How does Korea's official development assistance (ODA) affect its foreign direct investment (FDI) in Africa?

다국어 초록 (Multilingual Abstract)

International development cooperation in order to foster economic growth in developing countries has traditionally relied on official development assistance (ODA) from major donor countries as its primary source of development finance. In recent years, however, foreign direct investment (FDI) inflows into developing countries have attracted increasing attention as an important source of private development finance, and the promotion of FDI has emerged as an additional role of ODA. This is particularly relevant for African countries, where the demand for international development cooperation is higher than other regions and ODA alone is insufficient to meet development financing needs, underscoring the potential catalytic role of ODA in mobilizing FDI inflows. Compared with traditional major donor countries such as the United States, Korea has a relatively short history of economic cooperation with African countries. Nevertheless, African countries accounted for 22.14% of Korea's cumulative ODA between 2006 and 2023, making Africa a major partner in Korea's international development cooperation. Over the same period, however, Africa accounted for only 0.72% of Korea's cumulative FDI, a very small share. Against this backdrop, this study uses statistical analysis to examine whether Korea's ODA to Africa has a catalytic effect on Korea's FDI to Africa, thereby contributing to the design of more efficient ODA policies. This study employs a dynamic panel-data model, following earlier work such as Kimura and Todo (2010), to analyze the impact of ODA on FDI, and applies two-step system generalized method of moments (GMM) estimation. The empirical analysis covers the period from 2006, when Korea's sector-disaggregated ODA statistics became available, to 2023, the most recent year for which final statistical data are available. The sample consists of 30 African countries that received FDI from Korea for at least three years during this period. The dependent variable is Korea's FDI outflows to each African country, and the main explanatory variable is Korea's ODA to each African country, measured both in total and by sector. The estimation results indicate that, overall, Korea's ODA to Africa does not have a statistically significant impact on Korea's FDI to Africa. By contrast, infrastructure-related ODA exhibits a statistically significant, albeit weak, negative effect on FDI, contrary to general expectations and much of the previous literature. When infrastructure-related ODA is further disaggregated into four sectors—social infrastructure and services, economic infrastructure and services, production sectors, multi-sector/cross-cutting—the combination of ODA for social infrastructure and services and ODA for production sectors has a statistically significant negative impact on FDI, partially consistent with the findings of Selaya and Sunesen (2012). In addition, Korea's lagged FDI to Africa has a statistically significant positive effect on subsequent FDI flows, consistent with previous studies emphasizing the persistence of In summary, the combination of ODA for social infrastructure and services and ODA for production sectors appears to exert a negative signaling effect on Korea's FDI to Africa. A range of robustness checks— including specifications with up to five lags, one-step system GMM estimation, alternative definitions of ODA variables, and additional control variables—confirm that this combination of ODA has a statistically significant negative impact on FDI. Therefore, for Korea's ODA to Africa to more effectively attract Korean FDI in the future, it will be important to prioritize infrastructure ODA that directly improves the business and investment environment—such as support for the energy and transport sectors—which is more likely to generate a positive signaling effect for private investors.

목차 (Table of Contents)

• Chapter 1. Introduction 1
• 1.1. Background of the Research 1
• 1.2. Aim of the Study 12
• 1.3. Research Questions and Hypotheses 14
• 1.4. Scope of the Study 18
• 1.5. Structure of the Dissertation 19
• Chapter 2. Theoretical Framework and Literature Review 21
• 2.1. Theoretical Foundations of FDI and ODA 21
• 2.1.1. Eclectic Paradigm (OLI Paradigm) of FDI 21
• 2.1.2. Knowledge-Capital (KK) Model 24
• 2.1.3. Institutional Theory and FDI 27
• 2.1.4. Motivations of FDI 29
• 2.1.5. ODA and Economic Growth 31
• 2.2. ODA and FDI: Complement or Substitute? 33
• 2.2.1. Impact of ODA on FDI: Complementarity Effect 34
• 2.2.2. Impact of ODA on FDI: Substitute Effect 35
• 2.3. Literature Review of Empirical Studies on the ODA-FDI Relationship 38
• 2.3.1. Empirical Studies on Overall ODA-FDI Relationship 38
• 2.3.2. Empirical Studies on ODA-FDI Relationship in Africa 44
• 2.3.3. Empirical Studies indirectly addressing ODA-FDI Relationship 47
• 2.4. Research Gap and Positioning of the Study 55
• Chapter 3. Korea's ODA and FDI Trends in Africa (2006-2023) 58
• 3.1. Korea's Development Cooperation Strategy toward Africa 58
• 3.1.1. From an Aid-recipient Country to the DAC Member Country 58
• 3.1.2. Korea's Development Cooperation Strategy 60
• 3.1.3. Africa Development Cooperation Strategy (2023) 64
• 3.2. Trends of Korea's ODA: Global and Africa 67
• 3.2.1. Korea's Total ODA, 2006-2023 67
• 3.2.2. Sectoral Distribution of Korea's ODA, 2006-2023 68
• 3.2.3. Geographical Distribution of Korea's ODA, 2006-2023 71
• 3.2.4. Korea's Sectoral ODA to Africa, 2006-2023 72
• 3.3. Trends of Korea's FDI: Global and Africa 73
• 3.3.1. Korea's Global FDI, 2006-2023 73
• 3.3.2. Korea's FDI to Africa by Industries, 2006-2023 75
• Chapter 4. Research Design: Empirical Strategy and Estimation Methods 79
• 4.1. Analytical Framework: Extended Knowledge-Capital Model and Initial Gravity Specification 79
• 4.2. From Gravity to the Baseline Dynamic Specification for Examining Korea-Africa ODA-FDI Relationship 81
• 4.3. Estimation Methodology: Two-step System GMM 86
• 4.3.1. Why employ GMM estimation?: Issues of Endogeneity, Heteroskedasticity, and Autocorrelation 86
• 4.3.2. Two-step System GMM Estimation 87
• 4.3.3. Stata's Command for Two-step System GMM Estimation: xtabond2 90
• 4.4. Instrument Validity and Specification Tests 91
• 4.4.1. Arellano-Bond Autocorrelation Tests (AR tests) 92
• 4.4.2. Hansen-J Test of Overidentifying Restrictions 93
• 4.4.3. Difference-in-Hansen Test of Instrument Subsets Exogeneity 93
• Chapter 5. Data, Variable Construction, and Refinement of the Empirical Design 97
• 5.1. Data Sources and Coverage 97
• 5.2. Baseline Variables Definition 99
• 5.2.1. Dependent Variable: lnFDI krj 99
• 5.2.2. Key Explanatory Variable: lnODAkrj and its subsets 99
• 5.2.3. Rationale for the Inclusion of Multiple ODA Variables 102
• 5.2.4. Control Variables: L.lnGDPj, L.c_GDPpcGAPkrj, L.c_GDPpcGAPkrj2, L.WGI j, L.KOAFECj, year_trend 104
• 5.3. Common Data Transformations 110
• 5.4. Panel Structure and Time-Series Characteristics 112
• 5.5. Data Validity Tests 112
• 5.5.1. Cross-Sectional Dependence (Pesaran CD test) 112
• 5.5.2. Stationarity Test 113
• 5.5.3. Multicollinearity Test 115
• 5.5.4. Heteroskedasticity Test (Modified Wald Test) 116
• 5.5.5. First-order Autocorrelation Test (Wooldridge test) 116
• 5.5.6. Robust Hausman test with Heteroskedasticity and First-order Autocorrelation 116
• Chapter 6. Estimation Results and Robustness Check 119
• 6.1. Design of the Estimation Method for Hypotheses Testing 119
• 6.1.1. Subject Countries and Observation Period 119
• 6.1.2. Empirical Model Specification 119
• 6.2. Determining the Endogeneity of Explanatory Variables and Constructing the GMM Estimation Equation 123
• 6.2.1. Determining the Endogeneity of Explanatory Variables 123
• 6.2.2. Constructing the GMM Estimation Equation in Stata 136
• 6.3. Empirical Results 138
• 6.3.1. (Model 1) Key Explanatory Variable = L.lnODAkrj 138
• 6.3.2. (Model 2) Key Explanatory Variables = L.lnODA_INFkrj, L.lnODA_NonINFkrj 143
• 6.3.3. (Model 3-1) Key Explanatory Variables = L.lnODA_INFsockrj, L.lnODA_INFnonsockrj, L.lnODA_NonINFkrj 146
• 6.3.4. (Model 3-2) Key Explanatory Variables = L.lnODA_INFecokrj, L.lnODA_INFnonecokrj, L.lnODA_NonINFkrj 149
• 6.3.5. (Model 3-3) Key Explanatory Variables = L.lnODA_INFprokrj, L.lnODA_INFnonprokrj, L.lnODA_NonINFkrj 153
• 6.3.6. (Model 3-4) Key Explanatory Variables = L.lnODA_INFmulkrj, L.lnODA_INFnonmulkrj, L.lnODA_NonINFkrj 156
• 6.3.7. (Model 4-1) Key Explanatory Variables = L.lnODA_INFsocecokrj, L.lnODA_INFpromulkrj, L.lnODA_NonINFkrj 163
• 6.3.8. (Model 4-2) Key Explanatory Variables = L.lnODA_INFsocprokrj, L.lnODA_INFecomulkrj, L.lnODA_NonINFkrj 167
• 6.3.9. (Model 4-3) Key Explanatory Variables = L.lnODA_INFsocmulkrj, L.lnODA_INFecoprokrj, L.lnODA_NonINFkrj 171
• 6.4. Robustness Check 179
• 6.4.1. Various Lag Structures: Two-step System GMM Estimation 179
• 6.4.2. Various Lag Structures: One-step System GMM Estimation 185
• 6.4.3. Comparison of four estimations differing in GMM methods and variables 191
• Chapter 7. Conclusion and Policy Implications 199
• 7.1. Summary of Main Findings 199
• 7.2. Policy Implications for Korea's ODA Strategy 205
• 7.3. Importance of ODA's Catalytic Role: Additional Suggestions for ODA Policy Design in Africa 209
• 7.4. Limitations and Future Research Directions 212
• References 215
• Appendix 226
• Korean Abstract 242