THREE ESSAYS ON THE ECONOMICS OF PREFERENTIAL TRADE AGREEMENTS: FREE TRADE AREAS, RULES OF ORIGIN AND CUSTOMS UNIONS RENFENG XIAO

Similar documents
Elements of Economic Analysis II Lecture VI: Industry Supply

Price and Quantity Competition Revisited. Abstract

Uniform Output Subsidies in Economic Unions versus Profit-shifting Export Subsidies

UNIVERSITY OF NOTTINGHAM

THREE ESSAYS IN INDUSTRIAL ECONOMICS AND PUBLIC POLICY CHEAWANET BUNCHAI

Volume 30, Issue 1. Partial privatization in price-setting mixed duopoly. Kazuhiro Ohnishi Institute for Basic Economic Science, Japan

Privatization and government preference in an international Cournot triopoly

Tradable Emissions Permits in the Presence of Trade Distortions

- contrast so-called first-best outcome of Lindahl equilibrium with case of private provision through voluntary contributions of households

5. Market Structure and International Trade. Consider the role of economies of scale and market structure in generating intra-industry trade.

Interregional Trade, Industrial Location and. Import Infrastructure*

Taxation and Externalities. - Much recent discussion of policy towards externalities, e.g., global warming debate/kyoto

Ch Rival Pure private goods (most retail goods) Non-Rival Impure public goods (internet service)

Domestic Savings and International Capital Flows

Political Economy and Trade Policy

Problem Set #4 Solutions

Is Social Welfare Increased By Private Firm Entry. Introduction

Domestic Product Standards and Free Trade Areas

A MODEL OF COMPETITION AMONG TELECOMMUNICATION SERVICE PROVIDERS BASED ON REPEATED GAME

Quiz on Deterministic part of course October 22, 2002

Optimal policy for FDI incentives: An auction theory approach

Applications of Myerson s Lemma

Policy Choice: Theory and Evidence from Commitment via International Trade Agreements *

2. Equlibrium and Efficiency

The liberalization of Tariff Rate Quotas under oligopolistic competition 1. Margherita Scoppola

Online Appendix for Merger Review for Markets with Buyer Power

The Economics of Inter-City Competition in Financial and Distribution Markets

2) In the medium-run/long-run, a decrease in the budget deficit will produce:

Optimal Service-Based Procurement with Heterogeneous Suppliers

Special Interest Politics: Contribution Schedules versus Nash Bargaining

OPERATIONS RESEARCH. Game Theory

"Understanding the Welfare Implications of Preferential Trade Agreements" M. Ayhan Kose and Raymond Riezman

Prospect Theory and Asset Prices

/ Computational Genomics. Normalization

Problems to be discussed at the 5 th seminar Suggested solutions

NBER WORKING PAPER SERIES POLICY CHOICE: THEORY AND EVIDENCE FROM COMMITMENT VIA INTERNATIONAL TRADE AGREEMENTS. Nuno Limão Patricia Tovar

International technology licensing strategy under successive monopoly

In the 1990s, Japanese economy has experienced a surge in the unemployment rate,

NBER WORKING PAPER SERIES PRICES VS. QUANTITIES VS. TRADABLE QUANTITIES. Roberton C. Williams III. Working Paper

Equilibrium in Prediction Markets with Buyers and Sellers

Can Bilateral Trade Agreements Help Induce Free Trade?

Lecture 7. We now use Brouwer s fixed point theorem to prove Nash s theorem.

Benefit-Cost Analysis

Wages as Anti-Corruption Strategy: A Note

15-451/651: Design & Analysis of Algorithms January 22, 2019 Lecture #3: Amortized Analysis last changed: January 18, 2019

INTRODUCTION TO MACROECONOMICS FOR THE SHORT RUN (CHAPTER 1) WHY STUDY BUSINESS CYCLES? The intellectual challenge: Why is economic growth irregular?

Fall 2017 Social Sciences 7418 University of Wisconsin-Madison Problem Set 3 Answers

Economics 1410 Fall Section 7 Notes 1. Define the tax in a flexible way using T (z), where z is the income reported by the agent.

CHAPTER 9 FUNCTIONAL FORMS OF REGRESSION MODELS

Two Period Models. 1. Static Models. Econ602. Spring Lutz Hendricks

A Utilitarian Approach of the Rawls s Difference Principle

Incentives for Price Manipulation in Emission Permit Markets with Stackelberg Competition

Chapter 10 Making Choices: The Method, MARR, and Multiple Attributes

Voluntary Contribution to Public Goods: Preferences and Wealth

IND E 250 Final Exam Solutions June 8, Section A. Multiple choice and simple computation. [5 points each] (Version A)

Solution of periodic review inventory model with general constrains

4: SPOT MARKET MODELS

ECON 4921: Lecture 12. Jon Fiva, 2009

Money, Banking, and Financial Markets (Econ 353) Midterm Examination I June 27, Name Univ. Id #

ECONOMISING, STRATEGISING AND THE VERTICAL BOUNDARIES OF THE FIRM

Lecture Note 2 Time Value of Money

Allowing Firms to Choose Between Formula Apportionment. and Separate Accounting Taxation 1. Thomas A. Gresik. University of Notre Dame.

Allotment and Subcontracting in Procurement Bidding

Raising Food Prices and Welfare Change: A Simple Calibration. Xiaohua Yu

Global Environmental Problems and Actions Taken by Coalitions

ECE 586GT: Problem Set 2: Problems and Solutions Uniqueness of Nash equilibria, zero sum games, evolutionary dynamics

EXPORT SUBSIDIES AND CUSTOMS UNIONS

Global Optimization in Multi-Agent Models

Problem Set 6 Finance 1,

The liberalization of Tariff Rate Quotas under oligopolistic competition

Bilateral Trade Flows and Nontraded Goods

ECO 209Y MACROECONOMIC THEORY AND POLICY LECTURE 8: THE OPEN ECONOMY WITH FIXED EXCHANGE RATES

Consumption Based Asset Pricing

Least Cost Strategies for Complying with New NOx Emissions Limits

Flight Delays, Capacity Investment and Welfare under Air Transport Supply-demand Equilibrium

Internationally Mobile Factors of Production and Economic Policy in an Integrated Regional Union of States

MEMORANDUM. Department of Economics University of Oslo. Cathrine Hagem

c slope = -(1+i)/(1+π 2 ) MRS (between consumption in consecutive time periods) price ratio (across consecutive time periods)

Welfare Aspects in the Realignment of Commercial Framework. between Japan and China

Volume 29, Issue 1. Wage Subsidy and Sector-Specific Unemployment: A New Economic Geography Approach

ON THE COMPLEMENTARITY BETWEEN LAND REFORMS AND TRADE REFORMS

Wage-rise contract and endogenous timing in international mixed duopoly

Graduate School of Economics Waseda University

Keiretsu and relationship-specific investment: implications for market-opening trade policy

EDC Introduction

Second-Degree Price Discrimination on Two-Sided Markets

Tests for Two Ordered Categorical Variables

Macroeconomic Theory and Policy

Incentives for Price Manipulation in Emission Permit Markets with Stackelberg Competition

Microeconomics: BSc Year One Extending Choice Theory

Macroeconomic Theory and Policy

University of Toronto November 9, 2006 ECO 209Y MACROECONOMIC THEORY. Term Test #1 L0101 L0201 L0401 L5101 MW MW 1-2 MW 2-3 W 6-8

University of Toronto November 9, 2006 ECO 209Y MACROECONOMIC THEORY. Term Test #1 L0101 L0201 L0401 L5101 MW MW 1-2 MW 2-3 W 6-8

references Chapters on game theory in Mas-Colell, Whinston and Green

Advertising Policies: Comparative versus Informative Advertising in Oligopolistic Markets.

Alternative Approaches to Reciprocal Tariff Liberalization

Jeffrey Ely. October 7, This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.

CS 286r: Matching and Market Design Lecture 2 Combinatorial Markets, Walrasian Equilibrium, Tâtonnement

THE ECONOMICS OF TAXATION

Macroeconomic equilibrium in the short run: the Money market

Transcription:

THREE ESSYS ON THE ECONOMICS OF PREFERENTIL TRDE GREEMENTS: FREE TRDE RES RULES OF ORIGIN ND STOMS UNIONS by RENFENG XIO BS Huazhong Unversty of Scence and Technology 003 MS Huazhong Unversty of Scence and Technology 006 N BSTRCT OF DISSERTTION submtted n partal fulfllment of the requrements for the degree DOCTOR OF PHILOSOPHY Department of Economcs College of rts and Scences KNSS STTE UNIVERSITY Manhattan Kansas 010

bstract There have been consderable dscussons about why countres have nterests n formng preferental trade agreements (PTs) whch typcally take the forms of a free trade area () wth Rules of Orgn (ROO) and a customs unon () (orld Bank 005) Ths dssertaton contans three essays wth three dfferent models of trade under olgopoly to analyze varous ssues on preferental trade agreements The frst essay examnes welfare mplcatons of formng preferental trade arrangement (PTs) between two asymmetrc countres that dffer n ther market szes Key fndngs are as follows Frst when market sze asymmetry between two countres s not too large and ROO requrements are not too restrctve the formaton of an wth effectve ROO can be welfaremprovng to both members Second the formaton of a PT s more lkely to emerge between countres of smlar n ther market szes ceters parbus Thrd compared to the pre-pt equlbrum there are greater reductons n external tarffs under an than under a such that a non-member country s relatvely better off under the The second essay presents a three country model of trade under Bertrand prce competton to analyze dfferences n welfare mplcatons between an wth ROO and a customs unon () It s shown that the maxmum lmt of ROO requrements over whch there are welfare gans from trade for members depends crucally on the degree of substtutablty of fnal goods (or the ntensty of product market competton) It s also found that member countres and ther fnal-good exporters are better off n a than n an There are greater reductons n external tarffs under an than under a such that a non-member country s relatvely better off under the The thrd essay presents a three country model of wth Cournot quantty competton and derves the maxmum enforceable level of ROO over whch there are welfare gans from trade to each member country It s shown that ROO and external tarffs are strategc complements such that the hgher s the regonal nput restrctons the hgher s the external tarff necessary to nduce frms to fully comply wth ROO requrements It s also shown that an wth effectve ROO has a postve effect on the fnal-good trade But the trade-dvertng effect does not occur n the fnal-good sector

THREE ESSYS ON THE ECONOMICS OF PREFERENTIL TRDE GREEMENTS: FREE TRDE RES RULES OF ORIGIN ND STOMS UNIONS by RENFENG XIO BS Huazhong Unversty of Scence and Technology 003 MS Huazhong Unversty of Scence and Technology 006 DISSERTTION submtted n partal fulfllment of the requrements for the degree DOCTOR OF PHILOSOPHY Department of Economcs College of rts and Scences KNSS STTE UNIVERSITY Manhattan Kansas 010 pproved by: Major Professor Yang-Mng Chang

bstract There have been consderable dscussons about why countres have nterests n formng preferental trade agreements (PTs) whch typcally take the forms of a free trade area () wth Rules of Orgn (ROO) and a customs unon () Ths dssertaton contans three essays wth three dfferent models of trade under olgopoly to analyze varous ssues on preferental trade agreements The frst essay examnes welfare mplcatons of formng preferental trade arrangement (PTs) between two asymmetrc countres that dffer n ther market szes Key fndngs are as follows Frst when market sze asymmetry between two countres s not too large and ROO requrements are not too restrctve the formaton of an wth effectve ROO can be welfaremprovng to both members Second the formaton of a PT s more lkely to emerge between countres of smlar n ther market szes Thrd compared to the pre-pt equlbrum there are greater reductons n external tarffs under an than under a The second essay presents a three country model of trade under Bertrand prce competton to analyze dfferences n welfare mplcatons between an wth ROO and a It s shown that the maxmum lmt of ROO requrements over whch there are welfare gans from trade for members depends crucally on the degree of substtutablty of fnal goods It s also found that member countres and ther fnal-good exporters are better off n a than n an There are greater reductons n external tarffs under an than under a such that a non-member country s relatvely better off under the The thrd essay presents a three country model of wth Cournot quantty competton and derves the maxmum enforceable level of ROO over whch there are welfare gans from trade to each member country It s shown that ROO and external tarffs are strategc complements such that the hgher s the regonal nput restrctons It s also shown that an wth effectve ROO has a postve effect on the fnal-good trade But the trade-dvertng effect does not occur n the fnal-good sector

Table of Contents Table of Contents v Lst of Fgures v cknowledgements v Essay 1- Preferental Trade greements between symmetrc Countres: Preferental Trade greements between symmetrc Countres 1 11Introducton 1 The nalytcal Framework 5 13 Optmal Output Decsons of Frms wthout Market Integraton 7 14 Customs Unon wth Internal Market Integraton 11 15 Dfferent Effects and Implcatons between an and a 15 16 Concludng Remarks 7 Essay -Free Trade reas ( wth Rules of Orgn) vs Customs Unons: Internatonal Olgopoly under Bertrand Prce Competton 34 1 Introducton 35 The nalytcal Framework 38 3 Three lternatve Trade Regmes 41 4 Comparng the lternatve Trade Regmes 48 5 Concludng Remarks 55 Essay 3-Three Free Trade reas Rules of Orgn and Optmal Tarff Reductons under Internatonal Olgopoly: welfare nalyss 63 31 Introducton 64 3 the Model Set-Up 68 33 Optmal External Tarffs (Subject to Full Complance wth ROO) 73 34 The Maxmum Extent of the elfare-enhancng ROO n an 75 35 Economc Implcatons of the Model 79 36 Concludng Remarks 85 v

Lst of Fgures Fgure 11 Tarff Reductons under the formaton of a PT 17 Fgure 1 Producer surplus of countres and B 18 Fgure 13 Total consumpton n country and B 1 Fgure 14 Socal welfare of Countres and B 4 Fgure 15 orld welfare under the alternatve trade regmes 6 ˆ Fgure 1 The relatonshp between and 45 t Fgure The relatonshp between external and 49 S Fgure 3 The relatonshp between external and 53 Fgure 4 The relatonshp between G and 54 Fgure 31 Market structure of the fnal good and the welfare-mprovng content rato 78 v

cknowledgements I would lke to apprecate my major advsor Dr Yang-Mng Chang for hs gudance kndness assstance for dong my research I would also lke to thank Dr Phlp Gayle Dr Charles Braymen and Dr Tan Xa for ther valuable comments whch led to a sgnfcant mprovement n my research Specal thanks to my outsde charperson Dr Davd Yetter My sncere apprecaton contnues to the all the colleagues n Economcs Department KSU especally to Zjun Luo Yaseen lhaj-yaseen Eddery Lam Jenny u Dave Brown Shane Sanders Bhavneet ala Shu Zhou Burak Oneml Yuan Gao Yunyun Lv Casey bngton Yutng L Xn Xe Hana Peroutkova Specal thanks to my famles v

Essay One Preferental Trade greements between symmetrc Countres: Free Trade reas (wth Rules of Orgn) vs Customs Unons 1

1 Introducton ll bg countres wth a farly large market have economc ncentves to form preferental trade agreements (PTs) wth small countres? Under what condtons wll PTs be welfare-mprovng for both bg and small tradng partners? The last two decades have wtnessed an unprecedented prolferaton of trade agreements whch typcally take the forms of a free trade area () and a customs unon () (orld Bank 005) n nterestng observaton s that countres formng an or a are generally dfferent n the szes of ther markets In an member countres small or large n ther market szes collectvely elmnate barrers on certan goods traded among them But they ndvdually set ther own external tarffs toward non-members Ths consttutes a sgnfcant dfference between an and a the latter of whch requres member countres to set a common external tarff on mports enterng nto the unon (Krueger 1993; Krshna and Krueger 1995; Panagarya 000) There are other dstnctve aspects of an To prevent re-exportaton or trade deflecton from a country wth a lower tarff to another country wth a hgher tarff members sgn n preferental rules of orgn (ROO) under whch products cannot get duty-free access to a partner s market unless ROO requrements are met (Grossman and Helpman 1995; Rchardson 1995) 1 Several dfferent crtera may be adopted These nclude ROO requrements based on regonal nput or content restrctons a change n tarff headng partcular processes that should be performed wthn an ; and a substantal transformaton of a product Despte ther dfferences n crtera there generally nvolves an ROO-nduced extra cost n producng a fnal good elgble for preferental treatment under an ROO not only generate producton neffcency 3 they may also cause fnal-good markets wthn an to be segmented In a 1 Grossman and Helpman (1995) examne among other thngs the effects of wth ROO that prevent reexportaton from a lower tarff member to a hgher tarff member Rchardson (1995) shows explctly that there s no Nash equlbrum n settng external tarffs because all members of an compete to set the lowest tarff wth respect to non-members See eg Krueger (1993) Thoenng and Verder (004) Cadot Estavadeoral Suwa-Esenmann and Verder (006) and Krshna (006) The North mercan Free Trade greement (N) the European Unon and the SEN Free Trade rea agreement () for example all contan certan crtera of the ROO provsons whch provde preferental treatment for member countres 3 See Krshna and Krueger (1995) and Krshna (006) Falvey and Reed (1998) analyze the cases of nonpreferental ROO and ndcate that ROO may be used strategcally as polcy nstruments Ths s due to the potental arbtrarness n categorzng the geographcal sources of goods produced not n a sngle locaton Falvey and Reed (00) further show that producers may modfy ther producton processes and nput mx n response to content requrements n ROO

however there are no ROO provsons and arbtrage actvtes are relatvely costless so that the prces of products tend to be unform across members' markets (Mukunok 004) In other words there s nternal market ntegraton n a The objectve of ths paper s to examne dfferences n economc effects and welfare mplcatons between an and a when member countres are asymmetrc n ther market szes Based on a stylzed three-country model of trade under olgopoly we wsh to answer the followng questons that appear not to have been adequately analyzed or answered Under mperfect competton n fnal-good markets wll the formaton of an or a be more lkely to emerge between countres dssmlar n ther market szes? How wll market sze asymmetry affect ther economc ncentves of formng a PT (ether an or a )? Does an or a allow member countres to have a greater degree of market sze asymmetry? hat effects preferental ROO requrements have on the welfare of formng an between asymmetrc countres? ll the formaton of a be preferred to that of an wth ROO vewed from the perspectves of asymmetrc member countres a non-member country and the world as a whole? Frms nsde an are requred to comply wth ROO to be elgble for preferental treatments n trade between member countres ROO are effectve only when extra costs resultng from obeyng the rules are not too hgh; otherwse exportng frms wthn an smply choose to pay tarffs (Ju and Krshna 00 005) The formaton of a permts member countres to trade wth each other wthout payng tarffs One concern then s whether or member countres wll set hgh external tarffs to protect ther own frms whch make non-member countres worse off rtcle XXIV of the GTT tres to secure non-member countres from welfare losses by forbddng external tarffs set by members of a trade bloc to exceed ther pre-pt levels In the present paper we explctly take ths GTT requrement nto account when determnng optmal external tarffs for member countres Our analyss complements the contrbuton by Mukunok (004) n terms of welfare comparsons between an and a But the author does not take nto account ROO requrements and ther effects on ncreasng producton costs of fnal goods for trade wthn an Mukunok (004) adopts an olgopoly model of product dfferentaton and shows the case that a free trade area entals endogenous change from segmented to ntegrated markets for nternally produced goods Our model dffers from that of Mukunok s n several mport aspects 3

Frst we consder the decsons of exportng frms wthn an n complyng wth ROO provsons s such n analyzng the case of an ROO-nduced extra costs make nternal market ntegraton between member countres not a possblty Second our analyss allows for the case that two potentally partcpatng members are asymmetrc n terms of ther dfferent market szes Despte the dfferences n assumptons on economc condtons our analyss complements the contrbuton by Mukunok (004) and has some nterestng results common to both studes thout market ntegraton formaton makes non-members better off because of lower external tarffs set by countres Under plausble condtons welfare gan to each member country s hgher n a than n an Our analyss s closely related to two recent contrbutons on s One s Das and Ghosh (006) who develop a model of asymmetrc world economy n whch there are at least four countres The authors show that formaton s more lkely to emerge among smlar countres In our analyss we further take nto account effects of preferental ROO provsons n an e also examne dfferences n welfare mplcatons between an and a Our analyss also complements another contrbuton by Duttagupta and Panagarya (007) The authors show that ROO are not always harmful as they could make a prevously nfeasble feasble assumng that external tarffs reman at ther pre- levels In our analyss we further allow external tarffs to be endogenously set by members e fnd that when ROO requrements are not too restrctve an wth ROO can be welfare-mprovng to member countres of asymmetrc n ther market szes In hs contrbuton Krshna (1998) examnes the mpact of s on the nternal ncentves for multlateral lberalzaton The author challenges the contenton that s are superor to multlateral trade lberalzaton as a way of movng to free trade for all In hs model of mperfect competton n segmented markets Krshna (1998) looks at the ssues on preferental tradng arrangements from the poltcal economy perspectve and analyzes trade polcy as the result of lobbyng by nterest groups 4 In our analyss we pay attenton to the producton effcency perspectve of preferental trade agreements and compare dfferences n welfare mplcatons between an and a 4 Grossman and Helpman (1995) further examne the poltcal vablty of s when two countres negotate a freetrade agreement Based on a poltcal-economy framework n whch ndustral nterest groups attempt to nfluence ther government the authors show that an can be an equlbrum outcome For further contrbutons on formaton and nterest group poltcs see eg Krshna (1998) Magg and Rodríguez-Clare (1998) Mtra (00) and Ornelas (005) 4

The remander of the paper s organzed as follows Secton presents a smple framework of trade under nternatonal olgopoly In Secton 3 we frst dscuss the benchmark case wth no preferental trade agreements of any form e then examne condtons under whch two potental member countres fnd t benefcal to form an subject to the constrants that ther exportng frms nsde the meet ROO In Secton 4 we derve condtons under whch two asymmetrc countres form a by settng a common external tarff In Secton 5 we analyze dfferences n welfare mplcatons between an and a e also compare ther effects on tarff reductons consumers and frms n each member country as well as world welfare Secton 6 contans concludng remarks The nalytcal Framework Consder a smple world that s composed of three countres denoted as B and C 5 Countres and B are located n the same regon but are asymmetrc wth respect ther dfferent market szes In the absence of tradng agreements and B engage n a two-way trade n fnal goods wth each country mposng a specfc tarff on mports from the other (Brander and Spencer 1984) Countres and B consder the formaton of a PT (ether an or a ) Country C represents the rest of the world In the three-country world each country has a sngle frm called by ts own country s name and produces a homogeneous fnal good q Frms and B do not export ther fnal goods to country C but frm C exports ts fnal good to the markets n countres and B Denote q k as country 's consumpton of the fnal good produced by frm k where B and k B C e assume country s aggregate utlty functon to be U Q (1 / ) Q Y where s a postve parameter Q ( q q q ) represents the fnal-good consumpton n country B C and Y s the consumpton of a compettvely produced numerare good whch s freely traded The utlty functon mples that country s demand for fnal good s p Q where p s 5 lthough model settng and assumptons may dffer the use of a three-country model to analyze ssues on wth ROO can be found n several recent studes such as nson Cadot Estevadeordal de Melo Suwa-Esenmann and Tumurchudur (005) Ju and Krshna (005) and Ishkawa Mzoguch and Mukunok (007) 5

the fnal-good prce n the country and represents ts market sze The asymmetry of the two potental member countres s captured by ther dfference n market szes Specfcally we assume that B where ( 1) measures the degree of market sze asymmetry between the small country and the bg country B s n Ishkawa et al (007) we focus our analyss on how formaton affects the fnal-good markets by assumng that nput markets are perfectly compettve e assume that wthout ROO requrements of any form the average and margnal costs of producng the fnal good for all frms are constant For analytcal smplcty these costs are normalzed to zero Ths allows us to pay specal attenton to other types of costs that frms may ncur Denote c k as the extra cost that frm k ncurs n producng or exportng one unt of ts fnal good to country hen servng ts own domestc market c 0 for an nsde frm regardless of whether or not there s an or a Pror to formng a PT frm j( j B j ) s requred to pay tarffs for each unt of ts fnal good exported to country To frm j ts extra cost c j s equal to the tarff charged by country fter an s establshed a frm exportng ts fnal good wthn the ncurs an extra cost f t chooses to comply wth ROO In ths case c j represents ROO-nduced extra cost 6 and ts value s postve (e cj 0) Under a member countres trade wth each other wthout barrers In ths case the extra cost to frm j ts c j s zero s for frm C that serves the fnal-good markets n and B ths outsde frm s requred to pay tarffs The extra cost c C to frm C s equal to a specfc tarff the amount of whch depends on whether there s an a or wthout any form of a trade agreement e consder a two-stage game In the frst stage governments of countres and B decde whether or not to form an or and thereafter determne ther external tarffs In the second stage frms choose ther output levels and compete n the fnal-good markets n the regon e use backward nducton to solve for the sub-game perfect Nash equlbrum of the three alternatve trade regmes e begn wth the second stage at whch frms make ther producton decsons 6 In Secton 3 we wll dscuss n more detals the ncrease n costs resultng from complyng wth preferental ROO requrements 6

3 Optmal Output Decsons of Frms wthout Market Integraton Product markets n small and bg countres are segmented pror to the formaton of a PT Ths s because of mport tarffs under a two-way trade For the case n whch the two asymmetrc countres establsh an an nsde frm s requred to comply wth ROO to be elgble for tarff-free n exportng fnal good wthn the Product markets n the two countres reman to be segmented because of ROO-nduced extra costs For the pre-pt and regmes wthout market ntegraton frm k( B C) sells q k unts of fnal good to country ( B) proft of frm k s: Dependng on extra cost c k n dfferent stuatons dscussed earler the total ( p c ) q (1) k k k B e assume that each frm employs a Cournot strategy n ts producton decson takng as gven the quanttes of the fnal good produced by all other frms Based on equaton (1) we calculate the quantty of the fnal good exported to country ( B) by frm k( B C) as follows: ( cj cc) qk ck () 4 Total consumpton Q ( q q q ) of the fnal good and ts prce n country are gven respectvely as: B C (3 cj cc) Q ; p 4 The suffcent condtons for q k and ( cj cc) (3) 4 Q to be postve are when the market sze s large 1 enough such that 4ck cj cc and ( cj cc ) These condtons are assumed to hold 3 e then calculate consumer surplus and producer surplus for country as follows: (3 cj cc) ; S 3 where j B and j ( cj cc) ( j 3 cj cjc) ; (4) 16 Denotng t j and t C as the tarff rates that country charges on mports from country j 7

and country C we calculate total tarff revenue for country as follows: tj ( 3 cj cc ) tc ( cj 3 cc ) R tjqj tcqc (5) 4 Each potental member country s socal welfare whch s taken as the sum of consumer surplus producer surplus and tarff revenue s: S R (3 cj cc ) ( cj cc ) ( j 3 cj cjc ) tj ( 3 cj cc ) tc ( cj 3 cc ) 3 16 4 31 Pre-PT In the absence of trade agreements country ( B) charges a unform tarff to all mports of the fnal good from countres j( B j ) and C Denotng t as country 's PT tarff on each unt of ts mports we have c c t t t (7) C j C j (6) pre- Substtutng equaton (7) nto n equaton (6) settng the frst-order condton that d / dt 0 we derve the pre-pt optmal tarff rate: 7 3 t (8) 10 The pre-pt optmal tarff s postvely related to market sze n that the bg country sets a hgher tarff rate than the small one Usng equatons (7) and (8) we calculate the equlbrum quanttes prce and total consumptons of the fnal good n country as follows: q ; 5 q j 1 ; 10 q C 1 ; 10 3 p ; Q (9) 5 5 Substtutng the above equatons back nto equatons (4)-(6) we have consumer surplus producer surplus tarff revenue and socal welfare for country as follows: 4 1 9 j; S R 5 100 50 3 ; 50 1 j (10) 5 100 In what follows we use the pre-pt equlbrum as the benchmark to evaluate the 7 It s easy to verfy that the second-order condton for welfare maxmzaton s satsfed 8

alternatve trade regmes 3 wth Effectve ROO countres do not charge tarffs on mports from ther partners so that t = 0 for j B and j But they have ndependence n settng external tarffs on ther j mports Denotng the tarff rates as t for B we have t = t Dfferences n tarffs may cause a re-exportaton of the fnal good from a lower-tarff member to a hgher-tarff member To elmnate trade deflecton the countres agree that products cannot get dutyfree access to other countres n the unless ther productons satsfy the ROO requrements 8 Under preferental ROO each nsde frm s cost of exportng ts fnal good elgble for duty-free treatments ncreases th these constrants we denote ( 0) as the ROO-nduced extra cost That s c j Despte the formaton of an fnal-good exporters wthn the may or may not choose to comply wth ROO (Ju and Krshna 005) hether an frm decdes to meet ROO depends crucally on () mport tarffs set by members t and () the ROO-nduced extra cost For each unt of the fnal good exported to country by frm j f t (11) t s benefcal to the frm to comply wth ROO Otherwse the frm prefers to export the fnal good to country by smply payng tarff In ths case ROO becomes neffectve For the effcacy of ROO member countres set ther external tarff rates above Equaton (11) thus defnes the ROO-complyng condton For frm C outsde of the t pays the specfc tarff t when exportng ts fnal good to country In ths case c t Gven that c postve quanttes of the fnal good as shown n equaton () requres that C 3t 0 and 3 t 0 (1) e calculate the level of socal welfare for country by substtutng t j = 0 j C 8 There are dfferent crtera for ROO n terms of () regonal content requrements () a change n tarff headng () partcular processes that should be performed wthn an ; and (v) a substantal transformaton of a product See eg Thoenng and Verder (004) Cadot Estavadeoral Suwa-Esenmann and Verder (006) and Krshna (006) The North mercan Free Trade greement (N) the European Unon (EU) and the SEN Free Trade rea agreement () all contan ROO provsons whch provde preferental treatment for member countres 9

tc = t cj and c C t nto equatons (6) to obtan 1 3 [ 1( ) (14 6 )( ) 3 11 j t j ] ( ) t t (13) 3 4 Unless the formaton of an s based on noneconomc or poltcal objectves t s plausble to assume that each potental member s wllng to sgn n ROO provsons that mprove ts socal welfare Country decdes to jon an wth effectve ROO only when the welfare s hgher than the pre-pt welfare That s (14) Equaton (14) defnes the welfare-mprovng condton of formng an The next step s to determne the degree of market sze asymmetry and the range of the ROO-nduced extra cost that guarantee welfare gans from trade for each member country lso we wsh to determne each member's optmal tarff The problem facng each member country s to choose t that maxmzes n equaton (13) subject to the ROO-complyng condton n equaton (11) the welfare-mprovng condton n equaton (14) and the constrants that quanttes of the fnal good produced are postve as gven n equaton (1) For the bg country B ts soluton to the constraned welfare optmzaton problem exsts when the followng condtons are satsfed (see -1 n the ppendx): 1 198 and 0 ˆ where ˆ 4 7 5949 ( ) (15) 77 385 96450 For the small country ts soluton to the constraned welfare optmzaton problem also exsts f the above condtons are satsfed hen the degree of market sze asymmetry s not too large (1 198) and the ROO-nduced extra cost s suffcently small ( 0 ˆ ) formng an s welfare-mprovng to both the bg and small countres Nevertheless when the market sze asymmetry s too large ( 198 ) formng an wth effectve ROO may be welfaredeteroratng to at least one country (especally the bg country B ) Investgaton of equaton (15) reveals that for 1 198 the crtcal value of ˆ s negatvely assocated wth The fndngs of the analyses permt us to establsh PROPOSITION 1 Formng an wth ROO s welfare-mprovng to two partcpatng countres when the asymmetry n ther market szes s not too large and ROO requrements are not too restrctve The crtcal value of the ROO-nduced extra cost that makes formaton 10

welfare-mprovng s lower when the market sze asymmetry s greater The economc mplcatons of Proposton 1 are straghtforward Other thngs beng equal the lkelhood of formng a successful wth ROO between two asymmetrc countres s hgher when ther market szes are smlar Moreover ths lkelhood s hgher when ROO requrements are less restrctve ssumng that the condtons n equaton (15) hold we use the welfare functon n equaton (13) to solve for the optmal tarff rate for each member country as 1 1 t (16) 7 3 Ths ndcates that the optmal tarff s hgher the larger the market sze asymmetry or the hgher the ROO-nduced extra cost comparson between equaton (8) and equaton (16) ndcates that t < t (17) Ths result s consstent wth the requrement as specfed by the rtcle XXIV of GTT that an external tarff should not be set above the pre-pt level to avod a negatve mpact on a nonmember country Substtutng the optmal tarff from equaton (16) nto equatons (-6) takng nto account the cost condtons that tj 0 cj and t = c t we have equlbrum outputs prces C C consumptons consumer surplus producer surplus tarff revenue and socal welfare as follows: 1 ; q ; 1 qj qc ; (18a) 7 3 7 3 7 1 ; 5 1 p Q ; (18b) 7 3 7 3 1 7 3 7 3 ( ) ( j ) ; 1 5 1 S ( ) ; (18c) 7 3 1 1 1 11 8 5 4 R ( ) ; j j (18d) 7 7 3 18 1 14 49 4 Customs Unon wth Internal Market Integraton In a where member countres trade wth each other wth zero tarffs they collectvely set a common external tarff wth respect to non-member countres lso frms producng wthn 11

the are not subject to ROO requrements Denotng the external tarff rate as t we have from the cost condtons that cj 0 ( j B j) and c t ( 0) Because of free trade C the fnal good q may be resold between member countres and B untl ther prces are dentcal p p In the subsequent analyss we allow for such an nternal market ntegraton between the members 9 Insde and outsde frms treat the two markets n countres and B as an ntegrated one under a s such the total quantty of fnal good sold by frm k( B C) to the sngle market s q k where q k qk qbk Dependng on extra cost c k n dfferent stuatons dscussed earler the proft functons of nsde frm and outsde frm C are gven respectvely as and p q (19a) ( p t ) q (19b) C C ll frms ndependently determne ther total output levels of the fnal good q that maxmze ndvdual profts subject to the unform-prce condton e calculate the solutons as follows: 1 1 1 3 q (1 ) t ; qc (1 ) t (0) 4 Equlbrum market prce and total consumpton of the fnal good n a country are: 1 t (1 ) ; 1 p (7 Q t j ) (1) 8 4 8 e calculate consumer surplus and frm proft n country as follows: 1 (7 ) ; 1 S t j [(1 ) t ] () 18 3 Because of nternal market ntegraton n the we cannot determne tarff revenue collected by each member separately Gven a common external tarff ( t ) on mports from country C total tarff revenue collected by the s: k 9 See eg Ishkawa Mzoguch and Mukunok (007) 1

where 1 1 RB R RB t qc (1 ) t ( t ) (3) 4 Based on equatons (0)-(3) we calculate welfare for each country as (7 t j) ( j t ) S R R (4) 18 3 R depends on the dstrbuton of the total tarff revenue R It s plausble to assume B that two potental member countres form a when each one s welfare S s hgher than ts pre-pt welfare That s (5) e take nto account ths welfare-mprovng condton n determnng the common external tarff rtcle XXIV of the GTT requres that the post- tarff be no greater than the pre-pt tarff n order not to negatvely affect non-member countres In ths case we have t t Gven B and the pre-pt optmal tarff t follows that 3 (see equaton (9)) we have t tb It 10 t t (6) In settng a common external tarff the member countres jontly maxmze the sum of ther socal welfare ( ) subject to the welfare-mprovng condton n equaton (5) B and the GTT tarff-reducton condton n equaton (6) Makng use of equatons (3) and (4) we calculate overall welfare ( ) for the as B 19 5 9 3 B B ( t ) (1 ) t (1 ) 16 16 64 3 (7) The soluton to the welfare-maxmzaton problem exsts and the equlbrum quantty of the fnal good consumpton n the small country s postve Q 0 10 when market sze asymmetry falls nto the followng range: 1 64 Consderng ths condton and the GT/TO tarff requrement (see equaton (6)) we have two nterestng cases n solvng for the common external tarff by the : Case 1: hen 1 18 there s an nteror soluton e solve for the optmal tarff as: 10 hen 64 Q 0 e rule out ths case S 13

5(1 ) t (8) 38 Substtutng t from equaton (8) nto equatons (0)-(4) yelds the equlbrum outputs prce consumer surplus producer surplus socal welfare for each member and total tarff revenue n a : q 6 (1 ) ; q 19 16 3 Q j; 19 19 1 16 3 S 19 19 C ( j) ; 1 (1 ) ; (9a) 19 3(1 ) ; (9b) 19 p 18(1 ) ; (9c) 361 1 16 3 18 R (9d) 19 19 361 ( j) (1 ) 5 (1 ) RB (9e) 7 Case : hen 18 1834 or 35099 64 there s a corner soluton But when 1834 35099 there remans to have a corner soluton provded that the amount of tarff revenue collected by the bg country s large enough That s when ˆ R R B B where 3 3 3 R B ˆ ( ) Ths condton ensures that formng a s welfare-mprovng to 10 64 4 10 the bg country e For these two possbltes of a corner soluton we fnd that the B B optmal tarff s: 3 t t (30) 10 Substtutng t from equaton (30) nto equatons (0) yelds the equlbrum outputs q ( ) 1 ; q C ( ) ; (31a) 5 4 4 5 e further calculate total consumpton of the fnal good equlbrum market prce consumer surplus producer surplus socal welfare for each member country and total tarff of a as follows: 4 Q ( ) 7 1 ; Q ( 1 B ) ; p ( ) ; (31b) 5 8 8 5 5 8 14

(5 3) 300 S ; (35 8) 300 S B ; 1 1 ( ) ; (31c) 3 10 5 because (5 56) R 640 R B ( ) ; ; (65 48 64) B RB ; (31d) 640 3 1 (31e) 10 4 5 For the small country ts post- welfare s greater than ts pre-pt welfare Ths s even for R 0 But for the bg country B whether or not formng a s welfare mprovng cannot be determned unambguously hen the degree of market sze asymmetry s moderate (1834 35099) a welfare mprovement to the bg country requres a suffcent amount of tarff revenue from the non-member country C to cover the potental losses Ths s because ˆ R 0 when 1834 35099 Thus for B 1834 35099 wth an approprate dstrbuton of tarff revenue between the member countres there wll be welfare gans to each one under a e thus have PROPOSITION Formng a s always welfare-mprovng to small and bg countres when ther market sze asymmetry s suffcently low (1 1834) or suffcently hgh (64 35099) However when the asymmetry n market sze s moderate (1834 35099) t s not n the nterest of the bg country to jon a unless ts tarff revenue from the rest of the world s suffcently large 5 Dfferent Effects and Implcatons between an and a s dscussed n Secton 3 relatve to the pre-pt equlbrum there are welfare gans to both the small and bg countres when ther market sze asymmetry s not too large (1 198) and the ROO requrements are less restrctve ( 0 ˆ ) Further Proposton shows that there are welfare gans to two asymmetrc countres from formng an when ther market sze asymmetry s small enough ( 1 64 ) and when the bg country receves suffcent amount of tarff revenue from a non-member country In terms of dfferences n the degree of market sze asymmetry between an and a we have PROPOSITION 3 Other thngs beng equal formng a welfare-mprovng allows for a greater degree of market sze asymmetry than formng a welfare-mprovng 15

hen market sze asymmetry s not too large (1 198) ether formng an wth less restrctve ROO requrements ( 0 ˆ ) or formng a s welfare-mprovng to each member In what follows when comparng dfferences between an and a we assume that both the asymmetry condton and the ROO restrctons hold unless otherwse specfed 51 Effects on External Tarffs From t n equatons (8) and t n equaton (16) we fnd that the rato of the post- tarff to the pre-pt tarff for the bg country s smaller than that for the small country Ths mples that there s a greater tarff reducton by the bg country than by the small country after the formaton of an Moreover we fnd that for gven ROO-nduced extra cost t t decreases as the degree of market sze asymmetry ncreases Thus there s a greater reducton n the optmal tarffs for both member countres when the degree of market sze asymmetry s greater Equaton (16) ndcates that the optmal tarffs for both member countres are ncreasng functons of e then have t t 0 0 垐 t 1 (3) t t t Fgure 1 llustrates for each country the upper and lower lmts of the rato of tarff after formng an to ts pre- tarff fter formng a although both member countres set a common tarff t equatons (8) and (30)) the ratos of the related tarff to the pre-pt tarff for the small and bg countres are dfferent due to ther dfferences n the pre-pt tarffs t (see equaton (9)) s can be seen from Fgure 1 ths rato for the small country ncreases as ncreases untl 18 For the bg country the reducton n external tarff after formng a becomes relatvely greater when s larger (see 16

PT t 10 t t t 09 08 t t B 07 06 05 t t ˆ t B 18 100 105 110 115 10 15 130 135 140 Fgure 1 Tarff Reductons under the formaton of a PT t B ˆ s llustrated n Fgure 1 the optmal tarff s always hgher than the upper lmt of the optmal tarffs for both the small and bg countres e thus have PROPOSITION 4 Despte the fact that there are ROO-nduced extra costs when bg and small countres form an tarff reductons by the two members reman relatvely greater than those when they form a Interestngly the bg country lowers ts optmal tarff by an amount that s greater than that by the small country regardless of whether they form an or a 5 Comparng Profts of Insde Frms wthn a PT In an there are fundamentally no trade barrers between small and bg member countres other than ther preferental ROO provsons But the ROO-nduced trade cost s set at a level lower than the external tarff mposed by each member It s nterestng to see how ROO requrements would affect frms nsde the small and bg countres where market szes are dfferent s shown n equaton (18c) the small country s producer surplus s an decreasng 17

functon of It follows that follows that (33a) ˆ 0 ˆ 0 But for the bg country s producer surplus t s a ncreasng functon of It B 0 B 0 垐 B (33b) Equatons (33a) and (33b) ndcate that due to dfferences n market szes the small country frm prefers ROO requrements to be less restrctve whereas the bg country frm prefers them to be more restrctve PT 15 14 13 B 15 B 14 13 1 11 10 09 08 0 ˆ 10460 11505 198 100 105 110 115 10 15 130 135 1 11 10 09 08 B B B ˆ B B 0 B 1056 100 105 110 115 10 15 130 135 Fgure 1 Producer surplus of country Fgure Producer surplus of country B Fgures 1 and llustrate for each country the rato of frm proft or producer surplus before and after the formaton of an The fgures also llustrate the rato of frm proft before and after the formaton of a e fnd that the fnal-good producers n both the small and bg countres are better off under a than under an snce ˆ Note that despte market sze asymmetry between the small and bg countres frms 18

producng nsde the trade block are assumed to be dentcal n our analyss e fnd that the ratos of frm profts are decreasng functons of for the bg country but are ncreasng functons for the small country fter formng a trade agreement (ether an or a ) member countres wll share ther markets wth each other The small country frm s able to enter the large market n the bg country hen the market sze dfferental s greater the small country frm fnds t more proftable from accessng to the large market lthough the bg country frm shares the market n the small country the loss to the bg country frm n ts own large market may outwegh the gan to the frm n the small market Ths s especally true when the degree of market sze asymmetry s sgnfcantly large s presented n Fgure 1 the formaton of an always makes the fnal-good producers n the bg country worse off The effect on the fnal-good producers n the small country cannot be determned unambguously however It depends on the degree of market sze asymmetry as well as the ROO-nduced extra cost s shown n ppendx - we have the followng two possbltes: () when () 1 10460 or () 10460 11505 and 0 ; () when () 11505 64 or () 10460 11505 and It comes as not a surprse that the formaton of a always makes the fnal good producers n the smaller country better off The effect on the fnal-good producers n the larger country cannot be determned unambguously however It depends on the degree of market sze asymmetry s shown n ppendx -3 we have two possbltes: () B B when1 1056; () B when 1056 64 B e summarze the above fndngs n the followng proposton: PROPOSITION 5 In two asymmetrc countres that form a PT nsde frms unambguously make more profts under a than under an The bg country producer s hurt by an but may be better off n a when the degree of market sze asymmetry s suffcently low The small country producer s better off under a but may be hurt by an 19

53 Effects on the Non-Member Country In the model of trade under olgopoly the non-member country exports the fnal good to the regon where two asymmetrc countres may form an or a Makng uses of equatons (1) and (19b) we calculate profts of the outsde frm under dfferent trade regmes as follows: 1 (1 ) ; C (34a) 100 1 (1 ) 0 C when 1 198 and 0 ˆ ; (34b) 49 C It s straghtforward to verfy that 0 ˆ lso we fnd that 13 64 e thus have 1 (1 ) when 1 18; 7 (5 4) when 18 64 800 and C C C C C for 1 13 C (34c) when 1 198 and but C C for PROPOSITION 6 For the three trade regmes we consder non-member country fnds t most benefcal under an Nevertheless the non-member country may be negatvely affected by a when market sze asymmetry between member countres s small Formng a PT has two opposng effects on a non-member country One effect s postve n that countres reduce ther external tarffs The other effect s negatve n that the non-member country s hurt due to the fact that the outsde frm s trade costs reman to be hgher than those for the nsde frms Proposton 6 ndcates that the postve effect may outwegh the negatve effects under an or a when the degree of market sze asymmetry n member countres s large enough 54 Effect on Consumers n a Member Country From Q n equaton (18b) we know that the level of the fnal-good consumpton n each member country decreases wth Under ths crcumstance we have Q Q Q (35) 垐 0 0 0

Fgures 31 and 3 llustrate for the bg and small countres the lower and upper lmts of the fnal-good consumpton after formng an as compared to ther pre- consumpton levels e fnd that ths consumpton rato s strctly greater than one Ths mples that due to tarff reductons and lower trade costs between member countres total consumpton of the fnal good n each member country ncreases after formng an e expect the formaton to result n a decrease n the equlbrum market prce of the fnal good causng consumer surplus n each member country to ncrease QB 11 QB Q 119 Q 118 Q 0 Q 10 119 Q B 0 QB 117 116 115 Q ˆ Q 118 Q B Q B 114 113 11 117 116 Q B ˆ QB 111 110 109 Q Q 198 115 1353 10317 198 114 100 105 110 115 10 15 130 135 Fgure 31 Total consumpton n country B 108 107 106 105 100 105 110 115 10 15 130 135 Fgure 3 Total consumpton n country Fgures 31 and 3 also llustrate for each country the rato of the fnal-good consumpton after formng a to ts pre-pt consumpton For the bg country ths rato ncreases as the degree of market sze asymmetry ncreases Note that ths consumpton rato s strctly greater than 1 whch means that the fnal-good consumpton ncreases after the formaton a However for the small country ts consumpton decreases as the degree of market sze asymmetry ncreases e show n ppendx -4 that the consumpton rato for the small country s greater than 1 when 1 16 but s less than 1 when 16 These results ndcate that f the market sze asymmetry s crtcally hgh ( 16) consumers n the small country are 1

worse off under a If the degree of market sze asymmetry s not too large (1 198) the fnal-good consumpton n the small country s unambguously hgher under an than under a or wthout a PT s a result consumers n the small country prefer an over a For the bg county the relatonshp n the fnal-good consumpton between the case and the case cannot be determned unambguously s shown n ppendx -5 the levels of the fnal-good consumpton depend crucally on and hen () 1 10317 or () 10317 1353 and 9 51 ( ) ˆ consumers n the bg country are better off 19 133 under an snce Q Q But when () 10317 1353 and B B 9 51 0 ( ) or 19 133 () 1353 199 consumers n the bg country are better off under a snce Q B Q B e summarze the fndngs of the analyses n the followng proposton: PROPOSITION 7 Formng an wth ROO makes consumers n each member country better off If the member countres are dssmlar as ther symmetry n market sze s large consumers n the larger country prefer a over an lthough consumers n the smaller country are better off under an than under a they may be worse off n a when the market sze asymmetry s suffcently large 55 Effects on elfare of Each Member e show n the prevous secton (see Proposton ) that for each member county there s a welfare mprovement under a when 1 64 e now dscuss ths welfare ncrease after the formaton of a n more detal For the small country From Proposton 5 ts producer surplus s always greater than ts pre-pt producer surplus From proposton 7 ts consumer surplus decrease as market sze asymmetry ncreases and s greater than that wthout any PT when 1 16 For 16 the ncrease n producer surplus after formng a outweghs the decrease n consumer surplus and the loss n tarff revenue Ths explans why for the small country socal welfare under a s unambguously hgher

But for the bg country we fnd that the result s ambguously Under a producer surplus decreases as market sze asymmetry ncreases and s less than 1 when s sgnfcantly large The bg country s consumer surplus ncreases as ncreases hen market sze asymmetry s not too large (1 1834) or s large enough (35099 64) the ncrease n customer surplus outweghs the loss n both producer surplus and tarff revenue hen market sze asymmetry s moderate (1834 35099) the ncrease n customer surplus s more than offset by the decrease n producer surplus For the bg country to be able to mprove ts socal welfare from formng a t would need to collect a suffcent amount of tarff revenues from a non-member country Recall that after formng a the total tarff revenue to the countres s gven n equatons (9e) and (30e) Thus one member country s tarff revenue s maxmzed when the other one s s zero e know that each member country s welfare s an ncreasng functon of ts tarff revenue (see equatons (9d) and (31d)) Thus (36) R 0 R j 0 R e also know that for 1 198 the welfare s a decreasng functon of the ROO-nduced extra cost where ˆ 0 ˆ 0 0 ˆ (see equaton (19e)) In ths case we have (37) Fgures 41 and 4 llustrate the ratos of post-pt welfare of each member country to ts pre-pt levels For the bg country the lower lmt of the welfare rato under an s equal to one The bg country prefers a over an f 10807 198 For 1 10807 formaton s a preferred choce to the bg country f ts tarff revenue ( R B ) s sgnfcantly large For an acceptable value of we see from Fgure 4 that there s a greater welfare gan to the small country under a than under an provded that ts tarff revenue s also sgnfcantly large 3

PT B 1 B 10 118 116 114 PT 1 10 118 116 114 R B 0 11 110 108 106 104 10 B 0 B 10807 B R B R B 100 100 10 104 106 108 110 11 114 116 118 10 1 14 16 18 0 Fgure 41 Socal welfare of Country B B 0 B 11 110 108 106 104 10 0 ˆ 11550 R 100 100 10 104 106 108 110 11 114 116 118 10 1 14 16 18 0 Fgure 4 Socal welfare of Country But we cannot tell whether both the small and bg countres wll prefer a over an at the same tme It depends crucally on tarff revenue to the two countres th nternal market ntegraton we only know the overall tarff revenue collected by the (see equatons (9e) and (30e)) but not the amount to each ndvdual member It s nterestng to dentfy the condtons under whch both countres are better off under a than under an Ths leads to the followng proposton: PROPOSITION 8 hether there wll be a greater welfare mprovement for both asymmetrc countres under a than under an cannot be determned unambguously But when each of the small and bg countres collects a suffcent amount of tarff revenue formng a s a preferred choce over an PROOF: See ppendx -6 Proposton 8 mples that the dstrbuton of tarff revenue plays a role n affectng socal welfare when there s nternal market ntegraton n a 4

56 Effects on orld elfare Fnally we examne dfferences between an and a n terms of ther effects on the overall welfare of the three-country world In our settng world welfare s defned as the sum of socal welfare n the two member countres and producer surplus n the non-member country That s where m m m m (38) G B C m pre-pt e frst calculate the pre-pt level of world welfare by substtutng n equaton (10) m and (34a) nto G n equaton (38) to obtan 1 G B C 1 (39) 50 e then calculate the level of world welfare by substtutng equatons (18d) and (34b) nto m n equaton (38) to obtan 11 1 463 3 463 G B C 9 77 1078 539 1078 1 For the case that market sze asymmetry s not too large (1 198) the level of world welfare decreases as ncreases for 0 ˆ That s G 垐 G 0 G 0 (40) (41) Fgure 5 llustrates the lower and upper lmts of world welfare under an For 0 ˆ the level of world welfare ncreases wth the degree of market sze asymmetry Even though the level of welfare for the bg country decreases when s crtcally larger the ncrease n the sum of the small country s welfare and the non-member country s producer surplus outweghs the decrease n the bg country s welfare Next we calculate the level of world welfare by substtutng equatons (9d) (31d) and (34c) nto m n equaton (38) to obtan 343 360 343 when 1 18 7 G B C 775 80 736 when 18 64 1600 (4) 5

G 1 G 11 10 G 0 09 G ˆ 1036 08 100 10 104 106 108 110 11 114 116 118 10 1 14 16 18 130 13 Fgure 5 orld welfare under the alternatve trade regmes s llustrated n Fgure 5 world welfare ncreases wth market sze asymmetry when there s a Even though welfare for the bg country decreases when the value of s larger the ncrease n the sum of the small country s welfare and the non-member country s producer surplus outweghs the decrease n the bg country s welfare There s an overall welfare mprovement for the three-country world regardless of whether the two asymmetrc countres form an or a s for welfare comparson between an and a we fnd condtons under whch world welfare s hgher under an than under a That s G when () 1036 198 or () when 1 1036 G 1 1850 and 0 1019309 1 e thus have 77 8867 PROPOSITION 9 orld welfare mproves when small and bg countres form a PT (ether or ) other thngs beng equal orld welfare s greater under an than under a f ether one of the followng condtons s satsfed: () The market sze asymmetry s moderate (e1036 198); () The ROO-nduced extra cost s small It has been observed that there are more s than s (orld Bank 005) From the 6

perspectve of world welfare our analyss may help explan condtons under whch such an observaton would occur 6 Concludng Remarks In ths paper we have derved the condtons under whch there are welfare gans from formng a preferental trade agreement for countres dfferng n ther market szes These condtons are shown to depend on the degree of market sze symmetry between countres the form of trade agreement as ether a or a and ROO requrements n the case of an e have dscussed dfferences n welfare mplcatons between the two alternatve trade agreements In analyzng the welfare effect of an we explctly take nto account the decsons of exportng frms n complyng wth ROO Key fndngs are presented as follows () The formaton of an wth effectve ROO s welfare-mprovng when the asymmetry n market sze between member countres s not too large and the ROO requrements are less restrctve () Formng a s more lkely to emerge between countres of smlar n ther market szes () However formng a welfare-mprovng for partcpatng countres allows for a greater degree of ther market sze asymmetry than formng a welfare-mprovng (v) Compared to the pre-pt equlbrum tarff reductons are relatvely greater n an than n a whch makes a non-member country s relatvely better off under the (v) If the asymmetry n market sze s moderate and the ROO are not too restrctve global welfare s hgher under an than under a 7