The impact of environmental, social, and governance (esg) factors on foreign direct investment (fdi) a case from asean countries (khóa luận tốt nghiệp Đại học)

MINISTRY OF EDUCATION AND TRAINING STATE BANK OF VIETNAM HO CHI MINH UNIVERSITY OF BANKING GRADUATION THESIS THE IMPACT OF ENVIRONMENTAL, SOCIAL, AND GOVERNANCE ESG FACTORS ON FOREIGN

INTRODUCTION

Investment plays a crucial role in driving sustainable economic growth within a country It is vital for developing infrastructure, enhancing labor productivity, and fostering technological innovation While domestic capital accumulation can be a lengthy process, particularly for large-scale initiatives, attracting foreign direct investment (FDI) offers a quicker pathway for countries to access essential capital and advanced technology.

Foreign Direct Investment (FDI) plays a crucial role in enhancing economic growth by supplementing domestic capital, which fosters industrialization and modernization It introduces advanced production technologies and modern management practices, leading to improved labor productivity and product quality Additionally, FDI generates new job opportunities and increases workers' incomes, contributing to poverty reduction and enhanced living standards However, excessive dependence on FDI can lead to potential risks, as multinational corporations may focus on short-term profits at the expense of environmental and social well-being Consequently, it is essential for governments to implement effective policies that manage FDI inflows while safeguarding national interests Notably, FDI inflows into developing regions, particularly Southeast Asia, have experienced significant growth, with an 85% increase from 2010 to 2021, reaching a remarkable 208.5 billion USD.

In 2021, Singapore emerged as the leading recipient of foreign direct investment (FDI) in the ASEAN region, attracting an annual average of 72.7 billion USD (UNCTAD, 2021) Understanding the determinants of FDI is essential for the ASEAN-6 countries—Vietnam, Indonesia, Malaysia, the Philippines, Singapore, and Thailand—due to their socio-economic similarities, including a young population, a plentiful labor force, and advancing infrastructure By identifying these key factors, these nations can harness their significant potential for development.

2 help the Governments of the six member countries formulate appropriate policies to maximize their potential

Figure 1.1 FDI in Developing Regions

(Source: the author’s summary from UNCTAD data)

In the face of escalating climate change, investors are increasingly recognizing the significance of sustainable development and sustainable investment The rise of environmental, social, and governance (ESG) factors reflects a shift in focus for high-quality foreign direct investment (FDI), which now prioritizes not only financial returns but also adherence to environmental and social standards As highlighted by Mattila & Sasi (2023), ESG is crucial for evaluating and mitigating investment risks, thereby promoting sustainable development across economic, social, and environmental dimensions The Sovereign ESG score serves as a key metric for assessing a nation's sustainable development from an economic standpoint, as illustrated by the improvements in scores shown in Figure 2.

Between 2010 and 2021, the Sovereign ESG scores in the ASEAN-6 region showed notable changes, with scores in 2010 ranging from -0.06 for Vietnam to 0.12 for Singapore By 2021, these scores became more uniform, with Singapore scoring 0.33 and Indonesia and Malaysia reaching 0.44 Vietnam demonstrated the most significant improvement, increasing its score from -0.05 in 2010.

Central AmericaSouth AmericaCentral AsiaWest AsiaSouth AsiaSouth-East AsiaEast AsiaOther AfricaNorth Africa

3 to 0.42 in 2021 This indicates a growing interest in sustainable development among the governments of the six countries, thereby narrowing the ESG gap in the region

Figure 1.2 ASEAN-6 Sovereign ESG Score in 2010, 2021

(Source: the author’s summary from World Bank Data)

Figure 1.3 FDI - Sovereign ESG score of ASEAN-6

(Source: the author’s summary from World Bank and UNCTAD Data)

The author's initial observation indicates a correlation between Sovereign ESG and Foreign Direct Investment (FDI), suggesting that multinational corporations (MNCs) are more likely to invest in countries that demonstrate a strong commitment to sustainable development.

To understand how Sovereign ESG factors influence Foreign Direct Investment (FDI) inflows into ASEAN-6 countries, it is essential to conduct a research study titled "The Impact of Environmental, Social, and Governance (ESG) Factors on Foreign Direct Investment (FDI): A Case from ASEAN Countries." This study aims to clarify the relationship between ESG factors and FDI, while also proposing strategies to enhance FDI in the region.

RESEARCH OBJECTIVES

This research aims to explore the factors influencing Foreign Direct Investment (FDI) inflows, with a particular focus on the impact of Sovereign Environmental, Social, and Governance (ESG) criteria and its components—an area that has been largely overlooked in prior studies The study is designed to achieve specific research objectives that will provide deeper insights into the relationship between Sovereign ESG and FDI inflows.

(1) Systematize the theoretical basis of factors affecting FDI inflows for countries

(2) Develop and verify an empirical model based on the theoretical basis After that, analyze and evaluate the impact of each factor on FDI inflows through this model

(3) Provide recommendations to support economic policymakers and investors in making informed decisions to prioritize a more sustainable investment approach.

RESEARCH QUESTIONS

To achieve the research objectives, the study must answer the following questions:

(1) What is Foreign Direct Investment inflow? What are its impacts on a country? How to measure the factors affecting FDI inflows, including Sovereign ESG and its components?

(2) How to construct and test models to estimate the impact of each determinant on FDI inflows?

(3) Based on the findings, what policy recommendations can be made to attract FDI inflows while ensuring a country's sustainable development?

RESEARCH SUBJECT AND SCOPE

This thesis specifically studies the impact of Sovereign ESG and its components on FDI inflow of 6 countries in Southeast Asia

Study period: This study covers the period from 2010 to 2021 The starting year,

In 2010, a significant rebound in foreign direct investment (FDI) flows was observed, marking a recovery from the impacts of the 2008-2009 global financial crisis According to UNCTAD data, FDI inflows into Southeast Asia saw a remarkable increase, rising from $41 billion in 2009 to nearly $

In 2021, the total reached $113 billion, marking the highest level since 1990, as data limitations restrict the analysis to this year The World Bank's Sovereign ESG Data Portal primarily provides ESG data up to 2021 Additionally, the years 2010 to 2021 encompass the impact of the COVID-19 pandemic on six ASEAN countries during 2020 and 2021.

This study examines six Southeast Asian countries that are part of the Association of Southeast Asian Nations (ASEAN): Singapore, Malaysia, Indonesia, the Philippines, Thailand, and Vietnam These nations were chosen for their comparable economic scales, social development levels, and institutional frameworks.

RESEARCH METHODOLOGY

Based on theoretical basis, the author develops 3 quantitative research models to examine the impact of Sovereign ESG and its components on FDI inflows from the aggregate to detailed levels

The author utilized secondary data from the World Bank, incorporating resources such as the World Development Indicators Database (WDI), the Sovereign ESG Data Portal, and the Sovereign ESG Score Builder, with the analyzed data presented in a balanced panel format.

6 comprising Sovereign ESG and its components, as well as macroeconomic factors The sample includes 6 Southeast Asian nations over the period of 2010-2021

The research employs a quantitative analysis method utilizing three econometric models with panel data to evaluate various factors It estimates results through Pooled Ordinary Least Squares (OLS), Fixed Effects, Random Effects, Feasible Generalized Least Squares, and Efficient Generalized Method of Moments.

1.5.4 Summary and evaluation of research results

The research findings will be analyzed in relation to the existing theoretical framework to assess the influence of Sovereign ESG and its components on FDI inflows, ultimately supporting the thesis's overall objectives.

EXPECTED CONTRIBUTION

Research on the impact of national-level Environmental, Social, and Governance (ESG) factors on Foreign Direct Investment (FDI) inflows remains limited, particularly in Southeast Asia, highlighting a pressing need for comprehensive studies This knowledge gap necessitates further investigation to provide empirical evidence on the relationship between Sovereign ESG and FDI inflows By addressing this gap, this study aims to contribute to the existing body of research and offer valuable insights into the interplay between national ESG factors and FDI inflows in Southeast Asia.

The author aims to enhance policy making, assist multinational investors in making informed choices, and foster sustainable development through the research findings Additionally, this study establishes a foundation for future research, paving the way for new investigations into the connection between Sovereign ESG and Foreign Direct Investment (FDI).

THE STRUCTURE OF RESEARCH

Through Chapter 1, the author introduces an overview of the research topic, including reasons for choosing, research objectives and questions, research subjects and scope, research methods, and practical contributions

Chapter 2: Theoretical basis and Literature review

Chapter two explores the theoretical foundations and empirical data regarding the factors influencing Foreign Direct Investment (FDI) inflows, emphasizing the role of Sovereign ESG and its elements Following an extensive literature review, the author formulates hypotheses and pinpoints existing research gaps.

Chapter 3 delineates the research methodology, building on the theoretical framework and literature review from the previous chapter It details the research process, outlines data collection techniques, specifies the model, and defines the variables Additionally, the author discusses the methods employed for data analysis.

Chapter 4: Empirical results and discussion

Chapter 4 presents and interprets the formal quantitative results of the study, based on data analyzed using STATA software, including descriptive statistics, determining the correlation, testing and correcting all defects, thereby obtaining the best results for three research models Finally, the author interprets the impact of independent variables on dependent variable to confirm the correctness of the hypotheses that have been built

This concluding section summarizes the research findings, highlighting their implications and limitations The study offers recommendations for policymakers and investors, while also outlining potential directions for future research.

Chapter 1 provides an overview of the research topic (research rationale, objectives, questions, methodology, and anticipated contributions.) and serves as a foundation for subsequent chapters Based on the above contents, the study on "The impact of environmental, social, and governance (ESG) factors on foreign direct investment (FDI): A case study of ASEAN countries" will be carried out with both scientific and practical values

THEORETICAL BASIS AND LITERATURE REVIEW

THEORETICAL BASIS OF FOREIGN DIRECT INVESTMENT

Foreign Direct Investment (FDI) is essential for a nation's overall development, as it enhances economic growth by providing capital, facilitating technology transfer, and generating employment opportunities (Moosa, 2002) Defined by UNCTAD (1999), FDI involves long-term investments that signify an entity's control and interests in a foreign enterprise, comprising equity capital, reinvested earnings, and intercompany loans Various theories explain the factors influencing FDI, with location theory and the eclectic paradigm (OLI) being among the most significant (Agarwal, 1980).

Location theory highlights how the immobility of certain production factors, like labor and natural resources, creates variations in production costs across countries, influencing foreign direct investment (FDI) motivations Research by Culem (1988), Goldberg (1972), Moore & Driscoll (1997), and Riedel (1975) demonstrates that low wages in developing nations lower labor costs, boost profits, and enhance the competitiveness of multinational corporations (MNCs) Additionally, abundant natural resources serve as a significant attraction for FDI, supported by studies from Asiedu & Lien (2011), Campos & Kinoshita (2003), and Hailu (2010) Access to ample raw materials not only reduces transportation costs but also expedites production processes, further driving FDI inflows.

The Eclectic (OLI) Paradigm: According to Dunning (1979), a firm wishing to conduct FDI must satisfy three conditions Firstly, it must have ownership advantages

The OLI model outlines three key advantages that multinational corporations (MNCs) consider when engaging in foreign direct investment (FDI): ownership advantage, internalization advantage, and location advantage Ownership advantage involves proprietary technology that provides a competitive edge, while internalization advantage emphasizes the need for higher economic efficiency from utilizing these advantages internally rather than licensing them Location advantage pertains to the benefits offered by the foreign market, such as factor endowments, market size, education and training, and favorable government policies regarding macroeconomic factors like inflation and interest rates Numerous studies, including those by Barrell & Pain (1999) and Cheng & Kwan (2000), have utilized the OLI model to analyze the determinants of FDI.

Sustainable Foreign Direct Investment (FDI) seeks to achieve a balance between the profits of multinational corporations (MNCs) and the sustainable development of host countries, diverging from traditional FDI, which focuses solely on profit (Moosa, 2002) Although still lacking a formal definition, sustainable FDI is increasingly recognized as an investment approach that aligns with environmental, social, and governance (ESG) objectives This means that sustainable FDI not only aims to generate financial returns for investors but also prioritizes the well-being of the host country's environment and society, contributing positively to its long-term development.

Sustainable Foreign Direct Investment (FDI) is increasingly becoming a global trend, with the World Investment Report (UNCTAD, 2024) highlighting a notable rise in FDI projects aligned with 17 Sustainable Development Goals (SDGs) compared to 2023 Key sectors experiencing growth include renewable energy, which surged by 76%, transportation and energy infrastructure at 40%, and health and education at 22% Developing countries in Asia are emerging as prime destinations for these investments, driven by multinational corporations (MNCs) that prioritize projects with social and environmental benefits, as well as supportive government policies promoting green investment and sustainable development initiatives.

10 expects that sustainability will increasingly become an important factor in attracting FDI inflows.

FACTORS AFFECTING FOREIGN DIRECT INVESTMENT

2.2.1 Sovereign ESG and its components

Sovereign ESG is the first important factor to consider in attracting FDI into

Research has consistently highlighted the crucial role of ESG factors, from macroeconomic studies (Nordhaus, 1977) to microeconomic analyses (Bouye & Menville, 2021) This significance is especially pronounced during crises (Dietz et al., 2018; Lins et al., 2017) and following global disruptions like climate change and the COVID-19 pandemic (Chipalkatti et al., 2021; JP Morgan, 2020) As a result, foreign investors are increasingly recognizing the strong correlation between Sovereign ESG practices and long-term investment performance.

A universally accepted definition of Sovereign ESG is still lacking, highlighting the need for a holistic framework to assess country-level ESG performance (Gratcheva et al., 2021; Gratcheva & Gurhy, 2024) The ambiguity surrounding 'sustainability' poses significant challenges in this context Sovereign ESG pertains to the application of Environmental, Social, and Governance indicators to evaluate a country's sustainability, performance, and risk, providing multinational corporations (MNCs) with insights to align their investment decisions with the UN's 17 Sustainable Development Goals (D Zhang et al., 2022) Despite the recent emergence of the concept, empirical studies on the impact of Sovereign ESG on foreign direct investment (FDI) inflows remain limited, with existing research primarily focusing on individual ESG indicators, which yield varying effects on FDI From a financial perspective, Crifo et al (2017) found that higher Sovereign ESG ratings can reduce government borrowing costs, enhancing investment returns and attracting FDI, while Hübel & Scholz (2020) noted that countries with elevated sustainability levels may experience similar benefits.

A study of 109 countries by Wang et al (2020) highlights that improvements in Sovereign ESG factors positively influence economic growth and foreign direct investment (FDI) inflows Additionally, research by Pisani et al (2019) indicates that 'green cities' in China, characterized by improved air quality and wastewater treatment, are more successful in attracting FDI Therefore, the correlation between Sovereign ESG enhancements and increased FDI inflows is both significant and beneficial.

While some studies suggest that Sovereign ESG factors may not significantly influence FDI inflows, with Rezza (2014) finding no correlation between FDI and environmental regulations, and Singh & Jun (1995) and Wheeler & Mody (1992) indicating that political risks have minimal impact, this thesis aims to explore the relationship further Utilizing the World Bank's 'Sovereign ESG Score Builder', the author anticipates a positive correlation between Sovereign ESG factors and FDI inflows in six ASEAN countries.

Based on the discussion above, hypothesis (H 1 ) is proposed that Sovereign ESG has a positive impact on FDI inflows

The environment, a key pillar of Sovereign ESG, significantly impacts foreign direct investment (FDI) inflows, encompassing five critical categories: emissions and pollution, energy use and security, climate risk and resilience, food security, and natural capital management Research reveals diverse outcomes regarding the environmental influence on FDI, primarily driven by two hypotheses: the pollution haven hypothesis (PHH) and the pollution halo hypothesis (PHE).

The pollution haven hypothesis (PHH) posits that multinational corporations (MNCs) shift their environmentally harmful production processes to nations with less stringent environmental regulations This strategic relocation allows them to evade strict regulations and reduce pollution control expenses in their home countries Empirical evidence from studies by Copeland & Taylor (1994) and Spatareanu (2007) supports this theory, demonstrating that MNCs actively transfer polluting activities to regions with weaker environmental oversight.

Foreign Direct Investment (FDI) tends to increase in countries with less stringent environmental regulations Research by Dam and Scholtens (2008) indicates that there is a positive correlation between the strictness of environmental policies in a home country and the FDI a host country receives Consequently, host countries with higher levels of pollution are more likely to attract greater FDI inflows.

The pollution halo hypothesis (PHE) posits that foreign direct investment (FDI) inflows can promote pollution reduction in host countries by facilitating the transfer of green, energy-efficient technologies (Yilanci et al., 2023) Supporting this theory, Golub et al (2011) found that stricter environmental regulations in host nations boost "green FDI," which focuses on environmentally friendly goods, services, and technology investments Pisani et al (2019) further noted that FDI firms often prefer to invest in green cities to enhance their corporate image and ensure employee well-being, driven by increasing stakeholder awareness and pressure Consequently, the PHE suggests a strong positive correlation between a host country's environmental quality and its ability to attract FDI; the cleaner the environment and the stricter the regulations, the more likely it is to draw investment Thus, the author anticipates that enhancing all five categories of the environmental pillar will significantly attract FDI to Southeast Asian countries.

Based on the discussion above, hypothesis (H 2 ) is proposed that Environmental pillar positively correlates with FDI inflows

The social pillar encompasses interconnected, people-focused elements such as access to services, demographics, education and skills, employment, health and nutrition, along with poverty and inequality This framework is essential for sustainable development and plays a crucial role in attracting foreign direct investment (FDI).

High-quality human capital (education level, life expectancy, welfare) helps attract foreign direct investment (FDI) (Benhabib & Spiegel, 1994; Nelson & Phelps,

A high-quality workforce characterized by elevated education levels, long life expectancy, and good living standards significantly enhances labor productivity, making a country more attractive for foreign direct investment (FDI) Research by Borensztein et al (1998) emphasizes that a certain level of human capital is essential for maximizing FDI benefits Similarly, Li & Liu (2005) highlight the direct and indirect importance of human capital in fostering positive economic growth through FDI in developing nations Sharma & Gani (2007) found that human development indices are appealing to foreign investors in lower and middle-income countries Additionally, Kheng et al (2017) identified a reciprocal relationship between human capital and FDI inflows, suggesting that coordinated policies aimed at enhancing both FDI and human capital are crucial for optimal outcomes.

However, contrary to the aforementioned findings, when examining data from

A study by Iamsiraroj (2016) covering 124 countries over 40 years (1971-2010) revealed a negative correlation between primary education levels and foreign direct investment (FDI) inflows This finding implies that basic education may not serve as a crucial attraction for foreign investment Additionally, the author suggests that, akin to environmental factors, simultaneous investment in various areas may be necessary to enhance FDI potential.

6 categories, Southeast Asian countries can not only improve human capital but also strongly attract FDI

Based on the discussion above, hypothesis (H 3 ) is proposed that Social pillar positively impacts FDI inflows

Governance, the essential pillar of Sovereign ESG, encompasses six interconnected categories: economic environment, gender, government effectiveness, human rights, innovation, and stability & rule of law These elements collectively shape the framework for assessing governance quality within the broader context of ESG criteria.

Governance directly influences foreign investors’ decisions and FDI inflows A stable, transparent, and predictable business environment, built on a foundation of good governance, is the top attraction for MNCs

Extensive research highlights the importance of the Governance pillar in attracting Foreign Direct Investment (FDI) Canh et al (2020) found that uncertainty in domestic economic policies negatively impacts FDI inflows Studies by Buchanan et al (2012), Globerman & Shapiro (2003), Mengistu & Adhikary (2011), and Saidi et al (2013) emphasize that strong institutions, effective governance, and low corruption levels are essential for FDI Additionally, Bailey (2018) identified political stability and a robust legal framework as critical factors for attracting foreign investment, as a clear legal system protects property rights and minimizes investment risks, fostering trust among multinational corporations Corruption, as noted by Bailey (2018), can further impede FDI inflows.

Corruption poses a significant barrier to economic growth and sustainable development, highlighting the need for effective governance to foster a fair business environment for both domestic and foreign enterprises The author emphasizes the equal importance of all three pillars of Sovereign ESG, asserting that enhancing the six categories within the Governance pillar concurrently will generate a synergistic effect, ultimately boosting foreign direct investment (FDI) flows.

Based on the discussion above, hypothesis (H 4 ) is proposed that Governance pillar has a positive impact on FDI inflows

The World Bank categorizes emissions and pollution into five key indicators, representing various gases, including greenhouse gases (GHGs) such as carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), and fine particulate matter (PM2.5) CO2 emissions, primarily originating from fossil fuel combustion and cement production, are the leading cause of climate change (Opoku & Boachie, 2020) Since the Industrial Revolution, the rising concentration of CO2 in the atmosphere has resulted in severe consequences, including ocean acidification, sea-level rise, and extreme weather events.

RESEARCH GAPS

Since its introduction in the 2004 UN report "Who Cares Wins," the significance of environmental, social, and governance (ESG) factors has dramatically increased, particularly in response to global challenges like climate change and the COVID-19 pandemic.

Research on the impact of ESG factors has largely centered on their influence on portfolio investment at the firm level Nonetheless, there is a growing need to explore the connection between country-level ESG factors, known as Sovereign ESG, and their effect on foreign direct investment (FDI).

24 inflows remains under-explored Most previous studies have concentrated on the relationship between FDI and individual ESG components

This study addresses the research gap by thoroughly analyzing the relationship between Sovereign ESG and FDI inflows across three levels: from aggregate measures to detailed insights, examining the connection between Sovereign ESG, its pillars, and specific ESG categories.

This research targets Southeast Asia, a promising investment region where governments are prioritizing ESG factors to attract sustainable foreign direct investment (FDI) and meet the Sustainable Development Goals (SDGs) Despite the growing importance of this topic, studies specifically examining the six ASEAN nations remain scarce By elucidating the relationship between ESG and investment decisions made by multinational corporations, this research aims to enhance the theoretical framework surrounding this impact and offer valuable policy insights for countries in the region.

Chapter 2 provides a solid theoretical framework for the study, including an overview of theories and previous studies that explaining FDI, and an emphasis on the concept of sustainable FDI On this basis, the author presents specific research hypotheses to answer the research questions For factors affecting FDI inflows, this research uses both Sovereign ESG and macroeconomic factors as independent variables for the research model to ensure objectivity, avoid omitting variables, and provide a comprehensive analysis

Regarding the research model, variables that the author chose to study include

The article discusses three key groups related to ESG (Environmental, Social, and Governance) principles The first group emphasizes the importance of sovereignty in ESG practices The second group focuses on the three pillars of ESG: environmental sustainability, social responsibility, and effective governance The third group addresses critical factors impacting ESG performance, including emissions and pollution, natural capital management, education and skills development, employment opportunities, government effectiveness, human rights, trade openness, inflation, and real interest rates.

RESEARCH METHODOLOGY

RESEARCH PROCESS

To analyze the determinants of Foreign Direct Investment inflows, this thesis uses a quantitative research method based on specific processes:

 Step 1: Determining research objectives of the thesis

 Step 2: Finding out theoretical basis and reviewing relevant prior empirical evidence

 Step 3: Selecting variables to build quantitative research model from the theoretical basis and empirical evidence in step 2

 Step 4: Collecting secondary data as inputs for the research model developed in step 3

In Step 5, the author conducts a thorough data analysis utilizing the model from Step 3 and the data collected in Step 4 This analysis includes descriptive statistics, correlation analysis, and a multicollinearity test, followed by the selection of the appropriate regression model Additionally, the author performs tests for autocorrelation and heteroscedasticity, while also addressing potential model misspecification and endogeneity issues to ensure the robustness of the results.

In Step 6, the author evaluates the results to pinpoint statistically significant independent variables and their effects on the dependent variable This analysis is then compared to the theoretical framework outlined in Chapter 2 to determine whether the observed relationships align with the hypothesized expectations.

 Step 7: The author derives conclusions and proposes recommendations from the research results

The research process is visually presented in the diagram below:

RESEARCH MODEL

The author explores the relationship between Sovereign ESG and Foreign Direct Investment (FDI) inflow by employing a log-linear model, as outlined by Gujarati (2011) This model evaluates how proportional changes in FDI inflow correspond to absolute changes in key macroeconomic indicators, including trade openness, inflation, and real interest rates The analysis focuses on the natural logarithm of FDI inflow as the dependent variable, providing insights into the effects of these independent variables on investment trends.

Variables selection and research model building

Finding basic theory and empirical evidence

Additionally, to determine the impact of Sovereign ESG from aggregate to detail levels, the following independent variables are proposed:

 Using the Sovereign ESG score to examine the overall impact of ESG on FDI, first research model

 Using the Environment pillar score, Social pillar score, and Governance pillar score to examine the impact of Sovereign ESG on FDI inflows at the pillar level, second research model

This study analyzes the influence of Sovereign ESG factors on Foreign Direct Investment (FDI) inflows by assessing various scores, including Emissions and Pollution, Natural Capital Endowment and Management, Education and Skills, Employment, Government Effectiveness, and Human Rights The third research model specifically examines these category-level scores to determine their impact on attracting FDI.

Three research models are proposed as follows:

First research model: lnFDI i,t = 𝜶 𝟎 + 𝜶 𝟏 ESG i,t + 𝜶 𝟐 OPEN i,t + 𝜶 𝟑 INFLATION i,t + 𝜶 𝟒

Second research model: lnFDI i,t = 𝜷 𝟎 + 𝜷 𝟏 pENV i,t + 𝜷 𝟐 pSOC i,t + 𝜷 𝟑 pGOV i,t + 𝜷 𝟒 OPEN i,t + 𝜷 𝟓

Third research model: lnFDI i,t = 𝜸 𝟎 + 𝜸 𝟏 eEMI i,t + 𝜸 𝟐 eNAT i,t + 𝜸 𝟑 sEDU i,t + 𝜸 𝟒 sEMP i,t + 𝜸 𝟓 gGOV i,t + 𝜸 𝟔 gHUM i,t + 𝜸 𝟕 OPEN i,t + 𝜸 𝟖 INFLATION i,t + 𝜸 𝟗 INTEREST i,t + 𝜺 𝒊,𝒕

Where: i = 1, , 6; t = 1, , 12 lnFDIi,t FDI inflow of country i in year t

OPENi,t Trade openness of country i in year t

INFLATIONi,t Inflation of country i in year t

INTERESTi,t Real interest rate of country i in year t

ESGi,t Sovereign ESG score of country i in year t

Second research model: pENVi,t Environment Pillar score of country i in year t pSOCi,t Social Pillar score of country i in year t pGOVi,t Governance Pillar score of country i in year t

The third research model evaluates various category scores for country i in year t, including emissions and pollution (eEMIi,t), natural capital endowment and management (eNATi,t), education and skills (sEDUi,t), employment (sEMPi,t), government effectiveness (gGOVi,t), and human rights (gHUMi,t).

VARIABLE DEFINITION

LnFDI is used to measure the change in Foreign Direct Investment (FDI) inflows to a country over a specific period, typically 1 year (Nguyễn Xuân Hồng,

2019) This dependent variable is calculated by taking natural logarithm of the FDI inflow data obtained from World Bank Database

The Sovereign ESG scores, along with the scores for ESG Pillars and Categories, are derived from the World Bank's Sovereign ESG Score Builder Each country's scores are compiled annually, utilizing a wide range of comprehensive indicators.

The Sovereign ESG Data Framework consists of indicators organized into categories, which are further divided into three main pillars: Environment, Social, and Governance The scoring process for these indicators follows a systematic four-step approach.

 Step 1: Selecting and normalizing Indicators

This step focuses on choosing suitable indicators for scoring according to the Sovereign ESG Data Framework, utilizing Min-max normalization The indicators are divided into two categories: positive indicators that enhance Sovereign ESG and negative indicators that negatively affect it To maintain the comparability of results, the study applies Min-max normalization, a popular method for data normalization that enables data to fit within specified limits.

2015) This method is chosen due to its monotonicity and applicability to both positive and negative indicators The specific formula for Indicator Score is:

I : the indicator's value in the specific year used for scoring

I min : denote the minimum values of the indicator across all years

I max : denote the maximum values of the indicator across all years

In the "Higher is better" approach, the authors assigned "Yes" to positively correlated indicators, ensuring that the Indicator Score retains its sign during aggregation In contrast, for negatively correlated indicators, they chose "No," which means the Indicator Score will change its sign when aggregated.

The chosen countries are the ASEAN-6: Indonesia, Malaysia, Philippines, Singapore, Thailand and Vietnam

A weighted average approach was used to aggregate scores in this research The detailed calculation for the composite score is:

In which: n : the number of indicators used to calculate Composite Score w k : weighted average, ∑ 𝑛 𝑘=1 𝑤 𝑘 = 1

IS k : the i th indicator score

The Sovereign ESG score, ESG Pillars score, and ESG Categories score are calculated by aggregating corresponding Indicator Scores based on the Sovereign ESG Data Framework This comprehensive approach provides a thorough evaluation of each economy in ASEAN-6, covering a 12-year period By analyzing the data over an extended timeframe, the framework offers valuable insights into the long-term ESG performance of these economies.

2010 to 2021 These scores are used for variables in all three models below

This independent variable represents the Sovereign ESG score, which is a composite of 71 indicators, including 31 negative indicators ranging from -1 to 0, and

The Sovereign ESG Score, which ranges from -31/71 to 40/71, is calculated using 40 positive indicators that span from 0 to 1 As detailed in section 3.3.2.1 on Aggregating ESG Score, the author anticipates that ESG variables will significantly influence Foreign Direct Investment (FDI) inflows to the ASEAN-6, aligning with findings from Crifo et al (2017), Wang et al (2020), and Hübel & Scholz (2020).

Hypothesis H 1 : Sovereign ESG has a positive impact on FDI inflows

3.3.2.3 Second research model variables pENV

This independent variable represents Environment pillar score, which is a composite of 31 indicators, including 21 negative indicators, 10 positive indicators

The scoring method for all variables in second model have described in section

3.3.2.1 Aggregating ESG Score Similar to Golub et al (2011) and Pisani et al (2019), the author expects a positive coefficient on this variable

Hypothesis H2 posits a positive correlation between the environmental pillar and foreign direct investment (FDI) inflows The social pillar score (pSOC), serving as the independent variable, is derived from 22 indicators—eight of which are negative and 14 positive Consistent with the findings of Li & Liu (2005) and Kheng et al (2017), a positive coefficient is anticipated for this variable.

Hypothesis H3 posits that the social pillar has a positive impact on foreign direct investment (FDI) inflows The governance pillar score (pGOV), derived from 18 indicators—comprising 2 negative and 16 positive—suggests a favorable correlation This aligns with the findings of Bailey (2018) and Canh et al (2020), indicating that an improved governance score is likely to enhance FDI attractiveness.

Hypothesis H 4 : Governance pillar has a positive impact on FDI inflows

3.3.2.4 Third research model variables eEMI eEMI represents the Emissions & pollution category score that based on all five negative indicators The scoring method for all variables in the third model are detailed in section 3.4.2.1 Aggregating ESG Score Similar to the results of Omri et al (2014b); Opoku et al (2022), the author expects a positive coefficient for eEMI

Hypothesis H 5 : Emissions & pollution category exerts positive impacts on FDI inflows eNAT eNAT represents the Natural capital endowment & management category Score

It is calculated using 7 indicators, 5 of them are negative The author anticipates a positive relationship between eNAT and FDI inflows, similar to the findings of Hailu

Hypothesis H6 suggests that the management of natural capital positively influences foreign direct investment (FDI) inflows Conversely, the Education & Skills category Score (sEDU), which aggregates three positive indicators, is anticipated to negatively affect FDI flows into the ASEAN-6 region This expectation is consistent with the research findings of Narula.

Hypothesis H7 suggests that the Education and Skills category negatively impacts Foreign Direct Investment (FDI) inflows The Employment category score, denoted as sEMP, is derived from three indicators, two of which are negative Aligning with findings from Braun (2006) and Tabash et al (2024), the author anticipates a positive coefficient for this variable.

Hypothesis H8 posits that the Employment category positively influences FDI inflow Additionally, gGOV, representing the Government Effectiveness Score, is a composite of two positive indicators Consistent with the findings of Nizam & Hassan (2018) and Quang et al (2022), the author anticipates a positive coefficient for this variable.

Hypothesis H9 posits that the Government Effectiveness category positively influences Foreign Direct Investment (FDI) inflows Additionally, the Human Rights category Score (gHUM), which is derived from three positive indicators, is anticipated to yield a positive coefficient, aligning with the findings of Shan et al (2018) and Sabir et al (2019).

Hypothesis H 10 : Human Rights category has a positive impact on FDI inflow

3.3.2.5 Control variables for all three research models

OPEN, an independent variable sourced from the World Bank, assesses a country's trade openness and the barriers it imposes on foreign investment Following the findings of Sabir et al (2019), the author expects a positive coefficient for this variable The calculation formula for OPEN is provided below.

Hypothesis H 11 : Trade Openness positively correlates with FDI inflows

Inflation, sourced from the World Bank, is quantified using the consumer price index, indicating the annual percentage change in the cost of a fixed basket of goods and services Aligning with the research of Zaman et al (2012) and Kisto (2017), the author expects to observe a positive correlation for inflation.

Hypothesis H 12 : Inflation positively influences FDI inflows

RESEARCH DATA

This study leverages secondary data from the six largest economies in Southeast Asia, namely Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam, which have undergone significant integration and emerged as prominent destinations for foreign investment The dataset, sourced from the World Bank Data, spans a 12-year period from 2010 to 2021, providing a comprehensive insight into the region's economic trends and developments.

The study begins in 2010, a year when most ASEAN economies had bounced back from the global financial crisis of 2008-2009 It concludes in 2021, chosen due to the constraints in the availability of Environmental, Social, and Governance (ESG) data.

Data on Foreign Direct Investment (FDI), trade openness, inflation, and real interest rates were sourced from the World Bank's World Development Indicators Additionally, Sovereign ESG data and its components were gathered from the World Bank’s Sovereign ESG Score Builder, which calculates scores based on 71 indicators, as outlined in Table 3.2 below.

 The overall Sovereign ESG score was calculated based on all 71 indicators

The ESG Pillars scores are calculated based on three key dimensions: Environment, Social, and Governance The Environment score is derived from 31 indicators, the Social score from 22 indicators, and the Governance score from 18 indicators.

In this study, the ESG Categories score in the third model was focused on 6 out of the 17 available categories The environmental categories, including Emissions & Pollution and Natural Capital Endowment & Management, were assessed using 5 and 7 indicators, respectively Additionally, the social categories of Education & Skills and Employment were evaluated based on 3 indicators.

4 indicators; and the governance categories of Government Effectiveness and

Human Rights were calculated from 2 and 3 indicators respectively

Following Jiang et al.’s (2022) Sovereign ESG Framework, the author constructs expected signs between each indicator and the Sovereign ESG, as detailed in table 3.2 below:

Table 3.2 Sovereign ESG Data Framework

CO2 emissions (metric tons per capita) (-)

In the context of greenhouse gas (GHG) emissions, net emissions and removals from land use, land-use change, and forestry (LUCF) are measured in metric tons of CO2 equivalent Additionally, methane and nitrous oxide emissions are assessed per capita, expressed in metric tons of CO2 equivalent Furthermore, the mean annual exposure to PM2.5 air pollution is quantified in micrograms per cubic meter, highlighting the environmental impact of these pollutants.

Electricity production from coal sources (% of total) (-)

Energy imports, net (% of energy use) (-)

Energy intensity level of primary energy (MJ/$2017 PPP GDP) (-)

Energy use (kg of oil equivalent per capita) (-)

Fossil fuel energy consumption (% of total) (-)

Renewable electricity output (% of total electricity output) (+) Renewable energy consumption (% of total final energy consumption) (+)

Population density (people per sq km of land area) (-)

Level of water stress: freshwater withdrawal as a proportion of available freshwater resources (-)

Proportion of bodies of water with good ambient water quality (+)

Food Security Agricultural land (% of land area) (+)

Agriculture, forestry, and fishing, value added (% of GDP) (+)

Adjusted savings: natural resources depletion (% of GNI) (-)

Adjusted savings: net forest depletion (% of GNI) (-)

Annual freshwater withdrawals, total (% of internal resources) (-)

Forest area (% of land area) (+)

Terrestrial and marine protected areas (% of total territorial area) (+)

Access to clean fuels and technologies for cooking (% of population) (+)

Access to electricity (% of population) (+)

People using safely managed drinking water services (% of population) (+) People using safely managed sanitation services (% of population) (+)

Demography Fertility rate, total (births per woman) (+)

Life expectancy at birth, total (years) (+)

Population ages 65 and above (% of total population) (+)

Education & skills Government expenditure on education, total (% of government expenditure)

Literacy rate, adult total (% of people ages 15 and above) (+)

Employment Children in employment, total (% of children ages 7-14) (-)

Labor force participation rate, total (% of total population ages 15-64) (modeled

Unemployment, total (% of total labor force) (modeled ILO estimate) (-)

Cause of death, by communicable diseases and maternal, prenatal and nutrition conditions (% of total) (-)

Mortality rate, under-5 (per 1,000 live births) (-)

Prevalence of overweight (% of adults) (-)

Prevalence of undernourishment (% of population) (-)

Individuals using the Internet (% of population) (+)

Gender Proportion of seats held by women in national parliaments (%) (+)

Ratio of female to male labor force participation rate (%) (modeled ILO estimate) (+)

School enrollment, primary and secondary (gross), gender parity index (GPI) (+) Unmet need for contraception (% of married women ages 15-49) (-)

Human Rights Strength of legal rights index (0=weak to 12=strong) (+)

Economic and Social Rights Performance Score (+)

Research and development expenditure (% of GDP) (+)

Scientific and technical journal articles (+)

Political Stability and Absence of Violence/Terrorism: Estimate (+)

This study employs panel data, combining 12 years of time-series data from

From 2010 to 2021, a study utilizing panel data from six Southeast Asian countries highlights the advantages of this data type over time-series and cross-sectional data Panel data provides richer information, increased variability, reduced collinearity among variables, and enhanced efficiency, allowing for the repeated observation of cross-sectional units over time This approach effectively captures dynamic changes and is better suited for detecting and measuring unobserved effects that traditional methods may overlook (Gujarati, 2011).

This research tackles the issue of insufficient sovereign ESG data from the World Bank by employing a mean-filling method using panel data to fill in missing observations Specifically, the study replaces missing values with the time-series mean for each country prior to further analysis.

DATA ANALYSIS METHOD

Data for the research model was analyzed in STATA software, following a step- by-step process as illustrated in the diagram below:

Descriptive statistics play a crucial role in summarizing the fundamental features of a research dataset, providing a comprehensive overview of the data's characteristics By applying descriptive statistics, researchers can analyze key numerical values for each variable, including the number of observations, minimum and maximum values, and measures of central tendency such as mode, mean, and median Additionally, descriptive statistics enable the examination of measures of dispersion, including standard deviation and variance, as well as skewness and kurtosis, offering valuable insights into the data's distribution and shape.

Choosing regression model (POLS/ FEM/ REM)

Skewness is a statistical measure that indicates the symmetry of a distribution In a perfectly symmetric distribution, the skewness coefficient equals 0 When the skewness coefficient is negative, it suggests that the median exceeds the mean, resulting in a left-skewed distribution.

If the coefficient is positive (> 0), the median is usually less than the mean and the distribution is said to be skewed right (Lê Sĩ Đồng, 2013)

Kurtosis measures the peakedness or flatness of a distribution, indicating how closely data clusters around the mean A kurtosis value below 3 signifies a flatter distribution, suggesting that the data is more spread out than a normal distribution In contrast, a kurtosis value above 3 indicates a more peaked distribution, implying that the data is more concentrated around the mean (Lê Sĩ Đồng, 2013).

Correlation analysis is a statistical technique that measures the degree and direction of the linear relationship between two variables It is essential for evaluating the strength of relationships between independent and dependent variables, as well as identifying multicollinearity among independent variables The correlation coefficient (r) quantifies this relationship, with values ranging from -1 to 1 A coefficient close to 1 signifies a strong positive correlation, while a value near -1 indicates a strong negative correlation Conversely, an r value close to 0 suggests a weak or nonexistent correlation.

Multicollinearity arises when independent variables in a regression model exhibit high correlation, leading to unreliable estimates, inflated standard errors, and invalid hypothesis tests The Variance Inflation Factor (VIF) is a key tool for detecting multicollinearity, with a VIF value exceeding 5 typically signaling an issue (Hoàng Trọng & Chu Nguyễn Mộng Ngọc, 2008) To mitigate multicollinearity, researchers can utilize prior information or external data to accurately estimate individual coefficients In cases where the multicollinearity is not severe, simply removing the multicollinear variables may suffice.

42 variables or increasing the sample size can help mitigate the issue (Nguyễn Quang Dong, 2008)

The research model uses panel data, which can be estimated using three primary methods (Gujarati, 2011):

 Pooled Ordinary Least Squares (Pooled OLS): combines all observations to estimate the model It assumes that the coefficients remain consistent across time and cross-sectional units

The Fixed Effects Model (FEM) is a statistical approach that accommodates heterogeneity across units by assigning a unique, time-invariant intercept to each cross-sectional unit This model is particularly suitable for analyzing data where unobserved time-invariant factors have a significant impact on the dependent variable, allowing researchers to control for individual-specific effects that do not change over time.

The Random Effects Model (REM) operates under the premise that individual-specific effects are random and not correlated with the explanatory variables In this model, the error term includes both individual-specific and time-specific components, allowing for a comprehensive analysis of data variations across different individuals over time.

The process of selecting the optimal estimation method among Pooled OLS, FEM, and REM for panel data involves a two-step procedure (Gujarati, 2011)

Step 1: Choosing between Pooled OLS and FEM by F-test with the null hypothesis is that Pooled OLS is better than FEM If the F-test is statistically significant, there is evidence to confirm that Pooled OLS model is inappropriate and FEM is better

Step 2: Choosing between FEM and REM Comparing FEM and REM through

The Hausman test (1978) evaluates the consistency of both estimates under the null hypothesis that they are consistent, with the random effects model (REM) being the more efficient option The test results follow a Chi-squared distribution, with degrees of freedom equal to the number of explanatory variables If the calculated Chi-squared value exceeds the critical Chi-squared value at a given degree of freedom, the null hypothesis can be rejected, indicating that the REM is unsuitable and the fixed effects model (FEM) should be chosen Conversely, if the calculated value is less than or equal to the critical value, the REM model is deemed appropriate.

Heteroscedasticity refers to the situation in regression models where the variance of error terms is not constant, leading to unbiased yet inefficient estimates (Nguyễn Quang Dong, 2008) This inefficiency results in biased estimated variances, rendering t-tests, F-tests, and confidence intervals unreliable To detect heteroscedasticity in panel data, researchers often use the Pagan-Breusch test for Fixed Effects Models (FEM) and the Modified Wald test for Random Effects Models (REM), both of which operate under the null hypothesis of "no heteroscedasticity" (Drukker, 2003) A common method to address heteroscedasticity is the Feasible Generalized Least Squares (FGLS) model.

Autocorrelation refers to the correlation between consecutive observations in a dataset, which can cause biased variance estimates and undermine the reliability of statistical tests such as t-tests, F-tests, and R² values Although the estimates may remain unbiased, they lack efficiency This problem is especially significant in the context of time series and panel data analysis.

2008) Wooldridge's test is commonly used to detect first-order autocorrelation in panel data (Baum, 2001; Wooldridge, 2002) The Feasible Generalized Least Squares (FGLS) model is a common approach to address this issue

Feasible Generalized Least Squares (FGLS) is an effective method for estimating panel data regression by making specific assumptions about the correlation structure of errors across cross sections This approach is particularly beneficial when there are fewer cross-sectional observations relative to time observations (Kolev, 2014) Additionally, FGLS can accommodate models where error terms exhibit cross-sectional correlation, unconditional heteroskedasticity, and autocorrelation over time (Kolev).

2014) With the panel data structure of this research model, FGLS estimation is used to address the presence of autocorrelation within panels and heteroskedasticity across panels (Stata, 2021)

An endogenous regressor is a variable that correlates with the error term, leading to biased and inconsistent estimates This issue often stems from measurement errors in explanatory variables or omitted variable bias To address endogeneity, the instrumental variable technique can be employed, utilizing instruments that are correlated with the endogenous variable, uncorrelated with the error term, and excluded from the model's explanatory variables (Gujarati, 2011).

To effectively identify and address endogeneity in the study, the author employed a trial-and-error analysis for each explanatory variable Given the research's limitations—covering only six countries over twelve years with a mere 72 observations (N < T)—the traditional Arellano-Bond two-step GMM estimator proved unsuitable for examining endogeneity and long-term effects of FDI inflows Instead, the author recommends the Feasible two-step Efficient GMM (EGMM) estimator, which is better suited for static models and smaller sample sizes The process of testing and correcting for endogeneity involved a systematic two-step approach.

 Step 1: Examine the existence of endogeneity in the model

The suspected endogenous variables are tested with include instrumental variables are the remaining explanatory variables and dummy exclude instrumental variables are the lags of suspected endogenous variable

The difference-in-Sargan test, akin to the Durbin–Wu–Hausman test, evaluates suspected endogenous variables by providing a C-statistic, with the null hypothesis suggesting that these variables may be exogenous For instrumental variables, the Sargan-Hansen test assesses overidentifying restrictions and yields a Hansen J-statistic to determine the validity of the instruments, which should be uncorrelated with the residuals, maintaining the null hypothesis that the instruments are valid Ultimately, a tested variable is considered exogenous if all instruments are valid.

Conversely, the tested variable is endogenous (Baum et al., 2003, 2007)

 Step 2: Address the endogeneity issue (if any)

EMPIRICAL RESULTS AND DISCUSSIONS

DESCRIPTIVE STATISTICS

From 2010 to 2021, research data was collected from six Southeast Asian countries, processed using Microsoft Excel, and imported into the Stata database The dataset includes 72 balanced panel observations, featuring a time dimension of 12 years and a cross-sectional dimension of 6 countries Descriptive statistics, including mean, standard deviation, minimum, maximum, skewness, and kurtosis, are provided to summarize the findings.

Dev Min Max Skewness Kurtosis Dependence lnFDI 72 23.350 0.971 20.791 25.645 0.283 3.229

Second model pENV 72 -0.212 0.072 -0.39 -0.04 -0.050 2.528 pSOC 72 0.374 0.105 0.15 0.57 -0.129 2.155 pGOV 72 0.503 0.132 0.2 0.78 -0.008 2.616

Third model eEMI 72 -0.628 0.136 -0.96 -0.28 -0.095 3.172 eNAT 72 -0.122 0.161 -0.42 0.37 0.493 3.274 sEDU 72 0.629 0.183 0.07 0.95 -0.433 3.018 sEMP 72 0.166 0.159 -0.21 0.80 0.519 5.489 gGOV 72 0.584 0.194 0.16 0.97 -0.116 2.208 gHUM 72 0.495 0.227 0 0.99 0.365 2.677

(Source: the author’s summary from Stata)

Based on the descriptive statistics, the author draws several conclusions:

Table 3.1 displays the natural logarithm of foreign direct investment (FDI) inflows for six Southeast Asian countries from 2010 to 2021, encompassing 72 observations The average lnFDI is 23.350, translating to an annual FDI inflow of about 10.691 billion USD (e^23.350) The data indicates considerable variability in FDI inflows both among the countries and throughout the years, with the lowest recorded inflow being 1.814 billion USD from the Philippines.

In 2010, foreign direct investment (FDI) reached a minimum of 20.791 billion USD and peaked at 52.470 billion USD, with Singapore reporting 25.645 billion USD in 2021 Statistical analysis indicates that the natural logarithm of FDI (lnFDI) exhibits a standard deviation of 0.971, a skewness of 0.283, and a kurtosis of 3.229 This suggests a right-skewed distribution with a peak that exceeds that of a normal distribution, indicating that data is primarily concentrated on the lower end, while some observations, particularly from Singapore, show significantly higher values.

Table 3.1 shows that OPEN, INFLATION, and INTEREST have 72 observations, with mean values of (142.340, 2.930, 3.438) respectively, indicating that the average annual trade openness, inflation, and real interest rate of ASEAN-6 over

Over a 12-year period, trade openness, inflation, and real interest rates exhibit significant variability, with trade openness ranging from a minimum of 32.97% in Indonesia (2020) to a maximum of 379.1% in Singapore (2011) Inflation rates span from -1.14% in Malaysia (2020) to 18.68% in Vietnam (2011), while real interest rates fluctuate between -20.5% in Vietnam (2010) and 9.99% in Indonesia (2020) The statistical analysis reveals a standard deviation of 98.967 for trade openness, 2.836 for inflation, and 4.012 for real interest rates, alongside skewness values of 1.197, 2.597, and -3.050, respectively The kurtosis values of 3.357, 14.735, and 19.016 indicate that both trade openness and inflation have a right-skewed and leptokurtic distribution, suggesting data concentration on the left side of the mean, whereas real interest rates display a left-skewed and highly leptokurtic distribution, indicating concentration on the right side of the mean.

Table 3.1 presents an average variation across 72 ESG observations, revealing that the annual average Sovereign ESG score for six countries over twelve years is 0.169 The scores range from a low of -0.05 in Vietnam (2010) to a high of 0.50 in Indonesia (2021) With a standard deviation of 0.109, the data exhibits a right-skewed distribution (skewness = 0.821) and is more leptokurtic than normal (kurtosis = 3.687), indicating that the ESG data is less dispersed and concentrated to the left of the mean.

Over a 12-year period, the ASEAN-6 countries exhibit average annual scores in Environmental (pENV), Social (pSOC), and Governance (pGOV) pillars of -0.212, 0.374, and 0.503, respectively, as detailed in Table 3.1 Notably, pENV recorded a minimum score of -0.39, attributed to Vietnam.

The analysis of the three variables reveals a low to moderate range of variation, with pSOC ranging from a minimum of 0.15 in Malaysia (2010) to a maximum of 0.57 in Indonesia (2020), pGOV varying between 0.2 in Vietnam (2011) and 0.78 in Indonesia (2019), and pSOC showing a minimum of -0.04 in Malaysia (2017) The standard deviations for these variables are 0.072, 0.105, and 0.132, indicating a slight dispersion Additionally, the skewness values are -0.050, -0.129, and -0.008, suggesting a slightly left-skewed distribution, while the kurtosis values of 2.528, 2.155, and 2.616 indicate a platykurtic distribution, further confirming that the data is relatively dispersed compared to a normal distribution.

Table 3.1 presents the ASEAN-6 scores for Emissions & Pollution (eEMI) at -0.628 and Natural Capital Endowment & Management (eNAT) at -0.122, revealing a moderate to wide range of variation eEMI ranges from a minimum of -0.96 in Vietnam (2020) to a maximum of -0.28 in Singapore (2010), while eNAT varies from -0.42 in Indonesia (2012) to 0.37 in Thailand (2020) The standard deviations for eEMI and eNAT are 0.136 and 0.161, respectively, with eEMI exhibiting an approximately normal distribution and eNAT showing a slight right skew, as indicated by their skewness values of -0.095 and 0.493, and kurtosis values of 3.172 and 3.274.

Over a 12-year period, the ASEAN-6 countries recorded average annual scores of 0.629 in Education & Skills (sEDU) and 0.166 in Employment (sEMP) Notably, sEDU scores varied significantly, with a low of 0.07 in Singapore in 2021 and a high of 0.95 in Thailand during 2010 and 2016 Meanwhile, the minimum score for sEMP remains unspecified.

The correlation coefficient between sEDU and sEMP varies significantly across countries, ranging from a low of 0.21 in Vietnam in 2021 to a high of 0.8 in Thailand in 2015 Notably, the standard deviation and skewness values indicate that sEDU follows a left-skewed distribution, whereas sEMP is right-skewed Furthermore, the kurtosis values reveal that sEMP is more leptokurtic than a normal distribution, with data concentrated to the left of the mean.

Over a 12-year period, the average scores for Government Effectiveness (gGOV) and Human Rights (gHUM) among six countries were 0.584 and 0.495, respectively gGOV ranged from a low of 0.16 in Thailand (2010) to a high of 0.97 in Vietnam (2020), while gHUM varied from 0.00 in the Philippines (2021) to 0.99 in Indonesia (2018), reflecting significant variability in both metrics The standard deviations of 0.194 for gGOV and 0.227 for gHUM, along with skewness values of -0.116 and 0.365, indicate that gGOV is slightly left-skewed and gHUM is slightly right-skewed Additionally, both distributions exhibit platykurtic characteristics, with kurtosis values of 2.208 and 2.677, suggesting they are less peaked than a normal distribution.

FIRST MODEL ANALYSIS

Table 4.2 First model pairwise correlations

An analysis of the correlation between control variables and lnFDI reveals notable trends Notably, OPEN exhibits the strongest positive correlation with lnFDI at 0.706, while INFLATION displays the most significant negative correlation at -0.146 In contrast, ESG shows a relatively weak positive correlation with lnFDI at 0.274, and INTEREST has a near-zero correlation at 0.09 Additionally, the independent variables generally demonstrate low inter-correlation, with the exception of a moderate negative correlation between INFLATION and ESG at -0.398.

Performing multicollinearity test for the first model on Stata, the author obtained following results:

Table 4.3 First model Variance inflation factor

The analysis of the first model in Table 4.3 indicates no signs of multicollinearity, as evidenced by the highest Variance Inflation Factor (VIF) value of 1.37 for INFLATION and an average VIF of 1.20, both significantly below the accepted threshold of 5 (Hoàng Trọng & Chu Nguyễn Mộng Ngọc, 2008).

4.2.3 POLS/FEM/REM regression result summary

Estimating the first research model using all three methods POLS, FEM and REM on Stata, the author obtained the following results:

Table 4.4 First model F and Chi 2 result

H0: All independent variables are zero

Table 4.4 presents the results of three estimation methods: POLS, FEM, and REM, yielding F statistics of 25.63, 3.36, and 43.66, respectively, along with Chi-squared statistics The corresponding p-values are 0.000, 0.015, and 0.000, indicating statistical significance at the 10% level This evidence allows for the rejection of the null hypothesis (H0) that all independent variables are zero, confirming the suitability of the research model across all three estimation methods.

Table 4.5 Summarizing first model regression

Variable POLS FEM REM lnFDI lnFDI lnFDI

Under POLS estimation method, OPEN, INFLATION, and INTEREST significantly positively influences at 1%, 5%, and 10% levels, respectively However,

54 their effects were relatively small (0.007, 0.065, 0.033) In contrast, ESG had a notably stronger positive impact (2.842) and was significant at the 1% level

Fixed effects analysis revealed that OPEN, INFLATION, INTEREST are statistically insignificant However, ESG demonstrated a significant positive relationship at the 1% level, with a medium positive impact (1.810)

The REM model indicated that OPEN had a small effect on the outcome (0.007), but it was significant at the 1% level INFLATION and INTEREST were insignificant

In contrast, ESG had a significant positive impact at the 1% level, with a moderate effect size of 2.127

4.2.4.1 Pooled OLS and FEM regression

Estimating the first model using the Fixed Effects Model, the author obtained the following F-test results:

Table 4.6 First model F-test result

H 0 : POLS better than fixed-effect model

The F-test results in table 4.6 reveal a significant F-statistic of 12.42 and a p-value of 0.000, indicating strong evidence against the null hypothesis that POLS outperforms FEM Consequently, this suggests that FEM is a more suitable approach for the first model, providing a more accurate representation of the data.

REM estimates are compared to FEM estimates using the Hausman test (based on efficient estimator variances) for the first model With FEM being consistent and

REM efficient, the Chi 2 test results are as follows:

Table 4.7 First model Hausman FEM-REM test results

H 0 : difference in coefficients not systematic

The Hausman test in table 4.7 (Chi 2 (4) = 14.30, p-value = 0.006) strongly rejects the null hypothesis of no systematic differences in coefficients This suggests that

FEM is more appropriate for the first model than POLS

The author employed a FEM for the first model and then used a Modified Wald test to check for groupwise heteroskedasticity The results are as follows:

Table 4.8 First model Modified Wald test

The modified Wald test in table 4.8 (Chi 2 = 229.15, p-value = 0.000) strongly rejects the null hypothesis of homoskedasticity, indicating the presence of heteroskedasticity in the first model

Performing Wooldridge test for autocorrelation in panel data for the first model, the author obtained the following results:

Table 4.9 First model Wooldridge test for autocorrelation

The Wooldridge test results presented in Table 4.9 indicate an F-statistic of 0.258 and a p-value of 0.633, suggesting no evidence of first-order autocorrelation in the first model Consequently, the author concludes that while the first model exhibits heteroskedasticity, it does not demonstrate any signs of autocorrelation.

To address the heteroskedasticity in the first model, the Feasible Generalized Least Squares (FGLS) method was employed The results are as follows:

Table 4.10 First model Cross-sectional time-series FGLS regression lnFDI Coef Std.Err z P>|z| Sig

Number of obs 72 Wald chi 2 (4) 415.32

The Wald Chi 2 statistic is 415.32 with a p-value of 0.000, indicating that the FGLS estimation is suitable However, the presence of endogeneity in the model could lead to biased results Consequently, the following section will address testing and correcting for endogeneity in the initial model.

To address potential endogeneity issues, tests were conducted on each explanatory variable in the initial model, incorporating instrumental variables from the remaining explanatory variables and excluding dummy instruments based on the first and second lags of the instrumented variables The estimation process employed the Feasible Two-Step Efficient Generalized Method of Moments (EGMM), a suitable approach for static models and small samples that effectively accommodates heteroskedasticity.

Test results for the suspected endogenous variables:

Table 3.11 Difference-in-Sargan test regressor for endogeneity results

H 0 : specified endogenous regressor actually be treated as exogenous

Test results for the corresponding instrumental variables:

Table 4.12 Sargan-Hansen test overidentifying restriction results

Instrumented Instrumental variables Chi 2 (1) p-value

OPEN Include : INFLATION, INTEREST, ESG

INFLATION Include : OPEN, INTEREST, ESG

INTEREST Include : OPEN, INFLATION, ESG

ESG Include : OPEN, INFLATION, INTEREST

The endogeneity and instrumental variable tests reveal that while the instrumental variables for INFLATION and INTEREST are valid, with p-values of 0.8949 and 0.6733 exceeding the 10% significance level, the Difference-in-Sargan test indicates endogeneity, as evidenced by p-values of 0.0077 and 0.0158 falling below this threshold Consequently, there is substantial evidence to reject the null hypothesis, confirming that both INFLATION and INTEREST are endogenous variables.

To address the endogeneity issue in the first model, ESG was used as an instrumented variable; include instrumental variables were OPEN, INFLATION, and

INTEREST; excluded instrumental variables were L(1/2).ESG (lag 1, 2), L.OPEN

(lag 1) Using Feasible two-step Efficient GMM (EGMM) estimation method, the author obtained the following results in the final test:

Testing results for the instrumented variable ESG:

Table 4.13 Difference-in-Sargan test regressor for endogeneity result

Based on table 4.13, the difference-in-Sargan test shows that there is not enough evidence to reject the null hypothesis that ESG is exogenous (Chi 2 (1) = 1.062, p-value

= 0.3029 > 10%) Therefore, it can be concluded that variable ESG is exogenous

Testing results for the set of instrumental variables:

Table 4.14 Sargan-Hansen test overidentifying restriction for all instrumental variables result

ESG Include : OPEN, INFLATION, INTEREST

The Sargan-Hansen test results indicate a Chi-square value of 1.619 and a p-value of 0.445, which is above the 10% threshold, leading to the conclusion that the null hypothesis regarding the validity of the instruments cannot be rejected Consequently, both tests confirm that the instruments employed to analyze ESG are appropriate, establishing that ESG is exogenous in the first model.

Testing results for each instrumental variable:

Table 4.15 Difference-in-Sargan test overidentifying restriction for each instrumental variable results

Table 18 displays the Difference-in-Sargan test results for each instrumental variable, indicating that all Hansen J and C statistics are positive, with p-values surpassing 10% These results reinforce the null hypothesis of instrument validity, confirming that all instrumental variables are exogenous.

Duplicate and multicollinearity variable test results:

Table 4.16 Multicollinearity and duplicate variable test results

Statistical results for multicollinearity and duplicate variable (table 4.16) show that, there is no multicollinearity or duplicate variables when using instrumental variables in the first model

Estimation results of the first model:

Table 4.17 First model EGMM regression results lnFDI Coef Robust std err t-value p-value Sig

The EGMM regression results for the first model, as shown in Table 4.17, indicate a strong fit with an F-statistic of 33.14 and a p-value of 0.000, confirming the model's appropriateness The centered R² value of 0.6948 suggests that the independent variables account for approximately 69.48% of the variation in the dependent variable.

In next section, the author will conduct specification tests for the first model

The author conducts specification tests on the initial model derived from the estimation results, utilizing the Feasible Two-Step Efficient Generalized Method of Moments (EGMM) tailored for a static model and small sample sizes.

A Wald test was conducted to assess whether the explanatory variables, namely control variables (OPEN, INFLATION, INTEREST) and ESG, were redundant The findings were as follows:

The Wald test results presented in Table 4.18 indicate that both the control variables and ESG have a p-value of 0.000, providing strong evidence to reject the null hypothesis that the variables are unsuitable Consequently, it can be concluded that there is no issue of redundant variables in the first model.

A Ramsey test was performed to test for omitted variables The results were as follows:

Table 4.19 Ramsey/Pesaran-Taylor RESET test results

The Ramsey/Pesaran-Taylor RESET test results indicate a Chi 2 value of 0.00 and a p-value of 0.965, providing strong evidence to accept the null hypothesis (H0) that E(y|X) is linear in X This conclusion suggests that there is no issue with missing variables in the first model.

After performing a normality test on the residuals, following results were obtained:

Table 4.20 Skewness and kurtosis tests for normality result

H 0 : The residuals are normally distributed

The skewness and kurtosis tests for normality indicate that the residuals are normally distributed, supporting the null hypothesis With an adjusted chi-squared statistic of 0.61 and 2 degrees of freedom, the p-value of 0.736 exceeds the 10% significance level.

SECOND MODEL ANALYSIS

Table 4.21 Second model pairwise correlations

Correlation between independent variables and dependent variable

64 pENV and pGOV are positively correlated with the dependent variable lnFDI pGOV shows a moderate positive correlation (0.456), while pENV exhibits a lower positive correlation (0.259) In contrast, pSOC presents a negligible correlation

(0.003) with lnFDI Relationships between control variables OPEN, INFLATION, INTEREST and lnFDI are explained in the first model

The analysis of the correlation between the independent variables pENV, pSOC, and pGOV indicates that they exhibit moderate to weak relationships with the control variables Specifically, pENV demonstrates the strongest positive correlation with OPEN at 0.380, while it also reveals a significant negative correlation with INFLATION at -0.472.

Additionally, moderate to weak positive correlations exist among pENV, pSOC, and pGOV Relationships between control variables remain consistent with those observed in the first model

Conducting multicollinearity diagnostics for the second research model on Stata, the author obtained following results:

Table 4.22 Second model Variance inflation factor

Table 4.22 indicates that the second model exhibits no signs of multicollinearity, with the highest Variance Inflation Factor (VIF) coefficient recorded at 1.76 for pENV and an average VIF of 1.51 These values remain significantly below the threshold of 5, suggesting a robust model (Hoàng Trọng & Chu Nguyễn Mộng Ngọc).

2008), suggesting a minimal risk of multicollinearity

The second research model was estimated using POLS, FEM, and REM in Stata Resulting statistical F and Chi 2 values are as follows:

Table 4.23 Second model F and Chi 2 result

H 0 : All independent variables are zero

The analysis presented in Table 4.23 reveals that all three estimation methods yield F and Chi-squared values of 27.78, 4.39, and 166.66, along with p-values of 0.000, 0.001, and 0.000, respectively These results strongly indicate that the null hypothesis of an unsuitable model can be rejected.

Therefore, all 3 estimation methods are appropriate for the second model

Table 4.24 Summarizing second model regression

Variables POLS FEM REM lnFDI lnFDI lnFDI

The POLS estimation method reveals an R² value of 0.719, indicating that 71.90% of the variation in lnFDI is accounted for by the independent variables, while the FEM method shows an R² of 0.305, explaining 30.50% of the variation In the POLS model, OPEN is significant at the 1% level with a small positive effect (0.0084), and INFLATION is significant at the 10% level with a slight positive impact (0.0531), whereas INTEREST is not statistically significant At the category level, pENV and pGOV are both significant at the 1% level, with pENV exhibiting a negative effect (-3.615) and pGOV a positive effect (3.292) Additionally, pSOC is significant at the 5% level, showing a positive effect with a coefficient of 1.908.

According to the FEM estimation method, none of the control variables show statistical significance At the category level, pSOC demonstrates a significant positive effect at the 1% level (2.289), while pENV reveals a significant negative effect at the 5% level (-2.514) In contrast, pGOV does not exhibit statistical significance.

The REM method and the POLS method generate similar outcomes

By estimating the second model using Fixed Effects Model, the author obtained results of the F-test to compare POLS and FEM, as shown in the table below:

Table 4.25 Second model F-test result

F-test results in table 4.25 shows that F-statistic has a value of 8.17 and p-value of 0.000, providing strong evidence to reject H0 that POLS is better than FEM

Therefore, using FEM method is superior to POLS for the second model

A Hausman test was performed to compare the Random Effects Model (REM) and Fixed Effects Model (FEM) estimates for the second model Both (co)variance matrices utilized the disturbance variance estimate from the efficient estimator While FEM is known for its consistency, REM is recognized for its efficiency, with the Chi-squared test results indicating the differences between the two models.

Table 4.26 Second model Hausman test results

H 0 : difference in coefficients not systematic

The Hausman test results indicate a Chi-squared value of 75.92 with a p-value of 0.000, providing compelling evidence to reject the null hypothesis (H0) that the differences in coefficients are not systematic Consequently, the Fixed Effects Model (FEM) is deemed more appropriate for the second model.

The second model was estimated using FEM, followed by a Modified Wald test to assess groupwise heteroskedasticity The test results were as follows:

Table 4.27 Second model Modified Wald test

The results of the Modified Wald test, presented in Table 4.27, indicate a Chi-squared value of 145.81 and a p-value of 0.000 This strong statistical evidence leads to the rejection of the null hypothesis of homoskedasticity, confirming the presence of heteroskedasticity in the second model.

The author conducted a Wooldridge test for autocorrelation on the second model and obtained following results:

Table 4.28 Second model Wooldridge test for autocorrelation

The F-statistic of 0.001 and p-value of 0.977 indicate that the null hypothesis of no first-order autocorrelation cannot be rejected, suggesting no evidence of autocorrelation in the second model Consequently, while the second model is identified as heteroskedastic, it does not demonstrate any signs of autocorrelation.

Similar to the first model, the second model was estimated using the FGLS method to address heteroskedasticity and correlated errors The results are as follows:

Table 4.29 Second model Cross-sectional time-series FGLS regression lnFDI Coef Std.Err z P>|z| Sig

The FGLS estimation results in Table 4.29 indicate strong support for the second model, evidenced by a Wald Chi-square value of 498.42 and a p-value of 0.000 However, the presence of endogeneity in the model may result in biased outcomes The following section will focus on testing for endogeneity and implementing corrective measures for the second model.

Endogeneity tests were conducted for each explanatory variable in the second model The remaining explanatory variables served as include instrumental variables,

To tackle heteroskedasticity in small sample sizes, the model was estimated using the Feasible Two-Step Efficient Generalized Method of Moments (EGMM) for static models, while dummy exclude instruments were created based on the first and second lags of the instrumented variables.

Include : INFLATION, INTEREST, pENV, pSOC, pGOV

Include : OPEN, INTEREST, pENV, pSOC, pGOV Exclude : L.INFLATION, L2.INFLATION 0.028 0.8676 INTEREST

Include : OPEN, INFLATION, pENV, pSOC, pGOV Exclude : L.INTEREST, L2.INTEREST 0.982 0.3218 pENV

Include : OPEN, INFLATION, INTEREST, pSOC, pGOV

Include : OPEN, INFLATION, INTEREST, pENV, pGOV

Exclude : L.pSOC, L2.pSOC 0.016 0.8992 pGOV Include : OPEN, INFLATION, INTEREST pENV, pSOC 1.771 0.1832

The results of the endogeneity test (Table 4.30) and the instrumental variable test (Table 4.31) reveal that while the instrumental variables for INFLATION, INTEREST, and pGOV are valid, indicated by p-values of 0.8676, 0.3218, and 0.1832 respectively, the Difference-in-Sargan test shows evidence of endogeneity with p-values of 0.0251, 0.0418, and 0.0741, all below the 10% significance level This leads to the rejection of the null hypothesis that INFLATION, INTEREST, and pGOV are exogenous variables, confirming that all three are endogenous.

To tackle the endogeneity issue in the second model, the author employed pGOV as the instrumental variable, while incorporating OPEN, INFLATION, INTEREST, pENV, and pSOC as included instrumental variables Additionally, L.pGOV (lag 1), L.pSOC (lag 1), and L.pENV (lag 1) were utilized as excluded instrumental variables The author applied the EGMM estimation method, resulting in significant findings in the final test.

Testing results for the instrumented variable pGOV:

Instrumented variable Chi 2 (1) p-value pGOV 2.605 0.1065

The Difference-in-Sargan test results indicate that there is insufficient evidence to reject the null hypothesis, suggesting that the specified endogenous regressor can be considered exogenous Consequently, it can be concluded that the variable pGOV is exogenous, as evidenced by the Chi-square statistic of 2.605 and a p-value of 0.1065, which exceeds the 10% significance level.

Instrumented Instrumental variables Chi 2 (2) p-value pGOV

Include : OPEN, INFLATION, INTEREST, pENV, pSOC

The Sargan-Hansen test (table 4.33), with a Chi2(1) = 0.322 and a p-value 0.851 (greater than 10%), does not reject the null hypothesis of valid instruments

This indicates that the instruments used to explain pGOV are appropriate

Consequently, it is concluded that pGOV is exogenous in the second model

Table 4.34 Difference-in-Sargan test overidentifying restriction for each instrumental var i able results

The Difference-in-Sargan test indicates that all Hansen J and C-statistics are positive, with p-values exceeding 10%, supporting the null hypothesis of valid instruments Therefore, it can be concluded that all instrumental variables utilized in the model are exogenous.

Duplicate and multicollinearity variable test results:

Table 4.35 Multicollinearity and duplicate variable test results

Table 4.35 reveals that as both the number of multicollinear and duplicate variables are zero, therefore all instrumental variables in the second model are free from multicollinearity and duplicate

Estimation results of the second model:

Table 4.36 Second model EGMM regression results lnFDI Coef Robust std err t-value p-value Sig

Table 4.36 presents the EGMM regression results for the second model F(6,

59) = 48.75 and Prob > F = 0.000 provide strong evidence to conclude that the

73 second model is appropriate The centered R² of 0.732 suggests that the independent variables explain approximately 73.20% of the variation in lnFDI

Subsequently, the author will focus on specification tests for the second model

Following Feasible Two-Step EGMM estimation of a static model with small sample, specification tests were conducted on the second model The test results are shown below:

A Wald test was performed to assess whether the explanatory variables, namely control variables (OPEN, INFLATION, INTEREST) and the Sovereign ESG pillar group (pENV, pSOC, pGOV), were redundant

H 0 : variable(s) are not suitable (redundant)

OPEN, INFLATION, INTEREST 64.85 0.0000 pENV, pSOC, pGOV 29.50 0.0000

From Wald test results in table 4.37, all variables have p-value = 0.000, which strongly supports the rejection of null hypothesis that variables are not suitable

Therefore, the second model does not have redundant variables

Ramsey test results for omitted variable bias are as follows:

Table 4.38 Ramsey/Pesaran-Taylor RESET test results

THIRD MODEL ANALYSIS

Table 4.40 Third model pairwise correlations

Correlation between independent variables and dependent variable

Table 4.40 reveals that eEMI has a moderately positive correlation with lnFDI, indicated by a coefficient of 0.442 In contrast, eNAT, gGOV, and gHUM show low positive correlations of 0.173, 0.252, and 0.211, respectively Additionally, sEDU presents a low negative correlation of -0.210 with lnFDI, while sEMP reflects a nearly negligible correlation of 0.098 The initial model provides an explanation for the correlations between the control variables and lnFDI.

The analysis reveals that six independent category-level variables (eEMI, eNAT, sEDU, sEMP, gGOV, gHUM) exhibit moderate to low correlations with control variables Notably, eEMI shows a significant positive correlation with OPEN at 0.525, while eNAT demonstrates a negative correlation with INFLATION at -0.387 Overall, the pairwise correlations among these six variables are generally weak, with the strongest correlation observed between gGOV and gHUM at 0.304.

Correlations among control variables are noted in the first model

Conducting a multicollinearity test for the third research model on Stata, the author obtained following results:

Table 4.41 Third model Variance inflation factor

Table 4.41 indicates that OPEN has the highest VIF (1.71), while the average VIF remains low at 1.43 Given that all VIF coefficients and the mean VIF are below

5 (Hoàng Trọng & Chu Nguyễn Mộng Ngọc, 2008), there is no indication of multicollinearity in the third model

Table 4.42 Third model F and Chi 2 result

H 0 : All independent variables are zero

F(9, 62) 11.39 F(9,57) 2.96 Wald chi 2 (9) 102.53 Prob > F 0.000 Prob > F 0.0058 Prob > chi 2 0.000

Table 4.42 shows that the three regression methods—POLS, FEM, and REM—yield F and Chi-squared values of 11.39, 2.96, and 102.53, respectively, with corresponding p-values of 0.000, 0.0058, and 0.000, all of which are below the 10% significance level This provides sufficient evidence to reject the null hypothesis that the model is unsuitable, indicating that all three estimation methods are appropriate for the third model.

Table 4.43 Third model summarizing regression Variables OLS FEM REM lnFDI lnFDI lnFDI

The POLS estimation method reveals an R² value of 0.623, indicating that 62.30% of the variation in lnFDI is explained by the independent variables, while the FEM method shows an R² of 0.319, accounting for 31.90% of the variation Among the control variables, only OPEN is statistically significant at the 1% level, with a minimal positive effect of 0.00649 In terms of category-level variables, only sEDU and gHUM are significant; gHUM is significant at the 5% level with a low positive impact of 0.764, whereas sEDU is significant at the 10% level with a low negative impact of -0.917.

The FEM estimation method revealed that among the control variables, only OPEN exhibited a statistically significant effect at the 5% level, with a minimal positive impact of 0.00935 Among the category-level variables, eEMI, sEDU, and sEMP were found to be statistically significant Notably, sEDU demonstrated a significant negative effect at the 1% level, with an estimated coefficient of -1.166, while eEMI showed a negative effect at the 10% level, with a coefficient of -1.028 In contrast, sEMP had a significant positive effect at the 10% level, with an estimated coefficient of 0.686 Additionally, both REM and POLS produced similar estimation results.

The author utilized a Fixed Effects Model (FEM) for the third model and performed an F-test to compare POLS and FEM The F-test results are summarized in the table below:

Table 4.44 Third model F-test result

The F-test results indicate an F-statistic of 14.86 and a p-value of 0.000, providing strong evidence to reject the null hypothesis This suggests that the Fixed Effect Model (FEM) is superior to the Pooled Ordinary Least Squares (POLS) method.

A Hausman test was conducted to evaluate the estimates from the Random Effects Model (REM) and the Fixed Effects Model (FEM) for the third model This test utilized the (co)variance matrices based on the disturbance variance estimate from the efficient estimator The results indicate that while FEM is consistent, REM is efficient, with the Chi-squared test results providing further insights into the comparison.

Table 4.45 Third model Hausman test results

H 0 : difference in coefficients not systematic chi 2 (5) Prob > chi 2

As shown in table 4.45, Hausman test results indicate a Chi 2 value of 35.08 with

With 5 degrees of freedom and a p-value of 0.000, which is less than the significance level of 10%, there is substantial evidence to reject the null hypothesis This indicates that the differences in coefficients between the Random Effects Model (REM) and the Fixed Effects Model (FEM) are systematic Therefore, the Fixed Effects Model is deemed more suitable for the final analysis.

FEM method was estimated for the third model, followed by a Modified Wald test to assess heteroskedasticity The test results were as follows:

Table 4.46 Third model Modified Wald test

The Modified Wald test in table 4.46 strongly indicates the presence of heteroskedasticity in the third model The Chi 2 statistic of 186.16 and p-value of 0.000

< 𝛼 = 10% provide evidence against the null hypothesis of no heteroskedasticity

Performing the Wooldridge test for autocorrelation for the third model, the author obtained following results:

Table 4.47 Third model Wooldridge test for autocorrelation

The Wooldridge test for autocorrelation, as shown in Table 4.47, reveals an F-statistic of 8.446 and a p-value of 0.034, which is less than the significance level of 10% This evidence is adequate to reject the null hypothesis of no first-order autocorrelation, indicating that autocorrelation is present in the third model.

Given the aforementioned results, the author concludes that the third model exhibits both heteroscedasticity and autocorrelation

To address heteroskedasticity and autocorrelation in the third model, Feasible Generalized Least Squares (FGLS) method is employed The results are as follows:

Table 4.48 Third model Cross-sectional time-series FGLS regression lnFDI Coef Std.Err z P>|z| Sig

INTEREST 0.0095 0.007 1.37 0.172 eEMI 0.3343 0.327 1.02 0.307 eNAT 0.9024 0.232 3.89 0.000 *** sEDU -0.8559 0.294 -2.91 0.004 *** sEMP 0.9614 0.255 3.76 0.000 *** gGOV 0.4652 0.255 1.82 0.069 * gHUM 0.1445 0.230 0.63 0.529

The FGLS estimation results for the third model indicate a Wald Chi 2 statistic of 396.55 and a p-value of 0.000, confirming that the model is well-specified However, the presence of endogenous variables may lead to biased estimates Therefore, the author will perform endogeneity tests and apply necessary corrections to the third model in the following section.

Endogeneity tests were conducted on each explanatory variable of the third model, utilizing instrumental variables from the remaining explanatory variables The dummy exclude instruments consisted of lags 1 to 3 of the instrumented variables, with the exception of the variables sEDU and gHUM, which were derived from lags 1 to 5.

83 estimation was conducted using Feasible Two-Step Efficient Generalized Method of Moments (EGMM) for static models and small samples The results are as follows:

INTEREST 4.837 0.0279 eEMI 3.269 0.0706 eNAT 0.148 0.7003 sEDU 0.173 0.6779 sEMP 2.330 0.1269 gGOV 1.317 0.2511 gHUM 0.060 0.8063

Include : INFLATION, INTEREST, eEMI, eNAT, sEDU, sEMP, gGOV, gHUM Exclude : L.OPEN, L2.OPEN, L3.OPEN

Include : OPEN, INTEREST, eEMI, eNAT sEDU, sEMP, gGOV, gHUM Exclude : L.INFLATION, L2.INFLATION, L3.INFLATION

Include : OPEN, INFLATION, eEMI, eNAT, sEDU, sEMP, gGOV, gHUM Exclude : L INTEREST, L2 INTEREST, L3 INTEREST

Include : OPEN, INFLATION, INTEREST, eNAT, sEDU, sEMP, gGOV, gHUM Exclude : L.eEMI, L2.eEMI, L3.eEMI

Include : OPEN, INFLATION, INTEREST, eEMI, sEDU, sEMP, gGOV, gHUM Exclude : L.eNAT, L2.eNAT, L3.eNAT

Include : OPEN, INFLATION, INTEREST, eEM,I eNAT, sEMP, gGOV, gHUM

Exclude : L.sEDU, L2.sEDU, L3.sEDU, L4.sEDU, L5.sEDU

Include : OPEN, INFLATION, INTEREST, eEMI, eNAT, sEDU, gGOV, gHUM Exclude : L.sEMP, L2.sEMP, L3.sEMP

Include : OPEN, INFLATION, INTEREST, eEMI, eNAT, sEDU, sEMP, gHUM Exclude : L.gGOV, L2.gGOV, L3.gGOV

Include : OPEN, INFLATION, INTEREST, eEMI, eNAT, sEDU, sEMP, gGOV Exclude : L.gHUM, L2.gHUM, L3.gHUM, L4.gHUM, L5.gHUM

The Sargan-Hansen test results indicate that all sets of instrumental variables are valid, passing the test at the 10% significance level However, the difference-in-Sargan test reveals that the variables INFLATION, INTEREST, and eEMI have p-values of 0.0540, 0.0279, and 0.0706, respectively, which are below 0.1 This leads to the rejection of the null hypothesis, indicating that these variables are endogenous despite the validity of their instruments.

Perform endogenous mitigation for the third model with instrumented variables are OPEN, INTEREST, eEMI; include instrumental variables being variables

INFLATION, eNAT, sEDU, sEMP, gGOV, gHUM; exclude instrumental variables being variables L(1/3).INTEREST (lag 1 to 3); L(1/4).OPEN (lag 1 to 4); L(1/3).sEDU (lag 1 to 3); L(1/3).eEMI (lag 1 to 3); L(1/4).eNAT (lag 1 to 4);

The study utilized the Feasible two-step Efficient GMM (EGMM) method, specifically the Continuously-Updated approach for static model and small sample estimation, which effectively addresses issues of heteroscedasticity and autocorrelation The analysis focused on the first-difference lag of variables L(1/4)D.sEMP, L(1/5).gGOV, and L(1/4).gHUM The results from the final test reveal significant findings that contribute to the understanding of the model's dynamics.

Testing results for the instrumented variable OPEN, INTEREST and eEMI:

The difference-in-Sargan test results indicate that the variables OPEN, INTEREST, and eEMI have Chi 2 (1) values of 0.051, 0.980, and 0.517, respectively, with p-values of 0.8214, 0.3222, and 0.4720 Since these p-values exceed the 10% significance level, we cannot reject the null hypothesis, leading to the conclusion that OPEN, INTEREST, and eEMI are exogenous variables.

Include : INFLATION, eNAT, sEDU, sEMP, gGOV, gHUM Exclude: L(1/3).INTEREST, L(1/4).OPEN, L(1/3).sEDU, L(1/3).eEMI, L(1/4).eNAT, L(1/4)d.sEMP, L(1/5).gGOV, L(1/4).gHUM

The Sargan-Hansen test results indicate that the instruments used for OPEN, INTEREST, and eEMI are valid, as evidenced by a Chi 2 (1) value of 9.125 and a p-value of 0.9995, which is greater than 10% Consequently, these three variables are considered exogenous in the third model.

Table 4.53 Difference-in-Sargan test overidentifying restriction for each instrumental var i able results

CONCLUSIONS AND IMPLICATIONS

CONCLUSION

This thesis analyzes the factors influencing Foreign Direct Investment (FDI) inflows through three econometric models, focusing on key independent variables such as Trade Openness, Inflation, and Real Interest Rate Additionally, it explores the specific effects of Sovereign Environmental, Social, and Governance (ESG) factors on FDI by breaking down ESG into its fundamental pillars and components across the three models.

 First model: Sovereign ESG (ESG)

 Second model breaks down ESG into three pillars: Environmental pillar (pENV), Social pillar (pSOC), and Governance pillar (pGOV)

 Third model delves deeper into the specific components: Emissions & pollution (eEMI), Natural capital endowment & management (eNAT), Education & skills category (sEDU), Employment (sEMP), Government Effectiveness (gGOV), and Human Rights (gHUM)

The analysis utilizes secondary data from the World Bank's World Development Indicators Database, Sovereign ESG Data Portal, and Sovereign ESG Score Builder, focusing on six Southeast Asian countries over a 12-year period from 2010 to 2021, resulting in a total of 72 observations.

This thesis employed POLS, FEM, and REM methods, determining that the Fixed Effects Model (FEM) is the most appropriate regression technique for all three models To rectify issues of heteroscedasticity and autocorrelation, the author applied Feasible Generalized Least Squares (FGLS) Lastly, the author utilized the Efficient Generalized Method of Moments (EGMM) to tackle endogeneity and omitted variable bias across all models.

This study has contributed to the existing literature on factors influencing sustainable Foreign Direct Investment inflows The research outcomes are presented as follows:

Trade openness and inflation rate positively influence FDI inflows across all three models, while the real interest rate also shows a positive impact on FDI inflows in models 1 and 3.

 Sovereign ESG (ESG) has a positive correlation with FDI inflows in the first model

The Social (pSOC) and Governance (pGOV) pillars positively influence Foreign Direct Investment (FDI) inflows, according to the second model, while the Environmental pillar (pENV) exerts a negative effect on these inflows.

 All Sovereign ESG categories are statistically significant in model 3 Among these categories, Emissions & pollution (eEMI), Natural capital endowment

& management (eNAT), Employment (sEMP), Government Effectiveness (gGOV), and Human Rights (gHUM) have a positive impact on FDI inflows, while Education & skills (sEDU) has a negative impact

All independent variables in the models demonstrate statistical significance and exhibit positive relationships, with the exception of pENV and sEDU The findings indicate that Education and Skills (sEDU) do not have a positive effect on Foreign Direct Investment (FDI) inflows.

RECOMMENDATIONS

Research shows that enhancing Sovereign ESG scores is crucial for attracting foreign direct investment (FDI) in the six Southeast Asian countries With 71 indicators across 17 categories, countries must undertake comprehensive efforts to improve their scores It is essential for ASEAN-6 governments and citizens to understand the significance of Sovereign ESG in investment attraction Each nation should create a tailored roadmap that aligns with its capacities and sustainability objectives, while collectively working towards common goals.

97 roadmap towards the 17 sustainable development goals of the United Nations However, those ESG efforts must be communicated transparently to limit

‘greenwashing’ behaviors right in the national sustainability report

A standardized reporting framework for Sovereign ESG is crucial, encompassing clear definitions, evaluation criteria, scoring methodologies, and continuous monitoring to accurately assess each nation's ESG initiatives Countries such as Indonesia, Malaysia, the Philippines, Thailand, and Vietnam should actively collaborate with experts and draw insights from Singapore's successful implementation of Sovereign ESG solutions.

Research shows that the Environment Pillar, comprising 31 indicators across five categories, negatively affects Foreign Direct Investment (FDI) inflows, supporting the Pollution Haven hypothesis Multinational corporations (MNCs) frequently target areas with lenient environmental regulations to lower production costs, especially in heavy industries Nonetheless, this trend does not imply that attracting FDI necessitates compromising environmental standards.

MNCs should prioritize attracting green FDI in the ASEAN-6 region, focusing on investment projects that promote environmental and social protection Green FDI offers long-term economic advantages and supports sustainable development To achieve this, countries must establish transparent environmental policies, invest in green infrastructure like renewable energy and wastewater treatment, and collaborate with international organizations for access to green capital and technology These strategies will enhance environmental quality and boost regional competitiveness, making countries more appealing to responsible MNCs.

To attract high-quality foreign direct investment (FDI) and promote sustainable development in Southeast Asia, countries must enhance their social factors, focusing on healthcare, education, and social equity Investing in public health infrastructure in underserved areas is crucial for improving population health and fostering a productive workforce Additionally, expanding healthcare coverage and elevating healthcare quality through better-trained medical professionals and advanced equipment are vital steps Governments should prioritize quality education, including general education and vocational training, to equip citizens with essential skills for the modern economy, with particular emphasis on STEM fields and proficiency in foreign languages, especially English.

Reducing inequality and ensuring social security are essential for societal stability Key steps include addressing poverty, raising incomes, and safeguarding vulnerable populations Regional cooperation among Southeast Asian nations can enhance the effectiveness of these efforts By exchanging best practices, resources, and expertise, these countries can expedite advancements in healthcare, education, and social equity.

Research indicates that governance is the most crucial pillar of ESG in attracting foreign direct investment (FDI) Foreign investors prioritize stable and transparent investment environments to safeguard their returns Therefore, enhancing scores on the 18 governance indicators will be the key factor driving FDI inflows into the ASEAN-6 economies.

A robust national governance system is essential for the stability of macroeconomic policies that significantly affect multinational corporations (MNCs) Key factors include maintaining low inflation, stable interest and exchange rates, and a predictable economic environment Furthermore, transparency and efficiency in governance, achieved through streamlined administrative processes and reduced compliance costs, greatly enhance MNC operations.

High-quality governance is essential for safeguarding property rights and ensuring equitable treatment for all investors Countries such as Singapore demonstrate that enhancing the business and governance landscape can attract substantial foreign direct investment (FDI) and establish them as prominent global economic centers Consequently, advancements in these areas will facilitate easier business operations and lower investment transaction costs, ultimately encouraging foreign investment.

Improving performance in emissions and pollution significantly boosts foreign direct investment (FDI) in ASEAN-6 countries By mitigating greenhouse gas emissions such as carbon dioxide, methane, and nitrous oxide, countries can enhance their environmental protection efforts and improve their ESG scores This attracts sustainability-focused investors, particularly in green industries, highlighting the economic benefits of environmental responsibility.

To achieve significant GHG reductions, ASEAN-6 governments must implement strong legal frameworks and effective policies while supporting businesses in adopting cleaner technologies Promoting research and development, as well as facilitating renewable energy projects, is essential Establishing a carbon market will incentivize both multinational corporations and local businesses to lower emissions and create new revenue opportunities Additionally, enhancing cooperation with international partners and financial institutions is vital for securing resources and advanced technologies needed for a green transition Finally, increasing public awareness about environmental protection and the importance of GHG reduction is key to fostering a sustainable society.

Studies have shown that the abundance and effective management of natural resources can attract FDI inflows However, ASEAN Secretariat & EU-ASEAN

100 report (2021) reveals that despite its rich natural capital, the region is facing overexploitation and poor management

Raising awareness about the significance of sustainable natural capital is crucial for ASEAN, which must establish a clear roadmap to incorporate sustainability standards into trade and investment Governments should prioritize funding for ecosystem restoration projects, especially in coastal and mangrove areas, while fostering public-private partnerships to mobilize finance through mechanisms like green bonds The finance sector must redirect its focus from deforestation to supporting environmentally beneficial projects Ultimately, the active involvement of all stakeholders—individuals, communities, businesses, and governments—is essential for the protection and development of ASEAN's natural capital.

Southeast Asian countries possess abundant labor resources; however, inadequate education and skills hinder their ability to attract foreign direct investment (FDI), particularly in labor-intensive sectors This situation necessitates a reassessment of investment attraction policies, prompting governments to target capital-intensive industries that demand high-quality workers To cultivate such a workforce, it is crucial to enhance investment in education, particularly vocational training, ensuring that programs align with the needs of key industries Furthermore, fostering collaboration among educational institutions, multinational corporations (MNCs), and government entities is vital for establishing a continuous training system that adapts to the labor market's evolving demands By implementing these strategies, nations can not only draw FDI but also promote sustainable development through education.

A vibrant labor market with high participation rates and low unemployment is crucial for attracting foreign direct investment (FDI) in ASEAN-6 To leverage this advantage, governments should invest in education and vocational training to equip the workforce with the skills needed by FDI enterprises Promoting lifelong learning will help workers continuously upgrade their skills, enabling them to adapt to the dynamic environments of FDI companies To attract and retain talent, it is vital to enhance working conditions, ensure compliance with labor laws, and foster a safe and healthy work environment Additionally, governments should facilitate connections between job seekers and businesses, particularly FDI enterprises, to support employment opportunities.

Research indicates that Government Effectiveness significantly enhances Foreign Direct Investment (FDI) Specifically, advancements in the quality of public services and regulatory frameworks contribute to a more favorable business environment for private enterprises, thereby playing a crucial role in attracting FDI.

LIMITATIONS AND FURTHER RESEARCH DIRECTIONS

This study investigates the impact of Sovereign ESG, its subcomponents, and macroeconomic variables on foreign direct investment (FDI) inflows, revealing significant evidence of their influence Despite these findings, the research acknowledges certain limitations that should be considered.

This study focuses exclusively on six Southeast Asian countries, which may limit the applicability of the findings to other nations in the region or globally Additionally, the use of secondary data from the World Bank introduces a two-year time lag, as the data is not current with the research period ending in July.

2024), thereby reducing the timeliness of the results and potentially overlooking recent fluctuations in FDI inflows

The research model, while comprehensive, has limitations that may impact its accuracy in assessing the impact of sustainability factors on FDI inflows The inclusion of only three macroeconomic control variables may overlook other crucial determinants of FDI inflows, reducing the model's comprehensiveness Furthermore, the analysis is limited to six ESG categories, whereas the World Bank provides 17 categories, potentially compromising the model's explanatory power To enhance the model's effectiveness, it is recommended to expand it by incorporating additional macroeconomic control variables and the full set of 17 ESG categories provided by the World Bank.

The research methodology of the study is limited due to its basic approach, which overlooks essential unit root and cointegration tests This omission may lead to biased estimates and inaccurate conclusions Additionally, the model is static, focusing solely on the short-term impacts of various factors on foreign direct investment (FDI) inflows, without addressing potential long-term effects.

This thesis highlights key limitations and suggests future research directions focused on the influence of sovereign sustainable development on Foreign Direct Investment (FDI) and the overall macroeconomic landscape of nations.

Future research on foreign direct investment (FDI) should broaden its geographical scope by analyzing trends across various income groups of countries, which will help pinpoint the specific factors influencing FDI within each group Furthermore, extending the time frame of studies will yield a more thorough and precise understanding of the elements that impact FDI.

In addition to analyzing how sustainable development factors influence FDI inflows, future research should expand to encompass FDI outflows, total FDI, and other key macroeconomic indicators, including GDP, exports, and imports.

Future research should comprehensively evaluate the influence of sustainable development factors on foreign direct investment (FDI) and macroeconomic variables by analyzing all 17 categories or the 71 detailed sustainability indicators from the World Bank as of July 2024 To enhance accuracy, studies must rigorously test the assumptions of stationarity and cointegration in the data Additionally, employing dynamic models like Autoregressive Distributed Lag (ARDL) or Vector Autoregression (VAR) will allow for a thorough assessment of both short-term and long-term impacts of sustainable development factors.

The concluding chapter encapsulates the essential research findings and provides recommendations for the overall Sovereign ESG Score, along with insights for each pillar and category based on the results It also acknowledges the study's limitations and suggests avenues for future research.

Agarwal, J (1980) Determinants of foreign direct investment: A survey Review of

Agiomirgianakis, G M., Asteriou, D., & Papathoma, K (2003) The determinants of foreign direct investment: A panel data study for the OECD countries

Working Papers, Article 03/06 https://ideas.repec.org//p/cty/dpaper/03-

Ahmed E., U., & Gabriel, A M (2012) The Determinants and Impacts of Foreign

Direct Investment in Nigeria International Journal of Business and

Management, 7(24), p67 https://doi.org/10.5539/ijbm.v7n24p67

Akın, M Ş., & Vlad, V (2011) The Relationship between Education and Foreign

Direct Investment: Testing the Inverse U Shape European Journal of

Economic and Political Studies, 4(1) https://arastirmax.com/en/publication/european-journal-economic-and- political-studies/4/1/relationship-between-education-and-foreign-direct- investment-testing-inverse-u-shape/arid/bfb04110-60fd-47ce-8169

ASEAN Secretariat, & EU-ASEAN (2021) ASEAN Natural Capital Status Report:

Investing in Sustainable Natural Capital in ASEAN https://asean.org/book/asean-natural-capital-status-report-investing-in- sustainable-natural-capital-in-asean/

Asiedu, E (2002) On the Determinants of Foreign Direct Investment to Developing

Countries: Is Africa Different? World Development, 30(1), 107–119

Asiedu, E (2005) Foreign direct investment in Africa: The role of natural resources, market size, government policy, institutions and political instability (Working Paper 2005/24) WIDER Research Paper https://www.econstor.eu/handle/10419/63499

Asiedu, E (2013) Foreign Direct Investment, Natural Resources and Institutions

Asiedu, E., & Lien, D (2011) Democracy, foreign direct investment and natural resources Journal of International Economics, 84(1), 99–111 https://doi.org/10.1016/j.jinteco.2010.12.001

Aw, Y., & Tang, T C (2009) The Determinants of Inward Foreign Direct

Investment: The Case of Malaysia Monash University, Department of

Economics, Monash Economics Working Papers, 11

Awan, M Z., Khair Uz Zaman, & Khan, B (2010) Determinants of Foreign Direct

Investment in Services Sector of Pakistan an Econometric Approach Busines

Review: Leadership and Ideas for Tomorrow

Bailey, N (2018) Exploring the relationship between institutional factors and FDI attractiveness: A meta-analytic review International Business Review, 27(1), 139–148

Bajo-Rubio, O., & Sosvilla-Rivero, S (1994) An Econometric Analysis of Foreign

Direct Investment in Spain, 1964-89 Southern Economic Journal, 61(1), 104–120 https://doi.org/10.2307/1060133

Bank, I A D (2002) Inter-American Development Bank Annual Report 2001 IDB

Barrell, R., & Pain, N (1999) Trade restraints and Japanese direct investment flows European Economic Review, 43(1), 29–45 https://doi.org/10.1016/S0014-2921(98)00017-8

Baum, C (2001) Residual diagnostics for cross-section time series regression models Stata Journal, 1(1), 101–104

Baum, C., Schaffer, M., & Stillman, S (2003) Instrumental variables and GMM:

Estimation and testing Stata Journal, 3(1), 1–31

Baum, C., Schaffer, M., & Stillman, S (2007) Enhanced routines for instrumental variables/generalized method of moments estimation and testing Stata

Benhabib, J., & Spiegel, M (1994) The role of human capital in economic development evidence from aggregate cross-country data Journal of

Blanton, S L., & Blanton, R G (2007) What Attracts Foreign Investors? An

Examination of Human Rights and Foreign Direct Investment The Journal of Politics, 69(1), 143–155 https://doi.org/10.1111/j.1468-

Blanton, S L., & Blanton, R G (2009) A Sectoral Analysis of Human Rights and

FDI: Does Industry Type Matter?* International Studies Quarterly, 53(2), 469–493 https://doi.org/10.1111/j.1468-2478.2009.00542.x

Blaydes, L., & Kayser, M (2009) Counting Calories: Democracy and Distribution in the Developing World PSN: Other Political Economy: Development

(Topic), 55 https://doi.org/10.2139/ssrn.1521472

Blomstrom, M., Lipsey, R., & Zejan, M (1992) What Explains Developing

Country Growth? (NBER Working Paper 4132) National Bureau of

Economic Research, Inc https://econpapers.repec.org/paper/nbrnberwo/4132.htm

Bokpin, G A., Mensah, Lord, & Asamoah, M E (2015) Foreign direct investment and natural resources in Africa Journal of Economic Studies, 42(4), 608–

Borensztein, E., De Gregorio, J., & Lee, J.-W (1998) How does foreign direct investment affect economic growth?1 Journal of International Economics,

Bouye, E., & Menville, D D (2021) The Convergence of Sovereign

Environmental, Social and Governance Ratings Policy Research Working

Paper Series, Article 9583 https://ideas.repec.org//p/wbk/wbrwps/9583.html

Braun, S (2006) Core Labour Standards and FDI: Friends or Foes? The Case of

Child Labour (SFB 649 Discussion Paper SFB649DP2006-014) Humboldt xiv

University, Collaborative Research Center 649 https://econpapers.repec.org/paper/humwpaper/sfb649dp2006-014.htm Breusch, T S., & Pagan, A R (1979) A Simple Test for Heteroscedasticity and

Random Coefficient Variation Econometrica, 47(5) https://www.jstor.org/stable/1911963

Buchanan, B., Le, Q V., & Rishi, M (2012) Foreign direct investment and institutional quality: Some empirical evidence International Review of

Campos, N F., & Kinoshita, Y (2003) Why Does Fdi Go Where it Goes? New

Evidence From the Transition Economies International Monetary Fund

Canh, N P., Thanh, B N., Thanh, S D., & Schinckus, C (2020) Determinants of foreign direct investment inflows: The role of economic policy uncertainty

International Economics : A Journal Published by CEPII (Center for

Research and Expertise on the World Economy), 161

Chakrabarti, A (2001) The Determinants of Foreign Direct Investments:

Sensitivity Analyses of Cross-Country Regressions Kyklos, 54(1), 89–114 https://doi.org/10.1111/1467-6435.00142

Cheng, L K., & Kwan, Y K (2000) What are the determinants of the location of foreign direct investment? The Chinese experience Journal of International

Chipalkatti, N., Le, Q V., & Rishi, M (2021) Sustainability and Society: Do

Environmental, Social, and Governance Factors Matter for Foreign Direct Investment? Energies, 14(19), 1–18

Copeland, B., & Taylor, M S (1994) North-South Trade and the Environment The

Crifo, P., Diaye, M.-A., & Oueghlissi, R (2017) The effect of countries’ ESG ratings on their sovereign borrowing costs The Quarterly Review of

Culem, C G (1988) The locational determinants of direct investments among industrialized countries European Economic Review, 32(4), 885–904 https://doi.org/10.1016/0014-2921(88)90051-7

Dam, L., & Scholtens, B (2008) Environmental regulation and MNEs location:

Does CSR matter? Ecological Economics, 67(1), 55–65

Deichmann, J., Eshghi, A., Haughton, D., Sayek, S., Teebagy, N., & Topi, H

(1970) Geography Matters: Kohonen Classification of Determinants of Foreign Direct Investment In Transition Economies Journal of Business

Strategies, 20(1), 23–44 https://doi.org/10.54155/jbs.20.1.23-44

Deichmann, J., Eshghi, A., Haughton, D., Sayek, S., Teebagy, N., & Topi, H

(2003) Geography Matters: Kohonen Classification of Determinants of Foreign Direct Investment In Transition Economies Journal of Business

Strategies, 20(1), Article 1 https://doi.org/10.54155/jbs.20.1.23-44

Dev, S., & Nayak, B K (2003) Low Bargaining Power of Labour Attracts FDI in

India SSRN Electronic Journal https://doi.org/10.2139/ssrn.431060

Dietz, S., Gollier, C., & Kessler, L (2018) The climate beta Journal of

Dorożyńska, A., & Dorożyński, T (2014) The Role of Human Capital in Attracting

FDI the Case of the Lodz Region https://www.semanticscholar.org/paper/The-Role-of-Human-Capital-in- Attracting-FDI-the-of-Doro%C5%BCy%C5%84ska-

Doro%C5%BCy%C5%84ski/4d57394e1310b04354882c9ce01da0cafcd5135 Drukker, D (2003) Testing for serial correlation in linear panel-data models Stata

Dunning, J H (1979) Explaining Changing Patterns of International Production: In

Defence of the Eclectic Theory Oxford Bulletin of Economics and Statistics,

Dunning, J H (1998) Location and the Multinational Enterprise: A Neglected

Factor? Journal of International Business Studies, 29(1), 45–66 xvi

Ezirim, C B., Anoruo, E., & Muoghalu, M I (2006) The Impacts of External Debt

Burden and Foreign Direct Investment Remittances on Economic Growth: Empirical Evidence from Nigeria African Economic and Business Review,

Fazira, D R., & Cahyadin, M (2018) The Impact of Interest Rate, Corruption

Perception Index, and Economic Growth on Foreign Direct Investment in ASEAN-6 Jurnal Keuangan Dan Perbankan, 22(4), Article 4 https://doi.org/10.26905/jkdp.v22i4.2355

Gangi, Y (2012) The Impact of Governance on FDI flows to African Countries

World Review of Entrepreneurship Management and Sustainable

Globerman, S., & Shapiro, D (2003) Governance infrastructure and US foreign direct investment Journal of International Business Studies, 34(1), 19–39 Goldberg, M A (1972) The Determinants of U S Direct Investment in the E E

Golub, S., Kauffmann, C., & Yeres, P (2011) Defining And Measuring Green FDI:

An Exploratory Review Of Existing Work And Evidence OECD Working

Papers On International Investment https://doi.org/10.1787/5kg58j1cvcvk- en

Gratcheva, E M., Emery, T., & Wang, D (2021) Demystifying Sovereign ESG Gratcheva, E M., & Gurhy, B (2024) Sovereign Environmental, Social, and

Governance (ESG) Investing: Chasing Elusive Sustainability https://www.imf.org/en/Publications/WP/Issues/2024/05/17/Sovereign-

Environmental-Social-and-Governance-ESG-Investing-Chasing-Elusive- Sustainability-549165

Grosse, R., & Treviủo, L (1996) Foreign Direct Investment in the United States:

An Analysis by Country of Origin Journal of International Business Studies,

27, 139–155 https://doi.org/10.1057/palgrave.jibs.8490129 xvii

Gujarati, D (2011) Econometrics by Example McGraw-Hill https://www.academia.edu/43663954/Econometrics_by_Example

Hailu, Z A (2010) Impact of Foreign Direct Investment on Trade of African

Countries International Journal of Economics and Finance, 2(3), p122 https://doi.org/10.5539/ijef.v2n3p122

Hoang, H H., & Bui, D H (2015) Determinants of foreign direct investment in

ASEAN: A panel approach Management Science Letters, 5(2), 213–222 https://doi.org/10.5267/j.msl.2014.12.015

Hübel, B., & Scholz, H (2020) Integrating sustainability risks in asset management: The role of ESG exposures and ESG ratings Journal of Asset

Huu Cung, N (2021) Labor Force and Foreign Direct Investment: Empirical

Evidence from Vietnam Journal of Asian Finance Economics and Business,

8, 103–112 https://doi.org/10.13106/jafeb.2021.vol8.no1.103

Iamsiraroj, S (2016) The foreign direct investment–economic growth nexus

International Review of Economics & Finance, 42(C), 116–133

The International Labour Office's 2002 report, "A Future Without Child Labour," emphasizes the urgent need to eradicate child labor globally, aligning with the ILO Declaration on Fundamental Principles and Rights at Work This comprehensive report, prepared by the Director-General, outlines strategies and recommendations for achieving a future where children are free from exploitation and can enjoy their right to education and development.

Jiang, P.-C., Feng, G.-F., & Yang, H.-C (2022) New measurement of sovereign ESG index Innovation and Green Development, 1(2), 100009 https://doi.org/10.1016/j.igd.2022.100009

JP Morgan (2020) Why COVID-19 could be a major turning point for ESG investing https://www.jpmorgan.com/insights/global-research/esg/covid-19-esg-investing

Kheng, V., Sun, S., & Anwar, S (2017) Foreign direct investment and human capital in developing countries: A panel data approach Economic Change and Restructuring, 50, 341–365 https://doi.org/10.1007/s10644-016-9191-0 xviii

Kisto, M (2017) Determinants Of Foreign Direct Investment In Mauritius:

Evidence From Time Series Data 6(08)

Kolev, G I (2014) Robust variance estimation in panel data generalized least squares regression

Kucera, D (2002) The Effects of Core Workers Rights on Labour Costs and

Foreign Direct Investment: Evaluating the “Conventional Wisdom.” SSRN

Electronic Journal https://doi.org/10.2139/ssrn.313079

Lederman, D., Loayza, N V., & Soares, R R (2005) Accountability and

Corruption: Political Institutions Matter Economics & Politics, 17(1), 1–35 https://doi.org/10.1111/j.1468-0343.2005.00145.x

Li, X., & Liu, X (2005) Foreign Direct Investment and Economic Growth: An

Increasingly Endogenous Relationship World Development, 33(3), 393–407 https://doi.org/10.1016/j.worlddev.2004.11.001

Lins, K V., Servaes, H., & Tamayo, A (2017) Social Capital, Trust, and Firm

Performance: The Value of Corporate Social Responsibility during the

Financial Crisis The Journal of Finance, 72(4), 1785–1824 https://doi.org/10.1111/jofi.12505

List, J A., & Co, C Y (2000) The Effects of Environmental Regulations on

Foreign Direct Investment Journal of Environmental Economics and

Management, 40(1), 1–20 https://doi.org/10.1006/jeem.1999.1095

Martin, P., & I.P Ottaviano, G (1999) Growing locations: Industry location in a model of endogenous growth European Economic Review, 43(2), 281–302 https://doi.org/10.1016/S0014-2921(98)00031-2

Masron, T A., & Abdullah, H (2010) Institutional quality as a determinant for FDI

Inflows: Evidence from ASEAN World Journal of Management, 2(3),

Mattila, P., & Sasi, S (2023) Towards a Sustainable Future: An Examination of

Corporate Responses to the CSRD in the Real-Estate Sector https://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-522102

Mengistu, A A., & Adhikary, B K (2011) Does good governance matter for FDI inflows? Evidence from Asian economies Asia Pacific Business Review,

Montinola, G R., & Jackman, R W (2002) Sources of Corruption: A Cross-

Country Study British Journal of Political Science, 32(1), 147–170

Moore, K A., & Driscoll, A K (1997) Low-Wage Maternal Employment and

Outcomes for Children: A Study The Future of Children, 7(1), 122–127 https://doi.org/10.2307/1602584

Moosa, I A (2002) Foreign Direct Investment Palgrave Macmillan UK https://doi.org/10.1057/9781403907493

Mottaleb, K., & Kalirajan, K (2010) Determinants of Foreign Direct Investment in

Developing Countries: A Comparative Analysis [ASARC Working Paper]

The Australian National University, Australia South Asia Research Centre https://econpapers.repec.org/paper/pasasarcc/2010-13.htm

Narula, R., & Marin, A (2003) FDI Spillovers, Absorptive Capacities and Human

Capital Development: Evidence from Argentina https://research.cbs.dk/en/publications/fdi-spillovers-absorptive-capacities- and-human-capital-developmen

Nelson, R., & Phelps, E (1965) Investment in Humans, Technological Diffusion and Economic Growth (Cowles Foundation Discussion Paper 189) Cowles

Foundation for Research in Economics, Yale University https://econpapers.repec.org/paper/cwlcwldpp/189.htm

Nizam, I., & Hassan, Z (2018) THE IMPACT OF GOOD GOVERNANCE OF

FOREIGN DIRECT INVESTMENT INFLOWS: A STUDY ON THE SOUTH ASIA REGION https://doi.org/10.24924/ijabm/2018.04/v6.iss1.66.79 xx

Nordhaus, W (1977) Economic Growth and Climate: The Carbon Dioxide

Omri, A., Nguyen, D K., & Rault, C (2014a) Causal interactions between CO2 emissions, FDI, and economic growth: Evidence from dynamic simultaneous-equation models Economic Modelling, 42, 382–389 https://doi.org/10.1016/j.econmod.2014.07.026

Omri, A., Nguyen, K., & Rault, C (2014b) Causal interactions between CO2 emissions, FDI, and economic growth: Evidence from dynamic simultaneous-equation models Economic Modelling, 42, 382–389 https://doi.org/10.1016/j.econmod.2014.07.026

Opoku, E E O., Acheampong, A O., Dzator, J., & Kufuor, N K (2022) Does environmental sustainability attract foreign investment? Evidence from developing countries Business Strategy and the Environment, 31(7), 3542–

Opoku, E E O., & Boachie, M K (2020) The environmental impact of industrialization and foreign direct investment Energy Policy, 137(C) https://econpapers.repec.org/article/eeeenepol/v_3a137_3ay_3a2020_3ai_3ac _3as0301421519307645.htm

Pain, N., & Holland, D (1998) The Diffusion Of Innovations In Central And

Eastern Europe: A Study Of The Determinants And Impact O (National

Institute of Economic and Social Research (NIESR) Discussion Paper 137) National Institute of Economic and Social Research https://econpapers.repec.org/paper/nsrniesrd/137.htm

The study by Pao and Tsai (2011) investigates the multivariate Granger causality relationships among CO2 emissions, energy consumption, foreign direct investment (FDI), and gross domestic product (GDP) within the BRIC nations—Brazil, Russia, India, and China The findings reveal significant interactions between these variables, highlighting the complex dynamics of economic growth and environmental impact in emerging economies The research underscores the importance of understanding these relationships for effective energy and environmental policy formulation in the context of sustainable development.

Patro, S G., & Sahu, K K (2015) Normalization A Preprocessing Stage

International Advanced Research Journal in Science, Engineering and

Foreign direct investment (FDI) plays a crucial role in economic development, influencing exchange rate pass-through and volatility This study by Piriya Pholphirul (2002) utilizes spatial panel data to examine the relationship between FDI and exchange rate dynamics It highlights how variations in exchange rates can affect investment decisions and overall economic stability Understanding these interactions is vital for policymakers and investors aiming to navigate the complexities of global finance.

Pisani, N., Kolk, A., Ocelík, V., & Wu, G (2019) Does it pay for cities to be green?

An investigation of FDI inflows and environmental sustainability Journal of

Poelhekke, S., & van der Ploeg, F (Rick) (2013) Do Natural Resources Attract

Nonresource FDI? The Review of Economics and Statistics, 95(3), 1047–

Quang, P T., Rasoulinezhad, E., Linh, N N., & Thao, D P (2022) Investigating the determining factors of sustainable FDI in Vietnam China Finance

Review International, 12(2), 334–350 https://doi.org/10.1108/CFRI-10-

Rezza, A A (2014) A meta-analysis of FDI and environmental regulations

Environment and Development Economics, 20(2), 185–208 https://doi.org/10.1017/S1355770X14000114

Riedel, J (1975) The nature and determinants of export-oriented direct foreign investment in a developing country: A case study of Taiwan Review of World

Economics, 111(3), 505–528 https://doi.org/10.1007/BF02696445

Rodrik, D (1996) Labor standards in international trade: Do they matter and what do we do about them? Emerging Agenda for Global Trade : High Stakes for

Rodrik, D (2008) The Real Exchange Rate and Economic Growth Brookings

Papers on Economic Activity, 39(2 (Fall)), 365–439 xxii

Saadatmand, Y., & Choquette, J (2012) Neo-Liberal Policy and Foreign Direct

Investment in Africa International Journal of Business and Social Science,

Sabir, S., Rafique, A., & Abbas, K (2019) Institutions and FDI: Evidence from developed and developing countries Financial Innovation, 5(1), 1–20 https://doi.org/10.1186/s40854-019-0123-7

Sahoo, P (2006) Foreign direct investment in South Asia: Policy, trends, impact and determinants (Working Paper 56) ADBI Discussion Paper https://www.econstor.eu/handle/10419/53445

Saidi, Y., Ochi, A., & Ghadri, H (2013) Governance and FDI Attractiveness: Some

Evidence from Developing and Developed Countries Global Journal of

Saini, N., & Singhania, M (2018) Determinants of FDI in developed and developing countries: A quantitative analysis using GMM Journal of

Economic Studies, 45(2), 348–382 https://doi.org/10.1108/JES-07-2016-

Schneider, F., & Frey, B (1985) Economic and political determinants of foreign direct investment World Development, 13(2), 161–175

Shan, S., Lin, Z., Li, Y., & Zeng, Y (2018) Attracting Chinese FDI in Africa

Critical Perspectives on International Business, 14(2/3), 139–153 https://doi.org/10.1108/cpoib-11-2016-0055

Sharma, B., & Gani, A (2007) The Effects of Foreign Direct Investment on Human

Development Global Economy Journal, 4, 9–9 https://doi.org/10.2202/1524-5861.1049

Singh, H., & Jun, K W (1995) Some new evidence on determinants of foreign direct investment in developing countries Policy Research Working Paper

Series, Article 1531 https://ideas.repec.org//p/wbk/wbrwps/1531.html

Spatareanu, M (2007) Searching for Pollution Havens: The Impact of

Environmental Regulations on Foreign Direct Investment The Journal of xxiii

Environment & Development, 16(2), 161–182 https://doi.org/10.1177/1070496507302873

Stata (2021) Stata Longitudinal-Data/Panel-Data Reference Manual—Release 17

A Stata Press Publication https://www.coursesidekick.com/statistics/2465396

Tabash, M I., Farooq, U., Matar, A., & Al-Absy, M S M (2024) Role of governance in attracting foreign direct investment inflow: Empirical evidence from South Asia region Cogent Economics & Finance, 12(1),

UNCTAD (1999) World Investment Report 1999: Foreign Direct Investment and the Challenge of Development UN https://doi.org/10.18356/8e79f24c-en

UNCTAD (2019) World Investment Report 2019: Special Economic Zones UN https://doi.org/10.18356/8a8d05f9-en

UNCTAD (Ed.) (2020) International production beyond the pandemic (30th anniversary edition) United Nations

UNCTAD (2021) World Investment Report 2021: Investing in Sustainable

Recovery United Nations https://doi.org/10.18356/9789210054638 van der Ploeg, F (Rick), & Poelhekke, S (2010) The pungent smell of “red herrings”: Subsoil assets, rents, volatility and the resource curse Journal of

Wahid, A N M., Sawkut, R., & Seetanah, B (2009) Determinants of Foreign

Direct Investments (FDI): Lessons from the African Economies

Wang, J., Yu, J., & Zhong, R (2020) Country Sustainable Development and

Economic Growth: The International Evidence *

Wheeler, D., & Mody, A (1992) International investment location decisions: The case of U.S firms Journal of International Economics, 33(1), 57–76 https://doi.org/10.1016/0022-1996(92)90050-T xxiv

Wooldridge, J M (2002) Econometric Analysis of Cross Section and Panel Data

MIT Press https://mitpress.mit.edu/9780262232586/econometric-analysis- of-cross-section-and-panel-data/

World Bank (2016) Databank https://databank.worldbank.org/metadataglossary/world-development- indicators/series/NE.TRD.GNFS.ZS

World Investment Report 2024 (2024) World Investment Report

Wu, F (1999) Intrametropolitan FDI firm location in Guangzhou, China A Poisson and negative binomial analysis The Annals of Regional Science, 33(4), 535–

Yilanci, V., Cutcu, I., Cayir, B., & Saglam, M S (2023) Pollution haven or pollution halo in the fishing footprint: Evidence from Indonesia Marine

Pollution Bulletin, 188, 114626 https://doi.org/10.1016/j.marpolbul.2023.114626

Zaman, K., Shah, I A., Mushtaq Khan, M., & Ahmad, M (2012) Macroeconomic factors determining FDI impact on Pakistan’s growth South Asian Journal of

Global Business Research, 1(1), 79–95 https://doi.org/10.1108/20454451211207598

Zhang, D., Zhao, Z., & Lau, C K M (2022) Sovereign ESG and corporate investment: New insights from the United Kingdom Technological

Forecasting and Social Change, 183(C) https://ideas.repec.org//a/eee/tefoso/v183y2022ics004016252200422x.html Zhang, K H., & Markusen, J (1999) Vertical multinationals and host-country characteristics Journal of Development Economics, 59(2), 233–252

Hạ Thị Thiều Dao, Trương Tiến Sĩ, Nguyễn Xuân Đạo, Nguyễn Minh Sáng,

Nguyễn Xuân Trường và Nguyễn Thị Tường Vy (2016) đã xuất bản "Giáo trình Kinh tế xxv học quốc tế" tại Nhà xuất bản Kinh Tế TP.HCM Tài liệu này cung cấp kiến thức cơ bản về kinh tế quốc tế, giúp sinh viên và người học hiểu rõ hơn về các khái niệm và vấn đề liên quan đến kinh tế toàn cầu Để tìm hiểu thêm, bạn có thể truy cập vào [đường link này](https://

Hoàng Trọng và Chu Nguyễn Mộng Ngọc (2008) Thống kê ứng dụng trong kinh tế—Xã hội Nhà xuất bản Thống kê

Lê Sĩ Đồng (2013) Xác suất - Thống kê và ứng dụng Nhà xuất bản giáo dục Việt

Nguyễn Quang Dong (2008) Bài giảng kinh tế lượng Nhà xuất bản Giao thông vận tải

Nguyễn Xuân Hồng (2019) Các yếu tố ảnh hưởng đến dòng vốn đầu tư trực tiếp nước ngoài tại các nước Asean xxvi

ID COUNTRY lnFDI OPEN INFLATION INTEREST ESG pENV pSOC pGOV eEMI eNAT sEDU sEMP gGOV gHUM

APPENDIX 2 FIRST MODEL REGRESSION RESULT

Appendix 2.2 Multicollinearity test (Variance inflation factor)

Appendix 2.3 Pooled OLS regression xxx

Appendix 2.5 Random effects regression xxxi

Appendix 2.6 Summarizing POLS/FEM/REM regression

Appendix 2.8 Modified Wald test for Heteroscedasticity

Appendix 2.9 Wooldridge test for autocorrelation

Appendix 2.10 Feasible Generalized Least Squares regression xxxiii

Appendix 2.11 Efficient Generalized Method of Moments estimation xxxiv

Appendix 2.12 Wald test for independent variables

Appendix 2.13 Ramsey/Pesaran-Taylor RESET test

Appendix 2.14 Skewness and kurtosis tests for normality xxxv

APPENDIX 3 – SECOND MODEL REGRESSION RESULT

Appendix 3.3 Pooled OLS regression xxxvi

Appendix 3.5 Random effects regression xxxvii

Appendix 3.10 Feasible Generalized Least Squares regression xxxix

Appendix 3.11 Efficient Generalized Method of Moments estimation xl

Appendix 3.13 Ramsey/Pesaran-Taylor RESET test

Appendix 3.14 Skewness and kurtosis tests for normality

APPENDIX 4 – THIRD MODEL REGRESSION RESULT

Appendix 4.3 Pooled OLS regression xlii

Appendix 4.5 Random effects regression xliii

Appendix 4.10 Feasible Generalized Least Squares regression xlv

Appendix 4.11 Efficient Generalized Method of Moments estimation

Tiêu đề	The impact of environmental, social, and governance (esg) factors on foreign direct investment (fdi): a case from asean countries
Tác giả	Do Dieu Linh
Người hướng dẫn	Dr. Luu Thu Quang
Trường học	Ho Chi Minh University of Banking
Chuyên ngành	Finance – Banking
Thể loại	Khóa luận tốt nghiệp
Năm xuất bản	2024
Thành phố	Ho Chi Minh City

Định dạng
Số trang	152
Dung lượng	3,29 MB