intellectual capital and national innovation systems performance

10
1 3 Intellectual capital and national innovation systems performance 4 5 6 Wen-Min Lu a,Q1 , Qian Long Kweh b , Chia-Liang Huang a,1 7 a Department of Financial Management, National Defense University, No. 70, Section 2, Zhongyang North Road, Beitou, Taipei 112, Taiwan 8 b Department of Accounting, College of Business Management and Accounting, Universiti Tenaga Nasional, Sultan Haji Ahmad Shah Campus, 26700 Muadzam Shah, Pahang, Malaysia 9 10 12 article info 13 Article history: 14 Received 15 April 2014 15 Received in revised form 1 July 2014 16 Accepted 2 August 2014 17 Available online xxxx 18 Keywords: 19 National innovation system 20 Network DEA 21 Intellectual capital 22 Performance evaluation Q2 23 24 abstract 25 Innovation is a key resource for the well-being of national economies and international competitive 26 advantages. First, this study develops a network data envelopment analysis (DEA) production process 27 to evaluate the R&D efficiency and economic efficiency of the national innovation system (NIS) in 30 28 countries. Our findings show that the R&D efficiencies of the NIS are better than the economic efficien- 29 cies. Second, this study examines the effect of intellectual capital (IC) on the NIS performance through 30 truncated regression. Our findings indicate that IC does play an important role in affecting the NIS per- 31 formance. Finally, this study presents a managerial decision-making matrix and makes suggestions 32 through a performance improvement strategy map to help government and managers improve the NIS 33 performance. 34 Ó 2014 Elsevier B.V. All rights reserved. 35 36 37 38 1. Introduction 39 In the survey of 2007 global industrial technology innovation 40 policy development, the Science, Technology and Industry Score- 41 board of Organization for Economic Co-operation and Develop- 42 ment (OECD) finds that a large majority of OECD countries 43 emphasize technological system and innovation because these 44 two factors affect not only domestic and international competition, 45 but also economic growth [61]. A technological system is the 46 system that affects specific technology or industry. According to 47 Carlsson and Stankiewicz [13], technological system is a dynamic 48 network among economic entities in a specific technological area 49 that aim to create, transmit, and apply technology. In contrast, a 50 national innovation system (NIS) affects the overall development 51 of innovation. That is, a NIS is the network of institutions in the 52 public and private sectors whose activities and interactions initi- 53 ate, import, modify, and diffuse new technologies [28]. Technolog- 54 ical system differs significantly from NIS on the following three 55 perspectives. First, technological system emphasizes creations of 56 new technology, while NIS focuses on the transmission and appli- 57 cation of technology. Second, the strength of technological system 58 in each technological area differs significantly within a country. 59 Third, the classification of technological system is based on tech- 60 nology rather than national boundaries. 61 Since the seminal work of Freeman and Christopher [29], 62 which studies Japanese technological development, NIS research 63 has evolved into an important research agenda. NIS is formed 64 by the interaction among innovation creation, transmission, and 65 application of various institutions. In other words, NIS is an 66 important contributor for economic growth when the interac- 67 tions among various institutions are well organized. NIS not only 68 offers recommendations on innovation planning and competi- 69 tiveness enhancement on technological front to policy makers, 70 but also attracts the attention of researchers in the field of eco- 71 nomic and innovation. The establishment of NIS thus becomes 72 more important because the interaction among government, 73 research institution, academia, and industry could result in com- 74 mercialization of technological research outputs, which in turn 75 would increase national competitiveness. 76 The innovation of a country or an industry does not originate 77 from a single economic entity; rather, it is a combination of a com- 78 posite and systematic mechanism. According to OECD [60], the 79 development of technology and innovation is the outcome of the 80 interaction among corporations, academia, and government 81 research institution. The knowledge and technology possessed by 82 the personnel and institutions are the key factors in the innovation 83 process. The development of NIS is to improve the network relation- 84 ship among the system members, which leads to enhancement in 85 the overall national innovative capability [59]. As noted earlier, 86 the NIS performance is the primary determinant of national compet- 87 itiveness. A holistic assessment that considers overall industrial 88 effects would enable a greater understanding on the impact of NIS 89 on national competitiveness. Besides, this can be guidance for tech- 90 nological policies and research resource allocation in a country. http://dx.doi.org/10.1016/j.knosys.2014.08.001 0950-7051/Ó 2014 Elsevier B.V. All rights reserved. Corresponding author. Tel.: +886 2 2898 6600x604981; fax: +886 2 2898 5927. E-mail addresses: [email protected] (W.-M. Lu), [email protected] (Q.L. Kweh), [email protected] (C.-L. Huang). 1 Tel.: +886 2 2898 6600x604983; fax: +886 2 2898 5927. Knowledge-Based Systems xxx (2014) xxx–xxx Contents lists available at ScienceDirect Knowledge-Based Systems journal homepage: www.elsevier.com/locate/knosys KNOSYS 2918 No. of Pages 10, Model 5G 18 August 2014 Please cite this article in press as: W.-M. Lu et al., Intellectual capital and national innovation systems performance, Knowl. Based Syst. (2014), http:// dx.doi.org/10.1016/j.knosys.2014.08.001

Upload: chia-liang

Post on 09-Feb-2017

217 views

Category:

Documents


2 download

TRANSCRIPT

1

3

4

5

6 Q1

78

910

1 2

1314151617

1819202122 Q223

2 4

3738

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

Contents lists available at ScienceDirect

Knowledge-Based Systems

journal homepage: www.elsevier .com/ locate /knosys

Intellectual capital and national innovation systems performance

http://dx.doi.org/10.1016/j.knosys.2014.08.0010950-7051/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Tel.: +886 2 2898 6600x604981; fax: +886 2 2898 5927.E-mail addresses: [email protected] (W.-M. Lu), [email protected]

(Q.L. Kweh), [email protected] (C.-L. Huang).1 Tel.: +886 2 2898 6600x604983; fax: +886 2 2898 5927.

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and national innovation systems performance, Knowl. Based Syst. (2014),dx.doi.org/10.1016/j.knosys.2014.08.001

Wen-Min Lu a,⇑, Qian Long Kweh b, Chia-Liang Huang a,1

a Department of Financial Management, National Defense University, No. 70, Section 2, Zhongyang North Road, Beitou, Taipei 112, Taiwanb Department of Accounting, College of Business Management and Accounting, Universiti Tenaga Nasional, Sultan Haji Ahmad Shah Campus, 26700 Muadzam Shah, Pahang, Malaysia

2526272829303132333435

a r t i c l e i n f o

Article history:Received 15 April 2014Received in revised form 1 July 2014Accepted 2 August 2014Available online xxxx

Keywords:National innovation systemNetwork DEAIntellectual capitalPerformance evaluation

a b s t r a c t

Innovation is a key resource for the well-being of national economies and international competitiveadvantages. First, this study develops a network data envelopment analysis (DEA) production processto evaluate the R&D efficiency and economic efficiency of the national innovation system (NIS) in 30countries. Our findings show that the R&D efficiencies of the NIS are better than the economic efficien-cies. Second, this study examines the effect of intellectual capital (IC) on the NIS performance throughtruncated regression. Our findings indicate that IC does play an important role in affecting the NIS per-formance. Finally, this study presents a managerial decision-making matrix and makes suggestionsthrough a performance improvement strategy map to help government and managers improve the NISperformance.

� 2014 Elsevier B.V. All rights reserved.

36

61

1. Introduction Since the seminal work of Freeman and Christopher [29], 62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

In the survey of 2007 global industrial technology innovationpolicy development, the Science, Technology and Industry Score-board of Organization for Economic Co-operation and Develop-ment (OECD) finds that a large majority of OECD countriesemphasize technological system and innovation because thesetwo factors affect not only domestic and international competition,but also economic growth [61]. A technological system is thesystem that affects specific technology or industry. According toCarlsson and Stankiewicz [13], technological system is a dynamicnetwork among economic entities in a specific technological areathat aim to create, transmit, and apply technology. In contrast, anational innovation system (NIS) affects the overall developmentof innovation. That is, a NIS is the network of institutions in thepublic and private sectors whose activities and interactions initi-ate, import, modify, and diffuse new technologies [28]. Technolog-ical system differs significantly from NIS on the following threeperspectives. First, technological system emphasizes creations ofnew technology, while NIS focuses on the transmission and appli-cation of technology. Second, the strength of technological systemin each technological area differs significantly within a country.Third, the classification of technological system is based on tech-nology rather than national boundaries.

85

86

87

88

89

90

which studies Japanese technological development, NIS researchhas evolved into an important research agenda. NIS is formedby the interaction among innovation creation, transmission, andapplication of various institutions. In other words, NIS is animportant contributor for economic growth when the interac-tions among various institutions are well organized. NIS not onlyoffers recommendations on innovation planning and competi-tiveness enhancement on technological front to policy makers,but also attracts the attention of researchers in the field of eco-nomic and innovation. The establishment of NIS thus becomesmore important because the interaction among government,research institution, academia, and industry could result in com-mercialization of technological research outputs, which in turnwould increase national competitiveness.

The innovation of a country or an industry does not originatefrom a single economic entity; rather, it is a combination of a com-posite and systematic mechanism. According to OECD [60], thedevelopment of technology and innovation is the outcome of theinteraction among corporations, academia, and governmentresearch institution. The knowledge and technology possessed bythe personnel and institutions are the key factors in the innovationprocess. The development of NIS is to improve the network relation-ship among the system members, which leads to enhancement inthe overall national innovative capability [59]. As noted earlier,the NIS performance is the primary determinant of national compet-itiveness. A holistic assessment that considers overall industrialeffects would enable a greater understanding on the impact of NISon national competitiveness. Besides, this can be guidance for tech-nological policies and research resource allocation in a country.

http://

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

2 W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

To determine its own global positioning, a country can obtainthe recent NIS performance of other countries from the historicaldata of NIS development. However, it remains a challenge to eval-uate NIS, which is characterized by multiple dimensions, especiallyin a cross-country setting. To dates, there is no consensus about theNIS measurement. The common problems include the lack of con-sensus in the measurement model, ambiguous measurement indi-cators, and the lack of an effective holistic measurement model.These situations suggest that there is a critical need for a moreobjective and reliable measurement to express the relative effi-ciency of cross-countries NIS.

To ensure proper decision making, the performance of an insti-tution needs to be measured using multiple indicators [16]. Moststudies have recommended the use of multi-factor performancemeasure model to measure organizational operating performance[2,4,6]. Data envelopment analysis (DEA) is an approach thatsimultaneously evaluates multiple inputs and multiple outputsvia linear programming technique to determine the productionfrontier as a basis of efficiency measure [3,33]. Traditional DEAmodels, however, ignore intermediate measures or interconnect-ing activities in a production process. Therefore, this study doesnot apply the traditional DEA that ignores the connectivity of inter-nal economic activities. Set in a slightly different research setting,this study employs a network DEA model that is able to explore the‘black-box’ process of NIS in estimating the performance of NIS, theR&D and economic efficiencies.

The comprehensiveness of NIS cannot be assessed merelythrough the performance. Other measurement methods shouldbe employed to provide an overall assessment on NIS. In this glob-alized knowledge-based economy, knowledge, an important factorof production, has gradually replaced machinery, equipment, cap-ital, raw materials, and labors. The concept of intellectual capital(IC) dates back to the work of Galbraith [31], where IC is inter-preted as the difference between market value and book value.Stewart [72] suggests that IC can be deemed as a major source ofenhancing or creating competitive advantage for an institutionand it is also the most important resource of overall value creation.

Many scholars employ Tobit regression to examine the impactof a variety of external environmental factors on organizationaloperating efficiency [8,18,64,65,74,78]. Simar and Wilson [71],however, utilize truncated regression with bootstrapping approachand find that truncated regression provides greater validity thanTobit regression. Consistently, this study employs Simar and Wil-son [71] truncated regression to investigate the impact of IC onthe NIS performance of 30 countries.

The purposes of this study are threefold. First, we develop a net-work DEA model that comprises R&D efficiency and economic effi-ciency to measure the NIS performance. Second, we examine therelationship between IC and NIS performance using truncatedregression with the aim to provide recommendations for improv-ing NIS performance. Third, we construct a decision-making matrixand a performance improvement strategy map to assist countriesto enhance their competitiveness through the integration of thenetwork efficiency and IC dimensions.

The remaining sections of this study are organized as follows.Section 2 provides literature review. Section 3 describes data col-lection and research methodology. Section 4 presents the findings.Section 5 provides conclusions.

209

210

211

212

213

214

215

216

2. Literature review

2.1. National innovation system

The high-tech industry plays a vital role in a country’s globalcompetitiveness. Numerous studies have examined determinants

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

of industrial economy and global competitiveness (for example,[7,27,34,51,63,80]. Although there is no single accepted definitionof NIS, the importance of NIS has been emphasized in variousresearch such as business and industry [5,11,15,38,43,79], science[40], and policy-making institutions [1,19,22,42,46,58].

The institutional interactions related to innovation and theunderlying production systems are the basic characteristics ofNIS. The basic meaning of NIS is the national innovation effortinvested by each division, which is transformed into economicdevelopment and productivity growth that ultimately gear towardnational competitiveness. In summary, the definition of NIS in thisstudy encompasses the developmental concept of industry, gov-ernment, and academia to investigate the network operating rela-tionship among business, academia, and research systems. Fromthe innovation and knowledge creation perspective, NIS createstechnological innovation, stimulates the development of thenational economy [26,28,35,44,57], and ultimately the nationalcompetitiveness of a country [76].

OECD [60] analyzes the efficiency and interaction of variousinnovation elements with knowledge creation, proliferation, andapplication through a series of indicators. McKelvey [53] suggeststhat the adoption of NIS in new technology is dissimilar. WhileFreeman [28] emphasizes improvement on social and policy-related issues, Lundvall et al. [49] focuses on the interactive learn-ing between the manufacturer and the customer. Nelson [57]emphasizes the company capacity and the creation of innovativeroutine. Freeman [28] points out that the interaction and mutualfeedback among NIS members including enterprise, academia,and research institutions are critical to NIS performance. Specifi-cally, R&D policy is an important factor that affects national pro-ductivity. However, changes in national productivity cannot fullyexplain differences in R&D policy, which is still likely to distortmarket forces. Nasierowski and Arcelus [56] are the first scholarsto apply non-parametric DEA to unlock the relationship betweencountries’ commitments and both technological efficiency and pro-ductivity, in order to measure the input and output conversion pro-cess of NIS. The findings show that the interaction variable affectsthe overall efficiency.

2.2. Intellectual capital

The term ‘‘IC’’ was first proposed by the economist Galbraith[31]. He states that IC refers to the behavior of using brain insteadof just using knowledge and mere intelligence. In order to createvaluable IC, an organization should establish valuable organiza-tional network to link the interdepartmental working team, andalso to link the external parties like customers and suppliers toaccelerate the company’s value creation. Skandia led the way in1994 by developing and issuing the first IC report in addition totraditional financial report in order to convey supplementary infor-mation on its effort in measuring knowledge assets. Based on Skan-dia work, Edvinsson and Malone [24] categorize IC into humancapital and structural capital. The structural capital is further sub-divided into customer capital and organizational capital. They alsoprovide market structure and guidelines of IC. However, there is noclear and consistent definition of IC. In summary, IC can be definedas intangible assets, or the value whereby market value exceedsbook value, which in turn will enhance organizational value andnurture firms’ competitive advantage [25].

McElroy [52] proposes another theory, which highlights thatextant IC measurement neglects social capital. According to thesocial capital theory, the mutual trust, mutual benefit, commonvalues, networks, and social norms that exist in an organizationitself and other organizations can increase the organizational val-ues and inter-organizational values, which can speed up the trans-fer of information and the development of new knowledge. The

ional innovation systems performance, Knowl. Based Syst. (2014), http://

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx 3

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

theoretical social innovation capital is the firms’ collective capacityof innovation, which can be considered as the most valuable typeof IC that forms the basis of organizational learning, innovationand adaptation ability. Thus, he proposes to incorporate social cap-ital into Skandia Market Chart and suggests that IC should com-prise human capital, social capital, and structural capital.

This study summarizes the classification of IC done by priorscholars and finds that the definitions of IC differ according toorganizational culture and environment. Note that the focus ofour study is to assess the NIS efficiency at national level and to con-duct cross-countries comparison. Thus, we consider the impact ofthe characteristics of social culture and innovation. IC is dividedinto human capital, social capital, and innovation capital[21,52,56,75]. The relevant proxy measures of IC are summarizedin Table 1.

There are some other IC-related studies. For example, Lynn [51]investigates the cultural factors of different countries and the crit-ical success factors of IC management using six companies fromthe United States, Canada, and Sweden as case studies. Theresearcher adopts Hofstede [36] four cultural dimensions indica-tors, which comprises power distance, individualism, masculinity,and uncertainty avoidance, to compare the differential impacts oforganizational culture and national culture on IC management.The findings show that culture plays an important role in IC man-agement. Inkeles [39] claims that the indicators in Hofstede [36]fall under social capital.

Dakhli and De Clercq [21] analyze the comprehensive develop-ment of a country to test the relationship between human capitaland innovation. The researchers employ the conceptual social cap-ital trust, organizational activities, and the citizens’ code of conductto test the relationship between social capital and innovation indi-cators. Based on different social dimensions of 59 countries, theirempirical results offer a theory-based model in that innovationinvolves human capital and social capital. The value creation ofthese two forms of capital provides a global point of view. Thereis a positive relationship between human capital and innovation,as well as between human capital and the social capital trust,and innovation in organizational activities. Summing up, humancapital, innovation capital, and social capital are important intangi-ble assets that drive social network, create economic value, anddevelop national competitiveness.

3. Research design

3.1. Network production process for NIS

Evaluating cross-countries NIS performance is a complex pro-cess that needs to take into account multiple indicators. Employingmultiple inputs and multiple outputs, this study constructs a net-work production model [17,41,47,50,68,69,81] that comprisesR&D efficiency and economic efficiency to capture the linkages ofinternal economic activities effectively. Through evaluation of the

Table 1Intellectual capital indicators.

Intellectualcapital

Indicators (proxy variables)

Human capital The number of employees, percentage of managers who have highemployees, the average professional salary, annual staff training tim[9,12,24,66,72]

Social capital Universal trust, organizational trust, community activities, citizensindividualism, masculinity, uncertainty avoidance [36,55,56]

Innovationcapital

The ratio of R&D expenses to total revenue, the ratio of R&D resoumaintenance costs, effective use of the number of patents, incomeof R&D personnel [23,24,66,72,75]

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

cross-countries NIS efficiency, we identify countries that are rela-tively competitive and efficient as benchmarks for other countries.The inputs, outputs, and the network production process model areshown in Fig. 1.

3.1.1. R&D efficiency stageInput variables: Labor and capital are inputs that will directly

affect the current and future development of NIS. The human com-ponents include total R&D personnel nationwide (RDP) and totalpublic expenditures on education (EDU), while the capital compo-nents include the import of goods and commercial services (ICGS)and the total expenditures on R&D (EXRD). All of these input fac-tors are deemed important investments in NIS.

Human resources are important assets of a country’s economicand technological development. RDP such as well-trained andskilled engineers, the willingness to work hard and astutely, theprofessional skill and management style involved, and integrativecapabilities to make an energetic and risk-taking entrepreneurialfirm are the main indicators to evaluate basic R&D ability[56,62,77]. In terms of enhancement of R&D ability, EDU can trainand recruit talents in innovation to improve knowledge innovationand academic excellence. These initiatives can promote knowledgeaccumulation and create first-class research university to supportR&D. In this knowledge-based era, continuous innovation in tech-nological education will not only help industry upgrade and growin a sustainable way, but also serve as a key factor to enhancenational competitiveness [37,56,70].

In developed countries, the service industry evolves and gradu-ally replaces the traditional industries. With values of the innova-tive service industry increasing rapidly, the developed countrieshave become prominent players in the NIS global competitionthrough development in knowledge-intensive industries. ICGS isan important factor in fostering a country’s NIS [56,62], whileEXRD is a crucial input that can affect patents, intellectual prop-erty, and academic research. [37,56,62].

Output variables: The output variables of technological researchrepresent the skillfulness of a country in terms of investment invarious R&D aspects. There are three types of output that are gen-erally deemed as the final output in the R&D production process.First, published scientific articles (SAP) are concrete outputs oftechnological and academic research. SAP also serves as theoreticalbasis for R&D and strengthens a country’s development of techno-logical education. Therefore, this study uses journal articlesindexed in Science Citation Index, Social Sciences Citation Index,and Engineering Index as the proxy [56,62,77]. Second, whenR&D resources are constantly invested in large scale, the numberof patents received is the most representative output indicator.For any economic entity in any country, patents not only protectthe inventor of intellectual property but also show its unique andexclusive position in the market. Patents contribute significantlyto NIS internal network R&D diffusion behavior and the acquisitionof economic benefits. Thus, this study employs patents granted to

academic qualifications, employee turnover, average salary, the average age ofe, annual staff training costs, the scale of the project team, employee satisfaction

hip behavior norms [21]; Hofstede’s four cultural dimensions: power distance,

rces to total resources, patents, copyrights, trademarks, the amount of patentobtained via patent, the ratio of revenue to management expense, the proportion

ional innovation systems performance, Knowl. Based Syst. (2014), http://

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

Fig. 1. Network production process model of the national innovation system.

4 W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

residents (PAGR) as another output proxy in measuring R&D effi-ciency [37,56,67,77]. Third, patents secured abroad by country res-idents (PSA) should be utilized as a business tool in the globalmarket to ensure that a leading position can be secured in theintense competitive global business environment. Thus, to enhancethe rigorousness of our study, we employ effective PSA as outputvariable for R&D efficiency [37,56,58,62,77].

In summary, this first stage examines whether the ICGS andEDU improve the knowledge absorption capacity of citizens basedon the concept of R&D. Furthermore, through RDPN and EXRD, out-puts can be maximized, while input can be minimized for betterperformance. As a result, we name this stage as R&D efficiency.The input variables include ICGS, EXRD, RDPN, and EDU, whereasthe output variables include SAP, PAGR, and PSA.

3.1.2. Economic efficiency stageInput variables: SAP, PAGR, and PSA, the output from R&D effi-

ciency stage, are inputs for economic efficiency stage. The rationaleis that NIS is formed by the interactions and functions of innova-tion generation, diffusion, and application of various organizations.When there are proper interactions among these institutions, NISwill function effectively and can become a driving force for eco-nomic growth. The idea of the network model is that R&D capabil-ity should enhance national economic growth. Therefore, thisstudy treats the outputs of R&D efficiency stage, i.e. SAP, PAGR,and PSA as the connecting factors or intermediate variables fornational internal economic activities.

Output variables: Gross Domestic Product (GDP) is a measure ofa country’s overall territorial economic output. It is the marketvalue of all final goods and services officially made within a coun-try in a year. Gross National Product (GNP), on the other hand, isthe market value of all goods and services produced in one yearby labor and property supplied by the residents of a country. Itreflects the average income of a nation’s citizens. GDP and GNPare different. In its simplest terms, GDP is the value of goods andservices made in a country, while GNP is the value of goods andservices made by residents. As the world becomes connected, itis common that nations and industries have a multinationalhuman resources allocation. Since we evaluate national decision-making units (DMUs), i.e. the national innovation systems of 30countries in this study, we use GDP as an output proxy in measur-ing economic efficiency [55,56].

When comparing GDPs of different countries, the domestic cur-rency has to be converted. There are two types of conversion, thefirst is using the international exchange rate, and the second isusing purchasing power parity (PPP). The PPP is a concept thattwo countries’ exchange rates of currency equalize the differencein purchasing power between the two currencies. If the countries’GDP is obtained through the above two methods, the ranking willdiffer. If we use the international exchange rate, it is unfair to those

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

developing countries as the domestic purchasing power of con-sumers and producers could be underestimated. Therefore, weemploy PPP as another output proxy in measuring economic effi-ciency [55,56].

Productivity is a measure of output from a production process,per unit of input. The definition of productivity is using resourcecreatively and effectively, and increasing the values of productsand services. Productivity growth depends mainly on a country’snatural resources and accumulation of human capital. It enhancesthe technical level and improves the institutional environment. Asa result, productivity growth determines the expansion of variousresources, and affects the economic growth of a country [55,56].Productivity (PROD), measured as the difference between real out-put value and real input value, serves as another output variable.Summing up, this second stage investigates the economic out-comes of investment in SAP, PAGR, and PSA. The outputs areGDP, PPP, and PROD.

3.2. Data collection and descriptive statistics

In this study, all of the inputs and outputs data are extractedfrom National Science Indicators on Diskette, ISI Co., USA, WorldCompetitive Yearbook (WCY) produced by International Instituteof Management Development (IMD), and proxy statements. In con-ducting the performance measurement in this study, we employthe average values of variables for different countries over the per-iod 2007–2009, which have been adjusted properly by IMD. Therequirement to establish an international comparison imposes cer-tain restrictions in obtaining a transnational homogeneous sampleof national innovation system in terms of specialization. Anotherconcern is the laggedness of outcomes relative to certain inputs.For example, the effect of SAP on final outputs is not immediate.As noted above, we use average values over the sample period,which solve the problem of laggedness [30]. After data screening,we have 30 countries as sample.

According to Cooper et al. [20], the minimum number of DMUobservations should be three times greater than the number ofinputs and outputs used when using the DEA model. In this study,the number of DMU is 30, which is at least three times greater thanthe selected variables in each stage. The two models proposed inthis study conform to this requirement. Descriptive statistics ofthe inputs and outputs are provided in Table 2.

Descriptive statistics show only the distribution of variables,but not the isotonicity relationship between the input and outputvariables. As a result, we follow Golany and Roll [32] to analyzethe correlation between inputs and outputs. The results of correla-tion coefficients are shown in Tables 3 and 4. All of the coefficientsare positive. We hence conclude that the developed network DEAmodel of the R&D efficiency stage and economic efficiency stageholds high construct validity.

ional innovation systems performance, Knowl. Based Syst. (2014), http://

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432433

435435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453454

456456

457

458459

461461

462

463

464

465

466

467

468

469

470

471

Table 3Correlation coefficients between inputs and outputs in the first stage (R&D efficiency).

Input variables Output variables

SAP PASA PAGR

PUR 0.868 0.388 0.457GERD 0.699 0.784 0.902REM 0.919 0.547 0.475EDU 0.787 0.436 0.585

Table 4Correlation coefficients between inputs and outputs in the second stage (economicefficiency).

Input variables Output variables

GDP PPP PROD

SAP 0.898 0.927 0.858PASA 0.604 0.468 0.585PAGR 0.682 0.392 0.682

W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx 5

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

3.3. Slack-based measure in network DEA

The Tone and Tsutsui [73] network DEA model is employed inthis study. The model is based on the weighted slack-based mea-sure (SBM) approach that seeks optimal solution and accountsfor the importance of each stage. Solving the black-box issue, weare able to consider the link within an entity. Moreover, we canevaluate each stage of DMUs’ performance [48,54].

The traditional DEA models (such as CCR and BCC models) thatutilize the radial measure of efficiency are not appropriate in somecases because some of the inputs or outputs are substitution anddo not change proportionally. Tone and Tsutsui [73] propose a net-work SBM to measure organizational efficiency. The SBM is a non-radial method and is suitable for measuring efficiencies wheninputs and outputs may change non-proportionally. The SBM canbe divided into three kinds: input-oriented SBM, output-orientedSBM, and non-directional SBM.

Accounting for both input and output slacks, the non-orientedefficiency for the observed DMU (o = 1, . . . , n) can be formulatedby the following fractional programming problem:

TEo ¼ MIN1K

XK

k¼1

1� 1mk

Xmk

i¼1

sk�

i

xkio

!" #,1þ 1

pk

Xpk

r¼1

skþr

ykro

!" #

s:t:

xkio ¼

Xn

j¼1

xkijk

kj þ sk�

i ; i ¼ 1; . . . ;mk; k ¼ 1; . . . ;K;

ykro ¼

Xn

j¼1

ykrjk

kj � skþ

r ; r ¼ 1; . . . ; pk; k ¼ 1; . . . ;K;

Xn

j¼1

kkj ¼ 1; k ¼ 1; . . . ;K;

Xn

j¼1

zðk;hÞtj kkj ¼

Xn

j¼1

zðk;hÞtj khj ; t ¼ 1; . . . ; sðk;hÞ;8ðk; hÞ;

kkj P 0; kh

j P 0; sk�

i P 0; skþ

r P 0;8k;

ð1Þ

where xkij is the amount of input i to the DMU j at stage k; yk

rj is the

amount of output r to the DMU j at stage k; zðk;hÞtj is the amount of link-ing intermediate product t from stage k to stage h to the DMU j; n isthe number of DMUs, K is the number of stages (k = 1, . . . , K); 1/Krepresents is equal weight at each stage; mk represents the numberof inputs at stage k; pk represents the number of outputs at stagek; s(k,h) represents the number of items in link from stage k to stageh. The kk

j represents the weight an observed DMU references from

DMU j so as to achieve target efficiency at stage k. sk�i and skþ

r arethe slack values at stage k for inputs and outputs, respectively.

Table 2Descriptive statistics of the input and output variables.

Variables Mean Minimum Maximum Standard deviation

Input variablesPUR 314,904 1,173,167 36,466 273,764GERD 17,586 148,610 472 29,766REM 181,302 1,534,267 15,833 310,448EDU 38,097 142,375 4,230 41,478

Intermediate variablesSAP 48,181 246,180 6,222 53,695PASA 10,288 127,644 65 26,801PAGR 5,537 71,455 22 13,383

Output variablesGDP 953,425 4,549,800 118,400 1,128,516PPP 1,076,189 8,093,000 110,000 1,584,774PROD 1,988,285 9,093,405 229,652 2,269,625

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

This fractional program problem can be solved by being trans-formed into a linear program using the Charnes and Cooper trans-formation [14]. Let an optimal solution to Eq. (1) be (kk�

j ; s��i ; sþ�r ) for

the observed DMU. If TE�o ¼ 1 in Eq. (1), the observed DMU is calledoverall efficient. Meanwhile, a value of less than 1 on TE�o meansthat the observed DMU is overall inefficient.

The stage k efficiency score for the observed DMU can bedefined by

TEk�o ¼ 1� 1

mk

Xmk

i¼1

sk��i

xkio

!" #,1þ 1

pk

Xpk

r¼1

skþ�r

ykro

!" #; k ¼ 1; . . . ;K:

ð2ÞThe overall efficiency score is the equal weighted mean of the K

stage scores:

TE�o ¼1K

XK

k¼1

TEk�o ð3Þ

472

473474476476

477

478

479

480

481

482

483

484

485

487487

3.4. Truncated regression

Prior studies have applied Tobit regression to examine theimpacts of exogenous factors on DEA scores since efficiency valuesobtained range between 0 and 1. [8,18,64,65,74,78]. However,Simar and Wilson [71] verified that truncated regression was moreappropriate than Tobit regression. In other words, they illustratedthat Tobit regression was inappropriate to analyze the efficiencyscore under DEA. Their results indicate that truncated regressionis superior to the Tobit regression. Therefore, following Simarand Wilson [71], we use truncated regression model to investigatethe association between intellectual capital and the performance ofNIS. The model is as follows:

hj ¼ aþ Zjbþ ej; j ¼ 1; . . . ;n ð4Þ

In Eq. (4), the scalar a and the (d � 1) vector b are unknown parame-ters to be estimated; Zj is a (1 � d) vector of observation-specific vari-ables (intellectual capital) for NIS that we expect it to be related to theNIS overall efficiency score, which is proxied by hj. Since the distribu-tion of ej is restricted by the condition ej P 1 � a � Zjb, Eq. (4) is mod-ified to get Eq. (5), which assumes that the distribution beforetruncation is truncated with zero mean, unknown variance and atruncation point, which are determined by different conditions:

hj � aþ Zjbþ ej; j ¼ 1; . . . ;n

where

ej : Nð0;r2e Þ; such that ej P 1� a� Zjd; j ¼ 1; . . . ;n:

ð5Þ

ional innovation systems performance, Knowl. Based Syst. (2014), http://

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

Table 5National innovative system performance value.

DMU R&Defficiency

Economicefficiency

Overallefficiency

Area

Australia 1.000 0.513 0.757 OceaniaNew Zealand 1.000 0.175 0.587 Oceania

6 W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

The regression process of parametric bootstrapping is used to con-struct the bootstrap confidence intervals for the estimates ofparameters ðd;r2

e Þ, and to estimate Eq. (5) by maximizing the corre-sponding likelihood function, and give heed to the ðd;r2

e Þ. For thesake of brevity, the reader can refer Simar and Wilson [71] for thedetails of the estimation algorithm.

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

Brazil 1.000 1.000 1.000 Latin AmericaMexico 0.391 1.000 0.695 Latin AmericaAustria 0.548 0.306 0.427 EuropeBelgium 0.547 0.377 0.462 EuropeCzech Republic 1.000 0.237 0.618 EuropeDenmark 1.000 1.000 1.000 EuropeFinland 1.000 0.565 0.783 EuropeFrance 0.623 1.000 0.812 EuropeGermany 1.000 1.000 1.000 EuropeGreece 1.000 0.549 0.774 EuropeIreland 1.000 0.269 0.635 EuropeNetherlands 1.000 0.601 0.800 EuropeNorway 0.782 0.462 0.622 EuropePoland 1.000 1.000 1.000 EuropePortugal 1.000 0.434 0.717 EuropeSpain 1.000 1.000 1.000 EuropeSweden 0.647 0.412 0.529 EuropeSwitzerland 1.000 1.000 1.000 EuropeBritain 1.000 1.000 1.000 EuropeChina 1.000 1.000 1.000 AsiaHong Kong 1.000 1.000 1.000 AsiaJapan 1.000 1.000 1.000 AsiaKorea 1.000 1.000 1.000 AsiaMalaysia 1.000 1.000 1.000 AsiaSingapore 1.000 0.137 0.568 AsiaTaiwan 1.000 1.000 1.000 AsiaThailand 1.000 1.000 1.000 AsiaTurkey 1.000 1.000 1.000 AsiaAverage 0.918 0.735 0.827

Note: Overall Efficiency = 1/2 (R&D Efficiency + Economic Efficiency).

4. Empirical analysis

4.1. National innovation system efficiency analysis

Table 5 shows the results of the network DEA model. The meanvalues of R&D efficiency, economic efficiency, and the overall effi-ciency are 0.918, 0.735, and 0.827. The results show that the R&Dperformance is better than the economic performance, indicatingthat most countries have good R&D capabilities, but have failedto create economic value through their R&D capabilities.

In terms of the R&D efficiency stage, 24 countries are deemedefficient, accounting for 80% of the sample. Among the regions,the average R&D performance of countries in Oceania and Asia arethe best, reaching the value of 1. Meanwhile, countries in The Amer-icas have the lowest value of R&D efficiency (0.696). With respect tothe economic efficiency stage, 17 countries (approximately 57%) arefound to be efficient. The regional average performance of theAmerican national innovation system is the best (1.000), followedby Asia (0.940), Oceania (0.735), and Europe (0.659).

To understand whether significant differences on the efficiencyperformance exist among the different regions, this study uses theKruskal–Wallis test [10]. At the significance level of 5%, we showthat there is no significant difference on the performance of R&Defficiency stage among the regions. However, there is a significantdifference on the performance of economic efficiency stage amongthe regions. Details are shown in Table 6.

4.2. Intellectual capital and the national innovation systemperformance

IC sources of future benefits that lack physical embodiment canbe regarded as a contributor to national economic growth. A coun-try’s NIS aims at increasing the R&D and economic value. Having awell-developed system will be beneficial to all residents; therefore,enhancing national competitiveness so as to grow will be the pri-ority of governments. The best and perhaps only way for countriesto perform effectively and efficiently is through a strong combina-tion of IC and the development of NIS. To explore the relationshipbetween intellectual capital and the NIS performance, we use threevariables that are related to the dimensions of human capital, inno-vation capital, and social capital. They are described as follows,respectively. The IC-related data are extracted from the Greet Hof-stede Analysis2 database.

4.2.1. Human capitalThe number of employee is the substantial index of human cap-

ital [24,45]. Employees who have R&D talent play a core role in NISand they contribute to productivity growth and competitivenessby utilizing tangible assets and intangible assets in an organization.From the perspective of NIS, the proportion of R&D talents willaffect the innovation activities. Therefore, this study employs theratio of the number of R&D talents to working population as aproxy of human capital. This variable reflects the distribution ofinnovative human resources, and reveals the competitivenessand innovativeness of this country.

574

575

576

2 Greet Hofstede Analysis database are constructed by Geert Hofstede et al. thewebsite: http://www.geert-hofstede.com/hofstede_dimensions.php.

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

4.2.2. Innovation capitalVan Buren [75] provide a succinct definition of innovation cap-

ital in an organization. That is, three core proxies used to measurethe performance are the expenditure of R&D, the percentage ofinnovation in works, and products’ innovativeness. In essence, theyare used in the production process of NIS to provide a country witha competitive edge. Furthermore, all of the inputs in R&D expendi-ture, the number of R&D labors, the cost of training and education,and the import and export business could bring about ‘‘patent’’ asthe major output. Accordingly, we employ patent productivity asthe proxy of innovation capital. Successful innovation capital isin fact a reflection of the contribution by R&D employees. More-over, it also reveals whether the difference in the degree of innova-tion capital will affect the efficiency of the whole system.

4.2.3. Social capitalAccording to Inkeles [39], the ‘‘uncertainty avoidance’’ variable

that is developed as a proxy to measure cross-cultural differencesby Hofstede [36] can be classified as a part of social capital. In addi-tion, Nasierowski and Arcelus [56] also have confirmed that‘‘uncertainty avoidance’’ variable has played an intermediary rolein the NIS performance. Uncertainty avoidance is the level of riskaccepted by a culture of a nation, which can be gleaned from theemphasis on rule adherence, ritual behavior, and labor mobility.Culture of a nation with low levels of uncertainty avoidance hasgreater tolerance for ambiguity and a less need for formal rules.The opposite will hold true for high-uncertainty avoidance cul-tures. On the other hand, if a nation has a high level of uncertaintyavoidance, she will have more rigorous planning in the process ofinnovation system in order to avoid the occurrence of uncertain-ties. With rigorous planning, a nation can reduce the cost of non-essential spending on those uncertainties and the economics ofthe nation will be more stable and ultimately prospers. As a result,‘‘uncertainty avoidance’’ is utilized as the proxy of social capital.

ional innovation systems performance, Knowl. Based Syst. (2014), http://

577 Q3578580580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

Table 7Descriptive statistics of intellectual capital.

Variable Mean Standard deviation Minimum Maximum

HC 0.006 0.003 0.001 0.013InC 5.653 10.629 0.270 52.030SC 62.226 26.080 8.000 112.000

Note: HC and InC are measured in percentage, while SC is measured as an index.

Table 6The average performance of the network model by region.

Area Number of countries R&D efficiency mean Kruskal–wallis test (p-Value) Economic efficiency Mean Kruskal–Wallis test (p-Value)

Oceania 2 1 0.142 0.344 0.036*

Latin America 2 0.696 1Europe 17 0.891 0.659Asia 9 1 0.940Overall 30 0.918 0.735

* p-Value < 0.05.

W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx 7

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

This paper assumes and tests the regression model as follows:

TEj ¼ aþ b1HCj þ b2InCj þ b3SCj þ ej; j ¼ 1; . . . ;n ð6Þ

Using the truncated regression [71], we investigate whether IC hassignificant effects on the NIS performance. TEj is the DEA score. HCj

is the human capital; InCj is the intellectual capital; RCj is the socialcapital.

Table 7 presents the descriptive statistics of the IC variables,while the empirical results analysis is shown in Table 8. All ofthe independent variables are significantly related to the NISperformance.

In summary, in order to enhance the overall performance ofnational innovation systems, we suggest that governments recruithigh-quality R&D team to strengthen global competitiveness. Withrespect to innovation capital, governments should increase theopportunities of innovation and research exchanges, as well asenhance the quality and quantity of outcome of the patent andother intellectual property to expand its economic benefits. Wealso suggest that governments could enhance the social capitalby actively linking government agencies, business organizations,civil society, and academic research agencies to enhance nationalsynergy effect.

4.3. Discussion

4.3.1. Decision-making matrixThis study further explores the condition of R&D efficiency and

economic efficiency in NIS. The correlation between the two stagesis significantly positive (coefficient = 0.105). We construct a mana-gerial matrix for decision making. In Fig. 2, the horizontal axis rep-resents R&D efficiency and the vertical axis represents economic

Table 8The results of truncated regression models.

Variables Coefficient

Intercept –2.679HC (the number of R&D talents to working population) 4.360InC (patent productivity) 0.372SC (uncertainty avoidance) 1.123R2 0.224

*** p-Value < 0.01.** p-Value < 0.05.

* p-Value < 0.10.

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

efficiency. By assigning each DMU to one of the four zones, wecan clearly define the benchmark to be emulated by inefficientcountries. This provides an insight for governments to developappropriate strategies for enhancing their overall efficiency. Theanalysis result as follows.

Zone A: The countries in this zone exhibit better R&D efficiencyand economic efficiency than those in other zones. There are fif-teen countries in this zone: Brazil, Denmark, Germany, Poland,Britain, Switzerland, Malaysia, Turkey, China, Spain, Hong Kong,Taiwan, Thailand, Japan, and South Korea. They can be deemedas a role model for inefficient countries.

Zone B: This zone includes two countries, namely Mexico andFrance, which have lower R&D efficiency but relatively efficienteconomic performance. We suggest that these two countriesshould strengthen the foundation of energy and enhance industrialcooperation, and continue to interact with various internationalinstitutions in order to reduce a great waste of R&D resources.Remodeling and renewing the process of R&D will increase thebenefits to reach the synergy effect. Results from the uncertaintyabout the outcome of R&D suggest that technology is the prioritytask that would enable accurate and fast collection of information,and would protect the outcome of R&D patent and the intellectualproperty rights. In the aspect of human resource, they should hirehigh-quality technology professionals, and put emphasis on train-ing and education to improve the ability of R&D management.

Zone C: This zone contains nine countries, namely Australia,New Zealand, the Czech Republic, Finland, the Netherlands,Ireland, Greece, Portugal, and Singapore. Their R&D performanceis efficient but their economic efficiency is substandard. Theyshould focus on reducing operating costs and makingmarket-demand-oriented managerial strategies decision. Theycan select the key success factors for the development strategy,and commit to operating diversification in order to increase thevalue added and to maximize the economic value. In this case,technology that mainly lies in the innovation process are market-ing communication and channel expansion, which are meant forprofiting from R&D diffusion. As for the human resource, the gov-ernments should strengthen corporation among the industries,government, and academic institutions to increase the diffusionof R&D outcome to improve the overall efficiency of the nations.

Std. error Z-stat p-Value

1.924 –1.392 0.0812.117 2.059 0.019**

0.128 2.899 0.002***

7.767 1.464 0.072*

ional innovation systems performance, Knowl. Based Syst. (2014), http://

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

Fig. 2. Decision making matrix.

8 W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

Zone D: The four countries in this zone are Austria, Belgium,Sweden, and Norway. They are inefficient in R&D and economicperformances. We suggest that these four countries should fullyapply the improvement strategies in other three zones. For exam-ple, they should emphasize the strength of the R&D performance;managers should get rid of the concept that R&D creates greatercosts and accentuate the concept of unit efficiency. To increasethe economic efficiency, they should integrate the enterprise valuechain, and participate in industry cooperation and technologytransformation to strengthen the competitiveness of their NIS.

Fig. 3. Performance impro

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

4.3.2. Performance improvement strategy mapAs Section 2 points out, IC plays a key role in improving the NIS

performance. In this subsection, we discuss a performanceimprovement strategy map discussed. Continuing from decision-making matrix, we provide suggestions for countries in each quad-rant to improve, focus, and strengthen their NIS through IC. Fig. 3shows the map.

The performance improvement strategy for countries in quad-rant A: We recommend the countries (i) to continue strengtheninginternational exchanges, cooperation in science and technologypolicy projects, international intellectual property rights, and eco-nomic diffusion capacity (technology-related policy), (ii) tocontinue focusing on international trade and investment agree-ments, as well as regional innovation and the effectiveness of thenational innovation projects (industry-related policy), (iii) toachieve economic growth and to integrate innovation with envi-ronmental issues, in order to adapt to the fast-changing worldtrends (innovation-related policy).

For example, Japan is far ahead of other countries in innovatingnew technology because Japan depends not only on the scale ofR&D, but also the close linkage with the social and institutionalchange. The Department of the Government of Japan promoteseconomic modernization, with the Ministry of Economy, Tradeand Industry (METI) playing the role of to promote high-techdevelopment, long-term development of the world market for itsmain functions, and close cooperation between large domesticindustry and the Government (human capital strategy). The enter-prise-level R&D strategy and technology introduction and theinnovative changes viz. ‘‘factory is laboratory’’ bring about changesin innovation (innovation capital strategy). Training and education

vement strategy map.

ional innovation systems performance, Knowl. Based Syst. (2014), http://

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761Q4762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802

W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx 9

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

in science and technology have made the concept of quality andinnovation common in Japan, thus resulting in an environmentwhere Japanese pay attention to the uniqueness and innovative-ness of goods and services in Japan (social capital strategy), andultimately the Kingdom of innovation is formed [29]. Takentogether, Japan can serve as the benchmark for countries in otherquadrants to improve.

The performance improvement strategy for countries in quad-rant B: From the intellectual capital viewpoint, we recommendthat the countries should train and reward scientific and techno-logical personnel (human capital strategy). Note that countries inthis quadrant are inefficient in the R&D efficiency stage. Therefore,the countries should promote industry-university cooperation toenhance research and development. For example, they canimprove innovative capital structure by integrating innovationlearning into industry clusters (innovation capital strategy). More-over, through production technology and innovative strategic alli-ances, innovation infrastructure can be improved and theallocation of resources can be optimized. To enhance learning effi-ciency, cooperation between university and industry is necessary.

The performance improvement strategy for countries in quad-rant C: We recommend that countries in this quadrant to establisha better education and training system. Specifically, they shouldimprove applied science and technology project managementand innovative service quality (human capital strategy). The coun-tries should advocate strategic planning on technology commer-cialization, product commercialization, and technology transfer.They should also encourage entrepreneurial investment (innova-tion capital strategy). Organizational process re-engineering andre-engineering, innovation-active corporate culture and goodwill,and e-commerce function are all important tactics to shorten leadtime (social capital strategy).

The performance improvement strategy for countries in quad-rant D: We recommend countries in this quadrant to enhancethe research and development capability as well as the educationand training system. They should create supply and demand mar-kets. To achieve that, they may have to tighten their defenseresources and cooperate with others (human capital strategy).Engaging in strategic alliances, encouraging studies in innovation,integrating technology diffusion among industry clusters areamong the potential approaches for the countries to improve theirR&D and economic efficiencies (innovation capital strategy). Thegovernment should give tax incentives or incentive allowances.Institutional bodies also should support and promote innovation.Financial system support should be built (social capital strategy).

803804805806807808809810811812813814815816817818819820821822823824825826827828829

5. Conclusion

This study employs a network DEA model to investigate the effi-ciency of NIS. From the findings, we offer some managerial impli-cations through a decision-making matrix. Moreover, wedetermine the impact of intellectual capital on the overall perfor-mance using truncated regression. Our research findings may offersome suggestion to improve the NIS performance and resourceallocation. The findings are as follows. Generally, NIS is efficientin R&D context but needs improvement in economic efficiency.America NIS is the most efficient one in terms of both R&D effi-ciency and economic efficiency stages. The countries that areefficient in R&D efficiency stage are not necessary efficient in theeconomic efficiency stage. The countries that are efficient in bothstages include Brazil, Denmark, Germany, Poland, the United King-dom, Switzerland, Malaysia, Turkey, China, Spain, Hong Kong, Tai-wan, Thailand, Japan, and South Korea. The overall efficiency ofR&D inefficient countries can be improved through economic effi-ciency. For example, Mexico and France are inefficient in R&D

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

efficiency but their overall efficiencies are still satisfactory becauseof good efficiencies in economic performance. With regards to ourregression analysis, IC is significantly related to NIS efficiency. Thisresult shows that the recruit of high quality R&D talent, participa-tion in industry, government, academia, research institutions toincrease innovation and research exchange opportunities couldimprove the quality and quantity of R&D outcomes. Educationaltraining can be held frequently to strengthen organizational com-petitiveness and foster teamwork to further maximize the organi-zational economic benefits.

References

[1] L. Alcorta, New economic policies and the diffusion of machine tools in LatinAmerica, World Dev. 28 (2000) 1657–1672.

[2] E.I. Altman, Corporate Bankruptcy in America, Heath Lexington Books,Lexington, Massachusetts, 1971.

[3] T.R. Anderson, T.U. Daim, J. Kim, Technology forecasting for wirelesscommunication, Technovation 28 (2008) 602–614.

[4] J. Argenti, Corporate Collapse: The Causes and Symptoms, McGraw-Hill,London, 1976.

[5] E. Autio, A.-P. Hameri, The structure and dynamics of technological systems: aconceptual model, Technol. Soc. 17 (1995) 365–384.

[6] R.P. Bagozzi, L.W. Phillips, Representing and testing organizational theories: aholistic construal, Adm. Sci. Q. 27 (1982) 459–489.

[7] D. Bennett, K. Vaidya, Meeting technology needs of enterprises for nationalcompetitiveness, Int. J. Technol. Manage. 32 (2005) 112–153.

[8] J.N. Binam, K. Sylla, I. Diarra, G. Nyambi, Factors affecting technical efficiencyamong coffee farmers in Côte d’Ivoire: evidence from the centre west region,Afr. Dev. Rev. 15 (2003) 66–76.

[9] N. Bontis, Intellectual capital: an exploratory study that develops measuresand models, Manag. Decis. 36 (1998) 63–76.

[10] P.L. Brockett, B. Golany, Using rank statistics for determining programmaticefficiency differences in data envelopment analysis, Manage. Sci. 42 (1996)466–472.

[11] S. Brusoni, A. Geuna, An international comparison of sectoral knowledgebases: persistence and integration in the pharmaceutical industry, Res. Policy32 (2003) 1897–1912.

[12] P.N. Bukh, H.T. Larsen, J. Mouritsen, Constructing intellectual capitalstatements, Scand. J. Manag. 17 (2001) 87–108.

[13] B. Carlsson, R. Stankiewicz, On the nature, function and composition oftechnological systems, J. Evol. Econ. 1 (1991) 93–118.

[14] A. Charnes, W.W. Cooper, Programming with fractional functionals, Naval Res.Logis. Q. 9 (3–4) (1962) 181–186.

[15] A.K. Chakrabarti, Competition in high technology: analysis of patents of US,Japan, UK, France, West Germany, and Canada, IEEE Trans. Eng. Manage. 38(1991) 78–84.

[16] B.S. Chakravarthy, Measuring strategic performance, Strateg. Manag. J. 7(1986) 437–458.

[17] Y. Chen, J. Zhu, Measuring information technology’s indirect impact on firmperformance, Inf. Technol. Manage. 5 (2004) 9–22.

[18] H.-L. Chu, S.-Z. Liu, J. Romeis, C.-L. Yaung, The initial effects of physiciancompensation programs in Taiwan hospitals: implications for staff modelHMOs, Health Care Manage. Sci. 6 (2003) 17–26.

[19] D. Chudnovsky, J. Niosi, N. Bercovich, National innovation systems, learningprocesses and technology policy: comparison of Canada and Argentina,Desarro. Econ.–Rev. Ciencias Sociales 40 (2000) 213–252.

[20] W.W. Cooper, S. Li, L.M. Seiford, K. Tone, R.M. Thrall, J. Zhue, Sensitivity andstability analysis in DEA: some recent developments, J. Prod. Anal. 15 (2001)217–246.

[21] M. Dakhli, D. De Clercq, Human capital, social capital, and innovation: a multi-country study, Entrep. Region. Dev. 16 (2004) 107–128.

[22] M.V. Dijk, A. Szirmai, Industrial policy and technology diffusion: evidence frompaper making machinery in Indonesia, World Dev. 34 (2006) 2137–2152.

[23] R. Dzinkowski, The measurement and management of intellectual capital: anintroduction, Manage. Account. 78 (2000) 32–36.

[24] L. Edvinsson, M.S. Malone, Intellectual Capital: Realizing Your Company’s TrueValue by Finding its Hidden Roots, Harper Collins Publishers Inc, New York,1997.

[25] L. Edvinsson, P. Sullivan, Developing a model for managing intellectual capital,Eur. Manage. J. 14 (1996) 356–364.

[26] R. Ennals, Europe as a knowledge society: integrating universities,corporations, and government, Syst. Pract. Act. Res. 17 (2004) 237–248.

[27] M. Fransman, International competitiveness, technical change and the state:the machine tool industry in Taiwan and Japan, World Dev. 14 (1986) 1375–1396.

[28] C. Freeman, The ‘National System of Innovation’ in historical perspective,Camb. J. Econ. 19 (1995) 5–24.

[29] C. Freeman, F. Christopher, Technology Policy and Economic Performance:Lessons from Japan, Pinter Publishers, London, 1987.

[30] J.L. Furman, M.E. Porter, S. Stern, The determinants of national innovativecapacity, Res. Policy 31 (2002) 899–933.

ional innovation systems performance, Knowl. Based Syst. (2014), http://

830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886

887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943

10 W.-M. Lu et al. / Knowledge-Based Systems xxx (2014) xxx–xxx

KNOSYS 2918 No. of Pages 10, Model 5G

18 August 2014

[31] J.K. Galbraith, The Affluent Society, Hamilton, London, 1969.[32] B. Golany, Y. Roll, An application procedure for DEA, Omega 17 (1989) 237–250.[33] J. Guan, K. Chen, Measuring the innovation production process: a cross-region

empirical study of China’s high-tech innovations, Technovation 30 (2010)348–358.

[34] I. Gurkov, Business innovation in Russian industry, Post-Communist Econ. 16(2004) 423–438.

[35] T. Hayashi, Effect of R&D programmes on the formation of university–industry–government networks: comparative analysis of Japanese R&Dprogrammes, Res. Policy 32 (2003) 1421–1442.

[36] G. Hofstede, Cultures and Organizations: Software of the Mind, McGraw-Hill,London, 1991.

[37] M.-C. Hu, J.A. Mathews, National innovative capacity in East Asia, Res. Policy34 (2005) 1322–1349.

[38] S.-W. Hung, Competitive strategies for Taiwan’s thin film transistor-liquidcrystal display (TFT-LCD) industry, Technol. Soc. 28 (2006) 349–361.

[39] A. Inkeles, Measuring social capital and its consequences, Policy Sci. 33 (2000)245–268.

[40] R. Kaiser, H. Prange, The reconfiguration of national innovation systems–theexample of German biotechnology, Res. Policy 33 (2004) 395–408.

[41] C. Kao, S.-N. Hwang, Efficiency decomposition in two-stage data envelopmentanalysis: an application to non-life insurance companies in Taiwan, Eur. J.Oper. Res. 185 (2008) 418–429.

[42] S. Kenny, Defining a national system of innovation: implications for Irishindustrial development policy, Reg. Stud. 29 (1995) 692–697.

[43] N. Kumaresan, K. Miyazaki, An integrated network approach to systems ofinnovation–the case of robotics in Japan, Res. Policy 28 (1999) 563–585.

[44] M. Levin, Cross-boundary learning systems–integrating universities,corporations, and governmental institutions in knowledge generatingsystems, Syst. Pract. Action Res. 17 (2004) 151–159.

[45] J. Liebowitz, C.Y. Suen, Developing knowledge management metrics formeasuring intellectual capital, J. Intellect. Capital 1 (2000) 54–67.

[46] Y. Lim, Development of the public sector in the Korean innovation system, Int.J. Technol. Manage. 20 (2000) 684–701.

[47] S.-F. Lo, W.-M. Lu, Does size matter? Finding the profitability andmarketability benchmark of financial holding companies, Asia-Pacific J. Oper.Res. 23 (2006) 229–246.

[48] W.-M. Lu, S.-W. Hung, Exploring the operating efficiency of TechnologyDevelopment Programs by an intellectual capital perspective––a case study ofTaiwan, Technovation 31 (2011) 374–383.

[49] B.-Å. Lundvall, B. Johnson, E.S. Andersen, B. Dalum, National systems ofproduction, innovation and competence building, Res. Policy 31 (2002) 213–231.

[50] X. Luo, Evaluating the profitability and marketability efficiency of large banks:an application of data envelopment analysis, J. Bus. Res. 56 (2003) 627–635.

[51] L.H. Lynn, Technology competition policies and the semiconductor industriesof Japan and the United States: a fifty-year retrospective, IEEE Trans. Eng.Manage. 47 (2000) 200–210.

[52] M.W. McElroy, Social innovation capital, J. Intellect. Capital 3 (2002) 30–39.[53] M. McKelvey, Using evolutionary theory to define systems of innovation, in: C.

Edquist (Ed.), Systems of Innovation: Technologies, Institutions andOrganizations, Pinter, London, 1997, pp. 200–222.

[54] S.M. Mirhedayatian, M. Azadi, R. Farzipoor Saen, A novel network dataenvelopment analysis model for evaluating green supply chain management,Int. J. Prod. Econ. 147 (2014) 544–554.

[55] W. Nasierowski, F.J. Arcelus, On the stability of countries’ nationaltechnological systems, in: S.H. Zanakis, G. Doukidis, C. Zopounidis (Eds.),

944

Please cite this article in press as: W.-M. Lu et al., Intellectual capital and natdx.doi.org/10.1016/j.knosys.2014.08.001

Decision Making: Recent Developments and Worldwide Applications,Springer, Boston, 2000, pp. 97–111.

[56] W. Nasierowski, F.J. Arcelus, On the efficiency of national innovation systems,Socio-Econ. Plann. Sci. 37 (2003) 215–234.

[57] R.R. Nelson, National Innovation Systems, A Comparative Analysis, OxfordUniversity Press, New York, 1993.

[58] J. Niosi, P. Saviotti, B. Bellon, M. Crow, National systems of innovation: insearch of a workable concept, Technol. Soc. 15 (1993) 207–227.

[59] OECD, Review of National Science and Technology Policy, Paris, 1996.[60] OECD, National Innovation System: Background Report, Paris, 1997.[61] OECD, Science, Technology and Industry Scoreboard 2007: Innovation and

Performance in the Global Economy, Paris, 2007.[62] T.-W. Pan, S.-W. Hung, W.-M. Lu, DEA performance measurement of the

national innovation system in Asia and Europe, Asia-Pacific J. Oper. Res. 27(2010) 369–392.

[63] A.L. Porter, J.D. Roessner, X.-Y. Jin, N.C. Newman, Changes in nationaltechnological competitiveness: 1990, 1993, 1996 and 1999, Technol. Anal.Strateg. Manage. 13 (2001) 477–496.

[64] J. Raczka, Explaining the performance of heat plants in Poland, Energy Econ. 23(2001) 355–370.

[65] M. Resende, Regulatory regimes and efficiency in US local telephony, OxfordEcon. Pap. 52 (2000) 447–470.

[66] J. Roos, G. Roos, L. Edvinsson, N.C. Dragonetti, Intellectual Capital: Navigatingin the New Business Landscape, New York University Press, New York, 1998.

[67] S. Rousseau, R. Rousseau, Data envelopment analysis as a tool for constructingscientometric indicators, Scientometrics 40 (1997) 45–56.

[68] L.M. Seiford, J. Zhu, Profitability and marketability of the top 55 U.S.commercial banks, Manage. Sci. 45 (1999) 1270–1288.

[69] T.R. Sexton, H.F. Lewis, Two-stage DEA: an application to major leaguebaseball, J. Prod. Anal. 19 (2003) 227–249.

[70] J.Z. Shyu, Y.-C. Chiu, C.-C. Yuo, A cross-national comparative analysis ofinnovation policy in the integrated circuit industry, Technol. Soc. 23 (2001)227–240.

[71] L. Simar, P.W. Wilson, Estimation and inference in two-stage, semi-parametricmodels of production processes, J. Econom. 136 (2007) 31–64.

[72] T.A. Stewart, Intellectual Capital: The new Wealth of Organizations, BantamDoubleday Dell Publishing Group Inc, New York, 1997.

[73] K. Tone, M. Tsutsui, Network DEA: a slacks-based measure approach, Eur. J.Oper. Res. 197 (2009) 243–252.

[74] H. Turner, R. Windle, M. Dresner, North American containerport productivity:1984–1997, Trans. Res. Part E: Logis. Trans. Rev. 40 (2004) 339–356.

[75] M.E. Van Buren, A yardstick for knowledge management, Train. Dev. 53 (1999)71–78.

[76] C.-H. Wang, L.-Y. Lu, C.-B. Chen, Evaluating firm technological innovationcapability under uncertainty, Technovation 28 (2008) 349–363.

[77] E.C. Wang, W. Huang, Relative efficiency of R&D activities: a cross-countrystudy accounting for environmental factors in the DEA approach, Res. Policy 36(2007) 260–273.

[78] K.L. Wang, Y.T. Tseng, C.C. Weng, A study of production efficiencies ofintegrated securities firms in Taiwan, Appl. Financ. Econ. 13 (2003)159–167.

[79] C.-C. Yeh, P.-L. Chang, The Taiwan system of innovation in the tool machineindustry: a case study, J. Eng. Tech. Manage. 20 (2003) 367–380.

[80] E. Yglesias, Porter vs. Porter: modeling the technological competitiveness ofnations, Scientometrics 57 (2003) 281–293.

[81] J. Zhu, Multi-factor performance measure model with an application toFortune 500 companies, Eur. J. Oper. Res. 123 (2000) 105–124.

ional innovation systems performance, Knowl. Based Syst. (2014), http://