Đang tải... (xem toàn văn)
VIETNAM NETHERLANDS PROGRAMME FOR MASTER’S DEGREE UNIVERSITY OF ECONOMICS ISS INSTITUTE OF SOCIAL STUDIES HO CHI MINH CITY THE HAGUE VIETNAM THE NETHERLANDS VIETNAM NETHERLANDS PROGRAMME FOR MASTER’S[.]
UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ISS-INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS VIETNAM-NETHERLANDS PROGRAMME FOR MASTER’S DEGREE IN DEVELOPMENT ECONOMICS THE DETERMINANTS OF ECONOMIC GROWTH: A CASE STUDY OF SIX SOUTHEAST ASIAN COUNTRIES SUPERVISOR : PRESENTED BY : DR NGUYỄN MINH ĐỨC NGUYỄN KIỀU GIANG MDE-K16 HO CHI MINH CITY, DECEMBER 2012 ACKNOWLEDGMENT I’ve experienced the great time in this course with acquire academic knowledge, open-mind to approach new things, and be straightforward to express ideas I’ve guided by open-mind lecturers who always encourage us to express our opinions and discuss it This is really quality and extensive course We’re encouraged to be ourselves and be self-confident to discuss ideas, it’s really meaningful thing Sincerely, I would like to take this opportunity to express my honest thanks to the Vietnam-Netherlands Master Program for Economics of Development for the interesting and extensive curriculum, as well as sincere thanks to all of people who engage in this course such as management board of this program, lecturers, tutors To fulfill this course, I’ve received the help and support from many people such as director and vice director of this program, lecturers, tutors, my supervisor, classmates, friends, course administrators, librarians of MDE & Fulbright, and other people who give me the instructions, comments, advices, supports, sympathies, and encouragement during the course process Without these things, I could not fulfill the Thesis and finish the course First of all, I would like to express my gratitude to my supervisor – Dr Nguyen Minh Duc, who always remind me to finish the Thesis timely and make me determine to finish the Thesis with high effort, very patient and sympathy with me for delaying the time of submitting Thesis, spent his valuable time to help me, give comments and correct the Thesis mistake by mistake From bottom of my heart, I sincerely thank him for all I would like to express my special thank to Dr Nguyen Trong Hoai - Director of the Program, Vice Principal of UEH He fosters us to study and finish the course actively and efficiently, take care of us step by step And special thanks to Dr Pham Khanh Nam steps with us to overcome obstacles in thesis process honestly, give us the valuable advices and encouragement to finish the thesis i Thanks to TRD board defense give me valuable comments and advices to continue this topic Thanks to Dr Phan Minh Ngoc supplies me materials at my TRD time Thanks to Dr Le Dinh Truc gives me valuable comments and advice on panel technique in Eview And thanks to tutor Phung Thanh Binh for helping us honestly in the course Thanks to all classmates in MDE-K16, all of you support me unconditionally with discussions, talks, shared materials, team working playing, and other Thanks to friends Thu Huyen and Thanh Tien, your sharing materials are very useful for me Hope all of you fulfill this course and get success in your life Lastly, I would like to thank to my dear family, warm friends that supports and sympathies with me for spending most time on studying this course Thanks all for all, hope you will be happy and successful in life I pledge to take full responsibilities for mistakes, errors, omissions and shortcomings of the study ii ABTRACT This study based on the neoclassical growth theory, an extended version of this model As common trend, Cobb-Douglas production function is used to evaluate the robustness of determinants of economic growth in a dataset of six Southeast Asian countries from 1993 to 2009 The fixed effect model (FEM) is used to estimate this model The results indicate that the most important source of economic growth of these countries is capital accumulation and labor Keywords: Economic growth, determinants, Capital, Labor, Population growth, Government expenditure, Southeast Asian countries, Neoclassical Model, CobbDouglas production function, FEM iii TABLE OF CONTENT CHAPTER 1: INTRODUCTION 1.1 Problem statement 1.2 Research structure CHAPTER 2: RESEARCH OBJECTIVES & RESEARCH QUESTIONS 2.1 Research objectives 2.2 Research questions 2.3 Research scope 2.4 Research contribution CHAPTER 3: LITERATURE REVIEW 3.1 General Overview 3.1.1 Concept and Definitions 3.1.2 Traditional methods of examine determinants of economic growth 13 3.1.3 Measurement indicator group of development and growth 14 3.1.4 Overview of economic development theory 15 3.1.4.1 Classify theories by time 15 Classical theory Neoclassical model Endogenous Theory 3.1.4.2 Classify theory by four main classes 21 Linear stages of growth theory Structural change theory International dependence theory Neoclassical counter revolution theory 3.2 Theoretical Framework 26 3.2.1 Production function Cobb-Douglas 26 3.2.2 Harrod-Domar Growth Model 28 3.2.3 Neoclassical model-Solow model 29 3.2.4 Capital Fundamentalism 30 iv 3.3 Empirical Review 30 3.3.1 Development and growth of Vietnam and Southeast Asia 30 3.3.2 Review the empirical studies 33 CHAPTER 4: RESEARCH METHODOLOGY 39 4.1 Analyzed Framework 39 4.1.1 Model 1- Traditional neoclassical model 39 4.1.2 Model 2- Extend neoclassical model 40 4.1.3 FEM is selection for estimation 40 4.2 Variables 44 4.2.1 Dependent variable: GDP per capita 44 4.2.2 Independent variables 44 4.3 Data Description - Data collection – Data analysis 50 CHAPTER 5: DATA ANALYSIS AND DISCUSSION 52 5.1 Descriptive Statistics Analysis 52 5.1.1 Descriptive Statistics 5.1.2 Correlation Matrix 54 5.2 Econometric Analysis 54 5.2.1 Whether FEM or REM is more suitable 54 5.2.2 Model – Traditional model 58 5.2.3 Model - The Extened Neoclassical Model 59 5.3 The limitations of data, and modeling techniques 61 CHAPTER 6: CONCLUSION AND POLICY RECOMMENDATION 63 6.1 Main findings 63 6.2 Managerial Implications and Policies 64 6.3 Limitations and future research 65 REFERENCE 67 APPENDICES 71 v APPENDICES A: Data of GDP, Capital, Population growth, Labor, Government expenditure 70 B: Descriptive statistics of each country 77 C: Correlation Matrixes of each country 79 D: Residual Graph 81 E: Regression Results 82 F: Main Empirical studies Summary 92 E: Regression results 81 vi LIST OF TABLES Table 3.1: Sources of Growth in East Asia, by Country and Period 37 Table 4.1 Estimated Capital-Output ratio some Asian Countries 1980 45 Table 4.2 Range forecast of COR for countries 46 Table 4.3 Summary of variable 49 Table 5.1 Sample observations - Descriptive statistics - sample: 1993 2009 54 Table 5.2 Correlation on the sample observations 54 Table 5.3 Model - A comparison of results with FEM 55 Table 5.4 HAUSMAN TEST for MODEL 55 Table 5.5 Model - A comparison of results with FEM 57 Table 5.6 HAUSMAN TEST for MODEL 57 Table 5.7 Estimated results for FEM with tradition model (Model 1) 58 Table 5.8 Estimated results for FEM with extended model (Model 2) 60 vii LIST OF FIGURES Figure 3.1: Population of 1993 of countries Figure 3.2: Population of 2009 of countries Figure 3.3: GDP per capita of 1993 of countries 10 Figure 3.4: GDP per capita of 2009 of countries 10 Figure 4.1: Conceptual Framework of the study 39 Figure 5.1: GDP per capita of countries for the period of 1993-2009 52 Figure 5.2: Capital per capita GDP per capita for the period of 1993-2009 53 Figure 6.1: An overview of determinants of economic growth of countries 64 viii ABBREVIATIONS COR : Capital output ratio DCs : Developed (high-income) countries FEM : The fixed effects method GCF : Gross Capital Formation GFCF : Gross fixed assets formation GDP : Gross Domestic Product ICOR : Incremental capital output ratio LDCs : Less developing countries LICs : Low-income countries LSDV : The least-squares dummy variable OLS : Ordinary Least Square PCI : Province Competitiveness Index REM : Random Effect Model ix is capital’s depreciation rate i j is country j's steady-state investment rate j is country j's steady-state growth rate Set the steady-state growth rate of country j as follows: j j (1 ) w Where: j is country j’s growth rate w is the world growth rate, set w =0.04 is a parameter that governs the relative weight we place on the country’s own experience, set = 0.25 b) Perpetual inventory method, setting initial capital to zero This method uses the formula to estimate capital stock as follows: K t 1 j 0 j I t j (1 ) t K Where Ko is the initial capital stock The first estimate using the perpetual inventory method sets Ko=0 and accumulates forwards, then using this estimation for calculating COR The advantage of this approach is that it simply accumulates investment The disadvantages are this method does not produce a useful time series of capital stocks since the importance of the initial capital stock estimate diminishes to what may be considered negligible levels very slowly, and we can probably compute a better guess of the initial capital stock for countries than zero c) Perpetual inventory method: steady-state estimates of initial capital This method attempts to derive a better initial capital estimate than zero for all countries by modifying a method suggested by Harberger (1978) We will calculate initial capital by using the steady-state method then calculate an initial capital stock time-series Accordingly, the steps for calculating the capital stock series as follows: Firstly, compute the steady-state growth rate of country j by equation: j j (1 ) w (set w = 0.04 and = 0.25) 31 The next, assumption of COR is fixed: kj ij j And then, to calculate the initial capital stock for a country, we rewrite this formula as: K initial k j Yinitial (Harrod-Domar Growth Model) Finally, the perpetual inventory formula produces a capital stock time series for country j 3.3 EMPIRICAL STUDIES 3.3.1 DEVELOPMENT AND GROWTH OF VIETNAM AND SOUTHEAST ASIA There are many researches on economic growth of Asia Most of these conclude that the main determinants of economic growth are physical capital, labor, human capital, and TFP, in particular developing countries in this area The high economic growth in the emerging countries in East Asia influenced the conventional economic policy Nowadays there’re many researchers on the global economy recognize these successes demonstrate two propositions Firstly, there is a major spread of technology progress throughout the world, and Western nations are losing its traditional advantage and position No.1 Secondly, the center of the world's economic will shift to the Asian nations of the western Pacific Two conclusions are remarkable achievements of East Asian’s growth It matches with input growth so quickly that Asian economic growth, incredibly, ceases to be a mystery (Paul Krugman, 1994) Rapid increases in growth were facilitated by technological catch up in Asia Moreover, they also emphasize that capital accumulation can be increased by means of technology transfer, and in turn capital is considered as a necessary condition of economic growth Endogenous growth theory supports this view by papers such as Mankiw (2009) Kim and Lau (1994) pointed out that physical capital is the most important source of economic growth for developing countries, normally contributed 32 at 60% on average, next is human capital, then labor However, they supposed that technological progress plays no role in accounting for economic growth They explained that technological progress cannot affect output in case absence of new investment, the contribution of technological progress is also attributable to capital Hence, capital becomes the most robust of economic growth in developing countries Southeast Asia comprises 11 countries Brunei, Cambodia, Dong Timo, Indonesia, Lao, Malaysia, Mianma, the Philippines, Singapore, Thailand, and Vietnam All these countries have different history initial point as well as development conditions such as history development, political regime, national culture, religious trend, certain geographical location, available natural resource, etc And most these countries are developing ones, now on the road to catch-up economic development of the world Thus, these countries are learning development experience of developed countries and focusing on economic policies so as to achieve fast and sustainable development Vietnam is developing country and inevitable this trend as well As some researches on source of development of Vietnam, physical capital is still the robust source of growth, then labor and human capital Physical capital accounts 85% to 90% for GDP growth in period 1996-2005 (P.M Ngoc, 2006, p215) Another significant determinant is Labor, account 10% to 15% for GDP Technological progress is seem absent from economic growth from 1975-2005, there’s not significant contribution of technological progress to growth (P.M Ngoc, 2006) 3.3.2 REVIEW THE EMPIRICAL STUDIES There’re many researches and working papers with different applied models that research on economics growth for a sole country, cross-countries or cross-regions The researches on sole country such as Ab Wahab Muhamad (2004) for Malaysia, Achara Chandrachai et al (2004) for Thailand, Caesar B Cororaton (2004) for the Philippines, Hananto Sigit (2004) for Indonesia, Le Thanh Nghiep et al (2000) for 33 Vietnam, Nombulelo Duma (2007) for Sri Lanka, Phan Minh Ngoc (2006) for determinants of economic growth of Vietnam, Phan Minh Ngoc (2007) for sources of economic growth of Vietnam, Shandre Mugan Thangavelu (2004) for Singapore, Tran Tho Dat (2004) for Vietnam, Yan Wang et al (2001) for China… Other researches were conducted for cross-countries comprise Charles R Hulten et al (2007) for 112 countries, Laurits R Christensen et al (1980) for countries, N Gregory Mankiw et al (1992) for 121 countries, Robert J Barro (1989) for 120 countries, Robert J Barro (1994) for 100 countries, Robert J Barro & Jong Wha Lee (1993) for 116 countries, Robert J Barro (1996) for 100 countries, Robert J Barro (2001) for 10 Asian countries, Robert W Fogel (2004) for Asia, Steven N Durlauf et al (1998) for 122 economies, Susan M Collins et al (1996) for East Asian countries, Svetlana Ledyaeva et al (2008) for 74 Russian regions… We also can there’re many kinds of explanation variables into the panel data growth model to measure the contribution to growth in many empirical studies from traditional variables to new variables The traditional variables such as physical capital, Labor, human capital, TFP in researches of Laurits R Christensen et al (1980) for countries, N Gregory Mankiw et al (1992) for 121 countries In addition, there’re new other variables such as FDI to GDP ratio, Government expenditure on education, share of foreign equity ownership, and share of exports to GDP are used in paper working Shandre Mugan Thangavelu (2004); Robert J Barro (1989) measured investment in physical, and population growth; Robert J Barro (1994) measured Initial Level of GDP, Initial Level of Human Capital, Educational Spending, Fertility Rate, Government Consumption exclusive of education and defense, Investment Ratio, Terms of Trade, Democracy; Robert J Barro et al (1993) analyzed Black-market premium, Male secondary school, Female secondary school, Life- Expectancy; Robert J Barro (1996) measured Inflation rate; even Religion variables were measured the contribution to economic growth in paper of Robert J Barro & Rachel M McCleary ( 2003) such as investigate the effects of church attendance and religious beliefs on economic growth 34 Regarding to research Determinants of Economic Growth: A Cross-Country Empirical study of Robert J Barro (1996), it estimated for a panel data of 100 countries from 1960-1990, it strongly supports to general notion of conditional convergence With a given starting level of real GDP per capita, growth rate will get higher with higher initial schooling, higher life expectancy, lower fertility, better maintenance of rule of law, lower inflation, terms of trade improvement, and lower government consumption Similarly, with given values of other variables, growth rate will be negative related to the initial level of real GDP per capita In particular, at low level of political rights, an expansion of these rights will stimulate economic growth obviously But, when democracy is achieved at moderate level, if make additional expansion will reduce growth In contrast, there is a strong positive influence of living standard on a country’s trend to experience democracy The framework for the determination of growth follows as the extended version of the neoclassical model In general, the applied model has form: Log (GDP_per_capita) = C + a * Log(Initial_GDP) + b * Male_Secondary_&_HighSchool + c*Log(Life_Expectency) + d*Log(GDP)*Male_Schooling + e * Log(Fertility_rate) + f * Government_Consumption_to_GDP_Ratio + g * Rule_of_Law_Index + h * Term_of_trade_Change + Democracy_Index + i * Democraxy_Index_Squared + j * Inflation_rate + Region_Area_Dummy_variable Variable government consumption indicates a negative effect on economic growth The purpose of the measurement of government spending is estimate approximately the government expenditure that doesn’t improve productivity Thus, it implies that a higher volume of nonproductive government spending will reduce the growth rate with a given starting value of GDP In this aspect, a bigger government is worse for economic growth Regarding to researches on economic growth of researched countries in this paper includes Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam 35 Firstly, Indonesia has a research of Hananto Sigit (2004), with findings that the growth of the economy is mostly driven by capital accumulation (dependence on investment, especially on FDI), more capital and labor will produce more output, and education has a major influence on economic growth as well Secondly, the Philippines with finding is mostly negative TFP growth in the Philippines of Caesar B Cororaton (2004) Thirdly, Singapore with paper working of Shandre Mugan Thangavelu (2004), findings that labor quality in terms of skilled workers improves TFP growth, and the key component of the long-term growth is the quality of education of the labor force including skill and education forms an important component Fourthly, Thailand has research of Achara Chandrachai et al (2004) with finding that period 1977-1986 the main source of economic growth was the expansion of capital and labor, period 1987-1999 has main contribution to the economic growth came from capital, and TFP growth played a very insignificant role in the contribution to growth And Vietnam has some researches such as P M Ngoc (2006), P M Ngoc (2007), Le Thanh Nghiep et al (2000), with time series estimation for Cobb-Douglas production functions Ngoc analyze annual data 1975-2005 to measure the contribution of capital formation, labor, and technological progress to the growth of the Vietnamese economy, as well as measuring the impact of economic reforms “Doi moi” since the end of 1986, and the rates of returns to capital and labor The major findings are technological progress was statistically absent in the growth of the Vietnamese economy; and the most important source of economic growth is capital accumulation, accounting for 85% approximately to growth, next is Labor It shows that Vietnam’s growth has based mainly on foreign investment, thus if this heavy base is continuous, the economy on such financial source is going be unsustainable Briefly, in order to achieve sustainable growth in the coming decades, Vietnam should shift from rely much on foreign finance to more on productivity growth which has been absent so far Furthermore, Susan M Collins et al (1996) researched East Asia countries includes Indonesia, Malaysia, Philippines, Singapore, and Thailand, the 36 findings are Physical capital contribute most to economic growth in these countries Next to TFP on average, then to Human capital contribute to economic growth Meanwhile Philippines’ TFP has negative effect on growth in most times Table 3.1: Sources of Growth in East Asia, by Country and Period Source from ”Economic Growth in East Asia- Accumulation versus Assimilation” of Susan M Collins and Barry P Bosworth (1996) Summary This chapter gives an overview on theoretical background which measures traditional relationship between economic growth and capital, labor, and other 37 factors including population growth, government expenditure In addition, this part presents the computing method of physical capital series for each country is standard perpetual inventory method with steady-state estimates of initial capital The Neoclassical models have some limitations, but many empirical studies concluded that it is also meaningful for policy implications 38 CHAPTER 4: RESEARCH METHODOLODY 4.1 ANALYTICAL FRAMEWORK There are two chosen equations that to run regression in this paper as mentioned below The used estimation method is FEM Figure 4.1: Conceptual Framework of this study Model 1: Traditional Neoclassical model Physical Capital per capita Economic Growth (GDP per capita) Labor force Model 2: Extended Neoclassical model Physical Capital per capita Economic Growth Population Growth rate (GDP per capita) Government expenditure per GDP 4.1.1 MODEL 1: Traditional Neoclassical model bases on production function: ln(Yit) = 1 + 2 ln(Kit) + 3 ln(Lit) + uit As mentioned, there’re two reasons for looking at the numbers for output per capita rather than the numbers for total output The evolution of the standard of living is giving by the evolution of output per capita, not a country’s total output And, when comparing countries with different populations, output numbers must be adjusted to 39 take into account these differences in population size This is exactly what output per capita does Olivier Blanchard (2009) With assumption constant return to scale 2 + 3 = 1, we can re-write this equation as follow: ln(Yit) = 1 + 2 ln(Kit) + 3 ln(Lit) + uit With 2 + 3 = 1 3 = - 2 ln(Yit) = 1 + 2 ln(Kit) + (1 - 2) ln(Lit) + uit ln(Yit) = 1 + 2 ln(Kit) + ln(Lit) - 2 ln(Lit) + uit ln(Yit) – ln(Lit) = 1 + 2 ln(Kit) - 2 ln(Lit) + uit ln(Yit) – ln(Lit) = 1 + 2 [ ln(Kit) - ln(Lit)] + uit ln(Yit/Lit) = 1 + 2 ln(Kit/ Lit) + uit So, the re-write equation is: ln(Yit/Lit) = 1 + 2 ln(Kit/ Lit) + uit With assumption constant return to scale 4.1.2 MODEL 2: Extended Neoclassical model Ln(Yit/Lit) = 1 + 2 Ln(Kit/Lit) + 3*(GovExp_to_GDP it) + 4*Ln(Pop_growth) it + uit Where GovExp_to_GDP is ratio of Government expenditure to GDP; Pop_growth stands for population growth rate This model is mainly based on model that Barro’s research in 1996 for about 100 cross-countries 4.1.3 FEM IS SELECTION FOR ESTIMATION In general, panel estimation is applied for two equations, but which panel method for these, we will consider three methods as following: a) The simplest method – OLS Regarding the equations: 40 Model 1: Ln(Yit/Lit) = 1 + 2 Ln(Kit/ Lit)+ uit Model 2: Ln(Yit/Lit) = 1 + 2 Ln(Kit/Lit) + 3*(GovExp_to_GDPit) + 4*Ln(Pop_growth) it + uit With i indexes of the countries, and t indexes of the time In this situation, we ignore the sizes of sections (the number of nations), time period of panel data and estimates by Ordinary Least Square (OLS) We assume intercept of each country is the same (1), coefficients of independent variables of each country is the same each other and unchanged by times as well (2, 3, 4) This is so-called combination-regression The combination-regression can distort the real picture of relationship between dependent and independent variables in this model because of the limited assumptions Subsequently, this method is not applied for econometric analysis in this paper b) Fixed Effects Method (FEM) - Least-Squares Dummy Variable (LSDV) The fixed effects method (FEM) is applied for the balanced panel data of this paper In FEM, the constant is treated as section-specific The term Fixed Effect implies that intercept can be different between sections (countries) but unchanged by time (fixed by time) FEM is also known as the least-squares dummy variable (LSDV) estimator because it allows for different constants for each section, and dummy variable for each section (Damodar N.Guragati, 2004) That’s a method to consider the specific of each country that is allow for different constant for each country (1i) and unchanged by time, in addition assumes unchanged-coefficients (2, 3, 4) between countries and unchanged by time as well These different constant of each country maybe is due to specific characteristics of each country such as management level, culture, etc * Model – Traditional Solow model: Ln(Yit/Lit) = 1i + 2 Ln(Kit/ Lit) + uit We can rewrite the FEM in detail as follows: Ln(Yit/Lit) = α1 + α2D2i + α3D3i + α4D4i + α5D5i + α6D6i + 2 Ln(Kit/Lit) + uit 41 With D2i= if observations belong to Indonesia and D2i= if not; similarly D3i, D4i, D5i, D6i, for Malaysia, the Philippine, Singapore, and Thailand respectively In the other hand, α1 is representative for intercept of Vietnam Respectively, α2, α3, α4, α5, α6 are differential intercepts which show the difference of intercepts between Indonesia, Malaysia, the Philippine, Singapore and Thailand (Vietnam is root for comparison) This model can be known as the least-squares dummy variable (LSDV) or Covariance Model In other words, applying Solow model bases on Cobb-Douglass production function: ln(Yit/Lit) = 1i + 2 ln(Kit/ Lit) + uit Where i indexes the countries and t indexes the time (1) Y is real GDP of the country (2) K is physical capital stock (3) L is the number of labor (quantity of labor) (5) 1i is intercept of this model 1i = α1 + α2D2i + α3D3i + α4D4i + α5D5i + α6D6i Dummy variables D2i, D3i, D4i, D5i, D6i present for Indonesia, Malaysia, the Philippine, Singapore, and Thailand respectively With D2i= if observations belong to Indonesia and D2i= if not; similarly for D3i , D4i , D5i, D6i D2i=1 Indonesia D3i=1 D4i=1 D5i=1 Malaysia the Philippine Singapore D6i=1 Thailand * Model 2: Extended Solow model Ln(Yit/Lit) = 1 + 2*Ln(Kit/Lit) + 3*(GovExp_to_GDPit) + 4*Ln(Pop_growth) it + uit Similarly, we can rewrite the FEM in detail as follows: Ln(Yit/Lit) = α1 + α2D2i + α3D3i + α4D4i + α5D5i + α6D6i + 2 Ln(Kit/Lit) + 3 (GovEpx_to_GDPit) + 4 Ln(Pop_growth) + uit 42 Where: GovExp_to_GDP is ratio of governmental expenditure to GDP (%), Pop_growth is population growth rate (%) c) Random Effect Model (REM) Tải FULL (111 trang): https://bit.ly/3jXu2qV Dự phòng: fb.com/TaiHo123doc.net An alternative method of estimating a model is the random effects model The difference between the fixed effects (FEM) and the random effects method (REM) is that the latter handles the constants for each section not as fixed, but as random parameters This model has the following advantages: has fewer parameters to estimate compared to the FEM, allows for additional explanatory variables that have equal value for all observations within a group One obvious disadvantage of REM approach is that we need to make specific assumptions about the distribution of the The difference between the two possible ways of testing panel data models is this: the FEM assumes that each country differs in its intercept term, whereas the REM assumes that each country differs in its error term Usually, when the panel is balanced, one might expect that FEM will work best In other cases, where the sample contains limited observations of the existing crosssectional units, REM might be more appropriate (Dimitrios Asteriou and Stephen G.Hall, 2007) Some other suggestions can be use to consider between FEM and REM (Damodar N Guragati, 2004) are, firstly if T (the number of time series data) is large and N (the number of cross-sectional units) is small, there is likely to be little difference in the values of the parameters estimated by FEM and REM Thus, the choice can be based on computational convenience On this aspect, FEM may be preferable Secondly, If N is large and T is small, and if the assumptions underlying REM hold, REM estimators are more efficient than FEM In this paper, used panel data is balanced, thus FEM will work best In addition, T = 17 year and N is just = countries, T is greater than N, so FEM may be preferable because of convenient computation In summary, FEM is preferable in this paper 43 4.2 VARIABLES EXPLANATION: 4.2.1 Dependent variable: GDP per capita In this study, GDP per capita is proxy for economic growth + Yit symbols for GDP, is output or yield of country i at year t In this paper, annual real GDP at price 2005 is proxy for Y + Lit symbols for Labor, is the number of labor of country i at year t as well + Hence, Yit/Lit is output yield per capita of country i at year t As mentioned, when we execute a research for across countries, we should use data output per capita instead of data total output, as the evolution of the living standard is giving by the evolution of output per capita, not a country’s total output In addition, when comparing many countries with different populations, at that time total outputs must be adjusted to take into account these differences in population size Therefore, this is exactly what output per capita does (Blanchard, 2009:204) 4.2.2 Independent variables a) Capital per capita Tải FULL (111 trang): https://bit.ly/3jXu2qV Dự phòng: fb.com/TaiHo123doc.net + Kit symbols for physical capital stock of country i at year t + Hence, Kit/Lit is Physical capital per capita of country i at year t At present, capital stock data of Vietnam as well as other developing countries are not readily available from existing sources Accordingly, I have to compute annual capital stock to use in this paper There are two different methods used to compute capital in this paper Firstly, the method bases on working paper of Robert G King et al (1994), that is perpetual inventory method with steady-state estimates of initial capital Secondly, a other simpler method to calculate capital as Le Thanh Nghiep et al (2000) or Phan Minh Ngoc (2006) That is we assume the certain value for COR of each country appropriately, then compute the capital series Measuring capital stock by perpetual inventory method through some steps as: 44 Step 1: calculating the initial capital, then compute initial capital stock time-series Two methods of computing capital have difference together in just this step Method 1: bases on Robert G King et al (1994), Calculating the initial capital by using the steady-state method, then compute initial capital stock time-series + Firstly, compute the steady-state growth rate of nation as follows: J J (1 ) w (set w = 0.04 and = 0.25) J is the country j's growth rate over the research period w is the world growth rate over the research period (set 0.04) is a parameter that governs the relative weight we place on the country’s own experience (set = 0.25) + The next, assumption of COR is fixed, we compute the COR as follows: k i where: + i It : steady-state investment rate, Yt It is gross investment in year t and Yt is GDP in year t + is depreciation rate of capital stock + is the steady-state growth rate + Finally, calculate the initial capital stock value in a certain year by formula as: K initial k Yinitial (Harrord-Domar’s equation) Method 2: follow Le Thanh Nghiep et al (2000) or Phan Minh Ngoc (2006) The value of COR (k) of countries should in range as table of Nghiep (2002) below Table 4.1 45 6675172 ... ASIA There are many researches on economic growth of Asia Most of these conclude that the main determinants of economic growth are physical capital, labor, human capital, and TFP, in particular... constant capital-output ratio The advantage of this method is that we not have a assume anything about the initial capital stock; its weakness is that it assumes a constant capital-output ratio The. .. can be described as the growth in total factor productivity The last equation explains sources of the growth rate of output are growth rate of capital, growth rate of labor and growth rate of