โครงการ “TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST- WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM”

รายงานฉบับสมบูรณ

โครงการ “TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST-WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM”

เสนอตอ

สํานักงานกองทุนสนับสนุนการวิจ (สกว.)

คณะเศรษฐศาสตร มหาวิทยาลัยเชียงใหม

สิงหาคม 2559

สัญญาเลขที่ RDG5010021

รายงานฉบับสมบูรณ

โครงการ “TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST-WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM”

xxxxxxxxx xxxบุญจิตต และ ปเตอร คาลกินส คณะเศรษฐศาสตร มหาวิทยาลัยเชียงใหม

สํานักงานกองทุนสนับสนุนการวิจ (สกว.)

(ความเห็นในรายงานนี้เปนของผูวิจัย สกว.ไมจําเปนตองเห็นดวยเสมอไป)

EXECUTIVE SUMMARY

Of all the infrastructural investments a country or group of nations can make to serve as the platform for rapid socioeconomic development, expansion of the road network is the first and most essential. Without roads, the economic costs and construction time skyrocket for such other infrastructural projects as hydroelectric dams; electrification; sewer/drinking water/sanitation systems; importation of materiel for hotel, school and office building construction; and the provision of emergency health care networks and vehicles.

The East-West Economic Corridor system, connecting Vietnam to Myanmar and veering south through Hat Yai to Malaysia, is one of the most ambitious and carefully planned highway projects ever to be undertaken in Asia. As envisaged by the Asian Development Bank and the member nations of Southeast Asia, its impacts were to be strong economic growth, increased income per capita, inclusion of the poor in more egalitarian income distribution, the reduction of poverty, inclusive creation of jobs at all skill levels, the attraction of tourists, the facilitation of expanded trade, the specialization of production along lines of revealed comparative advantage, and the bridging of peoples and nations in general. But, as this report demonstrates, the real impacts of roads are only as strong as the weakest link in the underdevelopment profile of the city or province through which they run. This is the main message of this report. Its main recommendation is, therefore, that each individual province, with the help of its national government, must adopt concomitant policies to shore up those weak links so that the fullest advantage can be taken of those roads.

coefficients; the Xxxxxx-Xxxxx-Xxxxxxxxx indices of the incidence, depth, and intensity of poverty; and the proportion of value added deriving from business and trade. Tests of means were employed to detect significant differences in each study province between rural, semi-urban, and urban subpopulations.

The gaps in progress (b above) are explained by multiple regression equations, including seemingly unrelated regression and volatility models. The dependent variables are income, the levels of consumption of groups of key necessities, the intensity of poverty, the number of tourist arrivals, and the fluctuations in the prices of key export commodities like rubber.

Optimal policy recommendations (c above) were generated through the use of Social Accounting Matrices. Not only were the goods and services with the greatest value added identified for each study site; the matrices were optimized under realistic constraints on the availability of land, labour and capital to assure that the optimal plans so generated would be feasible.

Given the depth and scope of these interlocking methodologies, only four (4) study sites could be studied in detail. Selection criteria therefore had to be very rigorous. All sites had to lie along the EWEC, be economically or strategically important in their own right, but also represent a certain point along the spectrum of pre-EWEC development or under development. These selection criteria have greatly enhanced the external validity of the present research, and allowed for clear comparisons of the impacts of the road under a wide range of conditions. The final selection of the four study areas, from east to west, is as follows:

1. Da Nang, Vietnam, a well-developed port town that was already connected on a major North-South highway corridor with the rest of the country.

2. Savannakhet, Laos, a country-wide but landlocked city with untapped potential for tourist and trade going both east to Vietnam and west to Thailand, and a fledgling but emerging business sector.

3. Hat Yai, in Songkhla province, southern Thailand an already specialized border town with ideal conditions for rubber production and export, and the

geographical position to serve as the gateway for tourist arrivals from and through Malaysia. And

4. Mawlamyine, Myanmar, the fourth largest city, lying on the EWEC on a direct line between the capital Yangon and a proposed seaport. BUT, as this report will demonstrate, each of those favourable development conditions was reversed: the capital was moved inland, the seaport project was transferred to a different site, and ethnic warfare and rebellion prevented the EWEC from being completed.

Data for all sites was collected in the years 2010 and 2011 using standardized household and firm questionnaires and the integration of all available secondary and governmental data. Because of the unique characteristics of each site, the specific research objectives and testable hypotheses differed accordingly.

The analyses reported in this document demonstrate that the most observable impacts of the EWEC lies in sites typified by 2 above (Savannakhet, Laos). Basically, all of the non-transportation conditions were in place before the advent of the EWEC. Businesses were concentrated there, but still under-developed. Thus, the road became the blood vein for business, marketing and transportation across Laos, Thailand and Vietnam. The impact of the EWEC was all the greater because that section of the highway cut strategically through a mountainous border area between Laos and Vietnam. Key segments of the Savannakhet population that had been lagging behind in income, employment, or both because of relative isolation from jobs and markets were thus able to “catch up” with the rest of the population. Poverty has been reduced, the inequality in income distribution has been partially redressed, and the city is enjoying a surge of growth because the major development condition missing had been the road.

In the same vein, this report shows that, even when the road system is as yet incomplete, it can spark growth and job creation in the most backward provinces of type 4 (Mawlamyine). That this can happen before the road is fully operation demonstrates that direct job creation from the and indirect stimulation of economic activity can bring about positive impacts in terms of the eradication of poverty. It is well known that, since the data for the present report were collected, there has been a

massive political opening of Myanmar as a country. There is thus every hope that a fully operational EWEC can generate still greater employment and poverty-reduction impacts. Myanmar has woken up, has addressed her political issues, and is on her way to economic transformation. The recommendations in the last chapter of this report may therefore be of even greater interest to her leadership. In the same mold as Savannakhet, the EWEC may soon contribute massively to economic development of Myanmar on a much larger scale.

At the other end of the spectrum, two of the four study sites have not been able – either in fact or in potential -- to benefit as much from the EWEC as sites of types 2 and 4. Those are Hat Yai and Da Nang. For example, in sites of type 3 (exemplified by Hat Yai, Thailand), development was already outward-looking before the inauguration of the road project. Rubber could be produced with economies of size for export, while tourists could be “imported” from the south without creating a pan-Southeast Asian highway system. The main results noted in this report have been the enhancement of local business, and the attenuation of fluctuations in rubber prices.

Likewise, in provinces of type 4 (typified by Da Nang, Vietnam), conditions predating connection with the East-West Highway had already been enormously favorable. A seaport already lay available at the eastern edge of the province for both imports of raw materials and exports of agricultural and finished products. A major North-South road linked Da Nang to all other parts of the country, so that Da Nang was already faring as well as most other cities in Vietnam in terms of living standards, education, employment, and income distribution. Small east-west roads were adequate to drain the agricultural surplus and prospective workers from the interior. True, a wider, faster East-West connection to Laos and Thailand enhanced those effects, but it does not seem to have been essential to rapid economic development.

The policy conclusions from this report are therefore quite site-specific; and it would be both dangerous and wrong to generalize on what measures governments should put into place for the EWEC as a whole. Careful reading of this report will point to the remaining weaknesses in Mawlamyine and Savannakhet that could make the new road connections even more economically and socially productive. In contrast, large, well-linked and relatively developed urban centers like Da Nang

and Hat Yai would be better served by re-channeling their infrastructure budgets into the construction and upgrading of feeder roads, the reduction of urban traffic congestion, and the provision of public transportation.

บทคัดยอ

รหัสโครงการ: RDG5010021

ชื่อโครงการ: TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST-WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM

xxxxxxxxxxxx: xxxxxxxxx xxxxxxxxxxx และ ปเตอร คาลกินส

1คณะเศรษฐศาสตร มหาวิทยาลัยเชียงใหม E-mail Address: xxxxxxx.xxxx@xxxxx.xxx ระยะเวลาโครงการ: 1 มิถุนายน 2550 – 31 กรกฎาคม 2559

โครงการนี้มีวัตถุประสงคเพื่อxxxxxxxกระทบทางเศรษฐกิจและสังคมของเสนทางระเบียง เศรษฐกิจตะวันออก-ตะวันตก ดวยตารางเมตริกซบัญชีสังคม โดยมีพ้ืxxxxศึกษา 4 เมืองที่เปนตัวแทนของ แตละประเทศดังนี้ 1) เมืองดานัง ประเทศเวียดนาม 2) เมืองสะหวันเขต ประเทศลาว 3) เมืองเมาะละแหมง ประเทศเมียนมาร และ 4) เมืองหาดใหญ ประเทศไทย ท้งนี้เพื่อใหเกิดประโยชนตอภาครัฐของแตละ ประเทศที่เสนทางระเบียงเศรษฐกิจตะวันออก-ตะวันตกพาดผาน ในการใหขอเสนอดานนโยบาย การแขงขัน และอ่ืนๆ เชนการขจัดความยากจน การสรางxxx xxxทองเที่ยว และการพัฒนาอยางยั่งยืน ตลอดจนการรวมกลุมในxxxxxxxและประสิทธิภาพทางการคา

ผลจากการศึกษาพบวาผลกระทบของเสนทางตามระเบียงเศรษฐกิจตะวันออก-ตะวันตกเกิดขึ้น มากท่ีสุดที่เมืองสะหวันนะเขต ประเทศลาว ซึ่งแตเดิมเมืองสะหวันนะเขตแมวาจะนับเปนศูนยกลางธุรกิจ ในพื้นที่แถบนั้น แตยังxxมีการคมนาคมที่ยากลําบากและยังไมพัฒนาทั้งดานเศรษฐกิจและสังคมเทาที่ควร เสนทางของระเบียงเศรษฐกิจตะวันออก-ตะวันตกน้ี เปรียบเสมือนเสนเลือดใหญที่หลxxxxxxxทั้งธุรกิจ การคา การตลาด การคมนาคมขนสง ทั่วทั้งลาว ไทยและเวียดนาม การสรางเสนทางเชื่อมโยงระเบียง เศรษฐกิจตะวันออก-ตะวันตกน้ี สงผลกระทบในทางxxxxxเปนวงกวางอันเนื่องมาจากถนนตัดผานพรมแดน ระหวางลาวและเวียดนามซึ่งมีสภาพxxxxxxxxxxเปนภูเขา xxxxxxชวยลดความยากจนลงได

ในสวนของเมียนมารซึ่งการกอสรางยังไมแลวเสร็จนั้น หากเสนทางสายนี้กอสรางเสร็จสมบูรณ ยอมเปนสวนสําคัญตอการพัฒนาทางดานเศรษฐกิจในพื้นที่เมาะละแหมง อยางมาก

อยางไรก็ตาม การศึกษานี้พบวาอีกสองพื้นที่ คือ หาดใหญและดานังนั้น ไมไดรับประโยชนจาก โครงการระเบียงเศรษฐกิจตะวันออก-ตะวันตก มากเทาที่ควรเมื่อเทียบกับสะหวันนะเขตxxx xxxxละแหมง เนื่องจากท้งxxxxxxxxมีความเจริญมากอนหนาที่จะมีโครงการแลว การเชื่อมโยงระหวาง ลาวและไทยนั้น นับวามีความจําเปนแตก็ยังไมเรงดวนมากในแงของการพัฒนาเศรษฐกิจ

ขอเสนอแนะเชิงนโยบายจากการศึกษานี้คอนขางจะเฉพาะในแตละพื้นที่ ไมเหมาะท่ีจะนํามาใช เปนของเสนอแนะโดยรวมกับทุกพื้นที่ เนื่องจากบริบทของแตละพื้นที่มีความแตกตางกัน ทั้งนี้อาจจะแยก ไดเปนสองกลุม คือ สะหวันนะเขตxxxxxxxละแหมง ซึ่งxxxxxxxxxxxxยังไมพัฒนามากนัก การเปดเสนทาง คมนาคมใหมเชื่อมโยงระเบียงเศรษฐกิจตะวันออกตะวันตกจะทําใหxxxxxxxxxxxท้ังดานเศรษฐกิจและสังคม แตสําหรับดานังและหาดใหญ ซึ่งเมืองมีการพัฒนาไปมากแลว ไมมีความจําเปxxxxตองสรางถนนใหม แต ควรปรับปรุงถนนเสนเดิมใหมีความสะดวกสบายxxxxxxxขึ้นเพื่อลดปญหาการจราจรxxxxxและสนับสนุน การขนสงสาธารณะใหมีมากขึ้นดวย

คําหลัก: บัญชีเมตริกซสังคม ระเบียงเศรษฐกิจตะวันออก-ตะวันตก สะหวันนะเขต ดานัง หาดใหญ เมาะละแหมง

Abstract

Project Code: RDG5010021

Project Title: TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST-WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM

Investigators: Songsak Sriboonchitta1 and Xxxxx Xxxxxxx1

1Faculty of Economics, Chiang Mai University

E-mail Address: xxxxxxx.xxxx@xxxxx.xxx

Project Duration: 1 June 2007 – 30 July 2016

The purpose of the project is to build, evaluate and link a series of four sub- national Social Accounting Matrices (SAMs) chosen to follow the “East-West Economic Corridor” linking Myanmar, Thailand , Laos and Vietnam. The research sites of this project lies along this economic corridor within the Greater Maekong Sub-region and were selected to provide well-positioned provincial leaders with the information they need to take maximum yet sustainable advantage from the ongoing globalization and integration of the region which were Da Nang, Vietnam; Savannakhet, Laos; and Mawlamyine, Myanmar. In response to calls from national governments for greater competition policy, this study emphasized not only poverty alleviation, employment creation, tourism, and socially and environmentally sustainable development within each site; but also competition policy, regional integration, and trade efficiency along each international corridor.

The results of the study show the most observable impacts of the EWEC lies in sites typified by Savannakhet, Laos. Basically, all of the non-transportation conditions were in place before the advent of the EWEC. Businesses were concentrated there, but still under-underdeveloped. Thus, the road became the blood vein for business,

marketing and transportation across Laos, Thailand and Vietnam. The impact of EWEC was all the greater because that section of highway cut through a mountainous border area between Laos and Vietnam. The results found that poverty has been reduced, the inequality in income distribution has been partially redressed, and the road has promoted the growth of the city. In case of Mawlamyine, even the road system is yet incomplete; it can spark growth and job creation in province. It brings about positive impacts in terms of the eradication of the poverty. After the roads are fully constructed, the EWEC may soon contribute massively to economic development of Myanmar on the much larger scale. However, two of four study sites –Hat Yai, Thailand and Da Nang have not been able to benefit as much from the EWEC as sites of Mawlamyine and Savannakhet. Their developments were already outward-looking before the inauguration of the highway project. The main results of Hat Yai in this study have been the enhancement of local business, and the attenuation of fluctuations in rubber price. Likewise, in Da Nang, small east-west roads were adequate to drain the agricultural surplus and prospective workers from the interior. It is true that a wider and faster East- West connection to Laos and Thailand enhanced those effects but it does not seem to be essential to rapid economic development.

The policy conclusion from this study are quite site-specific; and it would be both dangerous and wrong to generalize on what measures governments should put into place for EWEC as a whole. The remaining weaknesses in Mawlamyine, Myanmar and Savannakhet, Laos that could make the new road connections even more economically and socially productive. In contrast, Da Nang, Vietnam and Hat Yai, Thailand would be better served by rechanneling their infrastructure budgets into the construction and upgrading of feeder roads, the reduction of urban traffic congestion, and the provision of public transportation.

Keyword: East-West Economic Corridor, Mawlamyine, Savannakhet, Hat Yai, Da Nang, Social Accounting Matrices (XXX)

CONTENTS

Page

EXCUSIVE SUMMARY	I
ABSTRACT IN THAI	II
ABSTRACT IN ENGLISH	III
LIST OF TABLES	IV
LIST OF FIGURE	V
CHAPTER I INTRODUCTION
1. Real world problem: inadequate links among nations and low living standards	1
2. Overview of the Thailand Research Fund project	2
2.1 Trade and Investment	5
2.2 Social and environment	6
2.3 Migration	7
2.4 Transport infrastructure development	9
2.5 Road Infrastructure and Socioeconomic Development	10
2.6 Tourism, Economic Development and Poverty Alleviation	12
3. Scientific problem : unmeasured impacts to date of the East West Economic Corridor	14
4. Goals and objectives of this study	16
5. Conceptual framework	18
6. Testable research hypotheses	22

CONTENTS (CONTS.)
	Page
CHAPTER II EWEC INITIATIVE AND BACKGROUND THE STUDY SITES
1. Initiative of the EWEC	25
2. Background of the EWEC in Da Nang	27
3. Background of the EWEC in Savannakhet	29
4. Background of the EWEC in Hat Yai	32
5. Background of the EWEC in Mawlamyine	34
CHAPTER III DATA COLLECTION AND METHODS OF DATA ANALYSES
I. Data collection	37
1. Data collection in Da Nang	38
2. Data collection and sampling in Savannakhet	39
3. Data collection and sampling in Hat Yai	40
4. Data collection and sampling in Mawlamyine	42
II. Methods of data analysis	43
1. Tests of means	43
2. Poverty analyses	44
2.1 The incidence of poverty: Head-Count Index – HCI	44
2.2 The depth of poverty or the Poverty Gap – PG1	44
2.3 The severity of poverty or the Squared Poverty Gap – PG2 2.4 The Lorenz curve and Gini Coefficient	44 45

CONTENTS (CONTS.)
	Page
3. Income and poverty regressions	46
4. Seemingly unrelated consumption expenditure regressions	48
5. Volatility models	49
6. Social accounting matrices (SAMs)	51
7. Sufficiency Economy Matrices (SEMs)	56
CHAPTER IV: ANALYSES OF POVERTY AND WELL-BEING
I. DA NANG	58
1. Income inequality and Gini coefficients	58
2. Econometric equations to explain poverty and income	60
II. SAVANNAKHET	63
1. Income inequality and Gini coefficients	63
2. Incidence, depth, and intensity of poverty	65
3. Econometric equations to explain poverty and income	66
III. HAT YAI	68
1. Income inequality and Gini coefficients	68
IV. MAWLAMYINE	69
1. Income inequality and Gini coefficients	69
2. Incidence, depth, and intensity of poverty	71
3. Econometric equations to explain poverty and income	72

CONTENTS (CONTS.) CHAPTER V THE CONSTRUCTION OF SOCIAL ACCOUNTING MATRICES (SAMs)	Page
1. Establishment of the matrix	76
2. Scenario 1	82
3. Scenario 2	84
4. Scenario 3	86
5. Scenario 4	88
6. Scenario 5	90
7. Scenario 6	91
8. Scenario 7	93
CHAPTER VI DISCUSSIONS OF THE RESULTS	95
1. Results of hypothesis testing in Da Nang	96
2. Results of hypothesis testing in Savannakhet	106
3. Results of hypothesis testing in Hat Yai	113
4. Results of hypothesis testing in Mawlamyine	118
CHAPTER VII: POLICY IMPLICATIONS
1. Policy implications for Da Nang	123
2. Policy implications for Savannakhet	124
3. Policy implications for Hat Yai	126
4. Policy implications for Mawlamyine	128
5. Scope and limitations of the research and suggestions for future research	129

LIST OF TABLES

TABLES	Pages
CHAPTER I 1.1 Basic Economic and Demographic Indicators for the Greater Mekong Sub region CHAPTER III	8
3.1 Household sample size in Savannakhet	40
3.2 Number of sample allocated, Tambon and status of the field survey	41
3.3 A Basic Social Matrix (SAM) CHAPTER IV	54
4.1 Trends in the inequality of income distribution in the Da Nang VHLSS survey samples	58
4.2 Comparative incidence, depth and intensity of poverty between the urban, semi-urban and rural samples	59
4.3 Regression analysis to explain income/capita by transportation, 2008	60
4.4 Regression analysis to explain the intensity of poverty	61
4.5 Regression to explain the dynamic change income between 2006 and 2008	62
4.6 Poverty and income distribution in Savannakhet	63
4.7 Gini Indices	65

LIST OF TABLES (CONTS.)
TABLES	Pages
CHAPTER IV 4.8 The ordinary least square regression results for household expenditures	66
4.9 Wald coefficient test result.	67
4.10 Descriptive statistics of the household level income, Songkhla (THB)	68
4.13 Poverty under per capita income and expenditure	71
4.14 Factors that explain income per capita	72
4.15 Multiple regression equation to explain the intensity of poverty per capita	73
4.16 The effect of poverty on other basic expenditures CHAPTER V	75
5.1 Summary table of the structure and impacts of the optimal solution under the 7 scenarios	78
5.2 Savannakhet SAM structure	81
5.3 The “Benchmark Scenario”	83
5.4 The “Neoclassical” labor-constraint model	85
5.5 The “Human Capital” labor-with-education constraint model	87

LIST OF TABLES (CONTS.)
TABLES	Pages
CHAPTER V 5.6 The “Growth with Equity” human capital plus income-to-poor constraint model	89
5.7 The “Sufficiency Economy” model with household spending transfers from “bads” to education	92
5.8 The “Sufficiency Economy” model with household spending transfers from “bads” to education but a budget limitation on social advertising	93
CHAPTER VI 6.1 Student t tests of significant differences in means of key variables between the 2006 and 2008 VHLSS surveys	98
6.2 One-way ANOVA analysis of significant differences by proximity to EWEC, 2008	99
6.3 Changes in significant differences between 2006 and 2008	101
6.4 Correlation matrix of significant correlations between Transportation-related, well-being and production variables	102
6.5 T-tests to compare the relative income of the three areas	106
6.6 Significant correlations of potential explanatory variables with key transportation-related variables	108

LIST OF TABLES (CONTS.)

TABLES Pages

CHAPTER VI

6.7 Quantile regression analysis to explain income per capita 112

6.8 T-tests to compare three measurements of poverty 113

6.9 Unit root test statistics for daily price of various categories of Rubber products

114

6.10 Estimated models and parameters and their respective standard errors

117

6.11 Factors that explain income per capita 118

6.12 Multiple regression equation to explain the intensity of poverty per capita

119

6.13 Some factors associated with high poverty intensity 120

6.14 Some factors associated with reduced poverty 122

LIST OF FIGURES
FIGURES		Page
CHAPTER I Figure 1.1	Map of Greater Mekong sub-region East-West transport corridor project	2
Figure 1.2	Conceptual pathways of impact of the EWEC on income, employment poverty, and consumption	20
Figure 1.3 Conceptual diagram of the level, causes, effects and policies to alleviate poverty		21
CHAPTER II Figure 2.1 East-West Economic Corridor in Da Nang		28
Figure 2.2 Map of Savannakhet		31
Figure 2.3	Map of Hat Yai	33
Figure 2.4: Map of Mawlamyine with EWEC		35
CHAPTER IV
Figure 4.1 The Lorenz curves of Savannakhet and its sub regions		64
Figure 4.2	Consumption- and income-based Lorenz curves of Mawlamyine	70

CHAPTER I: INTRODUCTION

1. Real world problem: inadequate links among nations and low living standards

The major cross cutting challenges facing the Southeast Asian region in this decade are: (1) economic integration; (2) infrastructure development; (3) rural development and poverty reduction; (4) development of markets and a strong private sector; (5) policy and institutional reform; (6) human resource development; and (7) environmental protection. Of all the infrastructural investments a country or group of nations can make to serve as the platform for rapid socioeconomic development, expansion of the road network is the first and most essential. Without roads, the economic costs and construction time skyrocket for such other infrastructural projects as hydroelectric dams; electrification; sewer/drinking water/sanitation systems; importation of materiel for hotel, school and office building construction; and the provision of emergency health care networks and vehicles.

The research study on the impact of the EWEC road network in various aspects and dimensions of the socioeconomic development agenda is crucial to accelerating economic growth on the right track. It is also vital to formulating and the reforming national level policy as well as enhancing pan-regional cooperation among participating countries. Only then can their economies hope to achieve mutual benefits by formulating friendly trade and economic policy based on the lessons learned.

Figure 1.1: Map of Greater Mekong sub-region East-West transport corridor project

2. Overview of the TRF project

The United Nations Economic and Social Commission for Asia and the Pacific (ESCAP) was established in 1947. It serves as the main social and economic development centre for the Asia and the Pacific. The mandate of ESCAP is to foster the cooperation between its member countries and associate members to overcome the challenges of socioeconomic development by launching results oriented projects; technical assistance; and capacity building focusing on macroeconomic policy, trade, transport, social and environmental issues, and information technology. ESCAP provides the strategic links between global and country level programs, and it supports government of the member countries to overcome the socioeconomic challenges for the sustainable economic development in global context.

In this connection, the Asian Highway network is a regional transport cooperation initiatives of ESCAP aimed at enhancing the efficiency and development of the road infrastructure in Asia, supporting the development of Euro-Asia transport linkages and improving connectivity for landlocked countries. The concept of Asian Highway project was administered in 1959 with the aim of promoting the development of international road transport in the Asia pacific region for social progress and better standards of life.

The Greater Mekong Subregion (GMS) development project was established in the year 1992 by the Asian Development Bank (hereafter, ADB). The purpose of the GMS project is to spur the development and economic integration of the sub-regional Mekong economies: Thailand, Laos, Vietnam, Myanmar, and Cambodia. By connecting and linking the region's countries, it should be possible to select new resources and pursue specialization for trade with the highest economic efficiency. The Economic Corridors project is one of the efforts to achieve these purposes.

The concept East-West Economic Corridor was introduced in October, 1998 in Manila at the Eighth Ministerial Conference of the Greater Mekong River Sub-region realizing the need of the regional development considering regional competitive and comparative advantages of economic development and economic cooperation. Since then, there has been an official or unofficial economic transaction prevailing between the neighboring countries. The eight ministerial conferences agreed to develop and promote free Trade Area (FTA), Tax Alliance and Common Market (EU) in the Mekong sub regions and enhance the economic globalization maintaining national investment for international markets as well as attracting foreign direct investment in the participating countries. The East–West Corridor (EWC) Project is part of the wider East–West Economic Corridor linking Da-Nang in Vietnam and Mawlamyine in Myanmar—covering Lao People’s Democratic Republic (Laos PDR), Myanmar, Thailand, and Vietnam. As a flagship project of the Greater Mekong Sub-region (GMS) Program, it was designed to improve National Road 9 linking landlocked areas in northeast Thailand to the Vietnam coast via Laos PDR. It is the second cross-border road project in the program and is in line with ADB's thrust for regional cooperation in the transport sector (ADB, 2008).

The Intergovernmental Agreement on the Asian Highway Network was adopted on 18 November 2003 by an intergovernmental meeting held in Bangkok, was open for signature in April 2004 in Shanghai, and entered into force on 4 July 20051.

The economic corridor has three distinctive features. Firstly, it is considered a defined geographical region. Secondly, the corridor mainly focuses on bilateral initiatives, not multi-lateral ones. Thirdly, the economic corridor requires detailed physical and space planning for infrastructure development for the most effective results. About, 1,450 km long East-West Economic Corridor passes through four countries and the East-West Economic Corridor also connects with main north-south roads, including Yangon – Dawei and Chiang Mai – Bangkok of Thailand, Road 13 of Laos, and National Highway No.1A of Vietnam.

Moreover, the summary findings of “evaluation study on the east-west project report” of ADB (2008) mentioned that the project is ““highly relevant” to development needs at the regional and national levels. As a regional road, it addressed a strategic need to strengthen links between Laos PDR, Thailand, and Vietnam. The all-weather road aimed to expand the market for transit and bilateral trade. It also interconnects national transport networks to generate trade and efficiency

1 xxxxx://xxxxxxxx.xx.xxx/xxxxx/XxxxXxxxxxx.xxxx?xxxxXXXXXX&xxxxx_xxxXX-X-00&xxxxxxxx00&xxxxxxx

benefits. It was appropriately designed to support economic centers and to complement poverty reduction (via a rural road component in Laos PDR)”(ADB, 2008. p iii) and ranked as a successful project. Moreover, the report added that there is little evidence of EWC's transformation from a transport corridor into an economic corridor and the full realization of corridor-level benefits (i.e., economic activities, tourism, and trade facilitation) is limited by institutional constraints. In the case of Laos PDR, this is partly explained by inadequate complementary investments and slow private sector growth.

The lessons gleaned from the project mentioned by ADB Evaluation Study on the East-West Corridor Project Completed are: (i) The pace of the economic development along the corridors depends on the complementary interventions requiring policy and institutional changes that enable better opportunities for the private sector. (ii) Multimodal planning is important for enhancing the effectiveness of transport corridors, and (iii) ADB should use loan savings prudently by exercising due diligence at the approval and at completion stages of the project. Most of the road infrastructural development projects further alleged that these projects foster the economic development process by a) enhancing marketing linkages among the parties at local and regional levels to b) ease the mobility of products and factors based on market forces and c) generating employment opportunities through d) increased market accessibility for the rural people.

AbuzarAsra et al. (2006) introduced an approach to analyze the impact of the east-west corridor along Savannakhet route no. 9 to its neighboring countries using the accounting framework of the traditional single area I-O model. The findings of the study show the significance of the economic measurement tackles broad issues on export-import expansion, interdependence of industrial structure from Rest of Savannakhet (intra and interregional analysis), evaluating impacts or changes in final demand (production, income, and employment), and short term projections and forecast of the domestic economy (AbuzarAsra, 2006).

BounpongKeorodom et al. (2007) showed the thought of business persons that are positive toward their business investments after the completion of The Second Mekong Bridge; 20 % would increase imported goods in construction materials and machinery equipment, and 15 % would increase the export of domestic products. In addition, 17 % of the respondents saw the development of EWEC as cost and time saving in transportation and 20% would see strong competition among domestic firms as well as from foreign investment.

Benson Sim at el. (2007) constructed a modified interregional input–output (IRIO) table to link the economies of the province of Mukdahan in Thailand and the Lao PDR of Savannakhet using the Chenery-Moses model. The results showed that the value of trade of these provinces with the rest of the world was much higher than the trade between them. Industries in the services sector were found to have generally higher value-added multipliers than industries in the manufacturing sector. The agriculture and forestry industry in Savannakhet and the manufacturing industries in both provinces had high backward and forward linkages. Exports to the rest of the world and consumption were found

to have the highest employment multipliers in Mukdahan and Savannakhet, respectively. Mukdahan was also found to have higher net foreign exchange earnings, implying that the Thai province may be able to add greater value to its exports than the Lao PDR province (Benson Sim, 2007).

2.1 Trade and Investment

The aim of creating the EWEC corridor project was to help develop trade toward the north and the south to the major commercial centers of Bangkok and Ho Chi Minh and to bring practical and long-term benefits to member countries. The corridor will also aid cities and small towns along it to strengthen trans-border trade and investment attraction, to develop new economic activities by effectively using economic space, and to establish trans-national economic areas and products for entry onto global markets in Europe and America. However, most member localities are underdeveloped, highly populated and geographically remote. Their agriculture-based economy has not played an important role, and industry is not yet strongly developed.

The major driving forces for EWEC economic growth were assumed to be the differences in factor endowments among different national areas of the corridor. Such differences should allow firms in each area to exploit their comparative advantage in the production of those goods and services that are intensive in the factors with which the areas are abundantly endowed. At the sub-regional level, the driving forces identified for the EWEC were: (i) the efficiency and welfare gains from trade investment and liberalization; (ii) the firm-level economies of scale and resulting increases in competition; (iii) induced changes in foreign investment from both firms in the sub-region that are pursuing a sub- regional production strategy and foreign multinationals that are pursuing global production strategies;

(iv) productivity gains from trade resulting from the greater availability of capital, skilled and unskilled labor (ADB, 2008).

Integration into the world economy is the major challenge for economic development through promotion of trade and economic liberalization from the perspective of the export and import markets of the country. Lao PDR has gradually started to access world economy since 1989, with accession to ASEAN in 1997. The country has signed a number of agreements to promote trade. These include an investment promotion agreement with Japan, and an economic cooperation agreement with South Korea to promote cooperation in the areas of trade, investment and services. The objective of the country’s the Sixth Five-year Plan (2006-2010)2 is to fully implement the integration roadmap and meet the commitments to become a member of the WTO and multi-lateral regional bodies such as AFTA and GMS. The evaluation review report of 6th Plan (2006-2010) of the Lao PDR claimed that during the five years (2006-2010), the export value of Lao PDR’s goods was US$ 5.69 billion and shared 23.4% of GDP, which has exhibited an increasing trend every year. Lao PDR's key export markets include the ASEAN countries (Thailand, Vietnam, and Malaysia), the EU, China, and Australia. The main export markets were Thailand, accounting for 59.60% of the total exports (or

2The Sixth National Socio Economic Development Plan (2006-2010).

equivalent to US$650.78 million); Vietnam 13.37%; Australia 6.19%; and China 1.85%. Recently, Lao exports to the US have increased but the overall value is low (just 1% of total exports). Similarly, imports to Lao PDR during the same five years (2006-2010) had a value of US$6.61 billion, which covered 27.3% of GDP and showed an increasing trend. Imports into the country were also dominated by the ASEAN countries (about 76.26% from Thailand and 12.25% from Vietnam), China 8.3%, Japan 2.6%, South Korea 1.88% and Malaysia 0.6%. In the EU market, Germany covered 1.04%, France 0.7%, and other countries. Mining and agriculture and agro-processing remain important not only for meeting domestic demand but also for exports to a rapidly growing regional international market. Manufacturing and tourism firms continue to face significant constraints to investment, productivity, and growth, and these will have to be addressed (Committee for planning and investments, 2006).

Investments from the ASEAN countries in Lao PDR have increased each year since 2004. Thailand, Vietnam and China are the top three ranking countries from ASIAN region, respectively. Hydropower, mineral and tourism sectors are the most attractive sectors for foreign direct investment. Foreign direct investment (FDI) flows predominantly into the natural resource sector (particularly mining and hydropower), accounting for approximately 80% of all FDI in 2008. Investment is likely to remain a key element of Lao PDR’s growth strategies and to continue its focus on exploitation of the country’s natural resources over the coming years. The Government of Laos has continued to work on WTO accession (Committee for planning and investments, 2006).

2.2 Social and environment

The intention of such broad-based development strategies was to remove existing constraints to the promotion of trade and investment and to sustain economic growth in the corridor to help alleviate poverty (xxxx://xxx.xxxxxx.xxx.xx). The proposed project initiatives were specifically expected to improve the economic role and social status of women in the sub-region, as their burden of supporting the household and performing subsistence agricultural work should shift to economically productive and remunerative activities. Greater education associated with improved skills training must parallel the shift to higher value-added industries thus helping to improve the economic welfare of women in the Corridor. This poverty-reduction objective of the EWEC in the rural and emerging urban areas is in turn linked with alleviation. Improved regulation of trade and institutionalization of informal trade are expected to stem environmental degradation and equalize the forest-surplus and forest-deficit imbalances between countries. Countries should enhance this environmental potential by introducing policies that emphasize training, private sector participation, and data gathering in all agro-industry and manufacturing sectors.

Although authorities claim that there has been notable improvement in the overall indicators for gender equality and empowerment of women, a remaining challenge lies in identifying and addressing the causes of the great disparities that exist between urban and rural areas and between different regions.

Many studies from South Korea, Singapore, Malaysia and Taiwan such as Xxx (1994), Xxx (2009), SALT CENTRE, Hwa Chong Institution (2007), Xxxx (2008), Sim (2009), and XX and SU (2006) have concluded that economic development issues in any developing country should be focused on gender and equality and women empowerment realizing the role of women in economic development.

The lessons learned from those studies clearly depict the vital role of women’s education and involvement in economic development. It is crucial that the empowerment of women and contribution to public life be encouraged at all levels of the development agenda. The local or village level agenda and activities can accelerate higher progress, including regional and national efforts.

Currently, however, as in several other developing countries, Lao’s rural societies are trapped in stereotypical attitudes and beliefs. Such thinking continues to affect girls in the family, school and society. Most females experience an unfair distribution of labor inside the home, and men are usually considered the head of households and the decision-makers at the domestic and village level. Consequently, women and girls are restricted in their ability to access education, gaining less exposure and contributing less to village and community development despite playing a significant social and economic role (Committee for planning and investments, 2006).

2.3 Migration

The motive of migration is “the opportunity for people to obtain jobs and a better livelihood” (ADB, 2009). Migration of workers is an important factor in economic development in the GMS. Social and economic growth in the GMS over the last 10 years has been consistently strong, particularly the improvement of transport infrastructure which contributed the movement of goods as well as people in this sub-region. GMS members have witnessed significant migration of labor between Myanmar and China, Lao PDR and Vietnam, Vietnam and Cambodia, and China and Vietnam. “These flows are further facilitated by such factors as these countries’ proximity to one another and their long porous borders, cultural similarities, population diasporas, and construction of highways” (ADB, 2009). These factors can produce the changes in trade and connectivity within the sub-region. Although such cross-border mobility of labor is increasing with the development of new employment opportunities, social and economic disparities remain (ADB, 2009).

Table 1.1 Basic Economic and Demographic Indicators for the Greater Mekong Subregion

Primarily Sending Countries

Sending and Receiving Countries

Indicators

Lao

PDR

Myanmar

Vietnam

Cambodia

Lao PRC

Thailand

Population(million)

6.834

48.137

86.967

14.494

1,388.612

65.905

Population growth rates (%)

2.316

0.783

0.977

1.765

0.655

0.615

Growth rate of population aged 15–39 (%), 2005–2010

2.97

0.60

1.42

2.93

(0.95)

(0.61)

Total fertility rate

(children born/woman)

4.41

1.89

1.83

3.04

1.79

1.65

Net migration rate (‘000)

(3.4)

(2005)

(0.2)

(2005)

(0.4)

(2009)

0.1

(2005)

0.4

(2009)

0.6

(2005)

GDP growth (%)

7.5

0.9

6.2

6.8

9.8

3.6

GDP/ capita ($ at PPP)

2,100

1,200

2,800

2,000

6,000

8,500

Note: () = negative. The indicators were estimated by United Nations.

Prospects: The 2006 Revision (United Nations publication, Sales No. E.07.XIII.2).

Sources: CIA Fact Book and United Nations Department of Economic and Social Affairs. World Population. Refer to Asian Development Bank (2009).

Table 1.1 portrays the striking disparities between the primarily sending countries and the sending and receiving countries in the sub-region. The high range of Gross Domestic Product per capita (GDPC) between Myanmar at $1,200 and Thailand at $8,500 attracts increased migration. Thailand also has the lowest total population growth rate, fertility rate, and growth rate in the 15–39 age groups. In contrast, the characteristics of the primary sending countries such as Lao PDR, Myanmar, and to a certain extent Vietnam are higher fertility rates, younger populations, a larger share of working age population, lower economic growth rates, and lower per capita GDP. These factors can create the opportunities of employment at the low end of the wage scale by migrant labor from other GMS members (ADB, 2009).

2.4 Transport infrastructure development

Infrastructures have long been recognized as playing a central role in economic development and the amelioration of living standards. The term “infrastructure” covers such multiple subsectors road transport, and rail links, energy, telecommunications, power supplies, water, sanitation, ports and airports and low-income housing (NESDB and The World Bank, 2008). Barrios (2008) has classified the development intervention into four categories as follows: “economic infrastructure e.g. credit, production support; physical infrastructure e.g. roads, irrigation; capacity building e.g. training, information dissemination; and support services e.g. marketing services, facilitation of access to basic social services”.

Xxx (2010) mentioned that any infrastructural improvement has impacts on both the economic and social sectors in general and the poor in particular. What the poor need is an opportunity for a better education, a minimal standard of healthcare services, power supplies, roads, clean water for drinking and sanitation, entry into markets, access to loans and investment capital, etc. Otherwise, closed access to schools leads to poorly educated people, lack of access to healthcare services and clean water leads to poor health, lack of access to world markets leads to heightened costs of production and transportation; time spent on non-productive activities reduces worker efficiency, etc. Thus, greater access to such infrastructure is potentially a powerful tool for improving both poverty reduction and economic development.

Jacoby (1998) estimated the household level benefits from road projects in Nepal. The results showed that poorer households get more benefit from road improvement in the remote areas. Consistent with this finding, Stone and Strutt (2009) used a computable general equilibrium to trace the benefits of the development of economic transport corridors. Their results showed the potential benefits of improving land transport and trade facilitation, particularly intra-regional trade within the Greater Mekong Subregion. Additionally, Xxxxx (2006) analyzed the different qualities of road access of rural villages in Lao PDR. The results show that roads help cut the incidence of poverty because poor people have opportunities to get more income from road improvement which leads to reduced costs for the goods they purchase.

Today, someone could argue the Asian Highway is probably the biggest highway network of the world allowing to experience at first hand the accounts that were relayed by great explorers because the Asian Highway network is a network of 141,000 kilometers of standardized roadways crisscrossing 32 Asian countries with linkages to Europe (UNESCAP, 2012). At the beginning phase of the project (1960-1970) considerable progress was achieved. However, progress slowed down when financial assistance was suspended in 1975 and ESCAP claimed that regional political and economic changes during the period of 1980s-1990s stimulated new momentum for the Asian Highway Project. In 1992, it became one of the three pillars of Asian Land Transport Infrastructure Development project, comprising Asian Highway, Trans-Asian Railway and facilitation of land transport projects (UNESCAP, 2013).

Undeniably, transport connectivity plays a central role in regional and sub regional integration. It is critical for enhancing economic cooperation, closing development gaps and enabling sustainable development. Transport networks facilitate the movement of people, goods, labor, raw materials, finished products and ideas easily and contribute to the social, economic and environmental betterment of the region (UNESCAP, 2013). Additionally, transport is an essential element in the development of Asia and the Pacific, and has played a critical role in the region’s rapid economic growth. Based on the 4th Meeting of the Working Group on the Asian Highway and Expert Group meeting on “Progress on Road Safety Improvement in Asia and the Pacific” Bangkok, Thailand held in 27 – 29 September 201,1 a total of 5499 KM has been completed so far.

To sum up, the Asian Highway network provides the critical road links between countries of the Asian and the Pacific region. While the Asian Highway network represents less than one percent of the total length of all roads in the ESCAP region, it is estimated that it connects close to one billion people or 50 % of the total urban population in the participating countries (UNESCAP, 2013). Similarly, the East-west corridor links Thailand to the Lao PDR at East and Cambodia at the west and North- South Corridor links Thailand to the Myanmar and Lao PDR at North and Malaysia at south. These two corridors are strategic Asian Highway for international trade and regional cooperation.

2.5 Road Infrastructure and Socioeconomic Development

The exploration is an important phenomenon from the beginning of the human existence and the great explorers embarked by land and sea in search of new worlds and riches. Perhaps, the purpose of travel could be explore new horizons, learn from different cultures, trade, or simply to secure food, shelter and means of subsistence for families and communities (UN, 2003). Viewed in the broadest sense, infrastructures are indispensable services to enable and accelerate socio-economic development. Where such infrastructure development projects have been implemented in a rational, well-coordinated and harmonized path, they have played a very significant positive role in the growth performance of a territory or nation. Even though development of such social and economic infrastructures is invariably associated with a huge amount of investment, most studies conclude that there is a positive relation between the level of infrastructure and the level of economic growth.

This said, the empirical findings on the precise impacts of transport infrastructure are inconclusive. Xxxxxx (2009) stresses how additional transport capacity and transport improvements influence the economies of societies with both underdeveloped and developed road networks have long been debated. Although road projects were not the sole cause of social and economic change in rural Philippines, they did enable trade and investment, increase competition, and release a previously underutilized production potential through access to a larger market area and more attractive purchase prices of production factors (Xxxxxx, 2009).

Xxxxxxxxx et al (2010) argued that “accessibility” is the main ‘product’ of a transport system and is closely related to mobility, economic development, social welfare and environmental impacts.

Therefore accessibility can be considered as a proxy for a raft of related (economic, social, environmental) effects of the transport infrastructure (Xxxxxxxxx et al., 2010). Still, a caveat is in order. Since increased mobility of capital and skilled labor has substantially altered the possibilities of achieving social and economic objectives, allocational incentives and policies designed to promote these objectives may generate unintended distributional shifts as a result of induced factor flows (McCulloch & Yellen, 1997).

Cost–benefit analysis (CBA) is a long-established tool for the evaluation of infrastructure projects. When cost changes in transport markets lead to changes in transport demand, the pertaining welfare effects are measured by changes in consumer surplus based on the demand for transport (Xxx, Ommeren, & Xxxxxxxx, 2009). The social benefits differ from the transport benefits to the extent that there exists monopoly power on product markets and/or agglomeration economies. The direct welfare effect of an infrastructure improvement is measured by the user benefits using consumer surplus, while the economy-wide welfare effect is measured by social welfare (Xxx, Xxxxxxx, & Xxxxxxxx, 2009).

In the developing economies, time savings induced by highway improvements can yield a variety of broader economic effects. These include (i) ensuring market access for producers to stimulate local production; (ii) drawing cheap labor, production inputs, and customers from a larger catchment area, thereby lowering wages and input/output prices; and (iii) altering land prices and development patterns (Gunasekera, 2008). The perception of the role of productive public expenditures as an engine of economic growth has changed markedly over the last few years. It was recently re- examined in the framework of endogenous growth theory when, following the empirical work of Xxxxxxxx (1989), new growth theory models began to take account of public spending as a factor for self-sustaining productivity gains and long-term growth (Barro, 1990; Xxxxxxxx, 2001). Demurger (2001) followed a growth equation approach and found that differences in geographical location, transport infrastructure, and telecommunication facilities account for a significant part of the observed variation in the growth performances of Chinese provinces. Indeed, the transport variable appears as one of the most regularly differentiating factors in growth gap decomposition (Xxxxxxxx, 2001).

Along with supportive economic and financial policies, infrastructure has long been recognized as a key element of the enabling environment for economic growth. Recently, the World Bank has also emphasized that by promoting growth, reliable and affordable infrastructure can reduce poverty and contribute to the achievement of the Millennium Development Goals by providing and supporting the delivery of key services (World Bank, 2004).

A vast range of previous studies have traced the significant impact of roads on poverty reduction through economic growth. For example, Kwon (2000) found that a growth elasticity with respect to poverty incidence of - 0.33 is good for roads and -0.09 is bad for roads at the provincial in Indonesia (Kwon, 2000).

Poverty reduction is a primary goal of development policy. However, a considerable segment of the population has to live on meager incomes with limited access to infrastructure services (Parker et al., 2008). “Poverty reduction requires economic growth which, when accompanied by sound macroeconomic management and good governance, results in sustainable and socially inclusive development” (ADB, 1999). In provincial China as well, Fan et al. (2002) showed the effects of different types of government expenditures on growth and rural poverty and argued that roads significantly reduce poverty incidence through increasing agricultural productivity and nonfarm employment. The estimated elasticity with respect to road density are 0.08 for agricultural GDP per worker, 0.10 for nonagricultural employment, and 0.15 for wages of non-agricultural workers in rural areas (Fan et al., 2002). The summaries how infrastructure investment contributes to the poverty reduction by influencing the productive sectors of the economy and creating employment for the poor after the policy interventions depicted in Figure 3 and 4.

2.6 Tourism, Economic Development and Poverty Alleviation

The United Nations World Tourism Organization (UNWTO) is a specialized international institution in the field of tourism. It offers a global forum for tourism policy issues and technical aspects of the tourism to promote the responsible and sustainable development of tourism paying attention towards the universally accessible tourism and the interest of the developing economies by playing a decisive role (UNWTO, 2009a). The UNWTO operates with its 154 member countries and more than 400 affiliated members representing local governments, tourism associations, and private sectors (UNWTO, 2010). The fundamental aim of the UNWTO shall be the promotion and development of tourism with a view to contributing to economic development, international understanding, peace, prosperity, and universal respect for, and observance of, human rights and fundamental freedoms for all without distinction as to race, sex, language or religion (UNWTO, 2009b). The concern of the UNWTO is to push and execute the global code of ethics for tourism development with the aim to ensure all the member countries, tourist destinations and businesses maximizes their positive benefits in terms of economic, social, cultural and environmental sustainability by minimizing negative social and environmental impacts (UNWTO, 2010).

Noticeably, tourism is the most incredible industry in the 20th century which witnessed a steady increase in tourism and over the last two decades of this century, the tourism industry has evolved and modernized considerably all over the world in terms of institutional development and methods of transport, and the facilities available at destination points have enjoyed an accelerated pace of improvement (Xxx, 2003; Xxxxx et al., 2009; Xxxxxxxx X. X., 1998)

The WTO with United Nations Conference on Trade and Development (UNCTAD) held in 2001 urged that tourism is the driving forces in the refocusing of economic development strategies in the poorest nations (WTO/UNCTAD, 2001) . The World Summit on Sustainable Tourism Development held in Johannesburg in 2002 advocated that tourism is one of the few development

opportunities for the poor and called for concrete action (WTO, 2002). “The livelihood strategies of rural households vary enormously, but, most households rely on a range of natural resource uses and on off-farm income from employment or remittances” (Xxxxxx, 2000a). A poverty reduction program requires strategies on a variety of complementary fronts and scales, but a prerequisite of significant progress is “pro-poor growth – growth which benefits the poor. Indeed, in most countries with high levels of poverty, tourism is a significant (contributing over 2% of gross domestic product or 5% of exports) or growing (aggregate growth of over 50% between 1990 and 1997) part of the economy” (Xxxxxx et al., 2001, p. 1).

Tourism is even more significant for the world’s poorest countries where it is a mainstay of their economies, a key factor in employment and trade, poverty reduction, as well as a vital lifeline for their development to achieve millennium development goals (MDG’s). To ensure a strong likelihood of economic, political, and social benefits at the local level, full participation of communities in tourism is indispensable (Ashley & Roe, 1998). Tourism can clearly contribute to combating poverty in the less and least developed economies provided that it is closely linked to the accessibility of financing sources, especially for small enterprises and micro-businesses (Ashley & Roe, 2002). The importance of travel and tourism as a social and economic activity has reached unprecedented proportions at the international level. Tourism is therefore used in many economies as a priority sector for economic development, redistribution of wealth, poverty reduction, and employment creation (WTO, 2010).

The GMS Economic Corridors constitute a set of well-defined geographical pathways targeting the transport network for economic development and infrastructural enhancement as the basis for more effective policy and institutional interventions in the GMS region. The economic corridors are integrated with the development agenda of the concerned countries on the basis of regional cooperation to promote economic development in the GMS sub region as a whole. The key objectives of the project are to: connect centers of economic activity; facilitate trade, investment, and tourism; reduce transport costs; and facilitate mobility of people, goods, and resources across borders.

Against these goals, the East-West Economic Corridor (EWEC) program, one of the 11 flagship projects of the GMS program was led by the Asian Development Bank (ADB). The road networks from East to West cross the four countries of Myanmar, Thailand, Lao PDR and Vietnam. Recently, the supporting facilities of road networks were completed, connecting Mawlamyine, Myanmar at the far West to Da Nang port, Vietnam at the Far East. Infrastructure was constructed to support the physical linkages within the EWEC such as the 2nd Lao-Thai friendship bridge between Mukdahan (Thailand)-Savannakhet (Lao PDR) and the Hai Van tunnel in Da Nang, Vietnam.

The total length of the EWEC is about 1,450 km by the longest distance of the road, of which 950 km are located in Thailand. The route begins on the East from the port town Da Nang, Vietnam and passes through Laweh and Lao Bao, cities in special economic areas of Vietnam close to the border of Laos PDR. EWEC goes through Savannakhet in Laos PDR and then crosses the Mekong River to Mukdahan, Thailand by 2nd Thai-Lao Friendship Bridge (Mukdahan-Savannakhet). The road continues

through Kalasin, Khonkhaen, Petchaboon, Pitsanulok and Maesot, Tak, Thailand border, through Myanmar up to the Moh-Tama gulf in Mawlamyine. Mawlamyine in turns provides a connection from the South China Sea to the Indian Ocean, India, and other destinations throughout the Middle East. The total surface of roadways in Lao PDR is more than 32,000 km, of which 23% is national highways, 21% state highways, and the rest are local.

3. Scientific problem : unmeasured impacts to date of the EWEC

Thanks to recent initiatives of national governments and the Asian Development Bank throughout the GMS region, a series of North-South, East-West highways is poised to criss cross continental Southeast Asia. Although China, and to a lesser extent Thailand, may have plans to develop specialization for commercial development, trade, and migration and along these new growth arteries, countries like Vietnam, Myanmar, Laos and Cambodia remain essentially unprepared. This situation will lead to a further gap in the levels of GDP per capita and annual average growth rates among GMS countries, risking to destabilize the region politically, economically and environmentally. It will even reduce the benefits going to the lead economies, China and Thailand, because of reduced benefits from trade. In addition, virtually no one is in a position to predict the impacts on the shared natural resources of the region, or on the optimal relocation of labor through migration.

This study responds to and helps to prevent the exacerbation of these problems by building detailed provincial-level SAMs in four widely-ranging provinces in three countries. The project will provide and transfer working models for the 5-year planning to the four provinces, that, if faithfully applied, could generate improved income and employment for millions.

This study seeks to promote income improvement, reduce poverty, provide skilled employment, and catalyze sustainable development through enhanced trade along the new East-West growth corridor. Because GMS countries vary widely in their resource bases, human capital development, planning potential and readiness for trade, this research project will build, apply, compare and integrate open-economy planning models at the provincial level in Mawlamyine, Myanmar; Hat Yai, Thailand; Savannakhet, Laos PDR; and Da Nang, Vietnam.

The choice of the provincial level as the locus of research-development serves to a) reflect the unique local potential and constraints of typical localities along the highway and b) to formulate bottom-up, consistent, integrated, and quantifiable five-year plans. The project will also train the necessary human capital in the form of four Ph.D. economists (from Nepal, Myanmar, Vietnam, and Thailand) alongside local economists and planners in the four provinces. These personnel will become proficient in the construction, analysis, and updating of Social Accounting Matrices (SAMs) and Sufficiency Economy Matrices (SEMs) for practical strategic development. Four Ph.D. dissertations and many professional journal articles on a wide range of development issues will be published. At the

end of the project, working models will be transferred directly to the leaders and chief economists of the four provinces, and the Ph.D. students will present their results throughout Thailand and abroad.

Answers are urgently needed for the EWEC as a whole and its component provinces to the following questions policy-orienting questions:

● What is the current structure -- in terms of both absolute and percentage value added with Provincial Domestic Product (PDP) -- of 10 key sectors (tourism, agriculture, OTOP, other handicrafts, other knowledge economy, other manufacturing, transportation, energy, telecommunications, and finance-banking) within the local economies of 4 representative provinces in the Greater Mekong Subregion?

● What is the current trade situation in the GMS? What are the causes and obstacles to growth in that trade? What are the combined impacts of the factors linking production, export-imports, employment and income within and among each economy?

● How much net employment, household income by economic class and occupation, negative or positive environmental impacts, net export value and net migration does each region generate? How much of these effects are attributable to community-based decision-making, eco-tourism, migration, gender equity, and the knowledge economy?

● What are the absolute values and ordinal ranking of the value-added, employment-creation, income distributional, environmental, and export multipliers of each sector? Does one sector dominate the others in its multipliers in all dimensions, or will it be necessary to favor 2 or more sectors to achieve balanced development? Is the priority set consistent with the Sufficiency Economy philosophy of the King of Thailand?

● What are the priority weights that provincial leaders would like to assign to value added, employment creation, income distribution, environmental sustainability, and net exports in their future development priorities?

● Given the weightings in 4, what is the optimal development plan for each of the four provinces in terms of PDP growth per capita, poverty alleviation, employment creation, tourism, and socially and environmentally sustainable development? Which are the key sectors to be promoted, and with what percentage of government and private investment?

● What levels of imports, exports, in- and out-migration, and foreign direct investment for each of the four provinces? In which sectors of the economy will these occur?

● To what extent can those extra-provincial flows be best met by the 3 other provinces of the study through cross-border trade, intraregional migration, intra-regional investment, and other forms of regional integration among the six countries.

● What are the estimated welfare benefits in terms of PDP (provincial domestic product) of the optimal patterns for each province taken separately, and for the four provinces as a whole.

● Is there are convergence or a divergence in the PDP per capita through regional integration? If a divergence, what complementary policies might be most effective in avoiding income polarization within the GMS? Do these have to do with income redistribution, cross-site learning, competition policy, or enhanced trade efficiency?

● Do women, ethnic minorities and the inhabitants of fragile, poor ecosystems benefit proportionately less than other members of society from the new patterns? If so, what specific policies at the provincial and/or transnational level could best relieve the inequities inherent in otherwise sustainable growth patterns? Would policy priorities based on the Sufficiency Economy be a more appropriate means of aiding those target groups?

4. Goals and objectives of this study

This project was designed over a 48-month period to build, evaluate, and link a series of four

(4) sub-national SAMs (social accounting matrices, to be explained in detail below), chosen in such a way as to follow the "East-West Growth Corridor" linking Myanmar, Thailand, Laos, and Vietnam. This corridor is already being studied by other researchers, national and international planners, notably the Asian Development Bank (xxx.xxx.xxx/XXX/Xxxxxxxx), in response to calls from national governments and international bodies alike for greater “competition policy”3 and fuller implementation of the “Chiang Mai initiative”4 . The research sites of the present project (Mawlamyine, Myanmar; Hat Yai, Thailand; Savannakhet, Laos PDR; and Da Nang, Vietnam) lie along key economic corridors within the GMS and were selected to provide well-positioned provincial leaders with the information they need to take maximum yet sustainable advantage from the ongoing globalization and integration of the region.

This project submission is, in fact, the first of three to be submitted to the TRF, the other two seeking to fund similar doctoral studies by a total of six (6) other students coming from two further sites in Thailand; Yunnan, China; Savannakhet, Laos, Haiphong, Vietnam; and Phnom Penh, Cambodia. There are logistical and humanitarian reasons for favoring Myanmar at this time. Of all the GMS countries, Myanmar is by far the least prepared to fend off the onslaught of Chinese goods, factories, and people that will flow into the country once the roads are open. Also, by replacing Savannakhet with some point in Burma, the TRF will finance a broader geographical sweep with this first round of scholarships. Finally, Burma has a repressive regime that must be replaced in some way.

3 This term has recently come into vogue to reflect the need to a) remove tariff and non-tariff barriers to trade (i.e., quotas or quality standards), b) make the internal industrial organization of countries more competitive, and c) encourage investments in transportation and market infrastructures that would allow competition to flourish.

4 This last agreement, signed in May 2000, by the finance ministers of ASEAN+3 countries was the regional financing arrangement to enable countries to cope with disruptive capital flows and maintain exchange rate.

It is therefore urgently necessary that Myanmar cultivate a new generation of young economists to eventually take over the planning of the country in a more rational way.

The mandate to serve the decision-making clientele at the provincial, tambol, and village levels comes from the climate of decentralization that is sweeping Thailand and the other members of the GMS. This climate involves what was formerly called the “OTOP” program of having each village specialize in one product or service; the need for community-based tourism; increases in the role of civil society; the necessity of establishing production, marketing and savings cooperatives; the desire to train economists at the sub-regional level capable of understanding, using and updating planning tools; and the psychological reality that if people are not involved in planning their own destinies they can rarely become enthusiastic about working together to bring about sustainable growth. It is also consistent with the Tenth Five-Year Plan of Thailand, based upon the Sufficiency Economy Philosophy.

This said, the four target provinces were not chosen because they were representative, but quite to the contrary as the most likely places where such positive decentralization of planning could occur. The strategic reasoning behind this project is a fortiori: if specialized marketing, community based tourism, growth of civil society, cooperatives, and local management of planning cannot work in these most favored areas; they will have even less chance of success in provinces lying off the main economic corridors of the GMS region! For example, it seems to us that Moulmein (also called Mawlamyine or Mawlamyaing in Burmese) is the perfect locale in Myanmar to conduct this study. It is the largest town in its province, has excellent access to the sea, lies within 50 km of the new East-West highway, and is a growth center already, even before the completion of the road. That part of Myanmar has historically been a source of out migration into Thailand because of massive unemployment. Our study should be able therefore not only to track the flow of goods and services, ecological impacts and income distributional effects but also the reduction or reversal of migration flows.

Specific objectives of the study

i) To provide practical, rigorous, clear, consistent and quantifiable planning tools for key provinces along the East-West Growth corridor linking Myanmar with Vietnam.

ii) To find the single best strategy for environmentally sustainable and socially equitable growth through the innovative optimization of those tools.

iii) To compare the results of optimization of the SAMs of standard economics with innovative SEMs inspired by the King’s philosophy.

iv) To identify, for the greater benefit of all partner countries in the region, the comparative and competitive advantage of key target provinces.

v) To train 4 non-Thai Ph.Ds in the King’s Philosophy and the construction of regional planning models.

vi) To develop a pilot study that may be used throughout Thailand during the 2007-2011 10th Five- year Plan.

vii) To publish groundbreaking theses and top-level journal articles to expound these results

viii) To present the results of the research to local collaborators, in research workshops organized by the Thailand Research Fund, and in international conferences

5. Conceptual framework

Economic development theories have grown by accretion since the early 1950s when Nurkse and Xxxxx first described the possibilities of growth based upon supposedly limitless supplies of rural labor. Successively, apparently conflicting schools of thought, added such themes as controlled migration (Todaro), community development (Xxxxxxxxx), Green Revolution unimodal technology (Hayami-Ruttan), growth accounting (Solow), women in development (Boserup, import substitution (Bhagwati), export-led open economies (World Bank), environmental sustainability (Hardin), access to basic needs (Sen), the knowledge economy (Harvard), structural adjustment (IMF), market transition (World Bank), positive globalization (Singer), and the sufficiency economy (King of Thailand). In reality, these themes are all essential to balanced development, but their relative priority weightings, sequencing, and timing all differ by sub-region because of both a) the true needs of the target population and b) the policy choices of governments.

Most economists are familiar with the concepts of forward and backward linkages of consumption, production, and trade, particularly in terms of the linkages between agricultural and the industrial sector (Hirschman). This study will expand upon those linkages to reflect the impacts upon the physical environment, and the distribution of income across economic classes and household types. This is because a Social Accounting Matrix adds to the standard “economic” accounts of an input- output table (Xxxxxxxx) a whole set of “social” accounts. This allows for more complete analysis and often reverses the relative priorities that should be accorded to given sectors in future planning.

This project seeks to integrate all these themes within a single model for each of four meso- economic provincial economies spanning the gamut of needs, policies, and positions along economic corridors from landlocked areas to the sea within the Greater Mekong Subregion. Two models for Thailand will seek to compare agroclimate and market infrastructure within a single cultural and political unit. The model for Vietnam will be used to explore in depth the variants of market transition from central planning. The Burma model will reflect severe underdevelopment in a market-based economy.

The project chooses the meso-economic level of analysis because micro-level households or firms are too varied and have objectives that are too limited to inform public economics. Also,

the significant ethnic and agro-climatic variability of the GMS would make macroeconomic models of the national economy, whether of the structural econometric or computable general equilibrium sort, too general to capture local needs and possibilities, including the one-village-one-product avenue.

In addition to the many theories cited above, the new theme of regional integration provides the cornerstone of the project. Kim (2005) examines the optimal path of regional integration in the East Asian region from the perspective of Korea, as a country with an intermediate level of technology in the Asian region. Based on the welfare analysis of three types of FTA regimes between technologically asymmetric countries, this paper demonstrates that the optimal path of regional integration for the country with an intermediate technology level would be to form multiple bilateral Free Trade Agreements (FTAs), i.e., Hub and Spoke type FTA between technically asymmetric countries from the very initial stage. The second best regime is the multilateral Asia wide FTA. The worst case for a country with an intermediate technology would be to form an FTA with a country with advanced technologies, and extend the FTA with the less developed countries in the later stage. These results imply that the optimal strategy of regional integration for a country with an intermediate technology level, such as Korea, is to form a Hub and Spoke type FTA rather than a bilateral FTA with Japan followed by the participation of other Asian countries. The pre-requirement for the multiple bilateral FTAs is the fundamental industrial restructuring of the Korean industries, such as reallocating the human and economic resources from the sectors of comparative disadvantage to the sectors with comparative advantage. The successful FTAs assumes the nation-wide efforts involving the massive restructuring cost including the establishment of a social-safety net during the transition period. Moreover, an Asian type of fiscal federalism is required for a genuine economic integration in the region.

Within this framework, Plummer (2003) focuses on directional and structural change in developing Asian trade patterns, with a focus on economic interaction with the United States (and using the EU as a benchmark). Several important results from his analysis are that: (1) while private-sector- led regional integration in Asia has increased for most developing Asian countries (in some cases, impressively), the United States remains the region’s single most important export market; (2) the trade structure of developing Asian countries has changed significantly over the past decade, a result that resonates with the literature on structural transformation in developing Asia; (3) the economic dynamism of developing Asian exports is consistent with the changes in trade with the United States (and the EU); and (4) science and technology have become the most important sector for most developing Asian exports (and imports), and the US market has been a key protagonist in this process. In addition, the paper considers emerging policy challenges in developing the US-Asian economic relationship at various levels, using the Vietnam-US Bilateral Trade Agreement (BTA) as an ex post case study and considering a number of ex ante initiatives that are currently being developed.

Figure 1.2 and 1.3 present preliminary conceptual diagrams for our paper. In the top half of the diagram we see the downward pathways of impact of the East-West highway on the supply side in

terms of generating changes in production, income and employment. In the bottom half, we see the upward impacts of the EWEC on the demand side of prices and consumption expenditure shares (“Engel shares”). The arrows show the theoretical directions of impact that we can test with econometric hypotheses.

Figure 1.2: Conceptual pathways of impact of the EWEC on income, employment poverty, and consumption

Figure 1.3 Conceptual diagram of the level, causes, effects and policies to alleviate poverty

Why does consumption figure so prominently in the above frameworks? It is because, for developing countries, consumption quantities and expenditure values are important factors to be considered for the measurement and evaluation of welfare because they vary much less and are less subject to over- or under-reporting in household surveys than direct measures of income. Furthermore, household expenditure data directly reflects consumers’ purchasing power and lifestyles. Indeed, the “correlation between consumer demographics and expenditure on recreation, hobbies and travelling reveals are the driving factors of spending” (MGP Information Systems Ltd., 2012).Thus, it is common for consumption to be used as a proxy for overall expenditure and in turn as a superior measure of well- being than income (Committee for Planning and Cooperation, National Statistical Centre, 2011).

Taken together, the above facts summarize the dynamics of poverty creation, the resulting socioeconomic effects, and four entry points for policies that could potentially be used by government to alleviate the syndrome of poverty and inequality. The ovals show the five hypotheses that will be evaluated in the remainder of this article to test the validity of this diagram for Mawlamyine, Myanmar.

Put simply, poverty and inequality are hypothesized to be alleviated by educational and infrastructure development but reduced by natural disasters. Even absent such factors, poverty and inequality may be determined by household characteristics, economic status and employment

opportunities. Poverty should in turn induce changes in the savings rates and consumption patterns of households. Consumption patterns may be divided into food, other non-food necessities, and non- necessities or “luxuries,” although the latter term in economics seems out of place for the vast majority of the Myanmar population.

Government may intervene in four ways to offset the poverty syndrome. First, it may invest in enhanced infrastructure to create the employment potential and growth impetus that forestall poverty creation in the first place. Second, job creation for the unemployed poor in strategic sectors enjoying comparative advantage can result in greater growth with equity. Third, government can make transfer payments to poor households to help them survive in the short term. Fourth, taxes and subsidies can be used to cheapen the goods on which the poor rely, and tax the commodities consumed by the rich.

6. Testable research hypotheses

In the same way that the goals differed by province, so did the testable research hypotheses.

A sample of these is given below for each of the four study provinces:

Da Nang, Vietnam

1. From the national viewpoint, the road has conferred a clear advantage on Da Nang province when compared with other provinces, regions and cities within Vietnam.

2. The EWEC has led to improvements over the period 2006 through 2008 in income, output value, sources of income and expenditure patterns among rural, semi-urban and urban households

3. Within Da Nang province, proximity to the road makes people better off through job creation, specialization and reduced inequality.

4. The differences among rural, semi-urban, urban areas have become less significant over time.

5. The distribution of income (Gini) for the whole sample has grown more equal over time as economic opportunities have opened to all classes of workers

6. By subsample, however, incomes in the rural areas have grown more unequal (i.e. higher Gini).

7. The incidence, depth and intensity of poverty are highest in rural areas farthest from the road.

8. The incidence, depth, and intensity of poverty have gone down for the whole sample and each of the three subsamples over time.

9. The greatest reduction in the incidence, depth and intensity of poverty has been in the urban areas, followed by the semi-urban and rural areas.

10. Those who are closer to the road have obtained transportation employment or have jobs related to the road are better off.

In contrast, for Savannakhet, Laos, we tested the following hypotheses:

1. Household income is significantly improved by proximity to the EWEC or by involvement in the transportation-related sectors, especially in rural areas.

2. Distance from the EWEC and lack of involvement in transportation-related activities are the most statistically significant causes of the incidence, depth, and severity of poverty in the households of Savannakhet.

3. The distance from the household to the EWEC is a more significant determinant of income per capita and employment than education, health, credit, or socioeconomic category in Savannakhet province.

4. Distance from the EWEC and income per capita are the most statistically significant causes of the household’s education, health, other consumption and production expenditures.

5. The agricultural sector still accounts for at least 60% of the total value added of the Savannakhet provincial economy.

6a. The value-added and employment multipliers from road construction and road-related activities (including tourism) are higher than for any other sector of the economy in Savannakhet.

6b. The export multipliers from road construction and road-related activities are higher than for any other sector of the economy in Savannakhet provinces

7a. A reallocation of investment patterns along optimal lines will result in at least a 30% increase in total value added for the Savannakhet economy.

7b. The optimal restructuring of the Savannakhet economy under resource and socioeconomic constraints is significantly different from that suggested by unconstrained multiplier analysis.

Meanwhile for Hat Yai, southern Thailand, we tested the following hypotheses:

1. Household level income is significantly different with location type of urban, semi-urban and non-municipal area.

2. The per capita household income is significantly increased by proximity to Asian Highway

3. Poverty is significantly higher in non-municipal area and also positively associated with the distance from the road head.

4. The price of natural rubber latex5 in Hat Yai is allied with the persistent risk

5 Here, latex refers to the natural rubber extract of type Ribbed Smoked Sheet

5. Volatility in Thai rubber prices has been more pronounced since the 2008 Hamburger crisis.

6. The conditional GARCH model provides a better fit to the data than ARIMA and other models, and can serve as a practical predictive tool to help predict and stabilize incomes in the Thai rubber industry

7. The volume of international tourist arrivals by land transport is associated with persistent volatility.

8. The conditional GARCH model captures the risk associated with international tourist arrivals by land transport than ARIMA and other models, and can serve as a practical predictive tool to help predict and stabilize incomes in the provincial tourism business in Songkhla.

Finally, for Mawlamyine, Myanmar, we sought to test the following

1. Proximity to the main road or part of the EWEC directly and significantly reduces the proportion of food consumption and transportation expenditure within overall household expenditure per capita.

2. Better transportation and proximity to highways significantly increases the knowledge people can use to focus on their income and health care.

3. There has been significant growth in equality of income in the Mawlamyine area.

4. Overall income per capita and the number of jobs in the Mawlamyine area could increase by at least 25%6 if the above constraints were removed.

5. The income, employment, and consumption multipliers of the road infrastructure and transportation-related economic sectors are significantly higher than for other sectors.

6Since this part of the analysis will use the Social Accounting Matrix rather than econometrics, we cannot strictly speak of a statistical test of this hypothesis. Instead, we have set a reasonably large increase (25%) as a measuring stick to represent a significant improvement in various dimensions of the economy.

CHAPTER II:

EWEC INITIATIVE AND BACKGROUND THE STUDY SITES

1. Initiative of the EWEC

The concept of an East-West Economic Corridor was introduced in October, 1998 in Manila at the Eighth Ministerial Conference of the Greater Mekong River Sub-region realizing the need of the regional development considering regional competitive and comparative advantages of economic development and economic cooperation. Since the past, there have been official or unofficial economic transactions prevailing between the neighboring countries. The eight ministerial conferences agreed to develop and promote free Trade Area (FTA), Tax Alliance and Common Market (EU) in the Mekong sub regions, and enhance the economic globalization maintaining national investment for international markets as well as attracting foreign direct investment in the participating countries. The East–West Corridor (EWC) Project is part of the wider East–West Economic Corridor linking Da-Nang in Viet Nam and Mawlamyaing in Myanmar—covering Lao People’s Democratic Republic (Laos PDR), Myanmar, Thailand, and Vietnam. As a flagship project of the Greater Mekong Sub-region (GMS) Program, it was designed to improve National Road 9 linking landlocked areas in northeast Thailand to the Vietnam coast via Laos PDR. It is the second cross-border road project in the Program and is in line with ADB's thrust for regional cooperation in the transport sector (ADB, 2008).

The economic corridor has three distinctive features. Firstly, it is considered a defined geographical region. Secondly, the corridor mainly focuses on bilateral initiatives, not multilateral

ones. Thirdly, the economic corridor requires detailed physical and space planning for infrastructure development for the most effective results. The nearly 1,450 km long East-West Economic Corridor passes through four countries, and the East-West Economic Corridor also connects with the main north-south roads, including Yangon – Dawei and Chiang Mai – Bangkok of Thailand, Road 13 of Laos, and the National Highway No.1A of Vietnam.

Moreover, the summary findings of “evaluation study on the east-west project report” of ADB (2008) mentioned that the project is “‘highly relevant’ to development needs at the regional and national levels. As a regional road, it addressed a strategic need to strengthen links between Laos PDR, Thailand, and Viet Nam. The all-weather road aimed to expand the market for transit and bilateral trade. It also interconnects national transport networks to generate trade and efficiency benefits. It was appropriately designed to support economic centers and to complement poverty reduction (via a rural road component in Laos PDR)” (ADB, 2008. p iii) and ranked as a successful project. Moreover, the report added that there is little evidence of EWC's transformation from a transport corridor into an economic corridor, and the full realization of corridor-level benefits (i.e., economic activities, tourism, and trade facilitation) is limited by institutional constraints. In the case of Laos PDR, this is partly explained by inadequate complementary investments and slow private sector growth.

The lessons gleaned from the project mentioned by ADB Evaluation Study on the East-West Corridor Project Completed are: (i) The pace of the economic development along the corridors depends on the complementary interventions requiring policy and institutional changes that enable better opportunities for the private sector. (ii) Multimodal planning is important for enhancing the effectiveness of transport corridors, and (iii) ADB should use loan savings prudently by exercising due diligence at the approval and completion stages of the project. Most of the road infrastructural development projects further alleged that these projects foster the economic development process by a) enhancing marketing linkages among the parties at local and regional levels to b) easing the mobility of products and factors based on market forces and c) generating employment opportunities through d) increased market accessibility for the rural people.

AbuzarAsra et al. (2006) introduced an approach to analyze the impact of the east-west corridor along Savannakhet route no. 9 to its neighboring countries using the accounting framework of the traditional single area I-O model. The finding of the study shows the significance of the economic measurement tackles broad issues on export-import expansion, interdependence of industrial structure from Rest of Savannakhet (intra and interregional analysis), evaluating impacts or changes in final demand (production, income, and employment), and short term projections and forecast of the domestic economy (AbuzarAsra, 2006).

BounpongKeorodom et al. (2007) showed the thought of business persons that are positive toward their business investments after the completion of The Second Mekong Bridge; 20 % would be

an increase of imported goods in construction materials and machinery equipment, and 15 % would be increase in export of domestic products. Additionally, 17 % of the respondents saw the development of EWEC as cost and time saving in transportation, and 20% would be strong competition among domestic firms as well as from foreign investment.

2. Background of the EWEC in Da Nang

Da Nang city has the privileged position of lying both on the sea, with a deep-water seaport, and at one of the two ends of the East-West Economic Corridor (EWEC). The project, initiated by the Asian Development Bank in the 1990s in partnership with the nations of Myanmar, Thailand, Laos, and Vietnam, was designed to provide the infrastructure for linking and boosting the economies of these countries. Without roads, it was reasoned, the movement of goods, people, and information is hamstrung, and the positive benefits from trade and globalization are erased by prohibitive transportation costs.

Compared with the three other countries linked by the EWEC, Vietnam has fared well overall. Myanmar’s repressive political regime has led to rebel uprisings between Mawlamyine and the Thai border at Mae Sot, such that only 18 of 100 kilometers of the road have actually been constructed. Laos, as a completely landlocked country, is facing problems of investing in attractions that will lure FDI and traveler expenditures within the country instead of transiting the country without investing or paying for anything. Thailand has done well largely because the main road helps to link the country with itself and with a second road network leading south from Kunming, China to Malaysia.

Vietnam shares some of those advantages as roads have also been improved, after decades of delay, linking North with South, and most critically, west with east. The EWEC falls into the latter category as it enters Vietnam from Laos, passes through Quang Tri and ThuaThien Hue provinces, enters Da Nang province and ends at Da Nang city. As such the road provides an enormous potential

for improving the well-being of Vietnamese throughout the country, but most particularly those living in all three provinces. The means of improving well-being are of course enhanced employment opportunities in transportation-related sectors, lower transportation costs for firms and exporters, improved upstream - downstream linkages for farmers, and reduced costs of consumption items for consumers as a whole.

It is clear that Da Nang, ThuaThien-Hue and Quang Tri, within the North Central and Central Coastal region, have increased at average or less than average rates. They are only the 6th, 3rd, and 11th largest out of 14 provinces in the zone, respectively, and have achieved a change rank of only 7 (Da Nang), 9, or 12 (ThuaThien-Hue and Quang Tri). The secondary data points to no clear impacts of the EWEC in terms of education, trade, job creation, or population growth when Da Nang is compared to other cities and provinces. But what of when primary data from the Vietnam Household Living Standards Survey (VHLSS) from 2006 and 2008 are used to analyze in depth the impacts on differing areas (rural, semi-urban and urban) on income, income distribution, and consumption within the Da Xxxxx sample? Here, we may find significant trends, as well as more detailed explanatory variables, that can shed further light on the impacts of the EWEC to date.

Figure 2.1: East-West Economic Corridor in Da Nang

3. Background of the EWEC in Savannakhet

Savannakhet Province in the centre of Laos lies at the junction of the East West Economic Corridor (Road no. 9) and the North-South Axis (Road no. 13). The Savannakhet population of 843 thousand represents one-tenth of the national population (2011) on 21,774 sq km. representing as one- ninth of national areas. This gives Savannakhet a population density of some 38.58. persq km which is much lower than that of neighboring points along the EWEC: Mukdahan, Thailand and Da Nang, Vietnam with 76.05 and 628.33, respectively. Almost all land in Savannakhet is underdeveloped. Given the low population density, most land is devoted to agriculture and forests; and households along the Laos section of the EWEC remain poor because jobs, resources, public infrastructure, clean water, and other facilities remain largely lacking. Hence, households feel little motivation to live or move to Savannakhet. “Thus, this province has a big potential for development of agro-industries, transportation, and international trade” (Warr, Menon, & Xxxxx, 2009). Nonetheless, the present Gross Provincial Product per capita (GPPC) of Savannakhet is USD 525, roughly 85% of Laos PDR’s national average and still lower than the corresponding figures for Mukdahan, Thailand and Hue, Vietnam.

The Mukdahan–Savannakhet border is an important gateway for trade between Lao PDR and the neighborhood countries of Thailand, Vietnam, southern China, and eventually Myanmar. Most trading companies and industry factories are located in downtown Savannakhet in order to facilitate trading activities between Lao PDR and Vietnam. This would make the use of roads for long-distance freight transport between Thailand and Vietnam more beneficial. The cross-border road network across the 2nd Lao-Thai friendship bridge allows landlocked Savannakhet, and by extension the rest of Lao PDR, to access the LaemChabang port in Thailand. LaemChabang is one of Asia’s leading ports to North American and European shipping networks (Jiwattanakulpaisarn, Ruangsawasdi, &Gaywee, 2010).

In terms of present levels of investment, there are no facilities such as gas stations, restaurants, hotels, tourist information centers, or gift shops along the EWEC because most areas it passes through are forests and agriculture lands. The houses and lifestyle of most residents along the road remain poor and public infrastructure and conveniences are still largely lacking. Most of the roads are inconvenient for communication, rough and damaged due to the failure to increase investment and maintenance (xxxx://xxxx.xxxxxx.xx.xx/xxxx/xxxxxx?xxxx00x00 474510e4b21).

This said, what investment there is along the EWEC in Laos is totally concentrated in the area of Savannakhet. So the district of Savannakhet has the potential to play a key role in investment. Moreover, a substantial number of enterprises are new generation businesses such as patrol stations and rental accommodations. If it fully lives up to its potential, the EWEC program will facilitate the transport of goods, migration of people, and stimulate economic growth, trade activities, and FDI in Savannakhet province. However, Laos still get less benefit from this program comparable with

neighboring countries because the large lack of a strong industrial base, and few products are suitable for export (Keorodom, Butphomvihane, &Vanhnalat, 2007).

The opening for use of the EWEC inaugurated the first Special Economic Zone of the country at Savannakhet. It sought to attract foreign investment because of its potential for drawing foreign currencies into the country. Investment in Laos along the EWEC and the routes connected to the remainder of the country aims to support economic exchanges along the East–West Economic Corridor. The EWEC is also connected to road No.13, the main internal route for making domestic transportation more convenient and distributing commodities to other sub-districts in the country. The EWEC also supports Laos PDR’s commercial extension through Vietnam.

Savannakhet Province’s economy has grown during 2005-2010 at a rate of 10.5% a year, significantly higher that of the whole country. GDP per capita has been increasing from USD 525 in fiscal years 2005-2006 to USD 897 in fiscal years 2009-2010, slightly surpassing the national level at USD 880 per person in 2009. Agriculture is the main economic sector in Savannakhet Province. In the years 2009-2010, this sector grew to 49.04% of GPP, followed by the service and industrial sectors, which reached 26.42% and 24.54%, respectively, of the total production value of the province. As a sign of the transition to a non-agricultural economy, the value of agriculture had declined gradually from 55.54% in 2005-2006 to 49.04% in 2009-2010, with corresponding increases in industry and the service sector.

Figure 2.2: Map of Savannakhet

4. Background of the EWEC in Hat Yai

Songkhla province is one of the Thailand's most important seaports, located at the coastal belt of the Gulf of Thailand bordering with the Kedah State in Northern Malaysia (Gov/Thai PRD, 2004). Natural Rubber plantations, ocean captures, and fisheries have been operating for years, including tourism businesses. The province has become one of the leading natural rubber growing provinces in the south, and land use patterns clearly show that more than 61% of total land coverage of the province

is covered by Para rubber plantations, followed by Forest resources (14.5%). Altogether, approximately 77 % of land cover is occupied by agriculture and fisheries1. Songkhla province is one of the leading natural rubber (NR) growing provinces in the south. Land use patterns show that almost more than 61%

of total land cover of the province is covered by Para rubber plantations . These and other natural

resource based occupations are the basis of income generation to local dwellers, but they are also a source of income inequality between the poor workers and the rich local and absentee land-owners of the region.

Additionally, the Hat Yai segment of Asian Highway No 4 (AH2) is a strategic road infrastructure investment for the economic development of the Songkhla as well as Thailand as a whole, and it links the North South Economic Corridor of the greater Mekong sub-region (GMS) with the Indonesia-Malaysia-Thailand growth triangle (IMT-GT) sub-region up to the Malaysian border. This may enhance product transport through reductions in transport cost and promote tourism, labor mobility, commercialization of competitive sectors, and economic globalization (ADB, 2007).

In this regard, the Asian Highway and/or its sub-branches directly or indirectly bind together the various agricultural production patches and agro-based industries of Songkhla including southern peninsular Siam, providing ample opportunities for national and international markets through land shipment. It is expected that the positive role in the economic activities of rural Songkhla has been played by the Asian Highway, and its sub-branches towards the economic prosperity thereby contributed to poverty reduction. Hence, we believed that economic and social objectives can only be attained when the rural dwellers of the province begin to attain educational levels, incomes, meaningful employment, and living standards that are comparable to those of the other Provinces. A special emphasis of this paper will therefore explore the poverty status and the possibilities for bottom-up economic planning to improve the wellbeing of the low income group of the people.

The economy of Hat Yai is influenced by tourism and businesses including plantation crops and fisheries. The province has a tropical climate, which is hot and humid, like other parts of Thailand. The economy of the province suffered a considerable negative shock in 1999 (-3.79%), 2009 (-5.63%), and 2012 (-2.85). This might be due to results of transmissions of the Asian Crisis at the provincial level in 1999, world economic downturn in 2008, and national political crises. The overall growth of GPP in the period 1995-2012 was 5.56 %. Songkhla province has played a greater role for socioeconomic development of southern Thailand. As an international marketing hub for natural rubber and other agricultural products, the province seems to be one of the driving forces or agents for socio- economic development of the peninsular Siam.

1The land covers is calculated from the GIS shape file. For details see chapter 5

2The land covers is calculated from the GIS shape file of Songkhla province and data received from the Faculty of Environmental Management, Prince of Songkhla University, Hat Yai.

Figure 2.3: Map of Hat Yai

5. Background of the EWEC in Mawlamyine

Mawlamyine is one of the two endpoints of the EWEC corridor (the other is Da Nang, Vietnam), which should act as a huge economic and trade advantage. However, civil unrest and administrative problems have meant that in ten years only 18 of 200 kilometers of the road inside

Myanmar have actually been completed. Mon state is a state of mountain range and flat land. Bordering Bago Division in the south of Sittaung River Mouth, Kayin State in the east, Thailand and Taninthayi Division in the south and Andaman Sea and Gulf of Mottama in the West, Mon state is situated between latitudes 14°52' north and 17°32 ' north and east longitudes 96° 51 ' east and 98° 13 ' east.

Mawlamyine Township is the administrative and commercial capital of Mon state and the third largest city in Myanmar after Yangon and Mandalay. It is constituted very properly and the buildings constructed in line are found everywhere. Its land area is 54080 acres (84.48 sq miles) with a population of more than 460000 in 2008. The widest area from the East to the West is 7 miles and the widest area from the South to the North is 17 miles.

Mawlamyine is surrounded by Kyaikmayaw township in the east, Chaungsone township in the west, Mudon and Ye townships in the south and Hpa-an in the north (See figure 2.3). It is made up of 22 wards, 19 village tracts, and 48 villages. Its land area is 54,080 acres (84.48 Sq. miles) with a population of more than 460,000 in 2008. This yields a population density of 2,162/Sq. Km in Mawlamyine Township which is comparable to that in Da Nang City, Vietnam (638 /Sq. Km) while Mawlamyine district has 279/Sq. Km which is more than Savannakhet Province, Laos (46/Sq. Km)3 but less than that in other Asian cities such as Bangkok, Thailand (3,628/sq. Km)4and Ho Chi Minh City, Vietnam (3,155/Sq. Km)5. The widest area from the East to the West is 7 miles and the widest area from the South to the North is 17 miles.

❖ Transportation Sector in Mon State

Myanmar has been taking the following measures undertaken with a view to facilitate economic and social development:

- extending roads and bridges to ensure smooth and secure transportation

- extending the construction of hospitals and schools

- extending the construction of railroad tracks, airstrips, international level airports, and port development works

- extending the construction of irrigation and embankments

- establishing industrials zones for industrial development

To ensure balanced development among regions, one of the major emphases has been placed on construction, renovation, and maintenance of roads and bridges in States and Divisions. Table 2.14

3xxxx://xxx.xxx.xxx.xx/Xxxxxxxx/Xxxxxxxxxxxxxxxx0000/xxxxxx%00xxxxxxx0/0.Xxxxxxx.XXX and xxxx://xxxxxxx-xxxxxxxxx.xxx/?xxxx_xxx00

4 xxxx://xx.xxxxxxxxx.xxx/xxxx/Xxxx_xx_xxxxxxxxx_xx_Xxxxxxxx_xx_xxxxxxxxx 5 xxxx://xxx.xxx.xxx.xx/xxxxxxx_xx.xxxx?xxxxxx000&xxxxxx0&XxxxXXx0000

exhibits road infrastructure extension in Mon State, comprising bituminous roads, metaled roads, surface roads, and earth roads.

❖ Feature of the EWEC in Mawlamyine

The EWEC firstly starts at Da Nang Port from Vietnam and ends at Mawlamyine Port of Myanmar. So Myawaddy-Mawlamyine road (almost 200 km) plays a vital role in the Myanmar side of the EWEC. Myawaddy-Mawlamyine road is on the way back to Phann and departs at Eindu, 24 km from Phan. Myawaddy and Mae Sot are connected by a friendship bridge.

KAlegauk Island

and deep sea port

Attayan Bridge (EWEC)

ThanLwin Bridge

EWEC

Old road ( Yangon

_ MLM through Phaann)

Figure 2.4: Map of Mawlamyine with EWEC

Eindu-MLM road is divided into 2 parts: Eindu-Zarthabyin and Zarthabyin-MLM route. Eindu-Zarthabyin is a 14 feet two-lane tar road up to Zarthabyinbridge, a 2900 feet Bailey Girder Bridge crossing the Gwaing River. The bridge allowance is 36 ton weight. Zarthabyin-MLM road is a

22 feet 2-lane new bituminous road which passes the MLM Industrial Zone and Attanyan Steel Girder Suspension Bridge near MLM.

Total number of bridges (above 180 feet) constructed by the Public Works increased to seven in 2008 from three in 1988. They are Attayan Bridge (Mawlamyine), Thanlwin Bridge (Mawlamyine), and Sittaung Bridge (Mupalin ) in Mon State. Thanlwin Bridge (Mawlamyine) is the longest bridge in Myanmar and connects the city of Mawlamyine with Mottama. Constructed at the confluence of the Thanlwin River, the Gwaing River and the Attayan River in Mon State, the bridge has a two-mile-long motor road and four-mile-long railroad as well as pedestrian lanes. The approach structure of the rail bridge on Mawlamyine bank is 1.22 miles (1,960 m) long, and on Mottama bank is 1.42 miles (2,290 m) long. The total length of the rail bridge is 4.75 miles (7,640 m) long. The bridge was designed and built by the Ministry of Construction. Attayan Bridge (Mawlamyine) seemed to be an important one before Thanlwin Bridge (Mawlamyine) as it is the route of EWEC. It can connect from Mawlamyine to Myawaddy (Border city) and to Phann which currently is likely to be more useful than Mawlamyine as Phann is situated in the way of the Asian Highway project. Its upper structure is of a steel frame type and the bridge can withstand 60-ton loads of all vehicles. Sittaung Bridge (Moppalin) was built four miles downstream of the existing Sittaung Bridge (Theinzayat) . The bridge is connected with two approach roads - a 224.1 feet long approach road constructed on Kyaikto bank and a 783.6 feet long road on Waw bank. The bridge is a 2,392.7 feet long, 28 feet wide motorway flanked by six feet wide pedestrian ways.

❖ Kalagauk Deep Sea Port In Mawlamyine

The key work for the EWEC is the ports at each end of the corridor. To the East there is the Da Nang Port which consists of two areas, Tien Sa Seaport and Han River Port, and is fully operational throughout a capacity of 4 million MT a year. Mawlamyine port is one of the coastal ports among the other eight coastal ports in Myanmar. The port authority of Myanmar has studied that Mawlamyine has been deemed inappropriate because the Maritime access for seagoing vessels to Mawlamyine Port is not deep enough for those vessels with more than 4.5 meters draft since it is 20 knots away from the mouth of Thanlwin River.

CHAPTER III:

DATA COLLECTION AND METHODS OF DATA ANALYSES

I. Data collection

Given the depth and scope of the interlocking methodologies to be presented in chapters III and IV of this report, only four (4) study sites could be studied in detail. Selection criteria therefore had to be very rigorous. All sites had to lie along the EWEC, be economically or strategically important in their own right, but also represent a certain point along the spectrum of pre-EWEC development or under development. These selection criteria have greatly enhanced the external validity of the present research, and allowed for clear comparisons of the impacts of the road under a wide range of conditions. The final selection of the four study areas, from east to west, is as follows:

1. Da Nang, Vietnam, a well-developed port town that was already connected on a major North-South highway corridor with the rest of the country.

2. Savannakhet, Laos, a country-wide but landlocked city with untapped potential for tourism and trade going both east to Vietnam and west to Thailand, and a fledgling but emerging business sector.

3. Hat Yai, in Songkhla province, southern Thailand, an already specialized border town with ideal conditions for rubber production and export, and the geographical position to serve as the gateway for tourist arrivals from and through Malaysia.

4. Mawlamyine, Myanmar, the fourth largest city, lying on the EWEC on a direct line between the capital Yangon and a proposed seaport. BUT, as this report will demonstrate, each of those favorable development conditions was reversed: the capital was moved inland, the seaport project was transferred to a different site, and ethnic warfare and rebellion prevented the EWEC from being completed.

Data for all sites was collected in the years 2010 and 2011 using standardized household and firm questionnaires and the integration of all available secondary and governmental data. Because of the unique characteristics of each site, specific research objectives and testable hypotheses differed accordingly.

The economic data and four SAMs spanning the GMS regions contain all of the dimensions important to the GMS project: transport, energy, telecommunications, trade, investment, SMEs, tourism, and the environment. Every effort is also being made to include the idea of Sufficiency Economy in the model. Although full-blown Sufficiency Economy analysis may lie beyond the scope of the current project, the SAMs are being expanded into Sufficiency Economy Matrices (SEMs) for the benefit of future research projects.

The bulk of the data used in this project derived from intensive field surveys of approximately 300 households and 75 business enterprises in each target province. They have been completed and augmented by secondary data generated for a wide variety of survey and econometric studies. The data have been standardized so that a) there will be 15 socio-demographic categories in each provincial model and b)the value of the

expenditures and the receipts of each of the sectors in the economy balance. Given the disparate sources of the data, this is the hardest part of validating a social accounting matrix. It will also take more time than the actual analysis of the results, which can proceed very quickly under a wide range of alternative scenarios.

In addition to the economic data going into the SAM, two other types of information will be essential. The first is technical data on the environmental impacts on the air, water, land resources, and disease incidence of each sector. These will be obtained through consultation with environmental engineers in each country. The second is focus groups with provincial leaders to generate the priorities that they assign to various development goals.

1. Data collection in Da Nang

We assembled data from the Vietnam Household Living Standards Survey for 2002, 2004, 2006, and 2008. Most of the analyses in this research focus on comparison of the last two years because the impacts of the road should be visible even over a 24-month period and because the percentage of paired observations in the unbalanced panel is the highest.. There are 115 observations in the household sample each year but only about two-thirds are the same between any two years. EXCEL spreadsheets were then prepared and transferred into separate SPSS files for:

- the unbalanced set of all observations in both years.

- the balanced set for the subset of identical households

- the "differences" between the two years for the balanced households. (For this step, we had to inflate the 2006 monetary values to convert them upward to real 2008 prices using the consumer price index and the producer price index for 2006 and 2008.)

We ran all regressions and statistical tests for all three datasets in order to find the most robust results and to make comparisons between the cross-section/static situation of each individual year and the dynamic evolution between 2006 and 2008. Only the most meaningful and significant results were used for the hypothesis testing.

2. Data collection and sampling in Savannakhet

Primary data for this study was collected in 2011 from households in Savannakhet, Lao PDR. Due to the incomplete nature of the LECSIV database for consumption expenditure analysis, an open ended questionnaire was designed and converted into a closed-form questionnaire and used for household level interviews of 243 households. We focused only on KaysonePhomvihane district in Savannakhet, the urban district that surrounds the EWEC road as it cuts through the province. Xxxxxxxxxxxx district is also included because it includes the intersection of the EWEC with the north-sound road no.13. The target population is thus limited to households who live along EWEC in both districts. This target population is divided into sub-groups, as well as rural, semi-urban, and urban categories at varying distances in time and kilometers from the EWEC. Data was collected on household characteristics, household revenue and consumption expenditures, and household expenditure on production process.

Savannakhet town in Laos PDR is situated on the banks of the Mekong River opposite Mukdahan in Thailand. The province shares borders with the neighboring countries, Thailand and Vietnam and the town is considered as a very active junction for trade among the three countries. Savannakhet province consists of 11 ethnic minorities including Lowland Lao, Phoutai, Thai Dam, Katang, Mongkong, Vali, Lava, Soui, Kapo, Kaleung and Ta-Oi. Located in the central part of the country, Savannakhet has a total area of 21,774 square kilometers and a population of 850,000 (2010). The province is subdivided into 15 districts namely: KaysonePhomvihane (previously known as Chanthabouly), Outhoumphone, Atsaphangthong, Phine, Sepone, Nong, Thapangthong, Songkhone, Champhon, Xonbuly, Xaybuly, Vilabuly, Atsaphone, Xayphouthong and Phalanxay. The capital of the province is KaysonePhomvihane. Table 3.1 show the household sample size in Savannakhet municipality is 0.17% of the population from 142,332 households in two districts.

Table 3.1 Household sample size in Savannakhet

	Total households in Savannakhet municipality	Urban	Semi urban	Rural
Universe	142,332	18,146	91,632	32,554
Sample	243	140	76	27
Sample %	0.17%	0.77%	0.08%	0.08%
Multiplying factor for averages		0.13	0.64	0.23

3. Data collection and sampling in Hat Yai

❖ Primary data

The field survey was scheduled in between July 15th – end of November 2010. PhD candidate Mr. Xxxx Xxxxxx Neupane with student ID code 511651801 participating from Nepal was assigned for this work. This field survey was part of his academic work. To facilitate the field survey and collection of the secondary information from various sources, the establishment of a field office at Hat Yai was provisioned and established accordingly. Mr. Xxxxxxx is a non-Thai speaker, and it is customary to hire assistants who speak both Thai and English language to facilitate and run interviews smoothly with local dwellers. Hence, necessary financial support and human resource management was also provisioned to complete the field work efficiently. Master’s students Mr. XxxxxXxxxxxxxxx, Miss XxxxxxxxXxxxxxxxxx, JanjiraSritawee, and PassarapornChaemmanee were hired granting full authority to the field office as per the financial condition of the TRF.

The household sector, firm sector, and hotel and tourism sectors were included in the field survey to assure inclusion of each category of the economy in the sample. Participatory Rural Appraisal, Rapid Rural Appraisal, and interview with detailed questionnaire were the main approaches adopted in the field survey. Out of 125 Tambons in Songhkla province, six Tambons were selected for household survey. For firm survey, the list of the firms established in the Songkhla province was collected from the Industrial Office Songkhla. With the help of this list, firms were contacted for information and the names of firms kept confidentially. The name of the Tambon, sample size in each Tambon and other sectors, and status of the field survey is presented in Table 3.2 below.

Table 3.2 Number of sample allocated, Tambon and status of the field survey

S.N.

Name of the Tambon/Sector

No. Sample Allocated

Remarks

KoYo

Completed

Boryang

Completed

Phang La

Completed

Tha Kham

Completed

KhongHae

Completed

Khuanglang

Completed

Hat Yai

Completed

Hotel and Tourism

data of completed survey of 30 observations has been lost because of heavy flood, and other work is postponed

Firm Survey

Flood adversely affected the area of data collection. Recognizing the people's’ sentiments after flood, the data collection work is postponed.

❖ Secondary data: Sources, Nature and Limitation

The secondary data was collected from the provincial and district level governmental and nongovernmental institutions, for instance, the Provincial Industrial Office, Provincial Trade office, Immigration office Songkhla, and their sub offices at the Malaysian border, Tourism Authority of Thailand, Songkhla, Sports, Tourism and Recreation office, Songkhla, National Rubber Research Centre, Hat Yai, Central Rubber Market Hat Yai, Hat Yai and Songkhla Municipality, Faculty of Environmental Management, Prince of Songkhla University, and data bank of Chiang Mai University. The World Wide Web page of these and other relevant institutions were also visited.

The daily tourist arrivals data set used in this study were gathered from the Songkhla Immigration Office. Other information about tourism was collected from the Songkhla Office of the Tourism authority of Thailand and the Songkhla Provincial Tourism Sports and Recreation Centre. Songkhla province has two immigration ports bordering with Malaysia, namely, Sadao and NarttayaPutchoo. Of these, Sadao is the main immigration port. Due to the unavailability of data on international tourists travelling by land to Songkhla province after arrival at Bangkok port, as well as the lack of disaggregated data between Thai and foreigners at the NarttayaPutchoo Immigration Office, only data from the Sadao Immigration port was considered.

Finally, the data used for the rubber price volatility analysis was gathered from various sources. The daily bid price of the NR products RSS3 and LAT for the year 2004-September to September 2010 (Thai calendar 2547-2553) was obtained from the webpage of Songkhla Provincial Agriculture Office (xxxx://xxxxxxxx.xxxx.xx.xx/). Data on world rubber consumption and production was collected from the webpage of Thai Rubber Association (xxxx://xxx.xxxxxx.xxx/xx) and from the web page of Indian NR (xxxx://xxxxxx.xxxxxxxxx.xxx/).

4. Data collection and sampling in Mawlamyine

Mawlamyine Township was chosen not because it is typical or representative of Myanmar as a whole, but precisely because of its above-average geographical and logistical potential for poverty alleviation. If the causes and effects of poverty cannot be determined there, and no workable policies can be applied, then a fortiori it will be virtually impossible to accomplish these aims in even more disadvantaged areas of Myanmar.

The unprecedented primary data used in this study was collected in May-July 2009 from 375 households in Mawlamyine Township, Mon State, Myanmar. A stratified random sample was selected based on the known percentages of the population living in rural, semi-urban, and urban areas. Interviews running between 50 and 90 minutes per household were conducted to collect income and expenditure data in a closed questionnaire format for minutely detailed subcategories of employment category, food, clothing, transportation, housing, health, insurance, communication, and education.

II. Methods of data analysis

To be implementable, the assessments and recommendations flowing from a study like this one must be based upon strong statistical evidence. This report is the fruit of a generous four-year research grant from the Thailand Research Fund that explains and applies a balanced set of statistical and economic modeling tools to a) measure the level of economic and social progress, b) explain gaps in that progress, and c) generate optimal policy recommendations. The levels of progress were measured by net income per capita; levels of education and employment; Gini coefficients; the Foster-Greer-Thorbecke indices of the incidence, depth, and intensity of poverty; and the proportion of value added deriving from business and trade. Tests of means were employed to detect significant differences in each study province between rural, semi-urban, and urban subpopulations.

The gaps in progress were explained by multiple regression equations, including seemingly unrelated regression and volatility models. The dependent variables are income, the levels of consumption of groups of key necessities, the intensity of poverty, the number of tourist arrivals, and the fluctuations in the prices of key export commodities like rubber.

Optimal policy recommendations were generated through the use of Social Accounting Matrices. Not only were the goods and services with the greatest value added identified for each study site; the matrices were optimized under realistic constraints on the availability of land, labour, and capital to assure that the optimal plans so generated would be feasible.

1 Tests of means

For each of the three target areas in each of the four study provinces, statistical analysis on household data was conducted to identify and correct outliers. Then three levels of analysis were performed:

a) average values for three geographic (rural, semi-urban, urban) and five income-class household categories (wealthy, upper middle, middle, lower middle, poor) for a total of fifteen separate categories were calculated. They were placed appropriately as dummy variables in the following two steps.

b) test for significant differences in income per capita as a result of distance to the EWEC or employment in transportation related sectors.

2. Poverty analyses

The Foster Greer and Thorbecke (FGT) Index is a set of widely-used poverty indicators. There are three measurements of poverty, which are the incidence of poverty or Head-Count Index (HCI), the depth of poverty or the Poverty Gap – (PG1), and the severity of poverty or the Squared of Poverty gap (PG2).

2.1 The incidence of poverty: Head-Count Index – HCI

(1)

Where N is the number of people in the sample/population; M is the number of poor people/households or those whose income levels are less than a given poverty line (z); and Yi is the

income level of person/household

This index measures the incidence of poverty. In other words, it shows the percentage of people/households in a region or a country living in poverty. Furthermore, given an income distribution, this index is a measure of that amount of poverty in that income distribution. It is the proportion of individuals or households (depending upon the survey unit) in the distribution whose incomes are below (or equal to) the poverty line. If the line is denoted by , then the HCI is the value of the cumulative density function (CDF) at .

2.2 The depth of poverty or the Poverty Gap – PG1

(2)

Is the Poverty gap corresponding to second order stochastic dominance for comparing poverty across two populations. It measures the depth of poverty, i.e., how poor on average are those in the subpopulation with inadequate incomes. It may be interpreted as the shortfall of the poor’s income from the poverty line expressed as an average for all people/households in the population.

2.3 The severity of poverty or the Squared Poverty Gap – PG2

(3)

This index reflects the third order of stochastic dominance for determining which of two populations is the poorest for eventual policy intervention. It measures the severity of poverty, which gives more weight to the poorest since they are farthest from the poverty line and the measure is based on the squares of that gap.

2.4 The Lorenz curve and Gini Coefficient

The Lorenz curve is a common tool for plotting inequality of wealth or income between two or more groups. The numbers of income recipients are plotted on the horizontal axis in cumulative percentages. The vertical axis shows the share of total income received by each percentage of population. It also is cumulative up to 100% making both axes equally long. The entire figure is enclosed in a square and a diagonal line is drawn from the lower left corner of the square to the upper right corner. This diagonal represents the perfect equality of income distribution since every point on it shows that the cumulative percentage of income received is exactly equal to the cumulative percentage of income recipients. The Lorenz curve shows the actual quantitative relationship between the percentage of income recipients and the percentage of the total income they did receive in a given year.

The farther the Lorenz curve lies from the diagonal of perfect equality, the greater the degree of inequality represented. The extreme case of perfect inequality would be represented by the congruence of the Lorenz curve with the bottom horizontal and right-hand vertical axes. The greater the degree of inequality, the greater the bend and the closer to the bottom horizontal axis the Lorenz curve will be. The Gini coefficient is a measure of inequality of income or wealth distribution that is intimately connected to the Lorenz curve. It is defined as a ratio with values between 0 and 1. A low Gini coefficient indicates more equal income, while a high Gini coefficient indicates a more unequal distribution. A zero Gini coefficient corresponds to perfect inequality (everyone has the same income); whereas a coefficient of one corresponds to perfect inequality where one person holds all income. The Gini coefficient requires that no one has a negative net income or wealth.

The Gini ratio is defined as the ratio of the different areas in the Lorenz curve diagram. If the area between the line of perfect equality and Lorenz curve is A, and the area under the Lorenz curve is B, then the Gini coefficient is A/(A+B). Since we know that the total area below the perfect equality line of the Lorenz curve diagram, A+B = 0.5, then, the Gini ratio, G = A/(0.5) = 2A = 1-2B. If the Lorenz curve is represented by the function Y = L(X), the value of B can be found by integration:

(4)

The simplified form of the formula for Gini coefficient is,

(5)

Where Xi = Cumulative percentage of population and Yi= cumulative percentage of income per capita under study. The Gini index is the Gini coefficient expressed as a percentage, and it is equal to the Gini coefficient multiplied by 100. The Gini index is more widely used. The status of the income and wealth distribution among the sample population will be constructed and the Gini ratio calculated so that the results may be generalized for policy suggestions.

3. Income and poverty regressions

Ordinary least-squares (OLS) regression analyses were exemplified by Xxxxxx (1967), who searched for the causes of poverty in the United States by using a multiple regression model. The percentages of families in poverty, living on farms, nonwhite, with no workers, and headed by those with less than eight years of schooling were hypothesized to be the key independent variables that explained poverty. An index of industrial structure and dummy variables for two states were also added to the equation. The results showed that a weak labor force participation rate, low quality workforce, lack of off-farm employment opportunities, and inadequate industrial structure explained the incidence of poverty.

We first conducted simple regressions to explain income per capita, the incidence of poverty and the dynamic change income as a function of distance and transportation-related employment as well as other economic and dummy variables.

In Mawlamyine, a double truncation regression model with poverty dummy variables for other basic (health, education and clothing) expenditure was run. Instead, the present article relies largely upon double-truncated regression analysis, the use of which is fairly innovative in poverty analysis. To our knowledge, only Takayama (1979) has employed a censored income distribution truncated from above by the poverty line to overcome the theoretical and practical limitations of the Gini coefficient and to transform the latter into a measure of poverty. Takayama felt that a similar approach could be applied to other measures of poverty. The truncated regression model is a kind of censored regression model in that either the low or “left” end, the high or “right” end, or both can be cut to the natural extreme values of the actual sample (Franses and Paap, 2004). This should theoretically provide a better fit and higher explanatory value to the model.

We also employed econometric models of the Logit and Probit classes to explore the conditions of inequality which determine which households are poor (1) vs. not poor (0) with respect to the poverty line, while identifying the characteristics of the poor and non-poor.

Probit and Logit models are the most widely used members of the family of generalized linear models in the case of binary dependent variables. Such models constitute an appropriate technique to determine the probability of a household remaining poor. This class of models explains the behavior of dichotomous dependent variables that can assume a value of one or zero depending upon the particular specification adopted; they compute the conditional probability of the dependent variable being equal to one (Xxxxxxx, 2010; Xxxxxxxx& Xxxxx, 2008).

The dependent poverty variable can be defined in multiple ways including the income, expenditure, and calorie intake methods. In this research, the dependent variable is a dummy variable with “poor” households living below 1.25 US dollars per day per person (1) distinguished from “non poor” households living above 1.25 dollars (0).

Lemieux (2012) introduced binary response models: Probit, Logit, and linear probability models following Wooldridge (2002). Consider the models:

𝐸(𝑦|𝑥) ≡ (𝑃𝑟(𝑦 = 1|𝑥) = 𝐺(𝑥𝛽) (6)

This is an index model as it constrains the effect of 𝑥 to go through the index𝑥𝛽. The function

G( ) in the linear probability model is the identity function. An obvious problem is the inconsistent

nature of the probability model whereby nothing can prevent from going below zero or above one. Therefore, Probit and Logit models are popularly employed to define probability models, where

G( ) becomes a cumulative distribution function that is standard normal in Probit models.

Lets start with the general model as follows:

(7)

Where: : denotes the dependent variable (the latent variable)

x: denotes the characteristics vector of household as binary variable (0 or 1).

β: denotes parameters vector

e: denotes the residuals (errors)

The binary variable (poor or non-poor) expression is defined as follows

(8)

The dependent variable captures a status of household as being either poor or non-poor that can be interpreted as the utility difference between choosing = 1 and 0. The probability that = 1 can be

derived from the dependent variable and the decision rule. Assuming e follows a standard normal distribution,

(9)

“The link function relating the linear predictor η= x to the expected value μ is the inverse normal cumulative distribution function, -1(μ)=η” (Hahn &Xxxxx, 2005). Therefore, equation (9) is a Probit model.

The specification of the Logit model assumes that e is a logistic cumulative distribution.

Therefore, the Logit model is given by

(10)

Unlike the Probit, the Logit approach became popular because of the closed form of the function

G( ) in the model. Thus, the Logit model assumes a logistic distribution, ln (μ/1-μ) = η. However, the estimate models are qualitatively similar in the case of Logit and Probit models.

Both the Logit and Probit are estimated by maximum likelihood estimation, which can be solved iteratively by numerical methods. Unlike the case of OLS, β in the binary choice model cannot be interpreted as the effect of x on the dependent variable y. Therefore, the marginal effects are computed to show the change in the probability when a unit changes in the independent variables. Additionally, Lemieux (2012) mentioned that marginal effects of Probit and Logit models are more likely similar and close to the linear probability model coefficients.

For selecting an acceptable model, we employed three different criteria of model adequacy: a high likelihood ratio, high McFadden R2, and high percentage of correct predictions. The likelihood ratio explains a significantly smaller deviance. The McFadden’s R2 is used to explain the percent of variation. The percentage of correct predictions is used to capture the goodness of fit on the observed data.

Using such Probit and Logit estimations, Xxxxxxx (2009) estimated the likelihood of a female household head being poor. He found that such households to be marginally poorer than male-headed households. In the same year, Xxxxxxxxx and Xx-xxxxxx (2009) determined the factors which affected the retirement preferences of Egyptian government sector employees as between early retirement and post-retirement employment. Using a Bivariate Probit model, they found that those who planned to work after their retirement tended to choose early retirement.

4. Seemingly unrelated consumption expenditure regressions

The ‘seemingly unrelated regression equations (SURE) model’ was developed by Xxxxxx Xxxxxxx in 1962. The model is a generalization of a linear regression model that explains a set of dependent variables when error variables are reasonably correlated. Separate analysis of univariate models may lead the covariate effect to inefficiently estimate the result (Xxxxx, Olubusoye, &Ojo, 2010). The model can be estimated through equation-by-equation using standard ordinary least squares (OLS) which is well known that the least squares estimators are best linear unbiased estimators by Xxxxx Markov theorem and maximum likelihood estimators when single equation normal likelihood functions are employed (Zellner, 2006). Thus, ‘seemingly unrelated regression equations (SURE)

model’ gives us the model that the estimated results are efficient in the coefficients and standard errors. Following Xxxxxx (2002), a seemingly unrelated regressions (SUR) model can be written as

𝑦𝑖 = 𝑋𝑖𝛽𝑖 + 𝜀𝑖 i=1,…,M

Where y is vector of dependent variables, X is K × K matrix, i is number of regressors, and M is the number of equations.

The disturbance formulation is

σ11I σ12I ⋯ σ1MI

E(είε′|X1, X2, … , XM) = σίjIT = Ω = [ σ21I σ22I ⋱ σ2MI] = ∑ ⊗ I

j ⋮ ⋮ ⋮

σM1I σM2I ⋯ σMMI

𝜎11 𝜎12 ⋯ 𝜎1𝑀

Where ∑ = [ 𝜎21 𝜎22

⋮ ⋮

⋱ 𝜎2𝑀 ]

⋮

𝜎𝑀1 𝜎𝑀2 ⋯ 𝜎𝑀𝑀

Σ is variance and covariance matrix. 𝛺 = 𝛴 ⊗ 𝐼with𝛺−1 = 𝛴−1 ⊗ 𝐼

The generalized regression model applies to the stacked model is written as followings,

y1 X1 0 ⋯ 0 β1 ε1

y = [ y2 ] = [ 0 X2 ⋯ 0 ] [ β2 ] + [ ε2 ] = Xβ + ε

∶ ⋮ ⋮ ⋱ ⋮ ∶ ∶

yM 0 0 ⋯ XM βM εM

The efficient estimator of SUR is the Generalized Least Squares (GLS). The GLS estimator is given as following;

𝛽̂ = [𝑋′𝛺−1𝑋]−1𝑋′𝛺−1𝑦 = (𝑋′(𝛴−1 ⊗ 𝐼)𝑋)−1𝑋′(𝛴−1 ⊗ 𝐼)𝑦

The important result is in the SUR model when the error terms are uncorrelated between the equations; truly unrelated and all equations have the same set of regressors, the efficient estimator is single-equation ordinary least squares; and OLS is the same as GLS. Hence, for linear SUR models “it is well known that the greater the correlation of the errors, the greater the efficiency gain when using SUR” (Alaba, Olubusoye, &Ojo, 2010).

5. Volatility models

Volatility models are extensively adopted to capture and predict risk in financial markets. However, these models are also successfully adopted in other sectors of the economy, for instance agricultural finance as in ref. (Xxxxx X.-X. , Xxxx, Xxxxx, &XxXxxxx, 2009). Such models are quite popular in the analysis and prediction of international tourist arrivals as in ref. (Xxxxxxx&XxXxxxx,

2005). The volatility in prices of some agricultural products is similar to that of financial volatility (Huang B.-W. , Xxxx, Xxxxx, &XxXxxxx, 2009). The volatility of prices of NR extract, namely Rubber Sheet and Ribbed smoked Sheet and Latex, also follows a similar pattern to financial volatility and is similarly influenced by petroleum prices. Uncertainties in the prices of NR adversely affect financing in NR based industries, as well as the livelihoods of small holders rubber growers.

Unit root test

It is customary to check the stationary behavior of the time series data before one works out other features of the prediction model. The economic literature widely agrees that many important time series economic variables display asymmetric adjustment paths, while the standard time series models assume linearity and symmetric adjustments. Moreover, if adjustment is approximately symmetric, the Dickey-

Fuller test is more powerful than any other test (34). A study variable series 𝑦𝑡 (RSS3 or LAT in our

case) is said to be stationary if the mean, variance, and covariance of the series remain constant over time (Xxx, XxXxxxx, & Xxx, 2009). If the 𝑦𝑡 is correlated at higher order lags, the assumption of white

noise disturbances, 𝜀𝑡 , is violated and the Augmented Dickey-Fuller (ADF) test allows us to perform a

parametric correction for higher-order serially correlated error processes. This is done by both assuming that the𝑦𝑡 series follows an 𝐴𝑅 (𝑝) process and adding 𝑝 lagged difference terms of the

dependent variable, i.e., ∆𝑦𝑡−𝑖 in the test regression model (Maddala, 1992; QMS, 2007).

The formulation of an ADF test using 𝑝 lags of the dependent variable is presented in equation (1), where ∆𝑦𝑡 is the first difference of 𝑦𝑡 and 𝑝 is the lag-length of the augmented terms for 𝑦𝑡 . The lags of ∆𝑦𝑡 capture any dynamic structure present in the dependent variables in order to ensure that 𝜀𝑡 is not autocorrelated (Xxxxxx, 2008). The unit root test is carried out under the null hypothesis 𝜑 = 0 against the alternative hypothesis of 𝜑 < 0. If the null hypothesis in the ADF test, 𝜑 = 0, then 𝑦𝑡 is said to be stationary.

∆𝑦 = 𝛼 + 𝛿𝑡𝑟𝑒𝑛𝑑 + 𝜑𝑦

+ ∑𝑝

𝜃 ∆𝑦 + 𝜀

(1)

𝑡 𝑡−1

𝑖=1

𝑖 𝑡−𝑖 𝑡

In this paper, we use ADF unit root tests for the daily price series 𝑦𝑡 of the respective variables (RS, RSS3 or LAT) over the full sample size of 1519 observations (September 2004 to September 31, 2010, excluding holidays), with and without a deterministic trend. The time trend is included in the test

in all the cases, i.e., level (𝑦𝑡), first difference (∆𝑦𝑡), and logarithm transformation of the respective variables (𝑙𝑜𝑔𝑦𝑡 and ∆𝑙𝑜𝑔𝑦𝑡 ). The ADF auxiliary regression in logarithmic form with deterministic trend is presented in equation (2).

∆𝑙𝑛𝑦 = 𝛼 + 𝛿𝑡𝑟𝑒𝑛𝑑 + 𝜑𝑙𝑛 𝑦

+ ∑𝑝 𝜃 ∆𝑙𝑛 𝑦

+ 𝜀

(2)

𝑡 𝑡−1

𝑖=1 𝑖

𝑡−𝑖 𝑡

where𝑙𝑛𝑦𝑡 is the logarithm of the respective series at time 𝑡, ∆𝑙𝑛 𝑦𝑡−𝑖 is the lagged first difference,𝜀𝑡 is the error term, and 𝛼, 𝛿, 𝜑 𝑎𝑛𝑑 𝜃 are the parameters in need of estimation.

We used Eviews package 6 for the ADF test of the unit root null hypothesis, 𝐻0 ∶ 𝜑 = 0,

against𝐻1 ∶ 𝜑 < 0. Alternatively, Xxxxxxxx and Xxxxxx proposed another non-parametric method for unit root test to overcome the problem of serial correlation in 1988. This ”PP” (Phillips-Perron) method

estimates the non-augmented DF test, and modifies the 𝑡-ratio of the 𝛼 coefficient so that serial correlation does not affect the asymptotic distribution of the test statistic. The test is based on the

statistic obtained from equation (3) (QMS, 2007)

𝑡𝛼

= 𝑡𝛼 (

𝛾0 1/2

)

𝑇(𝑓0 −𝛾0)(𝑠𝑒(𝛼⌃)

−

1/2

(3)

𝑓0

2𝑓0 𝑠

where𝛼̂ is the estimate, and 𝑡𝛼 is the 𝑡 – ratio of 𝛼, 𝑆𝑒(𝛼⌃)

is coefficient standard error, and 𝑆 is the

standard error of the test regression. 𝛾0 is the consistent estimate of the error variance defined as(𝑇 −

𝑘)𝑆2/𝑇, with 𝑘 number of regressors and 𝑓0 is an estimator of the residual spectrum at frequency zero.

Conversely, Xxxxxx and Xxxxxx (1979) have shown that regressions in first differences have little power against the alternatives of a stable near random walk model. Divino and McAleer (2010) have asserted that “it is well known that traditional unit root tests, primarily those based on the classic methods of Xxxxxx and Xxxxxx (1979, 1981) and Xxxxxxxx and Xxxxxx (1988), suffer from low power and size distortions.” Although such limitations deal with adopting other modified tests methods as suggested by Xxxxxx and Xx (1996), Xxxxxxx, Xxxxxxxxxx, and Xxxxx (1996), and Xx and Xxxxxx (2001).

These suggested modified unit root test methods are also subject to low power and size distortions under the short run persistence implied by GARCH component. However, such size distortions might be even greater for the traditional Dickey–Fuller test, despite the sensitivity of the modified tests to the degree of volatility in the GARCH process (41). Hence, we have adopted the modified ADFGLS (MADFGLS) test and the modified Phillips-Perron test, which both use generalized least square (GLS) de-trended data and the MAIC in order to choose the truncation lag (MPPGLS).

6. Social accounting matrices (SAMs)

We employed a Social Accounting Matrix (XXX) in gathering insights for development strategy formulation particularly when addressing the issues of growth and distribution. A Social Accounting Matrix is one of the most appropriate tools to analyze the current situation of an economy as well as for planning and policy making. The SAM provides a closed form, economy-wide accounting of linkages between activities (and/or commodities), factors, households, domestic institutions (e.g., investment, government), and foreign institutions in a tabular format that is transparent and amenable to multiplier analysis similar to that popularized by Xxxxxxxx (Xxxx, Xxxxx, & Xxxx, 2002).

SAMs have been broadly used in developing countries to assess the distributive effects of policies on households (Midmore& Xxxxxxxx-Xxxxxxxx, 1996). Xxxxx and Round (1985) documented several examples of SAM models that have been applied to the policy analyses of a wide range of countries.

Subsequently, Xxxxxx-Xxxxx (2000) used a SAM model to examine the aspects of labor migration from Mexico to the United States. Xxxxxxxx (2000) made use of SAM multipliers to assess the effects of agricultural growth on income and equity. Malan (2001) discussed the problem of income distribution in South Africa using a SAM. As predicted a quarter-century ago (Xxxxx 1986), SAM analysis has indeed proved useful in gathering insights for development strategy formulation, particularly when addressing the issues of growth and distribution.

SAMs also provide an ideal experimental environment to investigate the impact of infrastructure investment and have been applied in various research fields in different countries. In the US, Xxxxxxx and Robinson (1986) used a SAM for investigating the impacts of various exogenous factors on agriculture, with the focus on the relationship between the agricultural and non-agricultural sectors. Xxxxxxx (1992) used a SAM to investigate the roles of agriculture in the economic development of the UK economy. Reininga (2000) constructed a SAM for the Netherlands to examine its consistency and suitability as a database for economic policy analysis. Xxxx and Perdiz (2000) used SAM multipliers to measure the inequality among different groups of Spanish households. Xxxxxxx and Kola (2000) analyzed the effects of EU structural and agricultural policies on rural areas of different economic structures in Finland. SAMs are often built at the national level, less popular at the village level, not regularly used for the regional level, and rarely done at the provincial levels. The principal reason for this pattern is the lack of data availability. At the national level, SAMs are normally built top-down from input-output tables and national accounts, and then modified by national household and firm surveys. At the village level, SAMs are typically built bottom-up with data from micro-level household and production surveys. At the provincial level, official data are lacking because they have been aggregated into the national level while household surveys are much more costly than at the village level.

The use of optimized SAMs or XXXXx can be traced back to the linear programming method. Several researchers have used OSAMs to single out the best economic policies for various contexts. Xxxxxxx (2000) proposed a linear programming model linked to a SAM for Iran, set the equation of priorities of production activities as an objective function for planning, and specified the exact levels of resource and social constraints. Xx-Xxxxxxx and Schreiner (1993) constructed a general equilibrium model using a linear programming algorithm for the Saudi Arabian economy. They found that reducing agricultural subsidies would have a significant impact on commodity and factor prices. Xxxxxx et al. (2008) programmed the relationship between grazing permits and the rest of the economy to propose an optimal economic plan for Elko County, Nevada. These examples confirm that OSAMs are a good way to both help policy makers to manage an economy and indicate how to achieve maximum social welfare.

The basic SAM is a square matrix, the dimensions of which are determined by the institutional setting underlying the economy under consideration. Each account is represented by a combination of one row and one column with the same label. Each entry represents a payment to a row-account by a

column account. Thus, all receipts into an account are read left-to-right across the corresponding row while payments by the same account are recorded top-to-bottom down the corresponding column. A SAM which is developed from an input-output (Leontief) table, represents flows of all economic transactions that take place within an economy (regional or national). It is at the core, a matrix representation of the National Accounts for a given country, but can be extended to include non- national accounting flows, and created for whole regions or areas. SAMs refer to a single year, providing a static picture of the economy. According to Xxxxxxx (2007), SAMs add to the standard “economic” accounts of an input-output table the whole set of “social” and environment accounts; therefore SAMs allow for more complete analysis and future planning.

Table 3.3: A Basic Social Matrix (SAM)

Source: Round 2003 (Round, 2003

As further explained by Xxxxxxx (2007), before a SAM can be used for impact multiplier analysis or optimization, it must be converted from monetary terms to a macro-social linear SAM through a sequence of six steps.

Step 1 is to give a complete verbal and statistical description of the pre-policy or pre-project sectoral structure of domestic production and value added of the economy.

In step 2, the SAM is first converted into an “A” matrix of average expenditure propensities by dividing each endogenous element in the transaction matrix by its respective column sum.

Step 3 is to construct the SAM multiplier matrix to visualize the total effects of increasing the output of a given sector on all other sectors of the economy. The coefficients on the diagonal are the multipliers of each sector upon itself; the off-diagonal elements are induced production and sales through augmented Hirschman-style forward and backward linkages with the rest of the economy. The total multipliers of each column represent the sum of the production and commercial multipliers.

A SAM Multiplier is represented by the formula:

𝑌 = 𝐴𝑌 + 𝑋 (1)

whereY is a matrix of endogenous income; total value added, A is a matrix of average expenditure propensity; the proportion of expenditure of each column account to row account and X is a matrix of

exogenous accounts. In SAM structure, I is an identity matrix and(𝐼 − 𝐴)−1 is an inverse of the

matrix (𝐼 − 𝐴). Look at how an injection in one part of an economy affects other parts of the economy upon the total value added Y. So the model can be rewritten as:

𝑌 = (𝐼 − 𝐴)−1𝑋 = 𝑀𝑎𝑋 (2)

Where 𝑀𝑎is the SAM Multiplier matrix which shows how the incomes of the endogenous sectors of our matrix would be affected at the margin by a change in exogenous demand.

Step 4 is a preliminary visual interpretation of the multiplier matrix, with special attention to the sectors that interest planners most.

Step 5 is to simulate the economic impacts of spending, say, 1 million USD on each target sector as part of a general development program in the province.

A typical SAM models the interaction between production activities (sectors of an economy) and the commodities used (intermediate goods used for production). In our study, the transportation sector (trucks, cars, busses) and road infrastructure (EWEC, minor roads, bridges), will be broken out within the production activities in order to measure their multiplier effects: (1) Production factors; capital, land and labor, (2) Institutions; households, firms and government, (3) Capital accounts; financial side of an economy, and (4) Rest of the world; export, import, and other financial flows.

Last, a SAM incorporated with linear programming will be used to optimize the provincial economy subject to resource constraints and to generate the corresponding implementation plan.

The linear programming is used to solve optimization problems which minimize or maximize a linear function subject to linear constraints to determine how to achieve the best outcome. The standard form of the linear programming can be written as:

𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑉 = 𝑣 * 𝑥		(3)
𝑠𝑢𝑏j𝑒𝑐𝑡 𝑡𝑜 (𝐼 − 𝐴) ≥ 0	(4)
𝑟 ≤ 𝑏		(5)
𝑠 ≥ 𝑜𝑟 ≤ 𝑐		(6)
𝑥 ≥ 0		(7)

where V is total township value added, v is value added for each activity, x is the output level of each activity in million Kip, r is the resource constraint, s is sufficiency economy constraint, b and c are maximum or minimum permissible of each constraint. The simultaneous satisfaction of all constraints of all vectors of x in linear programming is called a feasible solution.

7. Sufficiency Economy Matrices (SEMs)

The proposed project serves as a test case of the application of an important new methodology, optimizable Social Accounting Matrices (SAMs), to the elaboration of coherent, detailed strategies for communities and entire provinces in anticipation of the opening of new roads. The SAMs will also pioneer the inclusion, not just of environmental and income distributional parameters, but also the three principles and two conditions of the King of Thailand’s sufficiency economy. Separate optimizations of the SAMs and Sufficiency Accounting Matrices (SEMs) are permitting, for the first time, the calculation of the exact benefits to be derived from the faithful application of the King’s vision to bottom-up regional development as opposed to standard international economics.

The innovations should not be underestimated. For example, it is not usual practice for SAMs to be optimized, to find the best development strategy for a community or province. In the 1970s, Thorbecke and his students at Cornell University developed social accounting matrices (SAMs); “social” matrices to “account” not only for the intersectoral impacts of growth in one productive sector but also the influence of such growth on income distribution, government revenues, and income to the holders of labour, capital, and land resources. The literature suggests, however, that it is rare to optimize either the Leontieffsubmatrix (LM) or the SAM. Vassileva and Calkinx xxxxxized a Leontieff matrix for Bulgaria`s progressive entry in the European community. Recently, Sivakuxxx xx xx (1999) published one of the few studies in the literature where the SAM is converted to a linear program and optimized to guide agro-industrial strategies for a community. Zhai axx Xang (2002) used the most recent version of the Chinese SAM to evaluate rural-urban migration and urban unemployment effects of China’s accession to the WTO. For Vietnam, Tarpet al(2002, 2003) used a SAM to analyze the structure of the Vietnamese economy in 2000, three years after the onset of the Asian financial crisis.

They found that agriculture and human capital investment through training are critical to improving national income. They also showed how reduction in trade barriers as Vietnam globalizes will affect income distribution. They relied, however, on detailed matrix decomposition and comparative multiplier analysis, rather than optimization, to arrive at their conclusions. None of the three other aggregate level SAMs constructed for Vietnam (Khan 1997, Nielsen 2001) or the Central Region (Bautista 2001) have been optimized. Only Calkins and Ngo (2004) proceeded to that final step.

These SAMs of this study are also being expanded to include ‘ethics’ and the other key concepts of the King’s Philosophy. This refers to the distribution of income, not only as Gini coefficients and decile ratios, but also the rural-urban income gap and differential returns to female and male labourers or household heads. In the rapidly shifting dimensions of the GMS economy, poverty, unemployment, Gini coefficients, exports, migration, and markets are all fragile and difficult to predict on an individual basis. Integrating all of these, a SAM can efficiently detect the most technically efficient and environmentally sustainable policies and projects driven by employment creation and poverty reduction (development); market growth (transition); and trade (globalization). More specifically, the geographical distribution of job creation and income must be specifically tracked. Furthermore, concerns for environmental ‘reasonableness’ lead us to include physical measures of environmental quality (engineering measures as cm of topsoil, BOD of waterways, and motorcycle and factory emissions) to ensure that communities do not simply barter one objective (increased PPP per capita) against another (sustainability).

CHAPTER IV:

ANALYSES OF POVERTY AND WELL-BEING

I. DA NANG

1. Income inequality and Gini coefficients

To determine poverty, we used measures of both absolute poverty and relative poverty. One way to measure the impact of the road and globalization over time that does not depend on the exact identity of households is to calculate the Gini coefficient of relative income and the Foster-Greer-Xxxxxxxxx (XXX) measures of absolute poverty.

Table 4.1: Trends in the inequality of income distribution in the Da Nang VHLSS survey samples

Year	Sample	Quintile ratio	Decile ratio	Gini coefficient
	Rural	10.50	20.51	0.385
	Urban	6.71	12.51	0.343
2002	Semi-urban	5.77	8.31	0.346
	Total	11.50	21.81	0.430
	Rural	2.85	3.83	0.198
	Urban	4.89	7.01	0.488
2004	Semi-urban	4.57	6.70	0.340
	Total	4.83	7.20	0.326
	Rural	4.28	5.34	0.278
	Urban	3.45	4.49	0.247
2006	Semi-urban	11.76	22.81	0.604
	Total	6.25	11.22	0.416
	Rural	6.04	8.34	0.385
	Urban	5.68	7.57	0.346
2008
	Semi-urban	3.21	5.11	0.228
	Total	5.44	7.45	0.336

The volatility of these ratios and the switching of relative severity between years seem very improbable. The data for each year may therefore have been collected from samples that changed too greatly. On the other hand, one can believe that the data does represent the real changes that have occurred; there are interesting implications for testing the Kuznets inverted U curve for the changes of inequality in the successive stages of the development process. The rural area goes down and then steadily upward, the urban area upward, downward and then back to its original level, and the semi- urban area is constant, undergoes a huge rise, and then drops to the lowest value of the three areas.

In terms of hypothesis testing, we may accept hypothesis five which states that “The distribution of income (Gini) for the whole sample has grown more equal over time as economic opportunities have opened to all classes of workers” Except for the year 2006, there has been a steady downward trend in both the quintile and decile ratios and the Gini coefficient for the sample as a whole. However, we must reject hypothesis 6, which holds that “By subsample, however, incomes in the rural areas have grown more unequal (i.e. higher Gini).” There are ups and downs for each subsample, and the area of inequality in the rural sample in 2008 is virtually the same as it was in 2002.

We turn now to hypotheses 7, 8, and 9 to the incidence (head count), depth (average level) and intensity (level squared) of the three subsamples. We are still calculating the values for these figures for 2008. Already however, it is clear from the data from 2006 (Table 4.2), that the incidence, depth and severity of poverty are all highest in the rural sample, followed by the semi-urban and finally the urban samples. This means that a higher percentage of rural dwellers are poor, they are on average poorer than in the other areas, and the cases of extreme poverty are also higher. Thus, at least for 2006, we may tentatively accept hypothesis 7, to the effect that “The incidence, depth and intensity of poverty are highest in the rural areas farthest from the road.”

Table 4.2: Comparative incidence, depth and intensity of poverty between the urban, semi-urban and rural samples

TOTAL SAMPLE for 2006	Incidence	Depth	Intensity
Sum across poor households	35	1512	2996439
Sum across all households	30.7%	13.3	26285
URBAN SAMPLE
Sum across poor households	9	297	622403
Sum across all households	16.7%	5.5	11526
SEMI-URBAN SAMPLE
Sum across poor households	14	465	701171
Sum across all households	38.9%	12.9	19477
RURAL SAMPLE
Sum across poor households	12	838	1917644
Sum across all households	50.0%	34.9	79902

2. Econometric equations to explain poverty and income

Table 4 3: Regression analysis to explain income/capita by transportation, 2008

Dependent variable: Income/capita, 2008	Unstandardized Coefficients		Standardized Coefficients
	B	Std. Error	Beta	t	Sig.
(Constant)	5484	2674		2.05	.043
% income from 83, 93, 91 in total income	-28043	35964	-.059	-.78	.437
% migration income in total hh income	1883	2544	.054	.74	.461
% expenditures in travel fees, excursions in Vietnam, and excursions outside of Vietnam in total expenditures	158106	38429	.292	4.11	.000
% expenditures on bicycles, automotive, motorbikes in total expenditures	-98392	189194	-.038	-.52	.604
No of hhlabour (aged 17-60)	-2170	656	-.253	-3.31	.001
Average education/hh member (= total years education / total hh members)	1818	312	.472	5.82	.000
Medical expenditures per capita	2		.305	4.29	.000
Rural, semi-urban, or urban	45	1073	.003	.04	.966
Adjusted R-squared = 0.456 F-statistic = 12.843 (.000)

The interpretation of this regression (Table 4.3) is as follows. Shares of transportation expenditures and transportation-related income are not significant determinants of the intensity of net income per capita. The distance from EWEC, as reflected in urbanity, is not significant in determining net income per capita. Nor is migration share in income a significant determinant of income. However, percentage expenditures on travel inside and outside Vietnam do contribute to net income per capita, perhaps because of knowledge and social bridging capital. As the size of the workforce (the number of workers in the household) increases, it is harder to increase income per capita; this suggests disguised unemployment. Education and medical expenditures increase income per capita, probably as a result of making labour more productive

Table 4.4: Regression analysis to explain the intensity of poverty

	Unstandardized Coefficients		Standardized Coefficients
Dependent variable: Intensity of poverty, 2008
	B	Std. Error	Beta	t	Sig.
Constant	.177	.025		7.19	.000
% expenditures on bicycles, automotive, motorbikes in total expenditures	2.492	1.941	.119	1.28	.202
% income from 83, 93, 91 in total income	.374	.343	.096	1.09	.278
% migration income in total income of hh	-.050	.025	-.176	-2.05	.043
Average education /hh member	-.014	.003	-.442	-5.21	.000
Semi_urban (0,1)	-.031	.018	-.151	-1.70	.092
Adjusted R-squared = 0.228; F-statistic = 7.688 (.000)

Already, we can see that the determinants of poverty are not the simple mirror-image or negation of the determinants of income (Table 4.4). Also, equation (c) is unstable. The results of this regression also point to three facts. First, transportation expenditures and transportation-related income are not significant determinants of the intensity of poverty. Secondly, migration, education, and living in the semi-urban area do however reduce the intensity of poverty. Finally, the intensity of poverty is significantly higher in the semi-urban areas relatively far from the EWEC.

Table 4.5: Regression to explain the dynamic change income between 2006 and 2008

Unstandardized

Coefficients

Standardized

Coefficients

Dependent variable:

Change in income, 2008/2006

Std.

	B	Error	Beta	t	Sig.
(Constant)	2.158	.315		6.84	.000
Total expenditure	.000	.000	-.312	-1.88	.067
Age of household head (years)	-.012	.007	-.277	-1.86	.070
Change_members	-.119	.041	-.488	-2.92	.006
% income from sales and marketing	-.933	.500	-.266	-1.87	.069
Total years of education of all hh members	.009	.006	.293	1.46	.151
Change_hh	.201	.141	.207	1.43	.161

Adjusted R-squared = 0.175 F-statistic = 2.631 (.030)

The equation shows that total expenditures are negatively correlated with the increase in income over a two-year period; implying that savings and investment are more important than consumption. Furthermore, the age of the household head and % income from sales a marketing are negatively correlated with improvements in income. Change in members is also negative; this shows that larger families tend to consume more than they produce - another sign of disguised unemployment. Meanwhile, total years of education promote a positive change in income, but this is not significant. Moving to the city (change in household) is not a significant determinant of the ratio in income between 2008 and 2006. The model is nonetheless unstable to taking out the non-significant variables, so we may be in the presence of multi-collinearity. We need to calculate and test other new variables.

II. SAVANNAKHET

1. Income inequality and Gini coefficients

This section of the paper analyses poverty and income distribution from the household data survey. The research employed the Gini coefficient to measure the inequality of income or wealth distribution, and the three Foster Xxxxx xxx Xhorbexxx (XXX) indices of absolute poverty to explore the relationship between household factors and income.

Table 6 shows a 243-household comparative survey in urban, semi-urban, and rural areas revealing that urban area has highest income per capita which has a wide range difference comparable with rural and semi-urban areas. This is the result of EWEC road and the growth of the business sector in the province. However, the poorest group is semi-urban households, whose average income is only slightly less than that of rural area. When considering the tercile ratios, the urban area seems to have a very bad situation, with high inequality and a wide range between groups. The rural area seems to have the lowest inequality. One possible reason could be that despite superior employment opportunities and infrastructures, the urban areas’ highly competitive labor market and population density cause economic inequality and other socio-economic problems. Policies to encourage populations to remain in urban areas might therefore constitute a viable option for reducing inter-household inequalities.

Table 4.6: Poverty and income distribution in Savannakhet

	Income/capita (million kip)	1Rank Income[1]	Tercile ratio	Rank tercile
Urban	878.81	3	5548.605	3
Semi-urban	312.55	1	3335.111	2
Rural	324.27	2	205.6	1
Total sample	505.21
2[1] We use “1” for the best or highest rank. For rank income, 1st is the highest rank which show lowest income and on rank tercile 1st is the best situation of income distribution.

To support the income distribution study, we constructed Lorenz curves for Savannakhet province overall and by sub-region

Figure 4.1: The Lorenz curves of Savannakhet and its sub regions

3a. Savannakhet 3b. Urban area

3c. Semi-urban area 3d. Rural area

We then calculated Gini coefficients to measure the inequality of income distribution overall and by sub-population (Table 4.7). When we consider the Gini index by area, the calculation points to the same direction with tercile ratio. The rural area has the best situation and the semi-urban area is the worst but the ratios of all areas are still extremely high and virtually identical. Therefore, this province clearly needs public policy to address the issue of income inequality, particularly in the semi-urban areas.

2. Incidence, depth, and intensity of poverty

Table 4.7: Gini Indices

Income	Overall	Urban	Semi-urban	Rural
Gini index	0.82	0.79	0.8	0.89
Incidence	55.56%	55.00%	60.53%	44.44%
Depth	42.86%	42.38%	47.55%	32.17%
Intensity	38.86%	38.80%	43.18%	27.02%

Source: Calculated from the survey data.

Table 4.7 also reports the FGT indices, which include the incidence (Head-count index), depth (Poverty gap), and intensity/severity (Poverty gap squared) of absolute poverty, used to gauge the extent of economic deprivation within the province with respect to the international poverty line. This equals 1.25$ per day (World Bank poverty criteria), or 3,687,503.75 Kip per year as of May 2010. The Lorenz curves are shown in figures 3a - 3d. It is clear that the semi-urban area has the highest incidence, depth and intensity of poverty, followed by the urban area. This is because the economic contribution of both industry and service sectors in Savannakhet province is growing gradually (figure 2). Population density and infrastructures in those areas cause employment opportunities to be highly competitive on labor markets, leading to income inequality and other socio-economic problems. This situation may have deepened the income gap between households whose primary occupation is agriculture compared with other occupations. While rural areas, however, the incidence, depth and intensity of poverty are the lowest. In other words, although inequality in the rural areas is high, they are not desperately poor.

3. Econometric equations to explain poverty and income

First of all, we employed the Ordinary least squares regression to see the household expenditure pattern on each expenditure while ignoring any correlation between the error terms of all four equations. Twenty variables were assumed to determine the share expenditure of food, healthcare, transportation and non-food items. In this step, we focused only on the absolute variables (income per capita, age of household head, distance of EWEC, total years of education of all members, household size, landholding per capita, living space per capita and total workers) and ignored dummy variables and share of income by source in order to reflect any significant impact on household expenditures. Moreover, we assumed the distance of EWEC played an important role in determining households’ food, health, transportation, and non-food item expenditures.

Table 4.8: The ordinary least square regression results for household expenditures

Variable		Independent variable
dependent variable	FOOD_EXP	HEALTH_EXP	TEANS _EXP	NON FOOD_EXP
C	0.59***	0.03*	0.27*	0.10
C	(0.14)	(0.09)	(0.16)	(0.08)
Y	0.000	0.000	0.000	0.000
Y	(0.000)	(0.000)	0.000	(0.000)
AGE	-0.001	0.000	0.002	-0.001
AGE	(0.003)	(0.000)	(0.003)	(0.008)
DISTANCE	-0.007**	-0.000	0.004	0.004*
DISTANCE	(0.003)	(0.001)	(0.004)	(0.002)
EDU	0.000	0.000	-0.000	0.001
EDU	(0.001)	(0.000)	(0.002)	(0.001)
HHSIZE	-0.04**	-0.006**	0.05***	-0.01
HHSIZE	(0.07)	(0.002)	(0.02)	(0.01)
LAND	0.01	-0.003	-0.01	0.001
LAND	(0.01)	(0.002)	(0.02)	(0.009)
SPACE	0.000	0.000	-0.000	0.000**
SPACE	(0.195)	(0.000)	(0.000)	(0.000)
TOTAL_WORKER	0.02	0.003*	-0.03**	0.02**
TOTAL_WORKER	(0.01)	(0.002)	(0.01)	(0.01)
SHARE_AGRI	0.03	-0.004	-0.009	-0.007
SHARE_AGRI	(0.05)	(0.008)	(0.064)	(0.034)
SHARE_BUSINES	0.04	-0.006	-0.10	0.08
S	(0.10)	(0.014)	(0.12)	(0.06)
SHARE_FIN	-0.07	0.03***	0.002	0.04
SHARE_FIN	(0.06)	(0.01)	(0.07)	(0.04)
SHARE_HHINDUS	-0.02	-0.009	-0.07	0.12**
SHARE_HHINDUS	(0.09)	(0.013)	(0.11)	(0.06)
SHARE_SERVICE	0.12*	0.005	-0.19**	0.06
SHARE_SERVICE	(0.06)	(0.009)	(0.07)	(0.04)
FEMALE	0.01	0.005	-0.05	0.04*
FEMALE	(0.04)	(0.005)	(0.04)	(0.02)
URBAN	-0.04	0.006	0.02	0.01
URBAN	(0.04)	(0.006)	(0.05)	(0.03)
RURAL	0.21*** (0.08)	0.009 (0.010)	- 0.24*** (0.09)	0.02 (0.05)
EM_AGRI	0.03	0.004	-0.05	0.01
EM_AGRI	(0.05)	(0.007)	(0.06)	(0.03)
EM_COMMERCE	-0.03	-0.004	0.10	-0.07
EM_COMMERCE	(0.08)	(0.010)	(0.01)	(0.05)
EM_HHINDUS	0.24**	0.001	-0.34**	0.10
EM_HHINDUS	(0.12)	(0.068)	(0.14)	(0.07)
EM_HIRED	0.13**	-0.01	-0.10	-0.04
EM_HIRED	(0.06)	(0.01)	(0.07)	(0.03)
EM_SERVICE	0.002	-0.02*	0.06	-0.02
EM_SERVICE	(0.092)	(0.01)	(0.11)	(0.06)
R-squared	0.18	0.15	0.19	0.14
Adjusted R-squared	0.10	0.07	0.11	0.06
S.E. of regression	0.26	0.04	0.31	0.16
Akaike info criterion	0.27	-3.72	0.59	-0.71
Schwarz criterion	0.59	-3.40	0.92	-0.39

Note: (1) the result show the coefficient and the standard error is in the parenthesis. (2) *significant at 1%, **significant at 5%, ***significant at 10%.

Table 4.8 shows the ordinary least square regression results of all expenditures. The variables that “significantly affect food expenditure” are the distance from the EWEC and household size. Healthcare expenditures show that household size and total workers “significantly affect healthcare expenditure.” The transportation expenditures show the same significant variables as for healthcare expenditures. The distance from the EWEC, living space per capita, and total workers are “significant in determining non-food items expenditure.”

As the next step, we used the Wald coefficient test to determine whether or not those variables that were not significant in the models should be taken out of the model. This test enables us to add a set of variables to an existing equation and to ask whether the set makes a significant contribution to explaining the variation in the dependent variable. We set up the coefficient of those insignificant variables equal to zero. The null hypothesis was that the additional set of regressors was not jointly significant; which means the coefficients would be equal to zero. Whenever those variables are not jointly significant (Quantitative Micro Software, LLC, 2004), we may omit them from each model. The results of the Wald coefficient test are shown in table 4.9.

Table 4.9: Wald coefficient test result.

Wald Test: Test Statistic	Valu e	df	Probability
Equation: FOOD_EXP F-statistic	0.75	(6, 216)	0.61
Chi-square	4.47	6	0.61
Equation: HEALTH_EXP F-statistic	0.33	(6, 216)	0.92
Chi-square	2.01	6	0.92
Equation: TRANS_EXP F-statistic	0.95	(6, 216)	0.46
Chi-square	5.73	6	0.45
Equation: NONFOOD_EXP F-statistic	0.23	(5, 216)	0.95
Chi-square	1.14	5	0.95

The Wald coefficient test results showed a high level of probability that prevented us from rejecting the null hypothesis that the insignificant variables were indeed equal to zero. The probabilities of the F-statistic and Chi-squared were not significant at the 5% level. As a result, we cannot reject the null hypothesis that the additional set of regressors are not jointly significant. Thus, we can omit those variables which are not significant in the models.

III. Hat Yai

1. Income inequality and Gini coefficients

In accordance with household survey, the monthly average household income is highest in the urban area (THB 39,188) followed by rural area (THB 38, 609), while lowest in the semi-urban area (THB 33,294). The average household size is 3.73 persons with monthly per capita income of THB 11,433 (Table 4.10). Interestingly, the monthly average household income is greater in the rural area than the semi-urban areas. This might be due to good price of Para rubber and other agricultural commodities complimented by high accessibility option of the Asian highway and its network.

Table 4.10: Descriptive statistics of the household level income, Songkhla (THB)

Source: Household Survey 2010

Additionally, the multiple comparison Tukey HSD test also confirms that the mean household income is significantly different in accordance with location type (Table 4.11). However, the mean income of households between rural and urban and rural and semi-urban area is not significantly different. Therefore, we can argue that good accessibility along the Asia Highway and its network might cause to fetch the higher prices of agricultural commodities in the rural area and consumer surplus due to easy mobility of factors of production in lower cost.

Table 4.11: Multiple Comparisons Tukey HSD test for household income by location

The official poverty line for the Songkhla province is THB 1,654 per person per month (55 THB/day, approx. US $ 1.75) in the year 2010 (Thai year 2553). This poverty threshold is quite higher than the international poverty measurement line, i.e., US $ 1 or 1.25 per day. As per the given poverty threshold, the incidence of the poverty in the survey household data is very low and the headcount poverty ratio is 1.17%. The official statistics of the proportion of poor in Songkhla is 2.01% in 20103. Therefore, our result is comparable and since we concentrated along the Asia High way, and the incidence of poverty might be lower than the other part of the province. However, the income inequality in the province is higher, and the income inequality ratio of the overall household sample is more than 36%.

Table 4.12: Income inequality among the households by location, Songkhla

Data source: Household Survey 2010

The income gap among the sample household is highest in urban areas followed by semi-urban areas. The Gini ratio for the former is more than 44% and more than 36% for the later. The income inequality in the rural area is considerably low as compared to the urban and semi-urban locations (Table 4.12). Therefore, policy needs to address the issue of income inequality in Songkhla.

IV. MAWLAMYINE

1. Income inequality and Gini coefficients

A given level of absolute poverty may or may not signal a problem of inequality in income distribution. In order to determine the level of inequality among our survey units, we have drawn the Lorenz curves (Figure 4.2) and calculated the Gini coefficient for our rural, semi-urban and urban subsamples. From these results, we may conclude that there is much greater inequality in income and consumption levels in the urban areas than in the other two areas of Mawlamyine Township. In contrast, rural areas have greater equality than urban areas under both income and expenditure criteria.

3Source: xxxx://xxxxxx.xxxxx.xx.xx/XxxxxxXxxx

Although inequality under the income criterion in rural areas outpaces the semi- urban areas, the reverse is true under the consumption criterion.

Figure 4.2: Consumption- and income-based Lorenz curves of Mawlamyine.

Semi-urban wards are just adjacent to the urban wards, yet poverty and income-based income inequality are much higher than in the rural areas. One reason might be that such semi-urban areas as ThiriMingala, Tharyaraye and Zayyarthiri quarters were first settled in 1989, Hlaingtharyar in 1991 and Zayyarmyaine in 2004 by families resettling from other areas of the country, leading to lower levels of economic development. As those semi urban areas have been only recently opened up, public utilities such as water and electricity systems are not yet well developed. This evidence fails to reject

hypothesis 2, which states that Income inequality is highest in the urban areas because of the influx of migrants from rural areas and lowest in the rural areas where it is difficult for the remaining population to amass wealth. The policy implication for this is that well-published relief programs should be aimed at the urban poor to reduce the income gap and social unrest in metropolitan areas.

2. Incidence, depth, and intensity of poverty

The estimated incidence in Mawlamyine (Table 6) of 33% of poverty under per capita income computation and 59% under per capita consumption are far worse than the official 2005 figures for Mon State. Assuming those government figures to be accurate and calculated using comparable methodologies across years, our 2009 findings demonstrate that the HCR has increased, as the ADB has surmised (ADB, 2010). This is all the more dramatic in that Mon State is one of the States and Divisions with less than average poverty.

Table 4.13: Poverty under per capita income and expenditure

Income-based poverty				Consumption-based poverty
Sample Population	Incidence	Depth	Intensity	Incidence	Depth	Intensity
Urban	22%	0.040	0.010	56%	0.143	0.052
Semi-urban	44%	0.114	0.040	67%	0.220	0.088
Rural	28%	0.088	0.042	47%	0.171	0.078
Entire	33%	0.083	0.03	59%	0.184	0.074

Source: Authors’ calculations for 2009.

Moreover, the semi-urban area seems to have the highest incidence of poverty, followed by the rural area for per capita income and urban for per capita consumption. Clearly, a differential dynamic in poverty incidence by area is at work. Nor is this differential dynamic new. In 2004-05 on average, only 22% of the population of Mon State (21 % rural, 23 % urban) was estimated to be under the poverty line. However, the countrywide rural poverty ratio had increased significantly from 22% to 32%.

We are now in a position to reject hypothesis 1, to the effect that Education and proximity to the EWEC have reduced the incidence, depth and intensity of poverty over time in rural, semi-urban, and urban areas despite the negative effects of the 2008 Hurricane Nargis. It seems rather that natural disasters and the unrealised potential of development have led to an increase in rural and semi-urban poverty levels. The policy implications from this rejection of hypothesis 1 are that the government should invest both monetary capital and management supervision to increase the road infrastructure and educational institutions. The semi-urban areas with the highest incidence of poverty should be given priority targeting for anti-poverty programs. In the meantime, in-kind aid such as a special food

scheme policy should be seriously considered for poor families with a high severity of poverty. Adequate caloric sources should be given first and then, if necessary, small protein supplements.

3. Econometric equations to explain poverty and income

Having measured the levels of absolute poverty and relative inequality in Mawlamyine, we now use our survey data to determine which variables and conditions most significantly determine income per capita and the intensity of poverty. Table 4.14 reports an OLS regression to identify the determinants of income per capita in the Mawlamyine area.

Table 4.14: Factors that explain income per capita

	B	Std. Error	Beta	t-statistic	Sign.
(Constant)	398405	200078		1.99	.047**
Percent income from small-scale merchandising and sales	1217452	221622	.28	5.49	.000***
Dependency ratio	223444	64264	.18	3.48	.001***
Household females	-207742	65815	-.18	-3.16	.002***
Total household education (years)	9364	3849	.14	2.43	.015**
Distance from EWEC (km)	30656	93880	.02	0.33	.744
Adjusted R-squared				.112
F-statistic				10.137	.000***
Degrees of freedom				359

*** = significant at the .01 level or better, ** at the .05 level and * at better than .10.

Income per capita in Mawlamyine area is significantly increased by the percent income from small scale merchandising jobs. Government could therefore consider promoting sales and jobs in that area. In contrast to the previous literature, however, the dependency ratio bears a positive effect on family income per capita. This is consistent with the Myanmar tradition whereby the family supports both elderly parents and children living with grandparents when parents are working at border areas and must send money back for their children’s school fees. In addition, the more females in the family, the less the income per capita is. This finding is consistent with those of Kyaw and Routary (2006) reported earlier. Apart from those females living in metropolitan areas and educated females, females do not usually travel to seek income but simply remain with parents or guardians. It is not surprising that having a greater number of females lessens family income per capita. Here our regression further shows that education is important to increasing family income, but that when all these other variables are taken into account, distance front the EWEC is no longer significant.

Table 4.15: Multiple regression equation to explain the intensity of poverty per capita

	B	Std. Error	Beta	t	Sig
Mawlamyine-employed females	.019	.004	.287	4.316	.000***
Migrant females	.031	.010	.160	2.993	.003***
Mawlamyine-employed males	.010	.004	.174	2.558	.011**
Migrant males	.011	.007	.098	1.668	.096*
Motorcycle ownership (1, 0)	-.020	.008	-.136	-2.304	.022**
Dummy variable for urban areas	-.020	.009	-.139	-2.241	.026**
Distance from EWEC (km)	.005	.005	.071	.993	.321
Adjusted R-squared				.186
F-statistic				13.181	.000***
Degrees of freedom				365

Of course, what causes some households to suffer extreme poverty is not always the opposite of what causes other people to have extreme wealth. We therefore estimate in Table 4.15 the causes of the intensity of poverty as the dependent variable in an OLS regression equation. Since the intensity-of- poverty dependent variable gives most weight to the poorest in society, it comes as no surprise that having to resort to migrant female workers significantly drives up the intensity of poverty. Surprisingly, the same problem affects female workers at Mawlamyine, male workers in Mawlamyine and, to a lesser degree, even migrant males. Since it is observed that most migrants from Mawlamyine are blue collar workers working at border and factories in Thailand, migrants of either gender should reduce, not intensify, poverty unless of course migration is unsuccessful.

Taken together, we can reject hypothesis 3, to the effect that the intensity of income poverty is increased by living in large female-headed households with a high dependency ratio, high distance from the EWEC, living in rural areas, and reliance upon casual jobs. In fact, none of these variables are significant in the poverty equation of Table 4.15 and only rurality is significant. Meanwhile, female head is significant only in the income equation (Table 4.14); the dependency ratio bears the wrong sign in the income equation, and employment, whether in migration or in Mawlamyine for both genders, seems paradoxically to increase the intensity of poverty.

Table 4.15 also leads us to reject hypothesis 4, that the intensity of income poverty is reduced by education, jobs in transportation, metal-working/blacksmithing, and living in urban areas. Only motorcycle ownership and urban residence significantly reduce poverty, and neither of these two variables were specified in the hypothesis. Motorcycles are not only a means of commuting, marketing, and migration; they are also the principal asset in operating a cycle taxi. This suggests a possible subsidy policy for motorcycle ownership in rural and semi-urban areas. Finally, the non-significance of the distance from EWEC variable strengthens our conclusion that the EWEC has not lived up to its job- and growth-creating potential.

Two types of policy implications flow from our analyses of hypotheses 3 and 4. First, transportation and mobility-related job creation should form the nucleus of the anti-poverty thrust in Mawlamyine. If people living in rural and semi-urban areas can be partially subsidised to afford motorcycles, they can undertake jobs in the business, sales, transportation and other services sectors. Otherwise, the entire positional advantage of Mawlamyine as a major transit and border trade city at the terminus of the EWEC will remain unrealised.

Secondly, adult education will also be necessary. Technical training for landing non- agricultural, non-manufacturing jobs could be a powerful government policy to reduce the intensity of destitution below the poverty line. Such a policy appears especially urgent since in the results of our firm and household surveys, the business and transportation sectors were far from flourishing in 2009. This is particularly true in the case of females, who must be equipped with knowledge to seek higher paying jobs outside the household, notably in the transportation, business and service jobs. Although metalworking jobs are not currently a significant force in reducing the food Engel coefficient, these and other traditional cultural employment within the handicraft sector should be maintained against the day when jobs in building the transportation infrastructure will no longer be available. A combination of wage-rate subsidies and low-interest loans may be envisaged to promote investment and job-seeking in these areas.

Three policy conclusions flow from our analyses of hypotheses 5. First, consistent with the writings of Xxxxxx and Xxxxx (2003), rural-based employment and welfare improvement programs should precede any effort to promote rural-urban migration. Families should be encouraged to remain in rural areas, and to unite larger family units under a single roof. Education and female migration can then be used to achieve a healthier balance among food and non-food basic necessity consumption.

Second, public healthcare institutions and personnel should be upgraded. Based on our survey experience, most of the younger generation from rural areas are paying inadequate attention to education and healthcare. Third, within non-food necessities, health care plans for dependents should be checked. The factor that a high dependency ratio induces families to spend more on education, health, and clothing is especially serious because, health institutions and personnel are already lacking in the public institutions to which the poor would normally go. Investments in the public health sector would therefore be a “pro-poor” policy. The above results can guide actions to reduce the twin problems of inequality and poverty in the near future.

Table 4.16: The effect of poverty on other basic expenditures

Dependent= Engel coefficient for basic needs		a) Before Wald test				b) AfterWald test
Variable	Coefficient	Std. Err.	z-Stat.	Prob.	Coefficient	Std. Err.	z-Stat.	Prob.
Constant intercept term	0.11	0.03	3.86	0.000	0.07	0.02	4.91	0.000
POVERTY
Intensity poverty in expenditure	-0.11	0.06	-2.09	0.037	-0.13	0.05	-2.66	0.008
OTHER ECONOMIC STATUS
Value of remittance (kyats)	0.00	0.00	2.49	0.013	0.00	0.00	1.66	0.097
Own motorcycle (1 = yes, 0 = otherwise	-0.02	0.01	-1.55	0.121
EDUCATION
Total years of education	0.00	0.00	5.52	0.00	0.00	0.00	5.09	0.000
Know about EWEC (Yes=1, otherwise =0)	-0.03	0.01	-2.06	0.04	-0.02	0.01	-1.84	0.066
EMPLOYMENT
Get tips from clients or superiors	0.11	0.04	2.91	0.004	0.11	0.04	2.87	0.004
Do health-care services (1,)	0.20	0.09	2.25	0.025	0.19	0.09	2.15	0.032
Grow fruit and vegetables (1,0)	0.06	0.03	1.96	0.05	0.07	0.03	2.10	0.036
Casual jobs	-0.02	0.01	-1.87	0.061	-0.02	0.01	-1.44	0.151
Migrant females	0.02	0.01	1.74	0.082	0.02	0.01	1.66	0.097
Migrant males	-0.01	0.01	-1.13	0.257
CONSUMPTION BEHAVIOUR
HH savings rate (% savings/total income)	-0.02	0.01	-1.62	0.105
Value of self-consumed products per/ yr(kyats)	0.00	0.00	-0.91	0.361
HH STRUCTURE/SOCIODEMOGRAPHY
Dependency ratio	0.02	0.01	3.56	0	0.02	0.01	3.98	0.000
Distance from EWEC	-0.02	0.01	-2.02	0.043	-0.02	0.01	-1.90	0.057
Rural	0.03	0.02	1.79	0.073	0.02	0.02	1.38	0.168
Age of household head	0.00	0.00	-1.29	0.198
Female household head (1,0)	0.01	0.01	0.49	0.628
Urban	0.00	0.01	-0.15	0.884
Log likelihood	375.57				370.49
Akaike information criterion	-1.9644				-1.975

CHAPTER V:

THE CONSTRUCTION OF SOCIAL ACCOUNTING MATRICES

1. Establishment of the matrix

In this part, we have constructed and employed a Social Accounting Matrix (XXX) to gather insights for economic development strategy and policy formulation. We applied and compared investment multipliers to visualize the total effects of increasing the output of a given sector on all other sectors of the economy and linear programming to optimize the provincial economy subject to resource constraints and provide the corresponding implementation plan. All these methods were combined to propose a strategic plan for maximizing the net developmental benefits from the EWEC for the Savannakhet economy. Table below summarizes structure and impacts of the optimal solution under the seven scenarios. Scenario 1 is The “Benchmark Scenario” that reproduced the actual total values of each activity column according to our 2010-

11 dataset. The value added of the improvement in provincial product in Scenario 1 was 9,506,019.13 million Kip. The value added of the transportation was 4,515,357.39 million Kip and the share of value added of the agriculture sector was 4.82%.

Next, for scenario 2 we added a labor constraint following Classical theory which assumed that the economy tends to full employment equilibrium. This optimization was called the “Neoclassical” labor-constraint model. The optimal level of this scenario was 9,512,681.09 m Kip, or a barely perceptible increase of 0.04% of current NPP; the value added of transportation and capital and financial sector increased slightly. The constraint’s shadow prices showed how much the objective function of net provincial product (NPP) would increase by one unit of amount of resources available. The shadow price reveals that adding one more laborer to the Savannakhet economy would give a marginal value of 2.90 m Kip.

Third, scenario 3 involved a “Human Capital” labor-with-education constraint model designed to increase labor productivity. The optimal level of this scenario was 12,366,475.17 m Kip, or an increase of 30.09% over current NPP. Others sector also went up 30% on average in optimal production level. The results were consistent with the King’s philosophy principle of “knowledge” because labor cannot be fully used without an equal increase (30%) in education or training programs. The labor constraint and educational investment constraint were binding, and their shadow prices reflected their marginal value to Savannakhet economy of adding one extra laborer and taking away 1 m Kips of education investment.

Scenario 4 -- the “Growth with Equity” human capital plus income-to-poor constraint model – was used to gauge the extent to which planners could help disadvantaged household groups to achieve a higher level of human development. Poor families focus on absolute cash-in-pocket rather than relative income. This social concern is consistent with the King’s principles of a) “ethics” and b) “balance,” not to mention c) “self-immunization” of the higher classes and government against social unrest. The optimal level was 12,366,475.17 m Kip, or an increase of 30.09% over current NPP. The total output of each sector also went up by 30% on average. The existing of 30% unemployed labor was still accompanied by an equal increase of 30% in education/training programs. The shadow prices display the result as same as previous scenario. Then, both labor and education contribute to economic product in Savannakhet province.

Scenario 5 was the “Sufficiency Economy” growth-with-equity plus reduction-of-“bads” constraint model. In this scenario we applied a subset of the King’s principles to the model. We required that the consumption of “bad” habits be reduced by 5%, which must be translated into a 5% increase in educational investment (now 35% instead of 30%). That constraint further was presumed to enhance worker health, leading to an increase in effective labor availability if 5%. This reasoning is consistent with the King’s philosophy principle of Self-immunization. Unfortunately, the searches for an optimal solution under these conditions proved infeasible because the SAM structure assumes that fixed proportion s of income are spent on each item.

Table 5.1: Summary table of the structure and impacts of the optimal solution under the 7 scenarios

Current values	Summary of changes in	Scenario 1		Scenario 2		Scenario 3		Scenario 4		Scenario 6		Scenario 7
Current values	Summary of changes in	Simulation	% error	Optimal value	% of current	Optimal value	% of current	Optimal value	% of current	Optimal value	% of current	Optimal value	% of current
IMPROVEMENT IN PROVINCIAL PRODUCT
9,506,019	Value added	9,506,019	0.00%	9,512,681	0.07%	12,366,475	30.09%	12,366,475	30.09%	13,308,220	40.00%	13,308,411	40.00%
EWEC DEVELOPMENT
4,515,357	Transportation	4,515,357	0.00%	4,516,920	0.03%	5,871,993	30.04%	5,871,993	30.04%	6,321,396	40.00%	6,321,492	40.00%
3,281,038	Employment	3,281,038	0.00%	3,281,038	0.00%	4,265,350	30.00%	4,265,350	30.00%	4,593,454	40.00%	4,593,454	40.00%
7,609,405	Capital &financial	7,609,405	0.00%	7,616,026	0.09%	9,900,825	30.11%	9,900,825	30.11%	10,652,824	40.00%	10,653,142	40.00%
INCOME REDISTRIBUTIONS
585,589	Rural hh	585,589	0.00%	585,614	0.00%	761,298	30.01%	761,298	30.01%	819,820	40.00%	819,824	40.00%
2,314,037	Semi-urban hh	2,314,037	0.00%	2,314,097	0.00%	3,008,326	30.00%	3,008,326	30.00%	3,239,621	40.00%	3,239,649	40.00%
915,977	Urban hh	915,977	0.00%	915,982	0.00%	1,190,777	30.00%	1,190,777	30.00%	1,282,359	40.00%	1,282,367	40.00%
678,197	Bottom 20% income	678,197	0.00%	678,197	0.00%	881,656	30.00%	881,656	30.00%	949,476	40.00%	949,476	40.00%
1,359,086	Bottom 40% income	1,359,086	0.00%	1,359,085	0.00%	1,766,811	30.00%	1,766,811	30.00%	1,902,720	40.00%	1,902,720	40.00%
KING'S BALANCE/SUFFICIENCY
5,052,343	Value of social "bads"	5,052,343	0.00%	5,061,239	0.18%	6,579,602	30.23%	6,579,602	30.23%	7,075,407	40.04%	7,073,440	40.00%
15,723	Donations & religious expenses	15,723	0.00%	15,713	-0.06%	20,427	29.92%	20,427	29.92%	22,012	40.00%	22,012	40.00%
	SOCIAL ADVERTISING (Each household group are limit on Engel expenditure transfer from tobacco/alcohol to education (mi het mi pon))
	RHH1									-190,367		-178,817
	RHH2									-6,871		-88,504
	RHH3									-265,662		-254,112
	RHH4									-6,654		-85,474
	RHH5									-11,632		-155,167
	SHH1									-42,198		-583,081
	SHH2									-409,694		-398,144
	SHH3									-805,355		-793,808
	SHH4									-63,192		-877,009
	SHH5									-541,386		-529,857
	UHH1									-11,472		-152,928
	UHH2									-443,496		-431,946
	UHH3									-16,365		-221,420

Table 5.1: Summary table of the structure and impacts of the optimal solution under the 7 scenarios (conts.)

SOCIAL ADVERTISING (Each household group are limit on Engel expenditure transfer from tobacco/alcohol to education (mi het mi pon))

UHH4

-202,339

-190,790

UHH5

-16,801

-227,534

Current Summary of changes Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 6 Scenario 7 values in Simulati Optimal % of Optimal % of Optimal % of Optimal % of Optimal % of

on % error value current value current value current value current value current

CHANGES IN SECTORAL BALANCE OF OPTIMAL PLAN

4.82%

% VA from primary

sector

4.82%

0.00%

4.82%

-0.05%

4.82%

-0.05%

4.82%

-0.05%

4.82%

0.00%

4.82%

0.00%

6.57%

% VA from industrial

sector

6.57%

0.00%

6.56%

-0.05%

6.56%

-0.05%

6.56%

-0.05%

6.57%

0.00%

6.57%

0.00%

4.61%

% VA from "bads"

4.61%

0.00%

4.70%

1.85%

4.70%

1.85%

4.70%

1.85%

4.61%

0.03%

4.61%

0.00%

2.88%

% VA from

transport/EWEC

2.88%

0.00%

2.87%

-0.06%

2.87%

-0.06%

2.87%

-0.06%

2.88%

0.00%

2.88%

0.00%

7.07%

% VA from tourist/services

7.07%

0.00%

7.06%

-0.06%

7.06%

-0.06%

7.06%

-0.06%

7.07%

0.00%

7.07%

0.00%

67.20%

% VA from construction

67.20%

0.00%

67.15%

-0.09%

67.15%

-0.09%

67.15%

-0.09%

67.20%

0.00%

67.20%

0.00%

1.74%

% VA from business

1.74%

0.00%

1.74%

-0.07%

1.74%

-0.07%

1.74%

-0.07%

1.74%

-0.01%

1.74%

0.00%

1.18%

% VA from financial

sector

1.18%

0.00%

1.18%

-0.07%

1.18%

-0.07%

1.18%

-0.07%

1.18%

-0.01%

1.18%

0.00%

3.10%

% VA from wholesale/retail

3.10%

0.00%

3.09%

-0.38%

3.09%

-0.38%

3.09%

-0.38%

3.10%

0.00%

3.10%

0.00%

0.79%

% VA from public

utilities

0.79%

0.00%

0.79%

-0.21%

0.79%

-0.21%

0.79%

-0.21%

0.79%

0.00%

0.79%

0.00%

0.05%

% VA from

health/education

0.05%

0.00%

0.05%

-0.07%

0.05%

-0.06%

0.05%

-0.06%

0.05%

-0.28%

0.05%

-0.02%

0.01%

%VA from welfare

sector

0.01%

0.00%

0.01%

-0.13%

0.01%

-0.13%

0.01%

-0.13%

0.01%

0.00%

0.01%

0.00%

Document Metadata

Table of Contents

โครงการ “TRADE, TOURISM AND SUSTAINABLE GROWTH IN 4 PROVINCES ALONG THE EAST- WEST CORRIDOR LINKING MYANMAR, THAILAND AND VIETNAM”

Document Metadata