Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống
1
/ 197 trang
THÔNG TIN TÀI LIỆU
Thông tin cơ bản
Định dạng
Số trang
197
Dung lượng
2,66 MB
Nội dung
MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY ***** NGUYEN THI HOANG OANH KNOWLEDGE SPILLOVER, SECTORAL INNOVATION AND FIRM TOTAL FACTOR PRODUCTIVITY: THE CASE OF MANUFACTURING INDUSTRIES IN VIETNAM PhD THESIS HO CHI MINH CITY, 2021 MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY ***** NGUYEN THI HOANG OANH KNOWLEDGE SPILLOVER, SECTORAL INNOVATION AND FIRM TOTAL FACTOR PRODUCTIVITY: THE CASE OF MANUFACTURING INDUSTRIES IN VIETNAM Major: Development Economics Code: 9310105 PhD Thesis ACADEMIC ADVISORS Dr Pham Khanh Nam Dr Pham Hoang Van HO CHI MINH CITY, 2021 i COMMITMENTS I hereby declare that this dissertation is my own work and it has not been previously submitted for a degree elsewhere All the references from works done by other authors have been explicitly cited To the best of my knowledge, I would like to certify that the above statements are true Ho Chi Minh City, January 2021 Nguyen Thi Hoang Oanh ii ACKNOWLEDGEMENT First of all, I would like to express my deepest gratitude to my academic advisers Dr Pham Khanh Nam and Dr Pham Hoang Van for their valuable advices and continuous support during my study They have given me autonomy on doing my research while continuing to provide me valuable feedbacks and encouragement Second, I would like to thank the School of Economics and University of Economics Ho Chi Minh City for providing me the best environment for my study and doing my research I also would like to thank the board of professors and independent external reviewers on giving me a lot of useful comments on my research All their comments enable me to improve and complete this version of my thesis Last but not least, I am extremely grateful to my family for their love, support and sacrifice Without them, the completion of my dissertation would not have been possible Ho Chi Minh City, January 2021 Nguyen Thi Hoang Oanh iii Table of Contents COMMITMENTS i ACKNOWLEDGEMENT ii ABBREVIATIONS vii LIST OF TABLES viii LIST OF FIGURES x Abstract Tóm tắt CHAPTER INTRODUCTION 1.1 PROBLEM STATEMENT 1.1.1 The significance of the research’s topic 1.1.1.1 The importance of the research the role of knowledge spillovers on innovation at sector level 1.1.1.2 The importance of the research on heterogeneity of firms’ TFP in considering both firms’ characteristics and spillover effects from sectors and regions 1.1.1.3 The importance of research on the role of knowledge spillovers on sectoral innovation and firms’ TFP in the manufacturing industries in Vietnam 1.1.2 The gaps and the new aspects in this thesis 1.1.2.1 The new aspects in theoretical framework 1.1.2.2 The new aspects of the methodology 10 1.1.2.3 The new aspects of the context 11 1.2 RESEARCH OBJECTIVES AND QUESTIONS 11 1.3 RESEARCH METHODOLOGY and RESEARCH SCOPE 12 1.4 RESEARCH CONTRIBUTION 13 1.5 STRUCTURE OF THIS STUDY 14 CHAPTER LITERATURE REVIEW 15 2.1 DEFINITION AND CONCEPTS 15 2.1.1 Knowledge spillovers 15 2.1.2 Innovation 17 2.1.3 Knowledge production function and the determination of innovation in this study 19 2.1.4 Sectoral Innovation System (SIS) and its determinants 21 2.1.5 Total Factor Productivity (TFP) 25 2.1.5.1 Definition of Total Factor Productivity 25 2.1.5.2 TFP measurement and its issue 27 iv 2.1.5.3 2.2 TFP measurement methods 28 THEORETICAL FRAMEWORK 29 2.2.1 Developing model on Knowledge Spillovers at sector-level 29 2.2.2 Channels of knowledge spillovers and the research hypothesis of the first objective 33 2.2.3 Theoretical framework of knowledge spillovers to firms 37 2.2.3.1 Debates on knowledge spillover of intra- sector to firms 37 2.2.3.2 Human capital externalities from the province to firms 39 2.2.4 Multilevel modeling on firms’ total factor productivity and the research hypothesis of the second objective 41 2.3 EMPIRICAL STUDIES 45 2.3.1 Empirical Studies on determinants of sectoral innovation 45 2.3.2 Empirical Studies on channels of knowledge spillover and applications of Spatial Regression Model 49 2.3.3 Empirical Studies on TFP 54 CHAPTER RESEARCH METHODOLOGY 62 3.1 THE RESEARCH MODEL ON SECTORAL INNOVATION 62 3.1.1 Model Specification 62 3.1.1.1 The spatial econometrics and test for model specification 62 3.1.1.2 Estimation Strategy of the Model 63 3.1.1.3 Measuring Direct and Indirect effects in the Model 65 3.1.1.4 Measurement variables 67 3.1.1.5 Hypothesis testing 70 3.1.2 3.2 Data 71 The research model of Cross-Classified Model 72 3.2.1 Measurement of Total Factor Productivity: Semi- parametric Approach 72 3.2.2 Data for measuring TFP 74 3.2.3 Application of Cross-Classified Multilevel Model on the study 76 3.2.4 The Research Model of Cross-classified Model on Firm Productivity 78 CHAPTER SECTORAL INNOVATION AND SPILLOVER EFFECTS: RESULTS FROM SPATIAL REGRESSION MODELS AND DISCUSSIONS 84 4.1 OVERVIEW OF RESEARCH AND DEVELOPMENT (R&D) ACTIVITIES AND PATENTS IN VIETNAM 84 4.1.1 R&D expenditure and R&D intensity in Vietnam 84 4.1.2 Funding sources and performance of R&D activities 86 4.1.3 The human resources in R&D activities 89 v 4.1.4 The overview of the registration and approval of patents in Vietnam 93 4.1.5 Overview of transfer channel of technology and innovation in Vietnamese manufacturing industries 95 4.2 DESCRIPTIVE STATISTICS 100 4.3 RESULTS OF THE MODEL ESTIMATION 104 4.4 DISCUSSION ON THE RESULTS 111 CHAPTER HETEROGENEITY IN TFP OF VIETNAMESE MANUFACTURING FIRMS: RESULTS FROM CROSS-CLASSIFIED MODELS AND DISCUSSIONS 114 5.1 THE MANUFACTURING INDUSTRIES AND FIRMS’ PRODUCTIVITY 114 5.1.1 The importance of manufacturing industries in Vietnam 114 5.1.2 The value added and contribution of capital, labor and TFP to economic growth by economic activities in Vietnam 120 5.1.3 Overview of capital, labor and total factor productivity (TFP) in manufacturing sectors in Vietnam 123 5.2 Summary on characteristics of Vietnamese manufacturing enterprises in the research data 126 5.3 THE RESULTS OF MODEL ESTIMATION 131 5.3.1 The Cross classified model with no predictors (Empty Model) 132 5.3.2 The Fixed effect models 133 5.3.3 The multilevel models: fixed region and random sector 135 5.3.4 The multilevel model: fixed sector and random province 136 5.3.5 The multilevel model with sector random effects and region random effects 140 5.3.6 The multilevel model with sector random effects and province random effects 142 5.4 SUMMARY ON RESULTS AND DISCUSSION 144 CHAPTER CONCLUSION AND POLICY IMPLICATIONS 148 6.1 CONCLUSIONS ON IMPORTANT FINDINGS 148 6.2 SOME POLICY IMPLICATIONS 149 6.3 LIMITATIONS AND FUTHER RESEARCH 151 REFERENCES i APPENDIX xx Table A1 Description of Sectors xx Table A2 The distribution of provinces by regions xxiii Table A3 Test on SDM versus SEM and SAR xxiv B Spatial Regression Model in analysis on Knowledge Spillover among Sectors xxix B1 General idea of Spatial Regression Model xxix vi B2 Knowledge Spillover among Sectors under Spatial Regression Model approach xxx B3 Robust Hausman Test xxxii vii ABBREVIATIONS Words APO ASEAN CCMM CIEM DERG FDI FE GDP GERD GSO GTFP IO LBE LP NIS NISTPASS OECD OP Ph.D QML R&D RE RIS Meanings Asian Productivity Organization Associate East Asia Nation Cross-Classified Multilevel Model Central Institute for Economic Management Development Economics Research Group Foreign Direct Investment Fixed Effects Gross Domestic Product Gross expenditure on R&D General Statistics Office Green Total Factor Productivity Input Output Learning by Exporting Levinsohn and Petrin National Innovation System National Institute for S&T Policy and Strategy Studies Organization for Economic Co-operation and Development Olley-Pakes Philosophy Doctor Quasi-Maximum Likelihood Research and Development Random Effects Regional Innovation System S&T SAC SAR SDM SEM SIC SIS SME TCS TFP VAT VES VPC VSIC Science and Technology Spatial Autoregressive Model with Auto Regressive Spatial Autoregressive Model Spatial Durbin Model Spatial Error Model Standard Industrial Classification Sectoral Innovation System Small and Medium Enterprise Technology and Competitiveness Survey Total Factor Productivity Value Added Tax Vietnam Enterprise Survey Variance Partitioning Coefficient Vietnam Standard Industrial Classification viii LIST OF TABLES Table 3.1 Description of variables in the models 69 Table 3.2 Description of variables in the model of the second objective 81 Table 4.1 Gross expenditure on R&D (GERD) and intensity of GERD 84 Table 4.2 Investment on Science &Technology and Research & Development in 2013 87 Table 4.3 Registration for patents by nationality of applicants 93 Table 4.4 Number of approved patents by nationality of the holders 94 Table 4.5 Descriptive Statistics of variables in the model of the first objective 100 Table 4.6 Correlation matrix of variables in the model 102 Table 4.7 Estimation results of the model by fixed and random effects 105 Table 4.8 Estimation results of the model by Spatial Regression with fixed year effects 106 Table 4.9 Estimation results of the model by Spatial Regression with random sector effects 108 Table 4.10 Estimation results of the model by Spatial Regression with fixed year effects 109 Table 4.11 Estimation results of the model by Spatial Regression with random sector effects 110 Table 5.1 The number of sectors per province 126 Table 5.2 The number of provinces that firms in a sector has located 127 Table 5.3 The descriptive statistics of variables in the model 128 Table 5.4 The correlation matrix between variables in the model 129 Table 5.5 The descriptive statistics of the logarithm of TFP in the data sample 130 Table 5.6 The result of Cross classified model with no predictors 132 Table 5.7 The result of firm fixed effect model 134 Table 5.8 The result of region Fixed effect and sector Random effect model 135 xviii Tuncay, S., 2015 Financial Openness and Total Factor Productivity in Turkey Procedia Economics and Finance 30(15), pp 848–862 doi: 10.1016/S22125671(15)01335-0 Tzokas, N., Kim, Y A., Akbar, H., & Al-Dajani, H., 2015 Absorptive capacity and performance: The role of customer relationship and technological capabilities in high-tech SMEs Industrial Marketing Management, 47, 134–142 doi:10.1016/j.indmarman.2015.02.033 UNESCO, 1968 Provisional guide to the collection of science statistics United Nations Educational Scientific and Cultural Organization, COM/MD/3, PARIS, 31 December 1968 Vu Hoang Duong & Le Van Hung, 2017 FDI Spill-Overs, Absorptive Capacity and Domestic Firms’ Technical Efficiency in Vietnamese Wearing Apparel Industry Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 65(3): 1075–1084 Waldkirch, A & Ofosu, A., 2010 Foreign presence, Spillovers, and Productivity: Evidence from Ghana World Development, Vol 38, No.8, pp 1114-1126 Wang, C C and Wu, A., 2016 Geographical FDI knowledge spillover and innovation of indigenous firms in China International Business Review Elsevier Ltd, 25(4), pp 895–906 doi: 10.1016/j.ibusrev.2015.12.004 Wang, C C., & Wu, A., 2016 Geographical FDI knowledge spillover and innovation of indigenous firms in China International Business Review, 25(4), 895–906 https://doi.org/10.1016/j.ibusrev.2015.12.004 Wang, Z et al., 2016 Analyzing the Space-Time Dynamics of Innovation in China: ESDA and Spatial Panel Approaches Growth and Change, 47(1), pp 111–129 doi: 10.1111/grow.12115 Wang, Z., Cheng, Y., Ye, X., & Wei, Y H D., 2016 Analyzing the Space-Time Dynamics of Innovation in China: ESDA and Spatial Panel Approaches Growth and Change, 47(1), 111–129 https://doi.org/10.1111/grow.12115 Wei, T., & Liu, Y., 2018 Estimation of resource-speci fic technological change Technological Forecasting & Social Change doi: 10.1016/j.techfore.2018.08.006 Westphal, L., Rhee, Y and Pursell, G., 1984 Sources of technological capability in South Korea In M Fransman and K King (eds), Technological Capability in the Third World London: Macmillan Wieser, R., 2005 Research And Development Productivity And Spillovers: Empirical Evidence At The Firm Level Journal of Economic Surveys, 19(4), 587–621 doi:10.1111/j.0950-0804.2005.00260 Wooldridge, J M., 2009 On estimating firm-level production functions using proxy variables to control for unobservable Economics Letters, 104(3), 112–114 doi:10.1016/j.econlet.2009.04.026 xix Wooldridge, J.M., 2005 On estimating firm-level production functions using proxy variables to control for unobservable Unpublished manuscript World Bank, 1993 The East Asia miracle Economic growth and economic policy New York: Oxford University Press Yang, G and Maskus, K E., 2001 Intellectual property rights, licensing, and innovation in an endogenous product-cycle model Journal of International Economics, 53(1), pp 169–187 doi: 10.1016/S0022-1996(00)00062-3 Yu, J., de Jong, R & Lee, L.F., 2008 Quasi-maximum likelihood estimators for spatial dynamic panel data with fixed effects when both n and t are large Journal of econometrics, 146:118-134 Yurtseven, A E., & Tandoğan, V S., 2012 Patterns of innovation and intra-industry heterogeneity in Turkey International Review of Applied Economics, 26(5), 657–671 https://doi.org/10.1080/02692171.2011.631900 Zhang, L., 2017 The knowledge spillover effects of FDI on the productivity and efficiency of research activities in China China Economic Review, 42, 1–14 https://doi.org/10.1016/j.chieco.2016.11.001 xx APPENDIX Table A1 Description of Sectors Code in Sector VSIC 2007 In I/O Description of Sectors Classification by Pavitt (1984)2 2012 1010 S_35 Processing and preserving of meat 1020 S_36 Fishery and processing and preserving of fishery product 1030 S_37 Processing and preserving of fruit and vegetables 1040 to S_38 S_40 Manufacture of grain mill products, starches and starch products and bakery products 1071 1072 to and fats 1050 1061 to Manufacture of vegetable and animal oils S_41 Manufacture of sugar S_43 Manufacture of coffee and tea S_45 Manufacture of macaroni, noodles, couscous 1073 1074 to 1075 1079 and similar farinaceous products; prepared meals and dishes and other food products 1080 S_46 Manufacture of prepared animal, fish, poultry feeds 1101 to S_47 Manufacture of wines 1102 The classification by Pavitt (1984) was based on sources of technology, requirements of users, and possibilities for appropriation in firms Supplier dominated firms (group 1) are mainly in traditional sectors of manufacturing like agriculture, housebuilding, informal household production Scale intensive producers (group 2) have principal activities in food products, metal manufacturing, shipbuilding, motor vehicles, glass and cement Science based firms (group 3) are to be found in the chemical and in the electronic/electrical sectors xxi 1103 to S_48 Manufacture of beers S_51 Spinning, weaving and finishing of textiles 1321 S_52 Manufacture of other textiles 1322 S_53 Manufacture of wearing apparel 1420 S_54 Manufacture of leather and related products 1511 to S_55 Manufacture of footwear S_56 Manufacture of wood and of products of wood 1104 1311 to 1313 1512 1610 to and cork, except furniture; manufacture of 1629 articles of straw and plaiting materials 1701 to S_57 Manufacture of paper and paper products S_58 Printing and reproduction of recorded media S_62 Manufacture of basic chemicals, fertilizer and 1709 1811 to 1820 2011 to nitrogen compounds, plastics and synthetic 2013 rubber in primary forms 2021 to S_65 Manufacture of other chemical products S_67 Manufacture of pharmaceuticals, medicinal 2030 2100 chemical and botanical products S_68 Manufacture of rubber products 2220 S_69 Manufacture of plastics products 2310 S_70 Manufacture of glass and glass products 2391 to S_71 Manufacture 2211 to 2212 2393 products of non-metallic mineral xxii 2394 to S_72 Manufacture of cement 2410 S_74 Manufacture of basic iron and steel 2420 to S_75 Manufacture of basic precious and other non- 2399 ferrous metals and Casting of metals 2432 2511 to S_76 S_77 Manufacture of computer, electronic and optical products 2680 2710 to except machinery and equipment 2599 2610 to Manufacture of fabricated metal products, S_81 Manufacture of electric motor, generators, transformers and electricity distribution and 2732 control apparatus; batteries and accumulators; wiring and wiring devices 2740 to S_84 Manufacture of electric lighting equipment; domestic appliances and other electrical 2790 equipment 2811 to S_87 and special-purpose machinery 2829 2910 to Manufacture of general-purpose machinery S_89 Manufacture of motor vehicles; trailers and semi- trailers and other transport equipment 3099 3100 S_94 Manufacture of furniture 3211 to S_95 Manufacture of jewelry, bijouterie and related articles; musical instruments; sports 3290 goods and games and toys 3311 to 3320 S_98 Repair and installation of machinery and equipment xxiii Table A2 The distribution of provinces by regions Region Province Region Province Region Province NorthWest Lào Cai Lai Châu Sơn La Điện Biên n Bái Hịa Bình Thái Ngun Hà Giang Cao Bằng Bắc Kạn Tuyên Quang Thái Nguyên Phú Thọ Lạng Sơn Quảng Ninh Bắc Giang North Central Thanh Hóa Nghệ An Hà Tĩnh Quảng Bình Quảng Trị TT-Huế Đà Nẵng Quảng Nam Quảng Ngãi Bình Định Phú Yên Khánh Hịa Ninh Thuận Bình Thuận South East Bình Phước Tây Ninh Bình Dương Đồng Nai BRVT NorthEast Red River Delta South Central Highlands Vĩnh Phúc Bắc Ninh Hải Dương Hưng Yên Thái Bình Hà Nam Nam Định Hà Nội Hải Phịng Ninh Bình Đắk Nơng Kon Tum Gia Lai Đắk Lắk Lâm Đồng Tp.HCM South West Long An Tiền Giang Bến Tre Vĩnh Long Đồng Tháp An Giang Kiên Giang Cần Thơ Hậu Giang Sóc Trăng Bạc Liêu Cà Mau Notes: This list includes the provinces in the data The number of provinces were reduced from 63 provinces to 62 provinces (no presence of Tra Vinh Province by merging the data in the period xxiv Table A3 Test on SDM versus SEM and SAR Model Test [Wx] S_RD_mean=[Wx]S_FDI_Supplier= [Wx]S_FDI_Customer = [Wx]S_InputImport= [Wx] S_export =0 Chi2 (5) 31.63 25.42 19.76 16.31 23.68 25.42 Prob>chi2 0.0000 0.0001 0.0014 0.006 0.0002 0.0001 testnl ([Wx]S_RD_mean = [Spatial]rho*[Main]S_RD_mean) ([Wx]S_FDI_Supplier = -[Spatial]rho*[Main]S_FDI_Supplier) ([Wx]S_FDI_Customer = [Spatial]rho*[Main]S_FDI_Customer)([Wx]S_InputImport = -[Spatial]rho*[Main]S_InputImport) ([Wx]S_export = [Spatial]rho*[Main]S_export) Chi2 (5) Prob>chi2 23.95 0.0002 25.29 0.0001 11.95 0.0355 9.81 0.08 23.78 0.0002 25.29 0.0001 Table A4 Test on SDM versus SAC Model Model Obs 190 190 190 190 190 190 Spatial Durbin Model ll(model) df -509.731 12 -257.86 15 -560.9 14 -560.21 17 -259.1 12 -257.86 15 Obs 190 190 190 190 Spatial Autocorrelation Model ll(model) df AIC -514.12 1044.2 -513.34 11 1048.7 -265.23 546.46 -262.53 11 547 AIC 1043.462 545.7 1149.8 1154.44 542.21 545.72 BIC 1082.426 594.43 1195.3 1209.6 581.18 594.43 BIC 1070.2 1084.4 572.44 582.77 Table A5 LR test to compare the model and the model in the Table 4.7 Likelihood-ratio test LR chi2(3) = 1240.89 (Assumption: model1 nested in model3) Prob > chi2 = 0.0000 Table A6 LR test to compare the model and the model in the Table 4.7 Likelihood-ratio test LR chi2(3) = 1293.79 (Assumption: model2 nested in model3) Prob > chi2 = 0.0000 xxv Table A7 Collinearity Diagnostics in the model with Innovation activities as the Dependent variable Vaiable Innovation Activities R&D activities Input from FDI suppliers Output to FDI customers Imported Input Exported Output Capital per worker Monopoly Concentration level Mean VIF VIF 1.38 1.56 SQRT Tolerance VIF 1.17 0.73 1.25 0.64 RSquared 0.27 0.36 1.27 1.12 0.8 0.2 1.9 2.44 1.92 1.4 1.21 1.3 1.59 1.4 1.56 1.38 1.2 1.1 1.14 0.53 0.41 0.52 0.7 0.83 0.77 0.47 0.59 0.48 0.29 0.17 0.22 Table A8 Collinearity Diagnostics in the model with Modification activities as the Dependent variable Vaiable Modification Activities R&D activities Input from FDI suppliers Output to FDI customers Imported Input Exported Output Capital per worker Monopoly Concentration level Mean VIF VIF SQRT VIF Tolerance RSquared 1.39 1.59 1.18 1.26 0.72 0.63 0.28 0.37 1.26 1.12 0.79 0.2 1.85 2.44 1.97 1.37 1.22 1.3 1.6 1.36 1.56 1.4 1.17 1.1 1.14 0.54 0.41 0.51 0.7 0.82 0.77 0.46 0.59 0.49 0.27 0.18 0.22 xxvi Table A9 Wooldridge test for autocorrelation in panel data with Innovation activities as the Dependent variable H0: no first-order autocorrelation F (1, 37) = 3.561 Prob > F = 0.0670 Table A10 Wooldridge test for autocorrelation in panel data with Modification activities as the Dependent variable H0: no first-order autocorrelation F( 1, 37) = 1.739 0.1953 4,000 Prob > F = 3309 1,000 Number of firms 2,000 3,000 3086 2202 2204 1601 2175 1577 1470 1327 857 568 625 645 545 462 743 559 321 130 144 87 59 97 598 455 422 346 419 410 315 266 104 686 196 71 880 785 833 35 36 37 38 40 41 43 45 46 47 48 51 52 53 54 55 56 57 58 62 65 67 68 69 70 71 72 74 75 76 77 81 84 87 89 94 95 98 Figure A1 Number of firms by sectors xxvii 2011 2012 10 20 30 40 2010 10 15 20 2014 10 20 30 40 2013 5 10 15 20 10 15 20 35 36 37 38 40 41 43 45 46 47 48 51 52 53 54 55 56 57 58 62 65 67 68 69 70 71 72 74 75 76 77 81 84 87 89 94 95 98 Figure A2 The correlation of Innovation activities and R&D activities in each year in the period In particular, the relationship between innovation activities and R&D activities is described in each year (Figure 3.19) There seems to be a slightly positive relationship between innovation activities and R&D activities This positive relationship may be more revealed in the year of 2010, 2011 and 2013 In comparison to the average level of innovation activities and R&D activities in the period, there may be more sectors having R&D activities in 2010, but more sector having innovation activities in 2014 xxviii 36 37 38 40 41 43 45 46 47 48 51 52 53 54 55 56 57 58 62 65 67 68 69 70 71 72 74 75 76 77 81 84 87 89 10 20 30 40 10 20 30 40 10 20 30 40 10 20 30 40 10 20 30 40 35 95 10 15 20 10 15 20 10 15 20 98 10 20 30 40 94 10 15 20 10 15 20 10 15 20 2010 10 15 20 2011 2012 2013 2014 Figure A3 The correlation of Innovation activities and R&D activities by each sector in the period The positive correlation between innovation activities and R&D activities seems to be clearer in consideration of each sector (Figure 3.20) Some sectors such as sector number 41 and number 67 had both R&D activities and innovation activities to be higher than the average level of all sectors in all years in the period The innovation activities in 2013 seem to be below the average level in all sectors, except the sector number 41 and number 45 xxix B Spatial Regression Model in analysis on Knowledge Spillover among Sectors B1 General idea of Spatial Regression Model Spatial Regression Model has become a prominent tool for measuring spatial spillover It is widely acknowledged that what occurs in one region may be related to what happens in neighboring regions Several economic and socio-demographic variables may be referred to spatial clustering or geographic-based correlation such as unemployment, crime rates, house prices, per capital health expenditures and the alike (Solle Olle, 2003; Moscone and Knapp,2005; Reveilli,2005; Kostov,2009; Elhorst and Freret,2009; Elhorst et al.,2010) It is obvious that unemployment, crime rates or house prices in this region may have some effects on that in other regions These regions could be countries, states, census tracts or zip codes As stipulated by Tobler (1979), everything is related to everything else, but closer things more so These relations may be observed and analyzed by spatial regression models In spatial relations, the variable change for a specific unit may have effect on the change of other units that is regressed by different typical spatial models The change in dependent variable of this unit may be correlated with that variable’s change of other units This case is appropriate with the application of Spatial Autoregressive Model (SAR) In other case, the change in unobserved factors of this unit may be affected by the change in these factors of other units that is solved by Spatial Error Model (SEM) Besides, when there exists the effect of an explanatory variable’s change for a specific unit on the unit itself and, potentially, all other unit indirectly, Spatial Durbin Model (SDM) is prominent method to explore these relations In general, spatial regression models determines the relations among units basing on the correlation matrix that indicates which regions are spatially related with a given region This is usually a square symmetric RxR matrix xxx with (i,j) element equal to if regions i and j are neighbors of one another, and zero otherwise The diagonal elements of this “spatial neighbors” matrix are conventionally set to zero Depending on the border between regions, LeSage and Pace (2009) pointed out four ways to construct such a matrix including linear contiguity, rook contiguity, bishop contiguity and queen contiguity Besides, distance-based criteria is also used to determine neighboring relations among regions This approach can be expanded in a lot of ways by different distances or weights The usage of different methods in determining the matrix depends on the context In practice, the spatial neighbors’ matrices are usually slightly transformed into spatial weights matrices The most common transformation method is to make the sum of each row in the neighbor matrix to be unity In this method, called “row-standardization”, each element in a row is divided by the sum of the elements in the row Therefore, a spatial weights matrix W, with element wij is defined by !H = z Œ• › ∑• › z Œ• (b1) z H is determined differently Under Depending on the method, the value of ! z H is equal to when two regions satisfy the criteria of the border approach, ! z H is distance-based same border Another typical criterion to define ! zH contiguity with dij to be the distance between (centroids of) regions i and j ! is defined to be if dij