(Luận văn thạc sĩ) determinants of provincial fdi in vietnam , a cross section data analysis

INSTITUTE OF SOIAL STUDIES UNIVERSITY OF ECONOMICS THE HAGUE HO CHI MINH CITY THE NETHERLANDS VIETNAM VIETNAM-NETHERLANDS PROGRAMME FOR M.AIN DEVELOPMENT ECONOMICS DETERMINANTS OF PROVINCIAL FDI IN VIETNAM: A CROSS SECTION DATA ANALYSIS A thesis submitted in partial fulfilment of the requirements for the degree of MASTER OF ARTS IN DEVELOPMENT ECONOMICS _ h By NGUYEN DAI HIEP ,_, 1•- Academic Supervisor: Dr NGUYEN VAN PHUC HO CHI MINH CITY, JANUARY 2011 ' ' DECLARATION I declare that 'Determinants of provincial FDI in Vietnam: A cross section data analysis' is my own work, that it has not been submitted for any degree or examination at any other University, and that all sources used or quoted are indicated and knowledge by complete references January 3, 2011 NGUYEN DAI HIEP h ACKNOWLEDGMENTS This thesis would never have been written if had not for the encouragement, support, and assistance which I received from large number of people Specially, I would like to thank my supervisor, Dr Nguyen Van Phuc, for his encouragement, kindness, patience and valuable advices, which helped shape and improve this thesis I would also like to thank a few anonymous referees and added explanatory variables who helped me finalize this work I would also like to sincerely thank Prof Dr Peter Calkins for his honest and valuable advices from I begin to choose the topic and TRD completion He is truly a noble teacher, who soonest feedback and detail instruction during TRD establishing h I would also like to thank science committee; all the members of the VietnameseDutch Project for MA programme in Development Economics, University of Economics-HCMC, VietNam for their support and goodwill, and to all the lecturers, and also to my friends in the class 15.Thank you a great time! Last, I want to thank my family members, friends Any errors and omissions in this thesis are my sole responsibility - - TABLE OF CONTENTS Declaration Acknowledgements - Table of contents Table list Abstract Chapter 1: Introduction 1.1 Problem Statement 1.2 Research Objectives 1.3 Research questions 1.4 Organization of the study Chapter 2: Theoretical Consideration and Literature Review ll 2.1 The regional development and competitive regionalism theory 11 h 2.2 FDI theories and its applicability 11 2.2.1 Capital Theory _ 11 2.2.2 The International Trade Arguments _ .12 2.2.3 Market Failures and Industrial Organization 13 2.2.4 The Eclectic Paradigm and International Investment Path 13 2.2.5 Agglomeration Effect 14 2.3 Empirical studies on the determinants ofFDI 17 2.4 Geographical literature on Vietnam, China and ASEAN countries l9 Chapter 3: Research Model, Data Collection and Variable Description 24 3.1 Model Specification 24 3.2 Data Collection 25 3.3 Variables description 26 Chapter 4: Empirical Estimation and Result 34 4.1 Correlation among explanatory variables 34 4.2 Empirical estimation and result .35 Chapter 5: Conclusion and Recommendation .42 5.1 Conclusion and recommendation 42 5.2 Limitation 43 References 45 Appendices 49 h • TABLE LIST Table 2.1: Theory summary 15 Table 2.2: Empirical Study Reading 22 Table 3.1: The implementation value of provincial FDI 26 Table 3.2: FDI capital of top ten provinces 26 Table 3.3: PCI result in 2009 27 Table 4.1: Matrix of Correlation among explanatory variables 34 Table 4.2: Regression Results 36 Table 4.3: Top five rank of attracting FDI in Viet Nam 38 Table 4.4: The rank of infrastructure in 2009 39 h ABSTRACT FDI is of essential importance for achieving economic growth for developing countries, especially for Vietnam, a country which has just opened more than twenty years There were too many researches about attracting FDI for developing countries However, there are still less researches related to regional competition of FDI Therefore, this paper examines the relationship between provincial FDI in VietNam and explanatory variables base on variable set of PCI project in VietNam and other traditional variables The purpose of thesis is finding why some provinces and cities such as Binh Duong, Dong Nai, Ba Ria Vung Tau, HCMC have had good FDI capital and others have not so From that the thesis suggests policy recommendation for provinces and cities enhancing regional system for developing economics I had a literature review on regional development, attracting regional FDI and across country, the estimated model was built with collected data and econometric analysis result, I had demonstrated that our hypotheses are right or wrong And then we h answered the research questions and objectives of this study Using data collected by the General Statistical Office of VietNam (GSO) and Provincial Competition Index (PCI) project, estimation result shows that gross industrial output, legal institution and infrastructures statistically significant to provincial FDI at the level 1% and 5%; business support service had significance to provincial FDI at the level nearly 10% Key words: PCI, FDI impact, Provinces in VietNam, cross section data analysis 1.1 Problem Statement The Provincial Competitiveness Index (PCI) is an effort to explain why some parts of the country perform better than others in terms of private sector dynamism, job creation and economic growth and attracting investment (FDI and local) Using new survey data from businesses that describe their perceptions of their local business environments as well as credible and comparable data from official and other sources regarding local conditions, the PCI rates provinces on a 100-point scale In 2005, the overall index is comprised of nine sub-indices that explain much of the variation in performance across provinces in Vietnam In 2006, new sub-indices were developed to capture other aspects of Provincial Government efforts to enhance the business environment h However, we have not found any empirical studies to show that which are independent variables of PCI and other traditional variables effect to provincial FDI and how to impact to provincial FDI in VietNam I also did not find any analysis related to the independent variables of PCI whether they have internal relation 1.2 Research Objectives The overall goal of this research is to investigate significant impacts of some independent variables of PCI and other traditional variables which affect provincial FDI inflows (Regional FDI) to help policy makers to focus on key points and the good points to improve their investment environment (by Provinces) and with higher level (by Government) There exist some previous studies related to attracting FDI to developing countries; most of these have found what factors ofthe country attracting FDI (across countries) However, the objectives of the thesis are to identify: (i) Independent variables of PCI and other traditional variables are significant impacts to FDI of Provinces in Vietnam; and (ii) Factors of PCI are highly correlated and we should revise PCI set (iii) PCI determinants out of the ten original factors should be included in a new, more significant subset base of PCI determinants (iv) Interaction effects between PCI improvement and FDI growth 1.3 Research questions The thesis focus on studying the determinants of provincial FDI in Viet Nam base on the independence variable set of the Provincial Competitiveness Index (PCI) and other some traditional variables could be attracting FDI of provinces in VietNam We found economic theory, and empirical studies related to FDI (chapter 2), the description of each independence variable which PCI project in Viet Nam use to h survey (chapter 3) We build research model (specification) and collect the data from PCI project (www.pcivietnam.org) and statistical yearbook of Vietnam from General Statistics Office (www.gso.gov.vn) to answer some research questions as following: (l)Which independent variables of PCI and other traditional variables are significant impacts to FDI of Provinces in Vietnam? (2)Factors ofPCI are highly correlated and we should revise PCI set? (3)Which PCI determinant out of the ten original factors should be included in a new, more significant subset base of PCI determinants? (4)Are interaction effects between PCI improvement and FDI growth? 1.4 Organization of the study This thesis has five chapters, while the chapter one has presented as above explain the purpose chose the theme The rest of this thesis is organized as follows: Chapter two briefly provides the regional development and FDI theory, and then we also have summarized the empirical researches related to attracting FDI, specially related to attracting FDI across to provinces of the country Chapter three is presented how to build the research model base on chapter two, the way to choose the data It is important to explain dependent and independent variables which PCI project have used to survey yearly, also including some traditional variables Chapter four is the econometric analysis and finding The last chapter will be conclusion and recommendation of the research h 10 APPENDICES Table 1: Sources of Table No Notation Name ofVariables Calculating Sources Depend Variable FDI Provincial Foreign Direct Investment- Average from pFDI (mill USD) 2006-2009 GSO Independent Variables Access to Land of each province, Average from PCI survey yearly, point= I to 10 2006-2009 projectVietNam AL Business Support Service of each Average from PCI province, survey yearly, point= to 10 2006-2009 project- BSS VietNam Entry Costs of each province, survey Average from PCI yearly, point=1 to 10 2006-2009 projectVietNam EC h Informal Charge of each province, Average from PCI survey yearly, point= I to 10 2006-2009 projectVietNam IC Infrastructure of each province, survey Value at 2008 project- yearly, point to 100 VietNam INF IP Gross output of industry at constant Average from 1994 prices by province ( Bil VND) 2006-2009 Legal Institution of each province, Average from PCI survey yearly, point=1 to 10 2006-2009 project- LI LT PCI GSO VietNam Labour Training of each province, 49 Average from PCI survey yearly, point= to 10 2006-2009 projectVietNam 10 PPL II TAl 12 TCRC 13 Proactivity of Provincial Leadership of Average from PCI each province, survey yearly, point=1 to 2006-2009 projectVietNam 10 Transparency and Access to Average from PCI Information of each province, survey 2006-2009 projectVietNam yearly, point=l to 10 Time Costs of Regulatory Compliance Average from PCI of each province, survey yearly, point=1 2006-2009 project- to 10 VietNam Retail sales of goods and services at Average from current prices by province 2006-2009 MS h 50 GSO Table 2: Expected effect of independent variables Name ofVariables Expected sign No Notation AL Access to Land of each province + BSS Business Support Service of each province, + EC Entry Costs of each province - IC Infonnal Charge of each province - INF Infrastructure of each province + IP Gross output of industry + LI Legal Institution of each province + LT Labour Training of each province + PPL Proactivity of Provincial Leadership of each + province 10 TAl Transparency and Access to Infonnation of each + province II TCRC Time Costs of Regulatory Compliance of each - province h 12 MS + Market Size 51 Table R3.1 OLS result: General Estimation Dependent Variable: FDI Method: Least Squares Date: 11/28110 Time: 07:52 Sample: 63 Included observations: 63 Coefficient Std Error t-Statistic Pro b c -549.2492 473.3872 -1.160254 0.2515 AL -58.81156 40.52070 -1.451395 0.1529 BSS -56.72349 42.86048 -1.323445 0.1917 EC 27.09780 45.87978 0.590626 0.5574 IC 23.66403 60.37637 0.391942 0.6968 INF 6.806537 3.462824 1.965603 0.0549 IP 0.011343 0.001623 6.990285 0.0000 LI 98.75612 42.98600 2.297402 0.0258 LT -33.07042 31.26290 -1.057817 0.2952 PPL -6.798794 27.04513 -0.251387 0.8025 TAl 22.33385 38.06656 0.586705 0.5600 TCRC 21.30427 35.70395 0.596692 0.5534 KEA -14.45187 77.72050 -0.185947 0.8532 R-squared 0.756475 Mean dependent var 109.6462 Adjusted R-squared 0.698029 S.D dependent var 277.7646 S.E ofregression 152.6369 Akaike info criterion 13.07559 Sum squared resid 1164902 Schwarz criterion 13.51782 Hannan-Quinn criter 13.24952 Durbin-Watson stat 1.524842 • Log likelihood h Variable -398.8810 F-statistic 12.94314 Prob(F -statistic) 0.000000 52 Table R3.2 Regression No.02 Dependent Variable: FDI Method: Least Squares Date: 11/28/10 Time: 07:55 Sample: 63 Included observations: 63 Coefficient Std Error t-Statistic Pro b c -480.5299 410.8795 -1.169515 0.2475 AL -60.00294 38.17154 -1.571929 0.1220 BSS -56.41570 41.12630 -1.371767 0.1760 EC 27.22179 44.93207 0.605843 0.5473 IC 14.95326 53.10300 0.281590 0.7794 INF 6.579613 2.760459 2.383522 0.0208 IP 0.011124 0.001394 7.978977 0.0000 LI 94.10527 37.40989 2.515519 0.0150 LT -32.71414 30.16600 -1.084470 0.2832 TAl 19.48059 35.97586 0.541491 0.5905 TCRC 21.93116 34.88220 0.628721 0.5323 R-squared 0.755947 Mean dependent var 109.6462 Adjusted R-squared 0.709013 S.D dependent var 277.7646 S.E ofregression 149.8351 Akaike info criterion 13.01426 Sum squared resid 1167428 Schwarz criterion 13.38846 Hannan-Quinn criter 13.16144 Durbin-Watson stat 1.508808 Log likelihood h Variable -398.9493 F-statistic 16.10682 Prob(F -statistic) 0.000000 • ~ 53 Table R3.3 Regression No.03 Dependent Variable: FDI Method: Least Squares Date: 11/28/10 Time: 07:57 Sample: I 63 Included observations: 63 Variable Coefficient Std Error t-Statistic Pro b c -447.5560 368.1538 -1.215677 0.2294 AL -47.88471 31.23973 -1.532815 0.1312 BSS -42.86376 33.25264 -1.289033 0.2029 EC 28.52452 43.85878 0.650372 0.5182 INF 6.318169 2.689276 2.349394 0.0225 IP 0.011003 0.001360 8.091305 0.0000 LI 94.53679 h 35.26903 2.680447 0.0097 LT -28.45041 28.98503 -0.981555 0.3307 TCRC 24.79147 33.98543 0.729474 0.4689 R-squared 0.754023 Mean dependent var 109.6462 Adjusted R-squared 0.717582 S.D dependent var 277.7646 S.E ofregression 147.6124 Akaike info criterion 12.95862 Sum squared resid 1176629 Schwarz criterion 13.26478 Hannan-Quinn criter 13.07904 Durbin-Watson stat 1.493211 Log likelihood -399.1965 F-statistic 20.69163 Prob(F-statistic) 0.000000 54 Table R3.4 Regression No.04 Dependent Variable: FDI Method: Least Squares Date: 11/28/10 Time: 07:58 Sample: 63 Included observations: 63 Variable Coefficient Std Error t-Statistic Pro b c -211.4629 203.2044 -1.040641 0.3025 AL -44.83035 30.53984 -1.467930 0.1477 BSS -37.33395 32.30533 -1.155659 0.2527 INF 6.596859 2.579005 2.557909 0.0133 IP 0.010739 0.001296 8.283895 0.0000 LI 101.1162 34.01125 2.973022 0.0043 LT -19.93587 27.10756 -0.735436 0.4651 h R-squared 0.750307 Mean dependent var 109.6462 Adjusted R-squared 0.723554 S.D dependent var 277.7646 S.E of regression 146.0434 Akaike info criterion 12.91012 Sum squared resid 1194406 Schwarz criterion 13.14825 Hannan-Quinn criter 13.00378 Durbin-Watson stat 1.467836 Log likelihood -399.6689 F-statistic 28.04590 Prob(F -statistic) 0.000000 55 Table R3.5 Regression No.05 Dependent Variable: FDI Method: Least Squares Date: 11/28/10 Time: 08:00 Sample: I 63 Included observations: 63 Variable Coefficient Std Error t-Statistic Pro b c -218.1488 202.1817 -1.078974 0.2851 AL -44.00827 30.39622 -1.447821 0.1531 BSS -48.13647 28.65678 -1.679759 0.0985 INF 6.159920 2.499510 2.464451 0.0168 IP 0.010838 0.001284 8.440153 0.0000 LI 95.47022 32.99978 2.893056 0.0054 h R-squared 0.747895 Mean dependent var 109.6462 Adjusted R-squared 0.725781 S.D dependent var 277.7646 S.E ofregression 145.4540 Akaike info criterion 12.88799 Sum squared resid 1205942 Schwarz criterion 13.09210 Hannan-Quinn criter 12.96827 Durbin-Watson stat 1.467968 Log likelihood -399.9717 F-statistic 33.81932 Prob(F -statistic) 0.000000 • 56 Table R3.6 Regression No.06 Dependent Variable: FDI Method: Least Squares Date: 11128/10 Time: 08:03 Sample: I 63 Included observations: 63 Coefficient Std Error t-Statistic Pro b c -425.0812 144.3463 -2.944870 0.0046 BSS -47.33038 28.92085 -1.636549 0.1071 INF 5.968698 2.519495 2.369006 0.0212 IP 0.011305 0.001255 9.009201 0.0000 LI 80.02072 31.52059 2.538681 0.0138 R-squared 0.738624 Mean dependent var 109.6462 Adjusted R-squared 0.720598 S.D dependent var 277.7646 S.E ofregression 146.8221 Akaike info criterion 12.89236 Sum squared resid 1250291 Schwarz criterion 13.06245 Hannan-Quinn criter 12.95926 Durbin-Watson stat 1.480202 Log likelihood h Variable -401.1093 F-statistic 40.97567 Prob(F -statistic) 0.000000 • 57 Table 4: The White Heteroskedasticity Test (No cross terms) White Heteroskedasticity Test: F-statistic 1.655384 Probability 0.080879 Obs*R-squared 31.12148 Probability 0.119838 Test Equation: Dependent Variable: RESID/\2 Method: Least Squares Date: 11128110 Time: 23:25 Sample: I 63 Included observations: 63 Variable Std Error t-Statistic Pro b c 1181520 3221604 0.366749 0.7158 AL 224840.1 310916.5 0.723153 0.4739 AL/\2 -18251.90 25041.06 -0.728879 0.4704 BSS 173714.7 171610.6 1.012261 0.3177 BSS/\2 -24185.85 18183.94 -1.330066 0.1912 EC -239308.0 704856.2 -0.339513 0.7360 EC/\2 15750.55 44351.78 0.355128 0.7244 IC 236931.9 439033.8 0.539667 0.5925 IC/\2 -16971.77 33208.46 -0.511068 0.6122 INF -15587.33 14130.50 -1.103099 0.2767 INP'2 162.7767 141.7105 1.148656 0.2577 IP 5.496573 2.822001 1.947757 0.0587 IP/\2 -l.84E-05 2.27E-05 -0.811153 0.4222 LI -258417.7 227961.7 -1.133601 0.2639 Ll/\2 32203.45 24910.96 1.292742 0.2037 LT 73823.70 103656.9 0.712193 0.4806 LT/\2 -7511.556 10190.74 -0.737097 0.4655 PPL 70854.75 91090.23 0.777852 0.4413 h Coefficient • 58 • PPL"'2 -8628.330 8108.898 -1.064057 0.2938 TAl -133598.0 126110.8 -1.059370 0.2959 TAI"'2 15280.56 11649.50 1.311693 0.1973 TCRC -496289.2 238223.7 -2.083291 0.0438 TCRC"'2 48021.71 21844.30 2.198363 0.0339 KEA -19626.68 51982.02 -0.377567 0.7078 R-squared 0.493992 Mean dependent var 18490.50 Adjusted R-squared 0.195577 S.D dependent var 94593.17 S.E of regression 84840.28 Akaike info criterion 25.81726 Sum squared resid 2.81E+11 Schwarz criterion 26.63369 Log likelihood -789.2437 F-statistic 1.655384 Prob(F -statistic) 0.080879 Durbin-Watson stat 1.501341 P-value = I I 9> 5% , No HET in this model h • 59 - Table 5: The Wald Test Wald Test: Equation: Untitled Value Test Statistic df Probability F-statistic 1.319975 (8, 50) 0.2555 Chi-square 10.55980 0.2279 Value Std Err Null Hypothesis Summary: Normalized Restriction(= 0) -549.2492 473.3872 C(3) -56.72349 42.86048 C(4) 27.09780 45.87978 C(8) 98.75612 42.98600 C(9) -33.07042 31.26290 C(IO) -6.798794 27.04513 C(ll) 22.33385 38.06656 C(l2) 21.30427 35.70395 h C(1) Restrictions are linear in coefficients P-value F = 0.2279 > 5% ~ Fail to reject Ho ~ Sellecting the simple model • 60 • Table 6: Correlogram Q-Statistics Date: II/29110 Time: 00:19 Sample: I 63 Included observations: 63 Autocorrelation *1 -I· I -I* -I*· I .I .*I -I· -I· I -I· *1 • ·I· I .I .I **1 .I .I .I I *I I -I· I I*· I 1* I I I I I -I* I •1 I ·I· I -I· I -I*· I -I* I I I ·I· I *1 I I I ·I· I I* I -I· I *1 I -I· I ·I· I I I I I **I I -I· I I I -I· I *1 I I I -I* I I I *1· I ·I· I ·I· I I h I I I I I I I I I I I I I I I I I I I AC Partial Correlation -0.085 PAC -0.085 ~ 61 0.4809 Prob 0.488 0.018 O.OIO 0.50I5 0.778 -0.004 -0.002 0.5028 0.9I8 0.110 0.110 1.3456 0.854 0.091 0.112 1.9306 0.859 -0.027 -0.012 1.9811 0.92I -0.046 -0.054 2.1372 0.952 -0.088 -0.115 2.7181 0.951 -0.008 -0.051 2.7230 0.974 10 0.013 0.006 2.7353 0.987 11 0.053 0.080 2.9587 0.991 12 -0.027 0.025 3.0153 0.995 13 -0.164 -0.150 5.2269 0.970 14 0.056 0.016 5.4903 0.978 15 -0.023 -0.044 5.5359 0.987 16 -0.036 -0.065 5.6495 0.991 17 -0.043 -0.011 5.8136 0.994 I8 -0.246 -0.235 Il.3I9 0.880 I9 -0.010 -0.060 11.329 0.912 20 -0.042 -0.044 11.500 0.932 21 0.004 -0.012 11.501 0.952 22 -0.142 -0.097 13.523 0.918 23 -0.037 -0.028 13.661 0.936 24 0.084 0.096 I4.408 0.937 25 O.D78 0.071 15.067 0.940 26 -0.033 -0.068 15.190 0.954 27 0.001 0.004 15.190 0.967 28 0.076 0.041 15.863 0.968 AR does not exist in this model • Q-Stat • • • Table 7: Matrix of Correlation among Explanatory variables BSS EC IC INF IP LI LT PPL TAl TCRC MS AL 1.000000 -0.060361 0.127156 0.587344 -0.023645 -0.279786 0.339452 0.018306 0.456193 0.232489 0.156726 -0.252539 BSS -0.060361 1.000000 0.214889 0.185477 0.755183 0.610939 0.302026 0.728165 0.509042 0.752013 0.543298 0.527710 EC 0.127156 0.214889 1.000000 0.155114 0.109491 -0.126805 0.338421 0.336528 0.289464 0.297717 0.082589 -0.161585 IC 0.587344 0.185477 0.155114 1.000000 0.156503 -0.113011 0.477916 0.280546 0.629927 0.370366 0.343537 -0.133767 INF -0.023645 0.755183 0.109491 0.156503 1.000000 0.543151 0.274197 0.651423 0.350388 0.532088 0.566604 0.442128 lP -0.279786 0.610939 -0.126805 -0.113011 0.543151 1.000000 0.012219 0.383837 0.196515 0.315223 0.272102 0.885623 LI 0.339452 0.302026 0.338421 0.477916 0.274197 0.012219 1.000000 0.392201 0.643586 0.316825 0.343543 -0.045219 LT 0.018306 0.728165 0.336528 0.280546 0.651423 0.383837 0.392201 1.000000 0.515889 0.652770 0.570061 0.324225 PPL 0.456193 0.509042 0.289464 0.629927 0.350388 0.196515 0.643586 0.515889 1.000000 0.628578 0.436839 0.188026 TAl 0.232489 0.752013 0.297717 0.370366 0.532088 0.315223 0.316825 0.652770 0.628578 1.000000 0.502590 0.281665 TCRC 0.156726 0.543298 0.082589 0.343537 0.566604 0.272102 0.343543 0.570061 0.436839 0.502590 1.000000 0.108097 MS -0.252539 0.527710 -0.161585 -0.133767 0.442128 0.885623 -0.045219 0.324225 0.188026 0.281665 0.108097 1.000000 h AL 62 - - - ~ - Table 8: log estimation between PCI and pFDI Dependent Variable: Log(FDI) Method: Least Squares Date: I 1/18/10 Time: 23:18 Sample (adjusted): 62 Included observations: 62 after adjustments Variable Coefficient Std Error t-Statistic Prob c -35.88196 8.563074 -4.190313 0.0001 Log(PCI) 9.593381 2.136220 4.490821 0.0000 R-squared 0.251567 Mean dependent var 2.555023 Adjusted R-squared 0.239093 S.D dependent var 2.380918 S.E of regression 2.076874 Akaike info criterion 4.331331 Sum squared resid 258.8044 Schwarz criterion 4.399949 Durbin-Watson stat -132.2713 0.451771 h Log likelihood F-statistic 20.16747 Prob(F -statistic) 0.000033 63