Tel Aviv University The Gershon H. Gordon Faculty of Social Science

The Eitan Berglas School of Economics

Migration of Firms, Migration of People and the Geographical

Distribution of Economic Growth

THESIS SUBMITTED FOR THE DEGREE "DOCTOR OF PHILOSOPHY"

by

Gilad Aharonovitz

SUBMITTED TO THE SENATE OF TEL-AVIV UNIVERSITY

June 2006

This work was written under the supervision of

Professor Daniel Tsiddon

אוניברסיטת תל אביב

גורדון. ע גרשון ה"הפקולטה למדעי החברה ע

ש איתן ברגלס"בית הספר לכלכלה ע

הגירה של אנשים וחלוקתה , פירמותהגירה של

הגיאוגרפית של הצמיחה הכלכלית

"דוקטור לפילוסופיה"חיבור לשם קבלת תואר

מאת

גלעד אהרונוביץ

א"הוגש לסנאט של אוניברסיטת ת

2006 יוני -ו " תשסוןסיו

עבודה זו נעשתה בהדרכתו של

פרופסור דניאל צידון

Migration of Firms, Migration of People and the Geographical

Distribution of Economic Growth

by

Gilad Aharonovitz

Abstract

The migration of firms and people between geographic units (cities, states or

countries) both influences and is influenced by the economic conditions in those

geographic units. This study examines economic growth and income distribution in

developed and less-developed geographic units and the interaction between them through

migration. This is accomplished by analyzing the interaction between economic growth

and various other factors, such as prices and wages, in each location and between the

different locations.

The first chapter analyzes the effect of the movement of firms between countries

on both the home and destination countries. Foreign Direct Investment (FDI) and the

activity of foreign firms can have a substantial effect on an economy. This chapter studies

the short and long run effects of the migration of locally-owned firms from a developed

country to an undeveloped country on the output and growth rate of each in the presence

of "home bias". The chapter analyzes the direction of firm migration over time, firm

ownership, GDP, GNP, wages and long run growth rates using a model in which the

source of growth is the increase in the number of firms (which produce with decreasing

marginal productivity). The opening-up of a less-developed country to the entry of

foreign firms leads to the movement of firms towards it and to a (possibly) faster growth

This chapter attempts to separate migration into two different streams (job-related and

housing-related) and examines the differences between the two, as well as the differential

effects of various factors on the more and less-educated populations within each stream.

Housing-related migrants tend to be less educated, and some of them tend to be positively

affected by the average wage in the origin state, whereas job-related migrants tend to be

better educated and some of them tend to be affected in the opposite manner by the origin

state’s average wage. Policy implications are also discussed.

The final chapter integrates the cost of commuting into a framework which is

close in spirit to that of the second chapter in order to analyze the effect of the land-

obstacle between Israel and the Palestinians on both sides. The short run effects of the

security land obstacle on both the Palestinian economy and on the level of terror are

obvious. This chapter analyzes the long run economic effects. Again, a theoretical model

of two cities with heterogeneous populations and two sectors (one of which is

characterized by an LBD process) is presented. The results of the model show that even if

both cities have the potential to independently develop in isolation from one another, the

introduction of commuting between the two cities can lead to a situation in which the

initially-less-developed city deteriorates while the initially-developed city prospers. The

relevance to the Israeli-Palestinian situation is established. The obstacle will force the

more talented Palestinians to work in their own cities and will therefore stimulate the

development of technology-oriented industry in the Palestinian economy through an LBD

process. In the long run, the Palestinian economy will benefit from the obstacle. Israel, on

the other hand, in the case of a conflict in the far future, might find itself facing a more

developed enemy.

rate in the future. Various government policies towards the entry of foreign firms are

examined and it is found that for the undeveloped country harsher policies towards

entering firms lead to better results in the long run. Counter-policies that can be adopted

by the developed country are briefly discussed.

The second chapter analyzes the effects of growth on migration between cities

with populations that are heterogeneous in ability and the effect of that migration on their

growth and explores the connection between the geographical distribution of the

population and that of earnings. A theoretical model of cities is presented in which the

city manufactures two goods – one simple and one complex - in which each city exhibits

its own Learning-by-Doing (LBD) process. The possibility of each city arriving at a

different steady state in which one city produces the simple good while the other

undergoes the LBD process and produces the complex good is demonstrated, even if each

city could have developed had it been isolated. The higher wage in the more developed

city will tend to attract the more talented residents of the less developed city, while less

talented residents in the developed city will tend to migrate to the less developed city in

search of lower real estate prices. The distribution of wages and mobility are also

discussed. Widely employed government policies to encourage development and reduce

inequality are shown to be largely ineffective. An effective policy alternative is

characterized.

The third chapter empirically analyzes population movements between states

within the USA, comparing them to the result of the model of the second chapter. Internal

migration is an important phenomenon that affects development at both the city and state

levels. Researchers have traditionally viewed gross migration as a homogeneous stream.

Government policy is analyzed throughout the dissertation. It is found that

policies currently in wide use probably lead to unfavorable long run results in the

situations described above. Alternative policies, which at first glance seem to be

ineffective, are analyzed and found to be successful in the long run.

הגירה של אנשים וחלוקתה הגיאוגרפית של הצמיחה הכלכלית, הגירה של פירמות

מאת

גלעד אהרונוביץ

תקציר

) 'מדינות וכו, מחוזות, ערים(שונים אנשים בין אזורים גיאוגרפיים הגירה של הגירה של פירמות ו

דיסרטציה זו בוחנת את הצמיחה . מושפעת מהם אך גם על התנאים הכלכליים באזורים אלו ותמשפיע

של האזורים אחד על הכלכלית וחלוקת ההכנסות באזורים מפותחים ושאינם מפותחים ואת השפעתם ההדדית

בין ליחה הכלכלית בין הצמשי ניתוח האינטראקציה "בחינה זו נעשית ע. מסוגים שוניםבאמצעות הגירההשני

.בין אזור אחד למשנהובתוך כל אזור בנפרד ו, ושכרמחיריםכגון , מגוון משתנים נוספים

וצאפירמות בין מדינות על מדינת המשל מעבר ההפרק הראשון של הדיסרטציה מנתח את השפעת

תית על הפעילות ופעילות של חברות זרות במדינה משפיעות בצורה מהו) FDI(השקעות זרות . ומדינת היעד

של הגירת חברות בבעלות ארוכות הטווח ההשפעות פרק זה מנתח את ההשפעות קצרות הטווח ו. הכלכלית בה

) Home Bias(בסביבה בה קיימת העדפה למדינת הבית , מקומית ממדינה מפותחת למדינה שאינה מפותחת

, התפתחות ושינויי הבעלות, פני זמןהפרק בוחן את כיווני ההגירה של החברות על . על התוצר וקצב הצמיחה

הוא גידול בכמות הכלכלית השכר וקצב הצמיחה תוך שימוש במודל בו מקור הצמיחה , ג"התמ, ג"התל

פתיחת . גידול המושפע הן מרווחיות החברות והן מכמות הידע, המתאפיינות בתפוקה פוחתת לגודל, החברות

גירה של חברות לכיוונה ולעיתים לקצבי צמיחה מהירים מפותחת לכניסת חברות זרות מביאה לה-המדינה הלא

מפותחת כלפי כניסת - יכולה לנקוט המדינה הלאםבה) Policies( מדיניות כלימגוון . בתקופות הבאותיותר

ואצל חלקם סיכויי ההגירה מושפעים בצורה חיובית מהשכר הממוצע , לצרכי מגורים הם פחות משכילים

בעוד מהגרי עבודה הם יותר משכילים וחלקם מושפעים בצורה הפוכה מהשכר הממוצע , במדינת המקור

.השלכות מדיניות של ממצאים אלופרק זה כן נדונות ב. מקורבמדינת ה

לתוך מסגרת ניתוח ) Commuting(בדיסרטציה משלב את עלות ומאפייני היוממות הפרק האחרון

בין ישראל הנבנית על מנת לנתח את השפעת גדר ההפרדה וזאת , הדומה ברוחה לזו המוצגת בפרק השני

על רמת הטרור הקצר על הכלכלה הפלשתינאית והשפעות גדר ההפרדה בטווח. ינאים על שני הצדדיםתלפלש

אורטי של שני אזורים ימודל ת. הפרק מתמקד בניתוח ההשפעות הכלכליות בטווח הארוךו, הן ברורות וידועות

עם למידה תוך ( טכנולוגי-טכנולוגי ולא, שני סקטורי ייצור,עם אוכלוסיות הטרוגניות ביכולות) שתי ערים(

המודל מראה כי אפילו אם שתי הערים היו . מוצג ומנותחואפשרות ליוממות ) וגיסקטור הטכנולכדי עשייה ב

אפשרות היוממות בין הערים יכולה , )ללא יוממות, הימבודדות אחת מהשני(יכולות להתפתח באופן עצמאי

לגרום למצב בו העיר שהייתה פחות מפותחת בתחילה מתדרדרת טכנולוגית בעוד העיר שהייתה יותר מפותחת

כאשר הגדר תכפה על , פלשתינאי מוצגת-רלוונטיות המודל למצב הישראלי . משגשגתכנולוגית ט

וכך תעודד את התפתחות הערים ) ולא בישראל(פלשתינאים מוכשרים יותר לעבוד בערי מגוריהם

. גדרהכלכלה הפלשתינאית תרוויח מה, בטווח הארוך, לכן. פלשתינאיות בתהליך של למידה תוך כדי עשייה

מבחינת מתקדם יותרצדעלולה למצוא את עצמה מול , במקרה של עימות בעתיד הרחוק, לעומת זאת, ישראל

.כלכלית וטכנולוגית

נמצא כי אלטרנטיבות מדיניות הנמצאות בשימוש . מדיניות ממשלתית מנותחת לכל אורך הדיסרטציה

למרות שהן , רצויות בטווח הארוךמובילות לעיתים קרובות לתוצאות לא במצבים המתוארים למעלה נרחב

מנותחות ונמצאות , שבמבט ראשון נראות לא אפקטיביות, אלטרנטיבות מדיניות אחרות. תורמות בטווח הקצר

.כיעילות בטווח הארוך

ונמצא כי מדיניות קשוחה יותר לכניסת חברות משיגה תוצאות טובות יותר בטווח , יםחברות זרות נבחנ

.נבחנת בקצרה, ה יכולה לאמץ המדינה המפותחתאות, מדיניות נגד. הארוך

הפרק השני מנתח את השפעות הצמיחה הכלכלית על הגירה בין ערים בהן האוכלוסייה הטרוגנית

אוגרפית יוכך בוחן את הקשר בין ההתפלגות הג, של הגירה זו על הצמיחההחוזרות את השפעותיה וביכולות

עיר יכולה לייצר שני כל . עריםשתי ג מודל תיאורטי של בפרק מוצ. של האוכלוסייה לחלוקת ההכנסות

, יכולה להשתפר) בנפרד(כל עיר כאשר במוצר המורכב , )טכנולוגי(ומורכב ) לא טכנולוגי( פשוט –מוצרים

מוצגת אפשרות בה כל עיר במסגרת המודל ).Learning by Doing(בתהליך של למידה תוך כדי עשייה

תייצר רק את המוצר הפשוט בעוד לא תתפתח טכנולוגית ועיר אחת , שונה) Steady State(תגיע למצב עמיד

כן . תעבור את תהליך הלמידה תוך כדי עשייה ותייצר את המוצר המורכב-תתפתח טכנולוגית עיר אחרת

כל אחת מהן הייתה ,במצב שבו אם לא הייתה אינטראקציה בין העריםאף מוכח כי דבר זה יכול לקרות

השכר הגבוה בעיר המפותחת יותר ימשוך את המוכשרים מהעיר הלא . לעיר מפותחת, אופן עצמאיב, הופכת

בעוד הפחות מוכשרים מהעיר המפותחת יהגרו לעיר הלא מפותחת בחיפוש אחרי מחירי דיור זולים , מפותחת

מגוון . פתחשני זרמי הגירה אלו מחזקים את המצב העמיד שכן הם מונעים מהעיר הפחות מתקדמת להת. יותר

שוויון ה-איין את דד פיתוח ולהקטולעבמטרה אופן נרחב נמצאים בשימוש במדיניות ממשלתית אשר סוגי

העיר התפתחות מקשים על ולעיתים אף , ים כלא אפקטיביים ונמצאיםבאוכלוסיה נבחנבחלוקת ההכנסות

מאופיינת ארוך באותה עלות אשר יכולה לסייע בטווח האלטרנטיבת מדיניות . בטווח הארוךהפחות מתקדמת

.בפרק

אוכלוסייה בין מדינות בתוך ארצות הברית התנועות את בוחן אמפירית בדיסרטציה הפרק השלישי

המשפיעה הן על רחבת היקףפנימית היא תופעה הגירה . קודםומשווה אותן לתוצאות המודל מהפרק ה

על זרמי ההגירה ברוטו רך כללכלה בדהספרות הסת. התפתחות ערים והן על התפתחות מחוזות ומדינות

הגירה לצרכי עבודה והגירה , פרק זה מנסה לפרק את זרמי ההגירה לשני זרמים שונים. כזרמים הומוגניים

ובוחן הן את ההבדל בין שני הזרמים והן את ההשפעות הדיפרנציאליות של גורמים שונים על , לצרכי מגורים

נמצא כי מהגרים באמצעות אמידת סיכויי ההגירה . אחד מהזרמיםמשכילים יותר ומשכילים פחות בתוך כל

Acknowledgements

I would like to thank Professor Daniel Tsiddon, my teacher and advisor, for his

professional and moral support.

I also wish to thank the Eitan Berglas School of Economics, Tel Aviv University, for the

financial, organizational and administrative support that enabled me to conduct this

research.

I wish to acknowledge my fellow doctoral students and workshop participants at Tel

Aviv University for their useful comments and suggestions.

Contents

1. Introduction 1

2. Migration of Firms, Home Bias and the Geographical Distribution

of Growth 12

2.1 Introduction 12

2.2 The Model 17

2.2.1 A Single Country Growth Model 17

2.2.2 The Complete Model 20

2.3 Entry Policies 24

2.3.1 Free Entry 25

2.3.2 Joint Ownership 27

2.3.3 Increase in Local Ownership over Time 32

2.4 Discussion 38

3. Inequality Within Cities, Inequality Between Cities and Patterns of

Urbanization 45

3.1 Introduction 45

3.2 The Model 53

3.3 Equilibrium 55

3.3.1 Steady State for an Individual City 56

3.3.2 Two Cities with Migration 61

3.3.3 Wage Inequality and Inter-Generational Earnings Mobility 64

3.4 The Model with Real Estate 65

3.5 Government Policy 72

3.5.1 Minimum Wage 73

3.5.2 Wage Supplement 73

3.5.3 Subsidizing Median Wages 75

3.5.4 Building a University in the Non-Technological City 76

3.5.5 Reducing Migration Costs Between Cities 76

3.5.6 Effective Government Policy 77

3.6 Discussion 79

4. Internal Migration in the United States – Who Moves and Why 89

4.1 Introduction 89

4.2 The Model 94

4.3 The Estimating Equations 96

4.4 The Data 98

4.5 Results 102

4.6 Discussion 111

5. Fences, Walls and the Development of Cities - The Long Term

Effects of the Israeli-Palestinian Land Obstacle 120

5.1 Introduction 120

5.2 The Model 125

5.3 Equilibrium 128

5.3.1 Single City Steady State 128

5.3.2 Two Cities with Commuting 132

5.4 Discussion 134

1

1. Introduction

The differences between geographical units create a potential for the migration of

both firms and people. While research has concentrated on the analysis of the

aforementioned migration on the one hand and migration's ongoing effects on various

locations on the other hand, unifying frameworks which analyze the mutual effects of

economic growth in each location and migration between locations and the mutual effect

of economic growth and other variables, such as prices and wages, in each location, are

less common.

Since the migration of firms and people can significantly affect economic

conditions, governments commonly adopt one policy or another to influence migration.

This study provides a framework for analyzing the effects of various policy alternatives.

The analysis finds several surprising long run results for some commonly adopted

migration policies. The first chapter analyzes the movement of firms between countries.

The second chapter analyzes the migration of people between cities using a theoretical

model while the third chapter provides an empirical characterization of migration. The

final chapter uses a model similar to that of the second chapter to analyze the effect of the

land-obstacle between Israel and the Palestinian territories.

Chapter 1: Migration of Firms, Home Bias and the Geographical Distribution of

Growth

Multinational firms are accounting for an increasingly large share of world trade

over time. At the same time, Foreign Direct Investment (FDI) is increasing rapidly,

especially in the developing countries, and is becoming the dominant factor in

determining a country’s economic performance. Both of these trends exist despite the

2

well known phenomenon of "home bias" in which preference is given to one's home

country over foreign countries in the decision to invest. The firm’s location decision has

previously been analyzed using a horizontal approach, in which firms can locate

production plants in more than one country, and a vertical approach which separates

headquarters from production.1 Grossman and Helpman (1991) explored an international

economy with monopolistic competition and sustained endogenous growth in quality or

variety of products in which firms can separate activities across countries.2

Though a great deal of research has been done on the multinational firm’s

decision making process and on the effect of multinational firms and international trade

on economic growth, little has been said about the effect of ownership on the firms in this

framework, particularly in the presence of home bias. This chapter analyzes the short and

long run effects of a (possible) movement of locally-owned firms from a developed

country to an undeveloped country on their output and growth rate in the presence of

home bias under several policy scenarios.

A model is constructed in which the production of a single good takes place in a

given number of firms, each with decreasing marginal productivity. Growth is the result

of an increase in the number of firms which is a positive function of both profitability

and the number of firms (which represents the stock of knowledge relevant to

establishing a new firm). In a framework with one developed country that has many firms

and one less developed country that has no firms, firms may migrate from the former to

1 See Markusen (1984, 1995) and Helpman (1984). 2 Grossman, Helpman and Szeidl (2003), Antràs and Helpman (2004) and Grossman and Helpman (2005) examined the firm's decision regarding the production and assembly of an intermediate good, outsourcing and the choice of outsourcing partner in a framework with different ex-post productivity for each firm. Effects of foreign investment and foreign firms on the hosting economies, relevant statistics and common policies can be found in UNCTAD (2002).

3

the latter in order to increase their profits, despite the home bias of their owners.

However, in later periods, firms may migrate back from the (now developed) country to

their former home.

Three policies towards the entry of firms are analyzed. The first, which allows

firms to freely enter, leads to the same long run growth rate in both countries, though the

wage and output in the less developed country always remain lower than in the more

developed country. The second, in which the undeveloped country demands partial local

ownership, eventually leads to the convergence of the wage, GDP and growth rates

(though not of GNP) in the two countries. The third policy demands an increasingly

larger share of local ownership over time. In the long run it results in the convergence of

the wage, GDP, GNP and growth rates of the two countries and may even enable the

undeveloped country to overtake the developed country in GNP. Overall, it is found that

harsher policies lead to better long run results for the undeveloped country.

Chapter 2: Inequality Within Cities, Inequality Between Cities and Patterns of

Urbanization

In recent decades, the Western world has experienced a continuous increase in

earnings inequality and at least part of the differences in earnings are due to

geographically related factors. Thus, in almost every country one finds poor cities which

generally lack technologically advanced industry and are characterized by lower-priced

local goods (such as real estate) alongside rich cities which are characterized by

technologically advanced industry, higher-priced local goods and a higher average wage.

Residents of the poor cities tend to migrate to the rich cities in search of higher wages,

while some residents of the rich cites migrate to the poor cities in search of lower prices.

4

Governments are sensitive not only to the overall level of inequality, but also to the

aforementioned inequality between regions and cities. As a result, government policy

often focuses on specific cities and includes a massive transfer of resources (relative to

total GDP) from rich areas to poor ones. However, in many cases this has no permanent

effect on economic conditions in the poor areas.

A great deal of research has been done on the divergence of countries to different

steady states, but divergence between cities has received little attention and is

significantly different from that between countries. Theories explaining the number and

location of cities were first suggested by Christaller (1933) and Lösch (1940). Fujita,

Krugman and Venables (1999) presented a series of models based on monopolistic

competition which showed that, depending on the level of transportation costs, cities can

either develop similarly or diverge towards one of two steady states, based on the market-

size effect.

Although much research has been conducted on the development of cities, very

little of it has focused on the connection between ability, migration and the technological

development of geographically distributed cities and how government policy affects this

connection. This chapter therefore analyzes the mutual effect of migration and the

distribution of the population (which is heterogeneous in ability) among cities on the

development of those cities and on the distribution of earnings and also analyzes the

effect of government policy on inequality between and within cities.

The model consists of multiple cities, each of which can produce a 'simple' non-

technological good and a 'complex' technological (or manufactured) good. The

production of the former is simply a function of the number of workers employed in that

5

sector while the production of the latter is a function of the number of workers employed,

as well as their ability which is uniformly distributed. The production of the technological

good involves a Learning-by-Doing process (LBD) which is specific to each city. Thus,

there can be a different steady state in each city, such that one city undergoes the LBD

process and primarily produces the technological good while the other produces the

simple good. Real estate is added to the model and the possibility is demonstrated of a

steady state with one developed city and one less developed city (with lower real estate

prices) and population movements of equal magnitude between them.

Wage inequality and mobility are analyzed and it is found that mobility is higher

in the technological city. Various government policies aimed at reducing earnings

inequality are examined and it is shown that though certain policies reduce inequality in

the short run, they prevent a poor city from developing and thus leave it poor, while other

policies fail to even reduce inequality in the short run. Only massive aid for a short period

of time (rather than the widespread policy of a low level of aid for a prolonged period)

can induce the long run development of a non-technological city.

Chapter 3: Internal Migration in the United States – Who Moves and Why

One of the results of the model presented in the previous chapter is the existence

of continuous migration in equilibrium, in which more talented individuals from the less

developed cities move to the more developed city, while less talented individuals from

the developed city move to the less developed city in search of lower real estate prices.

These two types of migration streams have different characteristics and different motives

and have opposite effects on the destination cities. Therefore, various policies may affect

those two streams differently and a different migration policy may be required for each.

6

Internal migration is an important phenomenon with substantial effects on the

development of cities and states. The high rate of migration and the fact that the

distribution of skills and education among migrants differs from that of the populations in

the origin and destination locations mean that migration has a significant impact on

economic development. Thus, migration patterns and changes in those patterns can affect

the geographical distribution of economic growth.

Empirical research into internal migration has looked at net migration (the net

change in the population of a given location due to migration; see, for example, Treyz et

al., 1993), gross migration (the number of individuals leaving or moving to a given

location; see, for example, Greenwood, 1975), characteristics of migrants (see

Greenwood (1997) for a survey) and other aspects of migration. Note that while it

appears that it is net migration that affects the development of states, migration policy

decisions must be based on gross migration streams, since the reasons for immigration as

opposed to emigration (at the state or city level) may differ.

Although there are substantial differences between those migrating in order to

find a new job and those migrating in search of cheaper housing, the current literature has

ignored this differentiation and does not break down gross migration streams according to

motive. This is perhaps due to the fact that the decision to migrate goes hand in hand with

a job change and the motives are difficult to separate.

The aim of this chapter is to separately analyze each one of the two migration

streams, job-related migration and housing-related migration. Moreover, since the

theoretical model presented in the previous chapter predicts that job-related migrants tend

7

to be more educated while housing-related migrants tend to be less so, migration will be

analyzed separately for each level of education.

The migration decision is analyzed using multi-year data on each migrant,

combined with data on wages, unemployment and other variables in the origin and

destination states according to level of education. The PSID dataset, which includes a

question on the reason for moving, is used for the analysis. This enables the separation of

migrants into two different streams for analysis, thus providing a better understanding of

the reasons for migrating and the characteristics of the gross migration streams.

The results confirm the differences between the two migration streams. The

probability of job-related migration is positively affected by the individual’s level of

education and, among some of the better-educated, is negatively affected by the origin

state's average wage. In contrast, the probability of housing-related migration exhibits the

opposite behavior. It was also found that a higher average wage in the destination state

(relative to the origin state) reduces housing-related migration, while it has no effect (and

even a positive one for one of the education groups) on job-related migration. The effect

of other variables, such as personal wage before and after migration, is also tested. If

policymakers ignore these differences between the streams, migration policy may achieve

the opposite effect from what was intended.

Chapter 4: Fences, Walls and the Development of Cities - The Long Term Effects of

the Israeli-Palestinian Land Obstacle3

Though Palestinian terror uses relatively simple means, it has been successful in

causing a significant number of casualties as well as severe damage to the Israeli

3 A paper based on this chapter has been accepted for publication in the Journal of Peace Research.

8

economy. Therefore, Israel has decided to build a security land obstacle4 between itself

and the Palestinians. Although it is generally agreed that it will decrease terror attacks

and will improve economic conditions in Israel, there is concern about its effect on

Palestinian workers and the Palestinian economy. Thus, even if Israel were to build the

obstacle on its own internationally recognized territory, there would still be opposition to

its construction for other reasons.

The Palestinian and Israeli economies are strongly interconnected. As Angrist

(1995) reports, during 1981-1990 about 40% of the Palestinian labor force worked in

Israel. Despite the many changes since then, a significant number of Palestinians still

travel daily to work in Israel. These ‘commuters’ are one of the main sources of income

for the Palestinian population. The obstacle will make it more difficult for Palestinian

workers to enter Israel (regardless of the prevailing political situation at that time) and

will therefore lower Palestinian income and welfare.

The immediate consequences of building the obstacle are obvious. However, little

has been said of the obstacle’s long term effects. The aim of this chapter is to analyze the

long term economic effects of an obstacle between Israel and the Palestinians, and, more

generally, to analyze the effect of prohibiting commuting between two cities when one is

more technologically advanced than the other.

In this chapter, a model is presented which is similar in spirit to that in the second

chapter (with the appropriate adaptations for commuting in place of migration) in which

there are two goods and two cities – a more advanced city (that can efficiently

4 In order to avoid sensitive political issues which are not the topic of this chapter, I use the term 'land obstacle' (hereinafter 'obstacle') to describe the fence or wall that Israel is building between the Palestinian territories and itself. The location of that obstacle is also not the issue of this study and the analysis is valid for every possible location, as long as the obstacle separates between the Palestinian territories and Israeli cities.

9

manufacture the technological good) and a less advanced city. The production of the

technological good depends on both the ability of the workers and the technological level

of the city. It is assumed that each city exhibits an LBD process in the production of the

technological good. The model shows that the possibility of commuting between the two

cities will result in the more talented workers of the less technological city working in the

more technological city. No one will work in the technological sector of the less

technological city and therefore it will not undergo any LBD process. The less

technological city will remain without any technological sector whatsoever.

The current situation of (partial) occupation and terror means that Palestinians are

periodically absent from work and therefore only unskilled jobs are available to them,

and has the effect of delaying development in the Palestinian territories. In the long run,

however, periods of relative calm are expected in which Palestinians will be able to work

more regularly in Israel and the Palestinian leadership will be able to focus on developing

its economy. In this situation, and if there had been no obstacle, more talented

Palestinians would have chosen to work in Israel and enjoyed a positive education

premium. Several sectors in the Israeli economy require highly skilled workers and many

of the factors that in the past prevented Palestinians from working in those sectors are no

longer relevant. Since the obstacle will hinder commuting from Palestinian to Israeli

cities, it will induce technological development in the Palestinian territories since the

more talented Palestinians will choose to work in Palestinian cities. It should be

mentioned that this conclusion is dependent neither on the existence of long periods of

relative calm nor on a substantial number of talented Palestinians working in Israel in the

absence of the obstacle. Thus, the obstacle will lower the number of talented Palestinians

10

working in Israel regardless of their original number (as long as it is positive) and will

therefore enhance Palestinian development. A long period of relative calm will increase

the number of talented Palestinians working in Israel and in that situation the obstacle

will have an even greater effect on future Palestinian development.

References

Angrist, J. (1995), 'The Economic Returns to Schooling in the West Bank and Gaza

Strip,' The American Economic Review 85(5), 1065-1087.

Antràs, P. and E. Helpman (2004), 'Global Sourcing,' Journal of Political Economy

112(3), 552-580.

Christaller, W. (1933), Central Places in Southern Germany. Jena, Germany: Fischer

(English translation by C. W. Baskin, London: Prentice Hall, 1966).

Fujita, M., P. Krugman and A. J. Venables (1999), The Spatial Economy. Cambridge:

MIT Press.

Greenwood, M. J. (1975), 'Research on Internal Migration in the United States: A

Survey,' Journal of Economic Literature 13(2), 397-433.

Greenwood, M. J. (1997), 'Internal Migration in Developed Countries,' in: Rosenzweig,

M. R. and O. Stark, editors, Handbook of Population and Family Economics 1B,

Elsevier.

Grossman G. M. and E. Helpman (1991), Innovation and Growth in the Global Economy.

Cambridge, Massachusetts \ London: The MIT Press

Grossman, G. M. and E. Helpman (2005), 'Outsourcing in a Global Economy,' Review of

Economic Studies 72(1), 135-159.

11

Grossman, G. M., E. Helpman and A. Szeidl (2003), 'Optimal Integration Strategies for

the Multinational Firm,' NBER Working Paper 10189

Helpman, E. (1984), 'A Simple Theory of International Trade with Multinational

Corporations,' Journal of Political Economy 92(3), 451-471.

Lösch, A. (1940), The Economics of Location. Jena, Germany: Fischer (English

translation New Haven, CT: Yale University Press, 1954).

Markusen, J. R. (1984), 'Multinationals, Multi-Plant Economies, and the Gains from

Trade,' Journal of International Economics 16(3-4), 205-226.

Markusen, J. R. (1995), 'The Boundaries of Multinational Enterprises and the Theory of

International Trade,' Journal of Economic Perspectives 9(2), 169-189.

Treyz, G. I., D. S. Rickman, G. L. Hunt and M. J. Greenwood (1993), 'The Dynamics of

U.S. Internal Migration,' The Review of Economics and Statistics 75(2), 209-214.

UNCTAD, 2002. World Investment Report: Transnational Corporations and Export

Competitiveness. New York and Geneva: United Nation Conference on Trade and

Development.

12

2. Migration of Firms, Home Bias and the Geographical

Distribution of Growth

2.1 Introduction

This chapter analyzes the effects of the choice to move of locally owned firms

from a developed country to an undeveloped country in an environment of home bias and

under various government policies that may cause a change in ownership over some of

the firms. It studies the short and long run effects of each policy and under which policy

could the undeveloped country overtake the developed one.

Multinational firms are accountable for an increasingly large share of world trade.

Foreign direct investment (FDI)1 is increasing more rapidly than income (especially in

the developing countries) and is thus becoming a dominant factor in determining a

country’s economic performance.2 Yet, Prasad et al (2003) analyzed financial

globalization and current account liberalization in developing countries by examining

both legal restrictions and actual capital flows, and compared their results to other

research. The general conclusion was that it is difficult to establish a causal relationship

between the degree of financial integration and output growth rates. The study presented

here analyzes three liberalization policies (policies toward entrance of foreign firms) and

shows under which policy does the aforementioned entrance contribute most to the host

country.

Movement of firms from one country to another is occurring despite the home

bias phenomenon in which preference is given to investment in one's home country over

1 An analysis of FDI and Foreign Portfolio Investment can be found in Goldstein and Razin (2005). 2 See Markusen and Venables (1998). Data on FDI and transnational corporations can be found at UNCTAD (2002).

13

investment in foreign countries, even when this is not the optimal decision.3 The

approach taken here, in order to account for the home bias, is a requirement for higher

profits when a firm is migrating out of its home country, when changes in ownership over

time induce a change in the difference required in profits accordingly.4 Most of the

North-South literature, that some of it is mentioned below, requires higher cost for

operation abroad and thus account in a certain way for home-bias.

The movement of firms between countries, which is analogous to labor migration

between countries, has a substantial effect on their economies. Countries react to the

entrance of foreign firms in a variety of ways, ranging from the encouragement of entry

to the discouragement of entry through various ways of taxation. Thus, while many

countries encourage the entry of firms through subsidies and tax exemptions (Ireland, for

example, used low tax rates for foreign firms), others, such as China and India, 'tax'

foreign firms through requirements for partial local ownership or the sharing of

knowledge.5 This chapter examines three policies – free entry (which is similar to

subsidy), a request for partial local ownership and a request for an increasingly larger

share of local ownership over time.

Multinational firms have previously been analyzed in horizontal as well as

vertical frameworks. The horizontal approach, in which firms locate production plants in

several countries, was adopted in Markusen (1984, 1995), who analyzed national and

multinational enterprises with firm-level knowledge. A similar approach appears in

3 French and Poterba (1991) analyze this phenomenon. See also Obstfeld (1998) and Lewis (1999). 4 Romer (1990a) present a growth model and analyzes capital productivity in different countries, finding that higher saving rates lead to higher investment and lower marginal product, implying a home bias in investment. This deviates from Ethier (1982), who is analyzing a model in which international trade in manufactures can lead to equalization in the price of immobile capital. 5 See UNCTAD (2002: 204-208) and Wei and Shleifer (2000: 306, 311-313).

14

Markusen and Venables (1998). A different approach, which separates headquarters from

production, appears in Helpman (1984). Grossman and Helpman (1991) explored an

international economy with monopolistic competition and sustained endogenous growth

in quality or variety of products, in which firms can separate activities across countries.

The approach taken here is closer in spirit to the first approach, in which firms can not

separate headquarter and production, but rather can relocate entirely. The increase in the

number of firms that occurs here resembles, though, to the increase in the number of

goods appearing in the second approach (see also Romer (1990b) for an endogenous

increase in the number of intermediate goods). It adds upon both approaches by analyzing

ownership over the firms and changes in the ownership, in an environment of home bias.

Both approached predict that an undeveloped country (South) usually gain from

interaction with the developed country (North). North-South trade literature dates back

way into the previous century, see for example Samuelson (1962), which also surveys

previous work. Krugman (1979) analyzes technology diffusion to the South. Coe and

Helpman (1993) and Bayoumi, Coe and Helpman (1996) found effects of foreign capital

stock and R&D spillovers through trade. Rodriguez and Rodrik (1999) found little

evidence that openness to trade is associated with economic growth. In the model

presented here the South gains from openness, through movement of firms that poses the

knowledge to manufacture, utilize world knowledge and grow, since only existing firms

and firms created from them know how to produce, utilize the knowledge and grow. The

magnitude of the gain depends on the government policy.

Grossman, Helpman and Szeidl (2003) examined the firm's decision regarding the

production and assembly of an intermediate good in a framework with different ex-post

15

productivity for each firm. In a similar framework, Antràs and Helpman (2004) analyzed

the firm's decision regarding integration vs. the outsourcing of intermediate goods

production in a global economy6 and Grossman and Helpman (2005) studied the firm’s

choice of outsourcing partner. Alvarez and Lucas (2004) analyzed gains from trade in an

economy with different levels of productivity.7 Firm's decision in this chapter is the

location of production, where no outsourcing is possible, location that may require a

change in ownership.

Though a great deal of research has been done on the multinational firm’s

decision making and on the effect of multinational firms, international trade and capital

flows on economic growth, little has been said about the effect of ownership of the firms

in this framework, particularly in the presence of home bias. The aim of this study is to

analyze the short and long run effects of a (possible) movement of locally owned firms

from a developed country to an undeveloped country on their output and growth rate in

an environment of home bias and under various government policies that may cause a

change in ownership over some of the firms.

In this chapter, a model is constructed in which the production of a single good

takes place in a given number of firms (which are using labor as the only production

factor), each with decreasing marginal productivity. Growth is the result of an increase in

the number of firms which depends positively on both profitability and the number of

firms (which represents the stock of knowledge relevant to establishing a new firm). In a

framework with one developed country that has many firms and one less developed

country that has no firms, firms may migrate from the former to the latter in order to

6 A similar question concerning ownership and control is analyzed in Feenstra and Hanson (2005). 7 See also Eaton and Kortum (2002).

16

increase their profitability, despite the home bias of their owners. However, in later

periods, firms may migrate back from the (now developed) country to their former home.

Three policies towards the entry of firms are analyzed. The first, which allows

firms to freely enter, leads to the same growth rate in both countries in the long run,

though the less developed country always remains with a lower wage and level of output

than the developed country. The second, in which the undeveloped country requires

partial local ownership, leads to a higher growth rate in the short run in the less-

developed country and eventual convergence of the wage, GDP and growth rates (though

not of GNP) in the two countries. The last policy to be analyzed requires an increasingly

larger share of local ownership over time. It results in a faster growth rate of the

undeveloped country and eventual convergence of the wage, GDP, GNP and growth rates

of the two countries, and may even lead to the undeveloped country overtaking8 the

developing country in GNP. The main intuition behind the result is that a request for local

ownership, though decreasing the number of firms that migrate upon opening of the

economy (which is later offset by larger profits and a faster growth rate), can eliminate

(and even reverse) the home bias effect, thus preventing out-migration of firms and

enabling overtaking.

The rest of this chapter is organized as follows: Section 2.2 presents the model.

Section 2.3 analyzes the equilibrium under various government policies. Section 2.4

concludes.

8 For an example of overtaking, see Brezis, Krugman and Tsiddon (1993).

17

2.2 The Model

In order to analyze the effect of the migration of firms, a simple growth model is

constructed. The model is first presented for the case of a single country in order to

demonstrate its characteristics and is then expanded to the two-country case.

2.2.1 A Single Country Growth Model

Assume a single country with a labor force of size L which remains unchanged in

every period. Output is produced by several firms which all manufacture the same

product9 using labor as the sole input. The number of firms operating in each period is

denoted by tn .

The production function, f(l), which is common to all firms, has positive but

decreasing marginal productivity and satisfies Inada's conditions10. For simplicity, the

following functional form is assumed:11

(1) αllf =)( , where 10

18

(3) α

απ

−=−=

tttttt n

Llwlfn )1()()(

Note that profit is decreasing in the total number of firms.

Since firms are continuously opening new facilities while closing old ones, it is

assumed that each firm exists for only one period and that at the end of the period its

production facilities become obsolete. At the end of the period the "parent" firm is

divided into several new firms that operate in the next period under the same ownership

(possible changes in the ownership of the new firms are discussed in Section 2.2). There

is no (free) entry of other firms into the market since the knowledge and facilities

necessary for production are obtainable only from the operation of an existing firm. The

number of new firms created from an existing one depends on the profit of the firm and

on the total number of firms. Thus, the more profitable a firm is, the more new firms it

can establish12 and the more firms that exist, the easier (though less profitable) it is to

establish new firms (since the number of firms represents the stock of knowledge

available for use in the establishment of new firms out of existing firms). This last

assumption is similar to the commonly made assumption that knowledge is a public good

(see for example Grossman and Helpman (1991:57-62)).13 Note, however, that only

existing firms can utilize the knowledge. Denote the number of firms created as

))(,( tt nng π , where 1),( >ttng π for positive n and π , and 0,0 >> πgg n though π is

12 Credit constraints and imperfect capital markets combined with financial investment required for establishing a new firm is a possible cause for such situation. 13 Note also that the increase in the number of firms assumed here resembles the constant increase in the number of products in their model.

19

decreasing in n.14 Moreover, it is assumed that the increase in g due to an increase in the

number of existing firms is exactly the same as the decrease in g due to the decrease in

profits (which is also a result of that same increase in the number of existing firms):

(4) knLngnngn =

−=∀ ))1(,())(,(,

α

απ , where k is a constant15.

The total GDP of the economy in period t is, therefore:

(5) αα LnlfnY tttt−== 1)(

The total payment to labor is ααα −= 1tt nLLw , which corresponds to a labor share of

output equal to α . The corresponding share of firms is )1( α− .

The consumers, who are also the workers and the owners of the firms, live for a

single period (L consumers in each period), without any possibility of saving. Therefore,

their demand for consumption equals their income (GNP) and therefore the market clears.

Note that Walras' Law allows for the analysis of the labor market's equilibrium without

analyzing the market for goods.

Proposition 1: The wage and GDP grow at the same constant rate.

Proof: The growth rate of output depends on the growth rate of the number of firms

which is constant:

(6) 111 111

11 −=−=− −−−

++ ααα

αα

kLnLn

YY

t

t

t

t

A similar equation for the wage yields the identical growth rate. QED

14 Another way of modeling this would have been to assume that an existing production facility continues to operate in the next period and that (g-1) new facilities are created from each existing firm. See Section 2.4 for further discussion of the differences between the two methods.

15 For example, α

α

αππ

Lnkng

)1(),(

−= .

20

Moreover, since the number of firms in period t is 11−tkn , one can easily find the

wage, profit and output for each period. Note that the total profit of all firms grows at the

same rate, but since the number of firms grows even faster, the profit of a single firm

decreases over time.16

So far I have analyzed the model for the simple case of a single country. I turn

now to analyze the case of two countries and examine the mutual influences between

them.

2.2.2 The Complete Model

Assume a world with two countries, H and F, each with a labor force of size L.

The output in each country is produced, as before, by several firms manufacturing the

same product using labor as the sole input. The number of firms in each period in each

country is denoted by itn , i=H,F. The production function is the same as before.

Each firm employs itn

L workers who are paid their marginal productivity:

(7) 1−

=

α

α itn

Lw .

Therefore the profit of a particular firm is:

(8) α

απ

−= i

t

it

it n

Ln )1()(

Each firm exists for only one period and at the end of the period it is divided into

g new firms, where g depends on the profit of the firm and on the total number of firms in

the world. Thus, a more profitable firm grows faster and therefore has a faster

16 As before, this result resembles the case of product variety; see, for example, Grossman and Helpman (1991:57-62).

21

multiplication rate. The more firms there are, the more knowledge is available for the

establishment of new firms and therefore the easier it is to do so. Since a free flow of

knowledge is assumed between the two countries, the multiplication rate is affected by

the total number of firms in the world.17 Therefore, one obtains ))(,( inng π with the

same assumptions as before. A necessary additional assumption (though only a technical

one) is that the multiplication rate is not excessively high:

(9) ))(,())(,(, jjijijiiji nnngnnnngnnn ππ +>+>∀

This assumption ensures that when the number of firms in one country is fewer than that

in the other, it will not exceed it in the next period but only move towards it, since

otherwise repeated overtaking might occur. A continuous model would not require such

an assumption. The example from the previous section is relevant here as well.

At the beginning of each period, the newly established firms choose where (i.e. in

which country) to locate their activity. This decision is made by the owners of the firms

and is affected by a home bias,18 where the ownership of a firm can be only H (citizens of

the home country), only F (citizens of the foreign country) or any mixture of the two. In

the case of mixed ownership, the share of ownership of investors from one country in the

parent firm of a firm operating in period t will be denoted by tB , and therefore the share

of ownership of investors from the other country will be )1( tB− . When no changes in

ownership occur, these shares also represent the ownership situation of the firm which is

actually operating in period t. The home bias effect causes owners to demand higher

17 This assumption is similar to that of international spillovers of knowledge; see for example Grossman and Helpman (1991:178). 18 See the Introduction of this chapter for references and some elaboration.

22

profits in order to locate the activity of a firm outside of their own country. Therefore, the

condition for locating a firm owned by investors in country i in country j is:

(10) itj

t γππ ≥

where 1>γ stands for the preference of the owners for an activity in their own country,

and for simplicity a symmetric preference between the two countries is assumed (i.e.γ is

equal for both countries). In a similar manner, when the share tB of a parent firm is

owned by investors from country j (and the rest by owners from country i), the condition

for locating a firm just created from it in country j will be:

(11) ittitt

jtt

jtt BBBB πγπγππ +−≥+− )1()1(

since each owner has a bias towards his home country.

Notice that the owners' decision is based solely on profits in the next period.

However, sufficiently high discount rates can yield the same results (see Section 2.4 for

discussion).

Ownership of the new firms remains the same as that of the parent firm,19 unless

the owners are willing to give up some of their ownership (see sub-Sections 2.3.2 and

2.3.3). In an identical situation to the one presented above, when owners are required to

give up a share β in order to operate in country j (while no such demand exists in

country i), the condition becomes:

(12) ittitt

jtt

jtt BBBB πγπγππβ +−≥+−− )1(])1)[(1(

19 Another possible interpretation is that the new firms are established by senior managers of the parent firm, who acquired the necessary skills and knowledge, and the origins of those senior managers are divided between the countries in the same ratio as the owners. Therefore the new ownership is, on average, divided between the two countries in the same ratio as the previous one, although the owners themselves are different.

23

where tB is the ownership of investors from country j in the parent firm, i.e. the

ownership prior to giving up β . The right side of (12) is the investors' profit from

locating in country i (adjusted for home bias) while the left side is the profit from locating

in country j adjusted for home bias and the requirement to give up a share of the firm.20

During the evolvement of the economy two types of ownership can be created -

direct or chained through other firms. However, we only need to keep track of the origin

of the owners and not the form of ownership. Consider, for example, a firm in the first

period owned by H investors. Now assume that in the second period it multiplies into two

companies, both of which operate in H. These two new companies can either be held

directly by investors in H (in this case there are two separate companies) or held by the

first period “parent” company, which serves as a holding company of the two new

companies (it has no other activity) and is itself owned by the investors in H (thus,

resulting in a multi-plant firm). In a similar manner, if one of the two new companies

operates in F, then under the former form of (direct) ownership this is simply an

investment abroad, while under the latter form of ownership the first period firm becomes

a multinational firm, with a subsidiary in H and a subsidiary in F. The division of

ownership between the countries affects the decision of where to locate as described

above, but the method of holding (i.e. directly or through another firm) does not, and as a

result we do not need to keep track of the form of ownership. Therefore, although both

multi-plant and multinational firms evolve in this model as part of the growth process and

the migration of firms between countries, it does not need to be analyzed.

20 Notice that the condition refers only to the profit of the current period, see sub-Sections 2.3.2 and 2.3.3 and Discussion of this chapter for an elaboration and for the case of the discounted stream of profits.

24

The specification of consumption behavior will complete the model. The

consumers in each country, who are also the workers and the owners of the firms, live for

a single period (L consumers in each period), without the possibility of saving. Therefore,

total demand equals their total income (i.e. GNP) and the market will clear.

The next section considers the situation when an undeveloped country decides to

allow the entry of foreign firms.

2.3 Entry Policies

Migration of firms has a significant effect on the economy, and therefore it is

accompanied by government policy. When a country decides to allow foreign firms to

operate within it, there are several (active trade) policies it can adopt, each leading to a

different result. One option is to encourage entry through subsidization of the foreign

firms, i.e. reduced tax rates and grants. At the other extreme is the policy of taxing

foreign firms which can take the form of demanding the sharing of knowledge or

requiring a share of the ownership for local citizens (possibly one that increases over

time) even when local owners will not be contributing to the activity of the firm. The

effects of allowing entry are analyzed under several possible policy regimes when one

country is developed and the other is undeveloped and the latter decides to allow foreign

firms to operate within it.

Assume two countries, H and F. In period T, prior to the possibility of entry, Tn

firms, owned by H owners, operate in H. There are no firms owned by residents of F and

therefore no firms operate in F prior to the possibility of entry. At the beginning of period

25

T, country F decides to allow foreign firms to operate within it, either with or without

restrictions.

2.3.1 Free Entry

One extent of policy is to encourage the entry of foreign firms through tax

exemptions, grants, restrictions on the import of similar products, etc.21 The policy

analyzed here is a more 'neutral' one, where country F simply allows foreign firms to

operate within it without restriction. Since all the firms are owned by residents of H, the

entry condition in (10) applies and firms will move to F until

αα

αγα

−=

− H

TFT n

LnL )1()1( , or :

(13) TH

T nn αα

γγ

/1

/1

1+= , T

FT nn αγ /11

1+

= .

Proposition 2: A free entry policy immediately improves the wage, GDP and GNP in F

and leads to continuing growth in those variables but F does not catch up to H in any of

these variables.

Proof: According to (13), the number of firms in F grows from zero to FTn . As a result,

the wage, GDP and GNP grow to 1−

α

α FTnL , αα Ln FT

−1)( and ααα Ln FT−1)( , respectively.

Notice that since FTHT nn > (since )1>γ and the companies located in F are owned by

residents of H, the wage, GDP and GNP will be higher in H.

Since firms can freely migrate in and out of F and ownership remains unchanged,

equation (13) holds for every period t, Tt ≥ and the same ratio of Hn to Fn will prevail

21 See UNCTAD (2002: 204-208) for a description of the various incentives and for detailed description of policy in Ireland and Malaysia.

26

in every period. The long run growth rate of the wage, GDP and GNP is 11

1 −

−

+α

tFt

F

nn

which is equivalent to 11

1 −

−

+α

t

t

nn

. Therefore, the wage, GDP and GNP grow at the

same rate in both countries and thus remain higher in country H.

QED

Note that since in every Tt ≥ , ))(,())(,( HttF

tt nngnng ππ > ,22 firms in F

multiply faster than those in H. However, since the proportion of firms operating in F

(out of the total n) remains the same, firms are continually migrating back to H. Note also

that since the demand for consumption equals GNP (see sub-Section 2.2.2), the goods

market in each country is cleared with the import (or export) of goods equal to the

difference between GNP and GDP.

The short run effects on F are straightforward. The wage, GDP and GNP of

country F rise and can rise even further if a policy of subsidization is implemented in that

period. In the long run, country F’s rate of growth depends on the growth in the number

of firms, which is between ))(,( Ftt nng π and ))(,(H

tt nng π , but F will always lag behind

H in all three variables.23 Figure 1.1, in which entry is permitted from period 6 onward,

demonstrates this. Thus, in period 6 there is an improvement in F's GDP and GNP and

continuous growth thereafter (at the same rate in both countries). Note that F's variables

are always lower than those of country H.

22 Ht

Ft nn < : therefore )()(

Ht

Ft nn ππ > and ))(,())(,(

Htt

Ftt nngnng ππ > .

23 It should be mentioned that a single period subsidy to foreign firms followed by a free entry policy yields similar long run results.

27

1

10

100

1000

1 2 3 4 5 6 7 8 9 10 11 12time

prod

uct (

loga

rithm

ic sc

ale)

H GDP H GNP F GDP F GNP

Figure 1.1 – GDP and GNP of H and F under Free Entry

Plotted for 1=L , )1(

2),(αππ

α

−= nng , 5.0=α , 10 =

Hn , 00 =Fn , 5.1=γ

2.3.2 Joint Ownership

Many countries demand a share of local ownership as a prerequisite for the

entrance of a foreign firm. This is done, among other reasons, in order to encourage local

industry, to better monitor the operations of foreign firms or can simply be the result of

corruption.24 Therefore, the second policy option to be analyzed is the requirement of a

one-time tax in the form of joint ownership in exchange for the right to operate in F.

When a fully H-owned firm wishes to operate in F, it must give up a share β of its

ownership to local owners. Firms established from that firm in the following periods can

operate in F (without further "taxation") or migrate out of F while retaining the same

24 See Wei and Shleifer (2000: 311-313).

28

division of ownership. They can, of course, freely return to F later since β of their

ownership is held by residents of F.

Under this policy regime, when country F begins to allow entry, the condition that

induces firms to enter is similar to (12) except that B=0 since all the firms are owned by

residents of H:25

(14) HTFT γππβ ≥− )1(

and firms will continue moving to F until αα

αγαβ

−=

−− H

TFT n

LnL )1()1)(1( , or:

(15) TH

T nn α

α

βγβ

γ

/1

/1

11

1

−

+

−

= and TF

T nn α

βγ

/1

11

1

−

+

= .

Proposition 3: A policy of joint ownership leads to an immediate improvement in F’s

wage, GDP and GNP and continuing growth in those variables in subsequent periods.

Catch-up in wage and GDP is possible for a sufficiently large β (i.e. share of local

ownership) but there can be no catch-up in GNP.

Proof: As in Proposition 2, the wage, GDP and GNP grow immediately in period T to

1−

α

α FTnL , αα −1)( FTnL , and

αααα αβα −− −+ 11 )()1()( FTF

T nLnL , respectively. Notice,

however, that under this policy regime the proportion FTn of Tn is smaller than under

the previous policy.

25 If firms discount future profits, the demand for local ownership also implies loss of some future profits, as well as loss of control. Therefore, firms request higher compensation (i.e. higher current profits) in order to migrate to F, resulting in lower (yet positive) number of migrating firms (and higher profits in F).

29

Since FTHT nn > (because 1>γ and 1 , firms

in F multiply faster than those in H and profits decrease at a faster rate. Furthermore, this

situation cannot be reversed to Ttnn FtHt >< , as can be seen from equation (9). As long

as this situation prevails, the entrance condition (14) will not hold for fully H-owned

firms and new firms will not enter country F. However, since firms operating in country

F are now partially locally owned, migration out of F needs to be carefully analyzed.

The condition for migrating back to H for firms with a share β of local ownership

is:

(16) HtHt

Ft

Ft 1111 )1()1( ++++ +−εε , as a desirable difference between Hn and Fn (ε can be

chosen as arbitrarily small), and note with ε>∆∆, , the current (actual) difference

between Hn and Fn . Every ε produces a difference in the multiplication rates of the

firms in the two countries which, for high enough 0t , reduces ∆ to ε within 0t periods.

30

Since the actual difference in the multiplication rates is higher (since the actual difference

in the number of firms is greater than ε ), the number of periods required to achieve ε is

smaller than 0t . The difference between Hn and Fn approaches 0 and therefore the share

of Fn in n approaches one half. As a result, the wage and GDP equalize between the two

countries while the GNP of H remains higher due to the ownership of firms operating in

F. The long run growth rate of n approaches ))5.0(,( nng π .

If21

31

economy’s growth rate, although it decreases over time. For 21≥β , F’s growth rate is

higher than that of H, but decreasing over time so that the share of firms operating in F

approaches one half and the wage and GDP of the two countries equalize over time. The

GNP of F remains lower, since the firms there are partially owned by H (see figure 1.2).

For 21

32

2.3.3 Increase in Local Ownership over Time

Firms operating in a foreign country may be subject to an ongoing increase in the

share of local ownership or an increase in the local activity in the same sector. This can

be the result of, among other things, a government policy to encourage local workers to

open their own firms or to encourage spillovers of knowledge to local firms (by not

protecting intellectual property) or may simply be the result of ongoing corruption.

Whatever the reason, the result is an ongoing increase in local ownership, which is

equivalent to a periodical 'tax' on the original owners. Therefore, the final policy to be

analyzed is the requirement of an increasing share of local ownership each period in

exchange for the right to operate in F. When a firm chooses to operate in F in a particular

period, the previous owners (the owners of the "parent" firm) are left with only β−1 of

the new firm, while a share of β is allocated to local owners. This is repeated in every

period in which the firm operates in F, whether its "parent" firm operated in H or F, and

regardless of its share of F ownership. I assume that β is not too large:

(18) γ

β 11−≤

Under this policy regime, since all firms are of H ownership, when F begins

allowing entry and since firms maximize the profit of only the current period28, the

entrance condition remains the same as in the previous case (i.e. equation (14)) and firms

move to F until αα

αγαβ

−=

−− H

TFT n

LnL )1()1)(1( , and HTn and

FTn equal their

values in equation (15).

28 See Section 2.4 for a discussion of the discounted stream of profits case.

33

Proposition 4: A policy of increasing local ownership over time leads to an immediate

improvement in F’s wage, GDP and GNP. F will catch up in these variables in the long

run. Whether it overtakes H depends on g.

Proof: As in Proposition 2, the wage, GDP and GNP grow immediately in period T to

1−

α

α FTnL , αα −1)( FTnL , and

αααα αβα −− −+ 11 )()1()( FTF

T nLnL , respectively. Notice that

all three variables are of the same magnitude as in the case of the previous policy (i.e.

constant local ownership). Since FTHT nn > (because 1>γ and 1 , firms multiply faster in F than in H and profits decrease

faster in F than in H. As long as this situation continues, the entry condition in (14) will

not hold for any fully H-owned firms and new firms will not enter F.

Firms that choose to migrate from F back to H will not return to F. Firms that

migrate out of F with local ownership of B (the ownership of their parent firms which

equals the ownership of the new firms if they leave F) satisfy:

(19) HHFF BBBB πγπγππβ +− ) and for the same profit ratio there will

34

not be any additional migration back to H, which would have increased the left side of

(20). The multiplication rate of firms in F is higher than in H and therefore without such

migration the ratio of the firm proportions (right side of (21)) decreases and therefore so

does the left side of (21) and (20). Therefore, the exit condition expressed in equation

(19) for firms with F ownership of B continues to be valid and there is no return

migration to F of those firms.

Notice that the share of F ownership in firms that decide to continue operating in

country F increases over time and eventually approaches 1. Since there are no firms

entering F in the periods following T, the share of F ownership in all firms that operate in

F approaches 1.

In the long run, since local F ownership approaches 1, the exit condition becomes:

(22) HF πγπβ − γβ , this last condition requires that HF ππ < , or

HF nn > , a condition that contradicts assumption (9). Therefore, there is no migration

from F back to H in the long run and the share of firms in each of the two countries

approaches one half (as in Proposition 3 with 21≥β ) and therefore the wage and GDP in

each of the two countries approach equality.

Overtaking in GNP requires that some firms (with a share of F ownership)

migrate out of F. The condition for this to occur is given in equation (12), which can be

rearranged as:

(23) )1)(1(

)(

1

1

tt

ttHt

Ft

BBBB

γβγγ

ππ

+−−−+

<+

+ (see also equation (20))

35

According to (21), for some g functions the left side of equation (23) is close enough to 1

(when B is small enough), such that migration from F back to H can occur. For example,

if the left side of (23) is close enough to 1 for t=T, (23) becomes )1)(1(

)(1γβββ

γββγ+−−

−+<

which is valid for every β as long as 5.11

36

1

10

100

1000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16time

prod

uct (

loga

rithm

ic sc

ale)

H GDP H GNP F GDP F GNP

Figure 1.3 – GDP and GNP of H and F under Increasing Local Ownership

Plotted for 1=L , )1(

2),(αππ

α

−= nng , 5.0=α , 10 =

Hn , 00 =Fn , 5.1=γ , 3.0=β

Proposition 5: Overtaking in GNP can be achieved for every g satisfying the previous

assumptions using a more complicated policy scheme.

Proof: Assume a policy scheme similar to the one analyzed in which a requirement for

increasing local ownership β over time that satisfies (18) is enforced in every period

except period S, S>T, where S is chosen so that a firms' profits in each of the countries

are similar in that period. Denote the requirement for local ownership in S as 1β , such

that:

(24) γ

β 111 −> ,

and therefore γβ )1(1 1−> . The condition for migration from F to H at the beginning of

period S is:

37

(25) HSFS BBBB πγπγβ ))1(()1)(1( 1 +−

38

2.4 Discussion

In the model that was presented, growth is generated by an increase in the number

of firms which depends on both the profitability of firms and the stock of knowledge, as

represented by the number of firms. In a two-country framework, where one country is

developed and has many firms and the other is undeveloped and has no firms, it was

shown that the migration of profit-maximizing firms from the developed to the

undeveloped country (despite the owners' home bias) exported the growth process to the

undeveloped country, thus putting the developed country at risk of losing its advantage.

Three entry policies were analyzed for the undeveloped country with the result that a

harsher policy yields better long-run results. While all policies improve the immediate

term situation30 and yield equal long run growth rates for the two countries, the free entry

policy leaves the undeveloped country with a permanent lag behind the developed

country. The one-time tax policy (i.e. joint ownership) leads to convergence in GDP and

a harsher policy of ongoing taxation (in the form of increasing local ownership over time)

results in convergence in GDP and overtaking in GNP. Therefore, according to this

model a policy of taxation is recommended as opposed to the policy of subsidization

adopted by many countries. All those policies are, of course, in a framework of current

account liberalization, since entry of foreign firms is required in each. They differ in the

manner in which the entrants are treated.

Notice, however, that this recommendation holds in the case in which foreign

firms, once they have begun operating in the undeveloped country, can produce and

grow. An appropriate education level among the population, protection of property rights,

30 Note, however, that the first policy analyzed yielded more favorable short run results.

39

a developed infrastructure and other similar conditions for this process to occur were

assumed to exist. Otherwise, the growth process does not occur.

The aim of the policy (i.e. catching up with or overtaking the other country) and

the policy recommendation with regard to the long run get an interesting interpretation

when the situation is different than two equally-sized countries. If the developed world is

large relative to a small undeveloped country, the latter does not affect the world output

and growth rate. In that case, overtaking also means achieving the highest output in the

long run. If there are several undeveloped countries, those countries might compete on

migrating firms by offering various benefits. Countries that will not offer benefits but

rather take a harsher policy would get fewer entering firms (but still, a positive number,

since the profits of a single firm increase when the number of firms decreases), and

therefore lower short-run output and wage, but they enjoy better long run result. Since

some firms do enter, the long run result of the harsher policy (increase in local ownership

over time) would still be overtaking.

Two simplifying assumptions need to be discussed here. The first is that during

the growth process a firm multiplies into g>1 firms and that all new firms are able to

migrate. An alternative to this assumption would have been that the firm continues to

exist and that a positive number of new firms are created (with only new firms being able

to migrate). Results in this case would be similar, but perhaps achieved at a slower pace.

Thus, when entry becomes possible, migration to the undeveloped country could continue

for more than one period since the share of migrating firms that equalizes profits may be

larger than the share of new (migration-able) firms. However, long run results remain

unchanged.

40

Another simplifying assumption of the model is that firms base their decisions on

profits only in the current period rather than discounting the entire stream of future

profits. While under the first policy discounting future profits changes nothing and under

the second one there is only a minor change in the first period (see sub-Section 2.3.2),

under the last policy analyzed, in which firms that migrate to F lose additional ownership

in every period, accounting for future profits could affect the return-migration decision

and therefore the model's result. Note, however, that since the relevant period in the

model is long, an assumption of a high enough discount rate is a natural one. In such a

case, when entry becomes permissible, the profit required to induce migration is higher

and therefore less firms migrate. However, the process that follows remains the same as

long as the home bias and the discount rate are sufficiently high relative to the periodic

'tax' (which is determined by the undeveloped country itself). Some firms will operate in

F, and those firms would become locally owned. Since the difference in profits in the

long run (as viewed by the firms in each country, i.e. adjusted for the home bias) between

the current country a firm operates within and the other country is strictly positive, those

firms would not migrate out and catching-up (and overtaking) will occur. Therefore

discounting the stream of future profits does not affect the result with regard to the long

run.

Finally, we turn briefly to the discussion of counter-policies that may be adopted

by the developed country. The movement of firms from the developed country H to the

undeveloped country F leads to F becoming the more developed country if it adopts the

appropriate policy and H allows the free entry of firms. This is a result of the behavior of

profit-maximizing firms that does not take into account the firm’s effect on the long run

41

balance between the two economies. There are number of counter-policies that H could

adopt including, for example, the reduction of out-migration once entry is allowed either

through direct restrictions on local firms, through subsidization or using a tax policy

which increases H's local ownership. Notice, however, that the former two policies

merely postpones the overtaking while the latter leads to convergence instead of

overtaking and does not enable the developed country to retain any lead that it once had.

The issue of counter-policies and the reaction to them requires additional research, since

it appears that liberal policies may not be recommended if a country wishes to maintain

its economic leadership.

42

References

Alvarez, F. and R. E. Lucas Jr. (2004), 'General Equilibrium Analysis of the Eaton-

Kortum Model of International Trade,' Working Paper, University of Chicago.

Antràs, P. and E. Helpman (2004), 'Global Sourcing,' Journal of Political Economy

112(3), 552-580.

Bayoumi, T., D. T. Coe and E. Helpman (1996), 'R&D Spillovers and Global Growth,'

NBER Working Paper 5628.

Brezis, E. S., P. R. Krugman and D. Tsiddon (1993), 'Leapfrogging in International

Competition: A Theory of Cycles in National Technological Leadership,' American

Economic Review 83(5), 1211-1219.

Coe, D. T. and E. Helpman (1993), 'International R&D Spillovers,' NBER Working Paper

4444.

Eaton, J. and S. Kortum (2002), 'Technology, Geography and Trade,' Econometrica

70(5), 1741-1779.

Ethier, W. J. (1982), 'National and International Returns to Scale in the Modern Theory

of International Trade,' American Economic Review 72(3), 389-405.

Feenstra, R. C. and G. H. Hanson (2005), 'Ownership and