasing functions.

86

REFERENCES

87

o Acemoglu, Daron, Pol Antra’s, and Elhanan Helpman (2005), "Contracts and
Technology Adoption", Mimeograph, Harvard University.

0 Antra’s, Pol (2005), "Incomplete Contracts and the Product Cycle", American
Economic Review, XCV , pp. 10544073.

0 Arestis, Philip, Panicos O. Demetriades, and Kul B. Luintel (2001), "Financial
Development and Economic Growth: The Role of Stock Markets." Journal of Money,
Credit and Banking, 33(1), pp 16—41.

0 Bacchetta, Philippe (1992), "Liberalization of Capital Movements and of the
Domestic Financial System." Econometrica, New Series, 59(236), pp 465-474.

0 Becker, Bo, and David Greenberg (2005), "F inancial Development and Interna-
tional Trade", University of Illinois at Urbana- Champaign mimeo.

0 Becker, Bo, and David Greenberg (2007), "F inancial Development, Fixed Costs
and International Trade" mimeo, University of Illinois at Urbana-Champaign.

0 Beck, T. (2002), "Financial Development and International Trade. Is There a
Link?", Journal of International Economics 57, p.107-31.

0 Beck, T. (2003), "Financial Dependence and International Trade." Review of
International Economics 11, pp 296-316.

0 Bekaert, G., Harvey, C. and C. Lundblad (2005), "Does Financial Liberalization
Spur Growth." Journal of Financial Economics 77, pp 3-55.

0 Bernard, A.B., Eaton, J., Jensen, J.B., Kortum, S., (2003), "Plants and Pro-
ductivity in International Trade", The Aemrican Economic Review, 93, 1268-1290.

0 Bernard, A. B. and Jensen, J. B., (2004), "Why Some Firms Export?", Review
of Economics and Statistics, 86(2), pp 561-569.

0 Bernard, A. B. and Wagner, Joachim (2001), " Export Entry and Exit by German
Firms ", Weltwirtschafliches Archiv, 137(1)

0 Braun, M. (2003), "Financial Contractibility and Asset Hardness." University
of California - Los Angeles mimeo.

88

o Caselli (2005), "Accounting for Cross-Country Income Differences." In Aghion,
P. and S. Durlauf (eds.) Handbook of Economic Growth.

0 Chaney, T. (2005), "Liquidity Constrained Exporters", University of Chicago
mimeo.

o Demirguc—Kunt, Aslie and Levine, Ross (1996), "Stock Markets, Corporate Fi-
nance, and Economic Growth: An Overview", World Bank Economic Review 10, pp.
223-39.

0 Demirguc-Kunt, Asli and Maksimovic, Vojislav (1996), "Financial Constraints,
Uses of Funds and Firm Growth: An International Comparison." World Bank Work-

ing paper.

0 Garcia-Vega, Maria and Guariglia, Alessandra (2007), "Volatility, financial con-
straints, and trade." University of Nottingham.

0 Gupta, N. and Yuan, K. (2004), "On the Growth Effect of Stock Market Liber-
alizations." University of Michigan mimeo.

0 Henry, Peter Blair (2000a), "Stock Market Liberalization, Economic Reform,
and Emerging Market Equity Prices." The Journal of Finance, 55(2), pp 529-564.

0 Henry, Peter Blair (2000b), "Do Stock Market Liberalizations Cause Investment
Booms?" Journal of Financial Economics 58, pp 301-334.

0 Hermalin, B., Rose, A. (1999), "Risks to lenders and borrowers in international
capital markets." In: Feldstein, M(Ed.), International Capital Flows. Chicago Uni-
versity Press.

0 Hur, J ., Raj, M. and Y. Riyanto (2006), "The Impact of Financial Development
and Asset Tangibility on Export." World Development 34(10), pp 1728-41.

0 Iacoviello, Matteo and Minetti, Raoul (2006), "Internation Business Cycles with
Domestic and Foreign Lenders" Journal of Monetary Economics 53, pp 2267-82.

0 Jappelli, Tullio and Pagano, Marco (1994), "Saving, Growth, and Liquidity
Constraints." Quarterly Journal of Economics, 109(1), pp 83—109.

89

o Jayaratne, Jith and Strahan, Philip E. (1996), "The Finance-Growth Nexus:
Evidence from Bank Branch Deregulation." Quarterly Journal of Economics, 111(3),
pp 639-70.

0 Ju, J. and S.-J. Wei (2005), "Endowment vs. Finance: A wooden Barrel Theory
of International Trade", CEPR Discussion Paper 5109.

o Kaminsky, Graciela Laura and Sergio L. Schmukler (2003), "Short-Run Pain,
Log—Run Gain: The Effects of Financial Liberalization." IMF Working Paper WP / 03
/ 34, International Monetary Fund.

0 King, R. G. and Levine, Ross (1993), "Finance, entrepreneurship and growth:
theory and evidence." Journal of Monetary Economics, 32(3), pp 513—42.

0 Kletzer, K. nad P. Bardhan (1987), "Credit Markets and Patterns of Interna-
tional Trade", Journal of Development Economics 27, p.57-70.

o Levchenko, Andrei (2004), "Institutional Quality and International Trade", IMF
Working Paper WP / 04 / 231, International Monetary Fund.

0 Levine, Ross (1997), " Financial Development and Economic Growth: Views and
Agenda", Journal of Economic Literature 35(2), pp. 688-726.

0 Levine, Ross and Zervos, Sara (1998), "Stock Markets and Economic Growth."
American Economic Review, 88(3), pp 537-58.

0 Manova, Kalina (2006), "Credit Constraints, Heterogeneous Firms, and Inter-
national Trade", Job market paper, Harvard University.

0 Manova, Kalina (2008), "Credit Constraints, Equity Market Liberalizations and
International Trade", Journal of International Economics 76, pp 33—47.

0 Matsuyama, K. (2005), "Credit Market Imperfections and Patterns of Interna-
tional Tiade and Capital Flows", Journal of the European Economic Association 3,
p. 714-23.

0 Melitz, Marc J. (2003), "The Impact of Trade on Intra—industry Reallocations
and Aggregate Industry Productivity", Econometrica, 71, 1695-1725.

90

o Nunn, N. (2007), "Relationship-Specificity, Incomplete Contracts, and the Pat-
tern of Trade", Quarterly Journal of Economics 122(2), pp 569-600.

0 Otsuki, Tsunehiro and Wilson, John S., (2004), "Standards and Technical Reg-
ulations and Firms in Developing Countries: New Evidence from A World Bank
Technical Barriers to Trade Survey", World Bank.

0 Rajan, R. and L. Zingales (1998), "Financial Dependence and Growth", Amer-
ican Economic Review 88, p.559-86.

o Roberts, Mark J. and Tybout, James R. (1997), "The Decision to Export in
Colombia: An Empirical Model of Entry with Sunk Costs", American Economic
Review, 87(4), pp 545-564.

0 Sachs, J. and A. Warner (1995), "Economic Reform and the Process of Global
Integration." Brookings Papers on Economic Activity 1, pp 1-113.

0 Shen, Chung-Hua and Lee, Chien-Chiang (2006), "Same Financial Development
yet Different Economic Growth—Why?" Journal of Money, Credit, and Banking,
38(7), pp 1907-1944.

0 Suwantaradon, Ruanjai (2008), "Financial Frictions and International Trade",
Job market paper, University of Minnesota.

0 Svaleryd, H. and J. Vlachos (2005), "Financial Markets, the Pattern of Indus-
trial Specialization and Comparative Advantage: Evidence from OECD Countries."
European Economic Review 49, pp 113-44.

0 Wacziarg, R. and K. Welch (2003), "Trade Liberalization and Growth: New
Evidence." NBER Working Paper 10152.

0 World Bank (1997), Ezpanding the Measure of Wealth: Indicators of Environ-
mentally Sustainable Development. World Bank, Washington, DC.

0 Korea Investors Service Incorporation (KIS), database on-line.

0 International Financial Statistics, International Monetary Fund, on-line
0 Standard & Poor’s, COMPUSTAT database on-line.

0 United Nations Statistics Division Database, on-line

0 World development indicator, on-line

91

1293 03062 5705

“
.Illll
H
ill
H
H
II”
“I
Em
ll"
"
Illi-

3