The dynamics of information and communications technologies infrastructure, economic growth, and financial development: Evidence from Asian countries

Technology in Society 42 (2015) 135e149

Contents lists available at ScienceDirect

Technology in Society journal homepage: www.elsevier.com/locate/techsoc

The dynamics of information and communications technologies infrastructure, economic growth, and financial development: Evidence from Asian countries Rudra P. Pradhan a, *, Mak B. Arvin b, Neville R. Norman c, d a

Vinod Gupta School of Management, Indian Institute of Technology Kharagpur, WB 721302, India Department of Economics, Trent University, Peterborough, Ontario K9J 7B8, Canada c Department of Economics, University of Melbourne, Victoria 3053, Australia d Department of Economics, University of Cambridge, Cambridge CB3 9DD, UK b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 September 2014 Received in revised form 3 April 2015 Accepted 14 April 2015 Available online

This paper investigates causal relationships between information and communications technologies (ICT) infrastructure, financial development, and economic growth in Asian countries over the twelve-year period 2001e2012. Using panel cointegration techniques, our empirical results show these variables are cointegrated, with a myriad of short-run and long-run causal links between ICT infrastructure and economic growth, between financial development and economic growth, and between ICT infrastructure and financial development. © 2015 Elsevier Ltd. All rights reserved.

Keywords: ICT infrastructure Financial development Economic growth Panel Granger causality tests

1. Introduction Information and communications technologies (ICT) are increasingly associated with more rapid economic growth, especially during the pronounced globalization era of the 1990s [28,45]. Today, ICT are vital components of modern infrastructure, with widespread applications through world economies. They initiate and reflect new technological developments, and foster widespread cost-reducing innovations, affecting innovative behaviour, economic restructuring and productivity performance in all sectors of the modern economy (see, for instance, [15,19,36,49,74,75,95,96,106,111,118,127,128,129]). The significance of ICT lies in the fact that the technologies are

* Corresponding author. E-mail addresses: [email protected] (R.P. Pradhan), [email protected] (M.B. Arvin), [email protected], nrn1v@econ. cam.ac.uk (N.R. Norman). http://dx.doi.org/10.1016/j.techsoc.2015.04.002 0160-791X/© 2015 Elsevier Ltd. All rights reserved.

general purpose [14]. The main features of these technologies are their fast path of technological improvement, their pervasiveness across the full spectrum of the economy, and their role as innovation-enablers. ICT allow closer links between firms, their customers, suppliers, and collaborative partners. They also lower geographical barriers. In addition, they help the creation of new knowledge and its faster diffusion through more efficient process of information transformation, both within and between firms and sectors [47,72,74]. Many contemporary theories of economic growth acknowledge the significance of ICT. The most prominent are Neo-Schumpeterian theories, building on Kondratiev's perception of long waves of economic ‘boom’ and ‘bust’ and Schumpeter's work from 1910 on the role of innovative entrepreneurs exploiting market imperfections to lay the foundations for future growth [105,112]. These theories relate to specific types of technologies to specific epochs of economic development. The ICT are characterised as

136

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

‘pervasive’ only if their applications affect almost all sectors of the economy [8]. The empirical literature provides significant support for a positive relationship between ICT infrastructure and economic growth (see for instance, [16,17,31,38,63,70,71,76,77,80,87,90,108,117]). However, these studies investigate solely the relationship between ICT infrastructure and economic growth without looking at the directions of causality. A main objective of this paper is to overcome this deficiency to examine more broadly the causes and consequences of the availability of ICT infrastructure. In particular, the paper investigates the short-run and long-run relationship between ICT infrastructure and economic growth for Asian countries, using a panel data approach. We also introduce a third variable, namely financial development, in the analysis. Financial development is defined in terms of the aggregate size of the financial sector, its sectoral composition, and the range of attributes of individual sectors that determine their effectiveness in meeting users' requirements. The evaluation of financial development should cover the roles of the key institutional players, including the central bank, commercial and merchant banks, saving institutions, development financial institutions, insurance companies, mortgage entities, pension funds, the stock market, and other financial market institutions [55,126]. Evidently, financial development includes both banking sector development and stock market development. Thus, two additional key hypotheses of this paper are that financial development is linked to both ICT infrastructure and economic growth. Our working hypothesis is that ICT infrastructure has contributed significantly to both financial development and economic growth. Our alternative hypothesis for testing is that the expansion of ICT infrastructure is simply a consequence of financial development and economic growth. We also examine the possible direct causal link between financial development and economic growth as a corollary. This paper is organized as follows. A review of the literature is given in Section 2. The following section outlines the contributions of the study. Data and the empirical model are explained in Section 4. The econometric analysis and empirical results are discussed in Section 5. The final section offers conclusions flowing from our analysis. 2. A brief overview of literature 2.1. ICT infrastructure and economic growth The first strand of literature considers the possible link between ICT infrastructure and economic growth. Theoretically, there is a direct association between ICT infrastructure and economic growth (see Fig. 1), treating ICT as an exogenous force emerging from innovation and government or corporate decisions, manifesting in multiplier effects on real economic activity.1 Significant research

1 For a recent study adopting a narrower definition of ICT, focussing on the link between broadband internet and economic growth, see Ref. [7].

supports this association [31,40,56,101,113]. All these studies (e.g., [21,48,110,124,130]) found a positive correlation between ICT infrastructure and economic growth. However, the causality between ICT infrastructure and economic growth has spawned considerable interest among economists since the seminal work of [35]; which supported a bidirectional causal relationship in the US. Subsequently, there have been many similar studies for both developed and developing countries. While most of these studies have confirmed the existence of a causal relationship running from ICT infrastructure to economic growth [21,27,40,101,102,131], there are a few cases where there is no evidence of causality from ICT infrastructure to economic growth [40,116,119]. Hence, the empirical studies on the relationship between ICT infrastructure and economic growth do not provide any definite conclusion and currently there is no consensus among economists about the nature of this relationship. In summary, there are four possible relationships have been emphasized in the empirical literature on the causal link between ICT infrastructure and economic growth: unidirectional ICT-led growth hypothesis, unidirectional growth-led ICT hypothesis, the feedback hypotheses, and the neutrality hypothesis. In this context [2,21,29,40,85,108,115,124], find results in support of an ICT-led growth hypothesis (i.e., a supplyleading hypothesis e SLH). By contrast, [10,102,119,132] support the validity of growth-led ICT hypothesis (i.e., a demand-following hypothesis e DFH). Moreover [22,23,33,34,73,103,107,116,124,125], find existence of a mutual causal relationship between the two variables (i.e., they support a feedback hypothesis e FBH). Finally, some studies find a complete lack of a causal relationship between the variables - or a null hypothesis (NLH). The absence of causality is supported by very few papers [40,116,119]. Table 1 presents a summary of these previous studies. 2.2. Financial development and economic growth The second strand of the literature considers the possible link between financial development and economic growth. In this context [26,37,42,44,50,58,69], establish the validity of financial development-led growth hypothesis (i.e., a DFH). At the same time [1,12,20,53,64,79,92,103], prove the validity of growth-led financial development hypothesis (i.e. a SLH). On the other hand [5,9,11,18,66,68,82,91,93,100,121], demonstrate the validity of a feedback hypothesis (that is, bidirectional causality). A summary of these previous studies is shown in Table 2. 3. Relevance and contribution of the study Since the end of the 1980s, most Asian countries have sought by official policy determinations to foster financial development, for example, by reducing governmental intervention in national financial sectors or by privatizing banks. Such policies have been designed to promote economic growth through, inter alia, a higher mobilization of savings or a rise in domestic and foreign investment [133].

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

137

Fig. 1. Information and communications technologies (ICT) infrastructure and Economic Growth: Possible Sub-Links. Source: [40].

However, the effectiveness of such policies requires the existence of a close causal relationship between financial and real sectors. Hence, this study assesses the contribution of financial sector development to economic growth by studying fresh evidence. We ask whether financial sector development has actually swayed economic growth in a sample of selected Asian countries and, whether the policy focus on financial development is appropriate for fostering economic growth and developing the ICT infrastructure. Since ICT infrastructure does not necessarily ensure economic growth, this study also examines the impact of financial development on ICT infrastructure. The study contributes to the literature of financial development by incorporating the issue of regional disparity in financial as well as ICT development indicators. To our knowledge, this is the first study to integrate the causal nexus between financial development, economic growth, and ICT infrastructure. The novel features of our study are: firstly, we meld two stands of the literature: one which examines the nexus between ICT infrastructure and economic growth; and one that links financial development to economic growth. In other words, we consider all three variables simultaneously and investigates the possible link between any two variables in the presence of the third variable. Secondly, we use a large sample of Asian countries over a recent span of time. Thirdly, we employ sophisticated econometrics e and certain empirical approaches heretofore not adopted in the literature e to answer questions concerning the nature of the causal relationship between the three variables both in the short and long run.

4. Data and empirical model Annual data spanning from 2001 to 2012 for 21 Asian countries were obtained from World Development Indicators of World Bank. Appendix 1 lists the countries included in the analysis.2 The multivariate framework encompasses per capita economic growth (PGDP), represented by percentage change in real per capita real gross domestic product; a composite index of information communications technologies infrastructure (ICTI), derived from five variables: telephone landlines, mobile phones, internet users, internet servers, and fixed broadband3; and a composite index of financial development (FIND), derived from seven variables: broad money supply, claims on private sectors,

2 The selection of 21 Asian countries is based on data availability in relation to all the variables we required for our econometric analysis and to follow a balanced data structure. Moreover, over the last two decades, Asian policymakers have devoted considerable effort to developing their financial markets and ICT sectors in order to improve their economic growth. Some of these efforts constitute improvements in market infrastructure, investment expansion, financial regulations, inter alia. Therefore, the study chose these regions to investigate formally whether the development of financial sectors and ICT sectors can be show rigorously to be causal factors of economic growth in the region and whether they also cause each other. Furthermore, countries in this Asian region have experienced considerable political and social turmoil over the past few years. In our judgement, these selected 21 countries offer both the most common and country-specific features that make them a suitable and sufficient panel for econometric analysis of these questions. 3 Our period of analysis (2001e2012) was dictated by data availability for ICT infrastructure, particularly for fixed-broadband operations.

138

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

Table 1 Summary of studies on the causal nexus between ICT infrastructure and economic growth. Study Group [40] [21] [124] [29] [115] [85] Group [10] [119] [41] [115] [102] Group [35] [120] [116] [125] [22] [73] [23] [103] Group [119] [115] [107]

Method

Study area

Inference

A 2 2 2 2 1 1

15DECs & 15 INCs 12 ASCs Korea Poland China 23 ASCs

SLH SLH SLH SLH SLH SLH

2&3 2 2 1 1

USA 10 LACs EECs China 34 OECD countries

DFH DFH DFH DFH DFH

2 2 1 2 2 1 2 1

USA USA 105 countries 23 countries DCs 105 countries 93 countries 34 OECD countries

FBH FBH FBH FBH FBH FBH FBH FBH

2 1 2

10 LACs China Malaysia

NEH NEH NEH

B

C

D

Note 1: Supply-leading hypothesis (SLH): if unidirectional causality is present from ICT infrastructure to economic growth; Demand-following hypothesis (DFH): if unidirectional causality form economic growth to ICT infrastructure is present; Feedback hypothesis (FBH): if bidirectional causality between ICT infrastructure and economic growth is present; Neutrality hypothesis (NLH): if no causality between ICT infrastructure and economic growth is present. Note 2: DECs: Developing Countries; ASCs: Asian Countries; INCs: Industrialized Countries; CEE: Central and Eastern Europe; EECs: East European Countries; USA: United States of America; OECD: Organization of Economic Cooperation and Development; LACs: Latin American Countries. Note 3: 1: Dynamic Panel Data Model; 2: Granger Causality Test; 3: Modified Sims Test.

domestic credit provided to the private sector, domestic credit provided by the banking sector, market capitalization, traded stocks, and turnover ratio. All the variables are defined in Tables 3 and 4. We use principal component analysis to derive our two indexes. This approach is explained under Appendix 2. Fig. 2 summarizes the possible relationships between the three variables (PGDP, ICTI and FIND). All these three variables are converted into their natural logarithms for use in the econometric analysis. Tables 5 and 6 present the descriptive statistics and the correlations matrix for these variables. A panel data analysis was most appropriate, given the more than-decade-long span of the data.4 We use the

Table 2 Summary of studies on the causal nexus between financial development and economic growth. Study Group [18] [46] [13] [89] [5] [1] [59] [123] [66] [12] [20] [104] [53] [86] Group [79] [4] [91] [97] [93] [66] [104] [86] Group [3] [32] [39] [121] [25] Group [88] [104]

Many panel data analyses, including those that perform Granger causality tests, use data that span at least 15e20 years. However, as mentioned in footnote 1, data on fixed-broadband operations for Asian countries is available only from 2001 to 2012. This means that our study could cover only 12 years. There are, however, other studies in economics with similar or even shorter spans of time (see, for example, [6,29,97]).

Study area

Inference

4 1 1 4 4 2 2 4 4 4 4 1 1 2

109 countries 9 ASCs Tunisia MENA region Malaysia Egypt China European Union 15 MENA countries Bolivia South Korea, Hong Kong, UK 15 ASCs 10 ASCs 21 AFCs

SLH SLH SLH SLH SLH SLH SLH SLH SLH SLH SLH SLH SLH SLH

4 4 2 4 4 4 1 2

China Malaysia Kenya 5 countries South Africa 15 MENA countries 15 ASCs 21 AFCs

DFH DFH DFH DFH DFH DFH DFH DFH

4 4 2 3 2

India, Pakistan, Sri Lanka Barbados Greece Kenya 69 countries

FBH FBH FBH FBH FBH

3 1

7 ASCs 15 ASCs

NEH DFH

B

C

D

Note 1: Supply-leading hypothesis (SLH): if unidirectional causality is present from financial development to economic growth; Demandfollowing hypothesis (DFH): if unidirectional causality form economic growth to financial development is present; Feedback hypothesis (FBH): if bidirectional causality between financial development and economic growth is present; Neutrality hypothesis (NLH): if no causality between financial development and economic growth is present. Note 2: ASCs: Asian countries; AFCs: African countries; MENA: Middle East and North Africa. Note 3: 1: Bivariate Granger Causality; 2: Trivariate Granger Causality; 3: Quadvariate Granger Causality; 4: Multivariate Granger Causality.

Table 3 Definition of variables used in deriving our composite index of information and communications technologies infrastructure. Variable Definition ICTI

TELL MOBP

4

Method A

INTU INTS FIXB

Composite index of information and communications technologies (ICT) infrastructure: This is derived using five indicators: telephone landlines, mobile phones, internet users, internet servers, and fixed broadband e derived through principal component analysis. The five indicators used to derive this index are defined more precisely below. Telephone landlines: Telephone landlines per thousand of population. Mobile phones: Mobile phone subscribers per thousand of population. Internet users: Internet users per thousand of population. Internet servers: Internet servers per thousand of population. Fixed broadband: Fixed broadband per thousand of population.

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149 Table 4 Definition of economic growth and variables used for deriving our composite index of financial development.

139

εit refers to independently and normally distributed random variables for all i and t with zero means and finite heterogeneous variances ðsi 2 Þ.

Variable Definition PGDP

FIND

BRMS

CLPS

DCPS

DCBS

MACA

TRAS

TURR

Per capita economic growth: Percentage change in real per capita gross domestic product, used as our indicator of economic growth Composite index of financial development: This is derived using seven financial development indicators: broad money supply, claims on private sector, domestic credit to private sector, domestic credit provided banking sector, market capitalization, traded stocks, and turnover ratio – derived through principal component analysis. These variables are defined below. Broad money supply: Broad money supply (expressed as a percentage of gross domestic product) is the sum of currency outside banks; demand and term deposits, including foreign currency deposits of resident sectors (other than the central bank); certificates of deposit and commercial paper. Claims on private sectors: Credit (expressed as a percentage of gross domestic product) refers gross credit from the financial system to private sector. It isolates credit issued to the private sector, as opposed to credit issued to government, government agencies, and public enterprises. Domestic credit to the private sector: This credit (expressed as a percentage of gross domestic product) refers to financial resources provided to the private sector, such as through loans, purchases of non-equity securities, and trade credits and other accounts receivable, that establish a claim for payment. Domestic credit provided by the banking sector: Domestic credit provided by the banking sector (expressed as a percentage of gross domestic product) includes all credit to various sectors on a gross basis, with the exception of credit to the central government, which is net. The banking sector includes monetary authorities, deposit money banks, and other banking institutions such as mortgage and building loan associations. Market capitalization: Percentage change in the market capitalization of the listed companies, used as a proxy for the evolution in the size of the stock market. Traded stocks: Percentage change in the total value of traded stocks, used as a proxy for the evolution in stock market liquidity. Turnover ratio: Percentage change in the turnover ratio in the stock market, used as a proxy for the evolution in stock market turnover.

Note: All monetary measures are in real US dollars.

following model to describe the long-run relationship between PGDP, ICTI, and FIND.5

PGDPit ¼ ait þ b1i ICTIit þ b2i FINDit þ εit

(1)

where, i ¼ 1, 2,…, 21 represents each country in the panel; t ¼ 2001, 2002,…, 2012 refers to the year; b1, b2 and b3 represent the parameters, relating the longrun elasticity estimates of PGDP with respect to ICTI and FIND respectively; and

5 Of course, other variations of this equation are also entertained so that PGDP is not always the dependent variable.

The task is to estimate the parameters in equation (1) and conduct some panel tests on the causal nexus between the three variables. It is postulated that b1 > 0 as an increase in ICT infrastructure will likely cause an increase in economic growth. Similarly, we expect b2 > 0, representing an increase in financial development leads to increase in economic growth.

5. Econometric analysis and empirical results The objective of this study is to determine whether there are short-run and long-run dynamic relationships between ICT infrastructure, financial development, and economic growth. The testing procedure consists of two steps: panel cointegration test and panel Granger causality test. In the panel data analysis, particularly for cointegration and Granger causality, an essential first step is to identify the stationarity properties of the variables. While there are a number of panel unit root tests available, in this study we use three panel unit root tests (Levine-Lin-Chu: LLC, and Fisher-types: ADF and PP tests) as proposed by Refs. [24,78,83] respectively. These tests are widely used in econometrics so we do not repeat their detailed descriptions here. The null hypothesis relevant to the above three unit root tests is that there exist unit roots in the series, i.e. the variables are non-stationary. Table 7 shows the results of the panel unit root tests for each variable. It can be seen that the level values of all series (PGDP, ICTI and FIND) are non-stationary. However, the first differences of all variables are stationary at the 5% significance level, i.e., all variables are integrated of order one [I (1)] since they achieve stationarity after being differenced once. This is true for all the individual sub-regions (i.e., East Asia, South Asia, South-East Asia and West Asia), as well as for Total Asia as a group. It thus follows that all the three variables are integrated of order one, which certainly meets the requirements of the cointegration test. The next step is to test whether there is a long-run relationship between these three variables. While there are a number of tests available for use, we choose that of [98]. To test the cointegration relationship among variables for our panels of countries, we deploy the following equation:

PGDPit ¼ ait þ di t þ b1i ICTIit þ b2i FINDit þ εit

(2)

where ait and di are the fixed effects for each country and deterministic trends respectively. The null hypothesis of no cointegration is examined, based on seven different test statistics [99], which includes four individual panel statistics [panel v-statistic, panel r-statistic, panel t-statistic (non-parametric) and panel t-statistic (parametric)] and three group statistics [group r-statistic, group t-statistic (non-parametric) and group t-statistic (parametric)]. A

140

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

Fig. 2. Proposed model and hypotheses.

Table 5 Summary statistics on the variables. Variable

Mea

Case 1: East Asia (EA) PGDP 1.46 ICTI 0.84 FIND 1.10 Case 2: South Asia (SA) PGDP 1.47 ICTI 0.78 FIND 1.10 Case 3: South-East Asia (SEA) PGDP 1.45 ICTI 0.78 FIND 1.07 Case 4: West Asia (WA) PGDP 1.38 ICTI 0.82 FIND 1.08 Case 5: Total Asia (TA) PGDP 1.44 ICTI 0.78 FIND 1.06

Med

Max

Min

Std

Ske

Kur

JB

Pr

1.46 0.86 1.10

1.59 0.93 1.15

1.29 0.72 1.06

0.06 0.06 0.02

0.24 0.06 0.94

3.00 2.34 3.32

0.43 3.15 6.64

0.81 0.21 0.04

1.48 0.76 1.08

1.53 0.95 1.15

1.39 0.71 1.06

0.04 0.05 0.03

0.43 1.28 0.54

2.58 4.58 1.92

1.67 16.3 4.16

0.43 0.00 0.12

1.47 0.77 1.09

1.58 0.93 1.12

1.33 0.71 1.04

0.05 0.06 0.03

1.01 1.06 0.27

4.37 3.28 1.37

15.8 12.2 7.88

0.00 0.00 0.02

1.40 0.81 1.08

1.59 0.93 1.11

1.07 0.73 1.03

0.11 0.05 0.02

1.17 0.30 0.27

3.99 2.64 2.28

12.7 0.95 1.59

0.00 0.62 0.45

1.46 0.76 1.04

1.59 0.96 1.16

1.07 0.70 1.01

0.08 0.06 0.03

2.06 0.74 1.18

9.51 2.78 4.46

48.4 18.2 62.8

0.00 0.00 0.00

Note 1: Mea: Mean; Med: Median; Max: Maximum; Min: Minimum; Std: Standard Deviation; Ske: Skewness; Kur: Kurtosis; JB: Jarque Bera Statistics; Pr: Probability. Note 2: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth. Note 3: Values reported here are the natural logs of the variables. We use natural log forms in our estimation. Note that since per capita economic growth had the potential to be negative for some countries in some years, we first normalized the data by adding a common factor to all data points and only then applied natural logarithm transformations.

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

statistics that are significant at 1% level. Thus, the null hypothesis of no cointegration can be rejected. This confirms that there is a long-run relationship among the three variables. This is again true for all the individual regions we examine as well as for Asia as-a-whole. On the basis of unit root results and the presence of cointegration among the variables, we investigate how these variables are tied in the long-run and whether there are short-run fluctuations in their relationships. The following error-correction models (ECMs) are used to uncover both short-run fluctuations and the nature of longrun equilibrium relationships among IsCT infrastructure, financial development, and economic growth.

Table 6 Correlation matrices. Variables Case 1: PGDP ICTI FIND Case 2: PGDP ICTI FIND Case 3: PGDP ICTI FIND Case 4: PGDP ICTI FIND Case 5: PGDP ICTI FIND

PGDP

ICTI

FIND

East Asia (EA) 1.00 0.61* 0.42* South Asia (SA) 1.00 0.15 0.44 South-East Asia (SEA) 1.00 0.14 0.12 West Asia (WA) 1.00 0.01 0.16 Total Asia (TA) 1.00 0.23 0.03

1.00 0.61

1.00

1.00 0.19

1.00

1.00 0.65*

1.00

DPGDPit ¼ q1j þ

1.00 0.25

p X

b1ik DPGDPitk þ

k¼1

1.00

þ 1.00 0.70*

141

r X

q X

l1ik DICTIitk

k¼1

m1ik DFINDitk þ h1i ECTit1 þ ε1it

(3)

k¼1

1.00

H0 : l1ik ¼ 0; m1ik ¼ 0; h1i ¼ 0 for k ¼ 1; 2; …; p=q=r for at least one k HA : l1ik s0; m1ik s0; h1i s0

Note 1: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth. Note 2: * indicates statistical level of significance at a 1% level.

DICTIit ¼ q2j þ

detailed description of this test, not given here due to space constraints, is available upon request. The null hypothesis of the above seven cointegration tests is that there exists cointegration in these three variables, i.e. they have an affirmed long-run equilibrium relationship. The results of panel cointegration test are shown in Table 8. It can be seen that, of seven statistics, we found two

þ

r P

p X

b2ik DICTIitk þ

k¼1

q X

l2ik DPGDPitk (4)

k¼1

m2ik DFINDitk þ h2i ECTit1 þ ε2it

k¼1

H0 : l2ik ¼ 0; m2ik ¼ 0; h2i ¼ 0 for k ¼ 1; 2; …; p=q=r for at least one k HA : l2ik s0; m2ik s0; h2i s0

Table 7 Results from panel unit root tests. Test statistics

PGDP LD

Case 1: East Asia (EA) LLC 0.45 ADF 4.24 PP 3.77 Inferences Case 2: South Asia (SA) LLC 0.08 ADF 6.63 PP 7.64 Inferences Case 3: South-East Asia (SEA) LLC 0.14 ADF 6.68 PP 6.67 Inferences Case 4: West Asia (WA) LLC 0.53 ADF 7.33 PP 2.73 Inferences Case 5: Total Asia (TA) LLC 0.19 ADF 25.0 PP 20.0 Inferences

ICTI

FIND

FD

LD

FD

6.86* 41.2* 62.2* I[1]

8.99 0.13 0.01

1.61** 14.7** 27.2* I[1]

5.20* 28.4* 52.9* I[1]

5.34 0.43 0.01

3.51* 15.7** 24.6** I[1]

8.82* 63.4* 95.6* I[1]

6.52 0.14 0.02

4.47* 25.6* 48.7* I[1] 12.5* 153* 248* I[1]

LD

FD

1.26 1.33 0.99

3.74* 23.4* 45.7* I[1]

0.06 9.24 19.2

4.20* 29.3* 26.7** I[1]

1.81** 18.6** 19.8** I[1]

1.01 3.39 2.75

5.15* 40.3* 83.3* I[1]

3.29 0.85 0.01

1.59** 18.8** 25.9* I[1]

0.87 8.04 7.06

2.81* 27.8* 58.9* I[1]

7.43 3.08 0.04

2.95* 44.1* 65.7* I[1]

2.07 14.6 21.1

6.89* 102* 225* I[1]

Note 1: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth. Note 2: LLC: Levine-Lin-Chu statistics; ADF: Augmented Dickey Fuller statistics; PP: Phillips and Perron statistics; LD: Level Data; FD: First differenced data; * indicates significance at a 1% level; and ** indicates significance at a 5% level.

142

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

Table 8 Results of pedroni panel cointegration test. Test statistics Case 1: East Asia (EA) Panel v-statistics Panel r-statistics Panel PP-statistics Panel ADF-statistics Group r-statistics Group PP-statistics Group ADF-statistics Case 2: South Asia (SA) Panel v-statistics Panel r-statistics Panel PP-statistics Panel ADF-statistics Group r-statistics Group PP-statistics Group ADF-statistics Case 3: South-East Asia (SEA) Panel v-statistics Panel r-statistics Panel PP-statistics Panel ADF-statistics Group r-statistics Group PP-statistics Group ADF-statistics Case 4: West Asia (WA) Panel v-statistics Panel r-statistics Panel PP-statistics Panel ADF-statistics Group r-statistics Group PP-statistics Group ADF-statistics Case 5: Total Asia (TA) Panel v-statistics Panel r-statistics Panel PP-statistics Panel ADF-statistics Group r-statistics Group PP-statistics Group ADF-statistics

No intercept & trend

With intercept only

With both intercept & trend

0.29 [0.39] 0.32 [0.37] 1.67 [0.05] 1.06 [0.14] 0.69 [0.75] 2.50 [0.00] 1.38 [0.08]

0.20 [0.42] 0.39 [0.35] 4.15 [0.00] 2.84 [0.00] 0.98 [0.89] 4.23 [0.00] 2.97 [0.00]

1.12 [0.87] 0.85 [0.80] 4.43 [0.00] 1.84 [0.03] 2.13 [0.98] 4.57 [0.00] 3.34 [0.00]

1.33 [0.09] 0.70 [0.24] 3.30 [0.00] 1.97 [0.02] 0.73 [0.77] 3.22 [0.00] 4.05 [0.00]

0.56 [0.29] 0.06 [0.52] 4.70 [0.00] 1.24 [0.10] 1.33 [0.91] 5.10 [0.00] 1.51 [0.06]

0.02 [0.51] 1.10 [0.06] 3.79 [0.00] 0.24 [0.41] 1.65 [0.95] 3.70 [0.00] 0.35 [0.64]

2.07 [0.81] 1.58 [0.06] 4.08 [0.00] 2.42 [0.00] 0.79 [0.21] 5.77 [0.00] 1.76 [0.04]

1.06 [0.14] 0.71 [0.24] 5.35 [0.00] 3.13 [0.00] 0.48 [0.68] 5.88 [0.00] 2.63 [0.00]

0.14 [0.44] 1.12 [0.87] 6.72 [0.00] 2.99 [0.00] 1.86 [0.97] 10.3 [0.00] 3.34 [0.00]

0.14 [0.44] 0.15 [0.44] 0.69 [0.24] 0.85 [0.19] 0.69 [0.76] 2.98 [0.00] 1.94 [0.02]

0.58 [0.28] 0.05 [0.52] 1.92 [0.03] 0.73 [0.23] 1.39 [0.91] 4.66 [0.00] 1.82 [0.04]

0.59 [0.72] 0.89 [0.81] 3.05 [0.00] 0.70 [0.76] 1.85 [0.97] 4.27 [0.00] 0.03 [0.49]

0.78 [0.22] 1.27 [0.10] 3.44 [0.00] 2.42 [0.00] 0.51 [0.69] 7.72 [0.00] 3.77 [0.00]

0.28 [0.39] 0.37 [0.35] 5.25 [0.00] 2.22 [0.01] 2.06 [0.98] 8.64 [0.00] 3.27 [0.00]

2.25 [0.98] 1.81 [0.96] 6.44 [0.00] 0.86 [0.19] 3.45 [0.99] 13.2 [0.00] 3.88 [0.00]

Note 1: Variables and regions shown above are defined in the text. Note 2: Figures in square brackets are probability levels indicating significance.

DFINDit ¼ q3j þ

p X

b3ik DFINDitk þ

k¼1

þ

r X

q X

l3ik DICTIitk

k¼1

m3ik DPGDPitk þ h3i ECTit1 þ ε3it

(5)

k¼1

H0 : l3ik ¼ 0; m3ik ¼ 0; h3i ¼ 0 HA : l3ik s0; m3ik s0; h3i s0

for k ¼ 1; 2; …; p=q=r for at least one k

In equations (3)e(5), p, q, and r are lag length for the differenced variables of the respective equations and can be determined by the Engle-Granger approach [43]. The lagged error-correction terms (ECTs) are derived from the cointegrating equations. The lagged ECTs represent the long-run dynamics, akin to an equilibrium process, while differenced variables represent the short-run adjustment dynamics between the variables. For short-run causal relationships, if the null hypothesis l1ik ¼ 0 (or l2ik ¼ 0) is rejected. There is Granger causality

running from ICT infrastructure to economic growth (or economic growth to ICT infrastructure). If the joint null hypothesis m1ik ¼ 0 (or m3ik ¼ 0) is rejected, there is Granger causality from financial development to economic growth (or economic growth to financial development). If the joint null hypothesis m2ik ¼ 0 (or l3ik ¼ 0) is rejected, there is Granger causality from financial development to ICT infrastructure (or ICT infrastructure to finance development). For long-run causal relationships, the null hypothesis (h1i ¼ 0, h2i ¼ 0, and h3i ¼ 0) needs to be rejected. Evidently, existence of long-run Granger causality hinges on the statistical significance of estimated coefficient of the lagged error-correction term. Table 9 summarizes the hypotheses we test in studying the relationships between ICT infrastructure, financial development, and economic growth. The results of panel Granger causality tests, for all individual regions and for Total Asia, are shown in Table 10 and are presented below. For East Asia: It can be seen that ICT infrastructure Granger-causes economic growth (ICTI 0 PGDP) in the

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149 Table 9 Hypotheses Tested in this Study. Causal flows

Short-run restrictions

Set 1: ICT-growth nexus ICTI 0 PGDP l1ik s 0 l2ik s 0 PGDP 0 ICTI Set 2: Finance-growth nexus FIND 0 PGDP m1ik s 0 m3ik s 0 PGDP 0 FIND Set 3: ICT-finance nexus ICTI 0 FIND m2ik s 0 FIND 0 ICTI l3ik s 0

143

Table 11 The summary of short-run granger causality. Long-run restrictions

Region

Causal relationships tested in the Models

East Asia (EA) South Asia (SA) South-East Asia (SEA) West Asia (WA) Total Asia (TA)

ICTI ICTI ICTI ICTI ICTI

ICTI vs. PGDP

h1i s 0 h2i s 0 h1i s 0 h3i s 0 h2i s 0 h3i s 0

Note 1: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth. Note 2: The definitions of all these variables are presented in Tables 3 and 4 respectively.

short run based on Wald-statistics. This is also true in the long run, as the lagged error-correction term is significant at the 1% level (see the estimated coefficient of ECT-1 in the first row). On the other hand, economic growth Grangercauses ICT infrastructure and financial development (PGDP 0 ICTI; PGDP 0 FIND) in the short run based on Wald-statistics. However, these relationships are not supported in the long run. Moreover, we find financial development causes ICT infrastructure (FIND 0 ICTI) in the short run, although this finding is not supported in the long run. For South Asia: Both ICT infrastructure and financial development Granger-cause economic growth (ICTI 0 PGDP; FIND 0 PGDP) in the short run based on Wald-statistics. This is also true in the long run, as the lagged error-correction term is significant at the 10% level (see the estimated ECT-1 coefficient in the fourth row). In addition, both economic growth and financial development

⇔ PGDP ⇔ PGDP * PGDP 0 PGDP ⇔ PGDP

FIND vs. PGDP FIND FIND FIND FIND FIND

⇔ PGDP 0 PGDP ⇔ PGDP * PGDP ⇔ PGDP

FIND vs. ICTI FIND FIND FIND FIND FIND

0 ICTI 0 ICTI * ICTI ⇔ ICTI ⇔ ICTI

Note 1: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth; and ECT-1: Lagged error-correction term. Note 2: ⇔: Existence of bidirectional Granger causality; 0 or *: Existence of unidirectional Granger causality.

Granger-causes ICT infrastructure (PGDP 0 ICTI; FIND 0 ICTI) in the short run based on Wald-statistics, while these links do not hold in the long run (as shown by the insignificance of the ECT-1 coefficient in the fifth row). For South-East Asia: In this case financial development causes economic growth (FIND 0 PGDP) both in the short run and the long run. However, other causal connections hold only in the short run in that: economic growth Granger-causes both ICT infrastructure (PGDP 0 ICTI) and financial development (PGDP 0 FIND) while ICT infrastructure Granger-causes financial development (ICTI 0 FIND). For West Asia: Here, ICT infrastructure Granger-causes both economic growth (ICTI 0 PGDP) and financial development (ICTI 0 FIND) in the short run based on Wald-statistics. The latter finding is also supported in the long run, as the lagged error-correction term is statistically significant at the 10% level in the twelfth row. In addition, financial development Granger causes ICT infrastructure

Table 10 Panel granger causality test results. Dependent variable

Case 1: DPGDP DICTI DFIND Case 2: DPGDP DICTI DFIND Case 3: DPGDP DICTI DFIND Case 4: DPGDP DICTI DFIND Case 5: DPGDP DICTI DFIND

Independent variables (possible sources of causation)

Lagged ECT term (for possible long-run causality)

DPGDP

DICTI

DFIND

ECT-1

e 4.59* 5.33*

3.81* e 1.39

0.60 2.44** e

2.00** 2.01 1.99

3.82** e 1.45

3.83** 2.86** e

2.19** 1.36 0.11

1.63 e 6.91*

6.68* 0.03 e

4.23* 0.63 0.66

3.74* e 4.23*

1.00 5.10* e

0.93 0.15 1.93**

2.93** e 7.03*

6.63* 6.98* e

3.20* 1.37 2.77

East Asia (EA)

South Asia (SA) e 12.6* 0.87 South-East Asia (SEA) e 2.96** 3.10** West Asia (WA) e 0.41 5.47* Total Asia (TA) e 7.60* 7.07*

Note 1: ICTI: Composite index of ICT infrastructure; FIND: Composite index of financial development; and PGDP: Per capita economic growth; and ECT-1: Lagged error-correction term. Note 2: * indicates significance at a 1% level; ** indicates significance at a 5% level.

144

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

Fig. 3. Granger causal relations between the variables in East Asia, 2001e2012.

(FIND 0 ICTI) in the short run while the relationship is not supported in the long run. For Total Asia: Both ICT infrastructure and financial development Granger-cause economic growth (ICTI 0 PGDP; FIND 0 PGDP) in the short run based on the Wald-statistics. This finding also holds in the long run, as the lagged error-correction term is significant at the 1% level (see the estimated coefficient of ECT-1 in the thirteenth row). We also find that economic growth Grangercauses both ICT infrastructure and financial development

(PGDP 0 ICTI; PGDP 0 FIND) in the short run based on Wald-statistics, while the relationship is not supported in the long run (see the coefficients of ECT-1 in the fourteenth and fifteenth rows, respectively). Lastly, we find the presence of bidirectional causality between financial development and ICT infrastructure (FIND ⇔ ICTI) in the short run, although this finding does not hold in the long run. The summary of these short-run results are reported in Table 11. From these results, it is evident that for the sake of generating long term economic growth, both ICT

Fig. 4. Granger causal relations between the variables in South Asia, 2001e2012.

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

145

Fig. 5. Granger causal relations between the variables in South-East Asia, 2001e2012.

infrastructure and financial development ought to be encouraged broadly in Asia and in most of the regions we have identified in our study. The only exception to this policy proposal appears to be West Asia. Based on our findings, ICT infrastructure appears to matter most for East Asian countries' long-run economic growth, while financial development is more fundamental for South-East Asian countries. In the case of South Asia both ICT infrastructure and financial development matter for the attainment of long-run economic growth.

Finally, in order to complement our analysis, we employed generalized impulse response functions to trace the effect of a one-off shock to one of the innovations on the current and future values of the endogenous variables. The generalized impulse responses offer additional insight into how shocks to ICT infrastructure and economic growth can affect and be affected by financial development. These results are graphed in Figs. 3e7. This analysis provides additional support for the finding that there is demonstrated causality among the variables in the vector error-correction model.

Fig. 6. Granger causal relations between the variables in West Asia, 2001e2012.

146

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

Fig. 7. Granger causal relations between the variables in total Asia, 2001e2012.

6. Conclusion and policy implications Although the short-run causal links are interesting, more remarkable are the long-run causal relationships. In all the samples we consider (except for West Asia) when DPGDP serves as the dependent variable, the lagged error correction term is statistically significant. This implies that PGDP tends to converge to its long-run equilibrium path in response to its statistically significant regressors. We can generally conclude that in most of our samples, and certainly for Asia taken as-a-whole, both ICT infrastructure and financial development matter in the determination of long-run economic growth. Slight differences in long-run results are observed due to heterogeneity across different regions within Asia. For West Asia which includes rich Arab oil producers and Israel, neither ICT infrastructure nor financial development plays a statistically significant role in the long-run growth of a country. This is not surprising given that these countries rely on oil revenues (other exports for Israel) for their long term economic growth. By contrast, ICT infrastructure and/or financial development play a key role in the long term economic growth of the other regions in Asia and therefore ought to be encouraged by policy-makers. Finally, from our analysis, future research on the economic growth of Asian countries would remiss if it omitted the possible role of the two growth-determining variables which we have identified in this study, namely ICT infrastructure and financial development.

Appendix 1. Sample of Asian countries The selected 21 Asian countries are Bangladesh, India, China, Hong Kong, Indonesia, Iraq, Iran, Israel, Japan,

Kuwait, Malaysia, Philippines, Pakistan, Qatar, Saudi Arabia, Singapore, South Korea, Sri Lanka, Thailand, the United Arab Emirates, and Vietnam. These are further grouped into four regions: East Asia (China, Hong Kong, Japan, and South Korea), South Asia (Bangladesh, India, Iran, Pakistan, and Sri Lanka), South-East Asia (Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam), and West Asia (Iraq, Israel, Kuwait, Qatar, Saudi Arabia, and the United Arab Emirates). Appendix 2. Principal component analysis (PCA) Modelling various indicators of ICT infrastructure and financial development in the same equation would lead to multicollinearity. Thus, we bring the five indicators of ICT infrastructure together by employing PCA. Analogously, we use PCA to integrate the seven indicators of financial development. PCA transforms the data into new variables (i.e., the principal components) that are not correlated. The use of PCA to construct indexes similar to ours is well-documented in papers using panel data (see, for example, [30,54,86,104,109,122]).6 To be clear, PCA is a special case of the more general method of factor analysis. The approach entails several steps: construction of a data matrix, creation of standardized variables, calculation of a correlation matrix, determination of eigen values (to rank principal components) and eigenvectors, selection of PCs (based on stopping rules), and interpretation of results [51,52]. The intent behind PCA is to transform the original set of variables into a smaller set of linear combinations that account for most of the variance of the original set. The aim is to construct from a set of variables, Xj's (j ¼ 1, 2,…, n)

6

[51,52,61,84,114] provide procedural detail on the use of PCA.

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149

new variables (Pi) called ‘principal components’, which are linear combinations of the X's. Representing it mathematically,

(6)

derived by PCA. Thus, ICTI captures the five indicators we mentioned earlier and which are summarized under Table 3. As is evident, the index is calculated for each country and for each year. Analogously, the following equation is used to create FIND, our composite index for financial development using the seven indicators that are summarized under Table 4.

FIND ¼ Here, P ¼ [P1, P2,…, Pm] are principal components; A ¼ [aij] for i ¼ (1, 2,…, m); and j ¼ (1, 2,…, n) are component loadings; and X ¼ [X1, X2,…, Xn] are original variables. The component loadings are the weights showing the variance contribution of principal components to variables. Since the principal components are selected orthogonal to each other, aij weights are proportional to correlation coefficient between variables and principal components. The first principal component (P1) is determined as the linear combination of X1, X2,…, Xn provided that the variance contribution is at a maximum. The second principal component (P2), independent from the first principal component, is determined so as to provide a maximum contribution to total variance left after the variance explained by the first principal component. Analogously, the third and the other principal components are determined as to provide the maximum contribution to the remaining variance and independent from each other. The aim here is to determine aij coefficients providing the linear combinations of variables based on the specified conditions. Note that the method of principal components could be applied by using the original values of the Xj's, by their deviations from their means, or by the standardized variables. The present study, however, adopts the latter procedure, as it is assumed to be more general and can be applied to variables measured in different units. It is important to note that the values of the principal components will be different depending on the way in which the variables are used (original values, deviations, or standardized values). The coefficients a's, called loadings, are chosen in such a way that the constructed principal components satisfy two conditions: (a) principal components are uncorrelated (orthogonal), and (b) the first principal component P1 absorbs and accounts for the maximum possible proportion of total variation in the set of all X's. Furthermore, principal component absorbs the maximum of the remaining variation in the X's (after allowing for the variation accounted for by the first principal component) and so on. There are different rules to define a high magnitude known as stopping rules. Here, variance explained criteria are implemented based on the rule of keeping enough principal components to account for 90% of the variation [51]. The following equation is used to construct ICTI, our composite index for ICT infrastructure:

ICTI ¼

5 X i¼1

aij

Xij SdðXi Þ

(7)

where ICTI is our composite index, Sd is standard deviation, Xij is ith variable in jth year; and aij is factor loading as

147

7 X i¼1

aij

Xij SdðXi Þ

(8)

Finally, it should be noted that a weakness of PCA is that it is a non-parametric approach. On the one hand, there are no parameters to estimate or to adjust; the estimation method is data-driven, and is not dependent on the predilections of the users. PCA does not take into account any a-priori knowledge, as alternative parametric algorithms do. Most importantly, PCA deals with linear relationships among the variables. In addition, PCA has some limitations, as every estimation method does: interpretation problems, scaling problems, and higher-order correlation problems. References [1] Abu-Bader S, Abu-Qarn AS. Financial development and economic growth: the Egyptian experience. J Policy Model 2008;30(5): 887e98. [2] Ahmed EM, Krishnasamy G. Telecommunications investment and economic growth in ASEAN 5: an assessment from UECM. N Z Econ Pap 2012;46(3):315e32. [3] Ahmed SM, Ansari MI. Financial sector development and economic growth: the South-Asian experience. J Asian Econ 1998;9(3): 503e17. [4] Ang JB, McKibbin WJ. Financial liberalization, financial sector development and growth: evidence from Malaysia? J Dev Econ 2007;84(1):215e33. [5] Ang JB. What are the mechanisms linking financial development and economic growth in Malaysia? Econ Model 2008;25(1): 38e53. [6] Arvin M, Choudhry S. Untied aid and exports: do untied aid disbursements create goodwill for donor exports? Can J Dev Stud 1997;18(1):9e22. [7] Arvin M, Pradhan R. Broadband penetration and economic growth nexus: evidence from cross-country panel data. Appl Econ 2014; 46(35):4360e9. [8] Avgerou C. How can IT enable economic growth in developing countries? Inform Technol Dev 1998;8(1):15e28. [9] Bangake C, Eggoh JC. Further evidence on finance-growth causality: a panel data analysis. Econ Syst 2011;35(2):176e88. [10] Beil RO, Ford GS, Jackson JD. On the relationship between telecommunications investment and economic growth in the United States. Int Econ J 2005;19(1):3e9. [11] Blackburn K, Hung VTY. A theory of growth, financial development and trade. Economica 1998;65(257):107e24. [12] Bojanic AN. The impact of financial development and trade on the economic growth of Bolivia. J Appl Econ 2012;15(1):51e70. [13] Boulila G, Trabelsi M. Financial development and long run growth: evidence from Tunisia: 1962e1997. Savings Dev 2004;28(3): 289e314. [14] Bresnahan T, Trajtenberg M. General purpose technologies: engine of growth? J Econ 1995;65(1):83e108. [15] Broersma L, Van Ark B. ICT, business services and labour productivity growth. Econ Innovtion New Technol 2007;16(5e6):433e49. [16] Brynjolfsson E, Hitt LM. Computing productivity: firm-level evidence. Rev Econ Stat 2003;85(4):793e808. [17] Brynjolfsson E, Yang S. Information technology and productivity: a review of the literature. Adv Comput 1996;43(1):179e214. [18] Calderon C, Liu L. The direction of causality between financial development and economic growth. J Dev Econ 2003;72(1): 321e34. [19] Carayannis EG, Popescu D. Profiling a methodology for economic growth and convergence: learning from the EU e-procurement

148

[20]

[21]

[22]

[23]

[24] [25] [26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36] [37]

[38]

[39]

[40] [41]

[42]

[43] [44]

[45] [46]

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149 experience for central and eastern European countries. Technovation 2005;25(1):1e14. Chaiechi T. Financial development shocks and contemporaneous feedback effect on key macroeconomic indicators: a post keynesian time series analysis. Econ Model 2012;29(2):487e501. Chakraborty C, Nandi B. Privatization, telecommunications and growth in selected Asian countries: an econometric analysis. Commun Strategies 2003;52(3):31e47. Chakraborty C, Nandi B. Telecommunication adoption and economic growth in developing countries: do levels of development matter. J Acad Bus Econ 2009;9(2):51e61. Chakraborty C, Nandi B. Mainline telecommunications infrastructure, levels of development and economic growth: evidence from a panel of developing countries. Telecommun Policy 2011;35(1): 441e9. Choi I. Unit root tests for panel data. J Int Money Finance 2001; 20(2):249e72. Chow WW, Fung MK. Financial development and growth: a clustering and causality analysis. J Int Trade Econ Dev 2011;35(3):1e24. Christopoulos DK, Tsionas EG. Financial development and economic growth: evidence from panel unit root and cointegration tests. J Dev Econ 2004;73(4):55e74. Chu N, Oxley L, Carlaw K. ICT and causality in the New Zealand economy. In: Proceedings of the 2005 international conference on simulation and modelling; 2005. Ciarli T, Rabellotti R. ICT in industrial districts: an empirical analysis on adoption, use and impact. Ind Innovation 2007;14(3): 277e303. Cieslik A, Kaniewsk M. Telecommunications infrastructure and regional economic development: the case of Poland. Reg Stud 2004;38(6):713e25. Coban S, Topcu M. The nexus between financial development and energy consumption in the EU: a dynamic panel data analysis. Energy Econ 2013;39:81e8. Colecchia A, Paul S. ICT investment and economic growth in the 1990s: is the United States a unique case? A comparative study of nine OECD countries. Rev Econ Dyn 2002;5(2):408e42. Craigwell R, Downes D, Howard M. The finance-growth nexus: a multivariate VAR analysis of a small open economy. Savings Dev 2001;25(2):209e23. Cronin FJ, Colleran EK, Herbert PL, Lewitzky S. Telecommunications infrastructure investment and economic development. Telecommun Policy 1993;17(6):415e30. Cronin FJ, Colleran EK, Herbert PL, Lewitzky S. Telecommunications and growth: the contribution of telecommunications infrastructure investment to aggregate and sectoral productivity. Telecommun Policy 1993;17(9):677e90. Cronin FJ, Parker EB, Colleran EK, Gold MA. Telecommunication infrastructure and economic growth: an analysis of causality. Telecommun Policy 1991;15(6):529e35. Dabinett G. Reflections on regional development policies in the information society. Plan Theory Pract 2002;3(2):232e7. Demetriades P, Hussein K. Financial development and economic growth: cointegration and causality tests for 16 countries. J Dev Econ 1996;51(2):387e411. Dewan S, Kraemer KL. Information technology and productivity: preliminary evidence from country-level data. Manag Sci 2000; 46(4):548e62. Dritsakis N, Adamopoulos A. Financial development and economic growth in Greece: an empirical investigation with granger causality analysis. Int Econ J 2004;18(4):547e59. Dutta A. Telecommunications and economic activity: an analysis of granger causality. J Manag Inform Syst 2001;17(4):71e95. Dvornik D, Sabolic D. Telecommunication liberalization and economic development in european countries in transition. Technol Soc 2007;29(4):378e87. Eller M, Haiss P, Steiner K. Foreign direct investment in the financial sector and economic growth in central and eastern Europe: the crucial role of the efficiency channel. Emerg Mark Rev 2006;7(4):300e19. Engle RF, Granger CWJ. Cointegration and error correction: representation, estimation and testing. Econometrica 1987;55(2):251e76. Enisan AA, Olufisayo AO. Stock market development and economic growth: evidence from seven Sub-Saharan African countries. J Econ Bus 2009;61(2):162e71. Farhadi M, Ismail R, Fooladi M. Information and communication technology use and economic growth. PLoS One 2012;7(11):1e7. Fase MMG, Abma RCN. Financial environment and economic growth in selected Asian countries. J Asian Econ 2003;14(1):11e21.

[47] Garcia-Muniz AS, Vicente M. ICT technologies in Europe: a study of technological diffusion and economic growth under network theory. Telecommun Policy 2014;38(No. 4):360e70. [48] Gilling E. Telecommunications and economic development: intercountry comparisons of the catalytic effect of telephone services on development. Master’s Thesis. McGill University; 1975. [49] Grimes A, Ren C, Stevens P. The need for speed: impacts of internet connectivity on firm productivity. J Prod Anal 2012;37(2): 187e201. [50] Hassan KM, Sanchez B, Yu J. Financial development and economic growth: new evidence from panel data. Q Rev Econ Finance 2011; 51(1):88e104. [51] Hosseini HM, Kaneko S. Dynamics sustainability assessment of countries at the macro level: a principal component analysis. Ecol Indic 2011;11(3):811e23. [52] Hosseini HM, Kaneko S. Causality between pillars of sustainable development: global stylized facts or regional phenomena. Ecol Indic 2012;14(1):197e201. [53] Hsueh S, Hu Y, Tu C. Economic growth and financial development in Asian countries: a bootstrap panel granger causality analysis. Econ Model 2013;32(3):294e301. [54] Huang Y. Determinants of financial development. New York, NY: Palgrave MacMillan; 2010. [55] IMF. Indicators of financial structure, development, and soundness. 2005 [Chapter 2]. Online available at: http://www.imf.org/ external/pubs/ft/fsa/eng/pdf/ch02.pdf. [56] Ishida H. The effect of ICT development on economic growth and energy consumption in Japan. Telemat Inform 2015;32(1):79e88. [58] Jalil A, Ma Y. Financial development and economic growth; time series evidence from China and Pakistan. J Econ Coop 2008;29(2): 29e68. [59] Jalil A, Feridun M, Ma Y. Finance-growth nexus in china revisited: new evidence from principal components and ARDL bounds tests. Int Rev Econ Finance 2010;19(2):189e95. [61] Joliffe IT. Principal component analysis. 2nd ed. New York, NY: Springer; 2002. [63] Jorgenson DW, Ho MS, Sitroh KJ. Lessons from the US growth resurgence. J Policy Model 2003;25(5):453e70. [64] Jung WS. Financial development and economic growth: international evidence. Econ Dev Cult Change 1986;34(2):336e46. [66] Kar M, Nazlioglu S, Agir H. Financial development and economic growth nexus in the MENA countries: bootstrap panel granger causality analysis. Econ Model 2011;28(1e2):685e93. [68] Khan A. Financial development and economic growth. Macroecon Dyn 2001;5(3):413e33. [69] King RG, Levine R. Finance and growth: schumpeter might be right? Q J Econ 1993;108(3):717e37. [70] Kraemer KL, Dedrick J. Payoff from investment in information technology: lessons from the Asia-Pacific region. World Dev 1994; 22(12):1921e31. [71] Kraemer KL, Dedrick J. Information technology and productivity: results and implications of cross-country studies. In: Pohjola M, editor. Information technology and economic development. Oxford: Oxford University Press; 2001. p. 257e79. [72] Kretschmer T. Information and communication technologies and productivity growth: a survey of the literature. OECD Digital Economy Papers; 2012. http://dx.doi.org/10.1787/5k9bh3jllgs7-en. No. 195. [73] Lam PL, Shiu A. Economic growth, telecommunication development and productivity growth of the telecommunication sector: evidence around the world. Telecommun Policy 2010;34(4):185e99. [74] Lee JW, Brahmasrene T. ICT, CO2 emissions and economic growth: evidence from a panel of ASEAN. Glob Econ Rev Perspect East Asian Econ Ind 2014;43(2):93e109. [75] Lee SH, Levendis J, Gutierrez L. Telecommunications and economic growth: an empirical analysis of Sub-Saharan Africa. Appl Econ 2012;44(4):461e9. [76] Lee ST, Gholami R, Tong TY. Time series analysis in the assessment of ICT impact at the aggregate level-lessons and implications for the new economy. Inform Manag 2005;42(7):1009e22. [77] Lehr B, Lichtenberg FR. Information technology and its impact on productivity: firm-level evidence from government and private data sources. Can J Econ 1999;32(2):335e62. [78] Levine A, Lin CF, Chu CS. Unit root tests in panel data: asymptotic and finite sample properties. J Econ 2002;108(1):1e2. 4. [79] Liang Q, Teng J. Financial development and economic growth: evidence from China. China Econ Rev 2006;17(4):395e411. [80] Lichtenberg FR. The output contributions of computer equipment and personnel: a firm level analysis. Econ Innovation New Technol 1995;3(3):201e17.

R.P. Pradhan et al. / Technology in Society 42 (2015) 135e149 [82] Luintel K, Khan M. A quantitative reassessment of the financegrowth nexus: evidence from a multivariate VAR. J Dev Econ 1999;60(2):381e405. [83] Maddala GS, Wu S. A comparative study of unit roots with panel data a new simple test. Oxf Bull Econ Stat 1999;61(1):631e51. [84] Manly B. Multivariate statistical methods: a primer. London: Chapman & Hall; 1994. [85] Mehmood B, Siddiqui W. What causes what? panel cointegration approach on investment in telecommunications and economic growth: case of Asian countries. Romanian Econ J 2013;16(47):3e16. [86] Menyah K, Nazlioglu S, Wolde-Rufael Y. Financial development, trade openness and economic growth in African countries: new insights from a panel causality approach. Econ Model 2014;37(2): 2014. [87] Motohashi K. ICT diffusion and its economic impact in OECD countries. STI Rev 1997;20(1):13e45. [88] Mukhopadhyay B, Pradhan RP, Feridun M. Finance-growth nexus revisited for some Asian countries. Appl Econ Lett 2011;18(16): 1527e30. [89] Naceur SB, Ghazouani S. Stock markets, banks, and economic growth: empirical evidence from the MENA region. Res Int Bus Finance 2007;21(2):297e315. [90] Norton SW. Transaction costs, telecommunications, and the microeconomics of macroeconomic growth. Econ Dev Cult Change 1992;41(1):175e96. [91] Odhiambo NM. Financial development in Kenya: a dynamic test of the finance-led growth hypotheses. Econ Issues 2008;13(2):21e36. [92] Odhiambo NM. Finance-growth-poverty Nexus in South Africa: a dynamic causality linkage. J Socio-Economics 2009;38(2):320e5. [93] Odhiambo NM. Finance-investment-growth nexus in South Africa: an ARDL bounds testing procedure. Econ Change Restruct 2010; 43(3):205e19. [95] OECD. Information and communications technologies (ICT) and economic growth: evidence from OECD countries, industries, and firms. Paris: Organization of Economic Cooperation and Development (OECD); 2003. [96] Osei-Bryson KM, Ko M. Exploring the relationship between information technology investments and firm performance using regression splines analysis. Inform Manag 2004;42(1):1e13. [97] Panopoulou E. Financial variables and euro area growth: a nonparametric causality analysis. Econ Model 2009;26(6):1414e9. [98] Pedroni P. Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxf Bull Econ Stat 1999;61(S1): 653e70. [99] Pedroni P. Panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis: new results. Econ Theory 2004;20(3):597e627. [100] Pradhan RP. Financial development, growth and stock market development: the trilateral analysis in India. J Quant Econ 2011; 9(1):134e45. [101] Pradhan RP, Arvin MB, Norman NR, Bele SK. Economic growth and the development of telecommunications infrastructure in the G-20 countries: a panel-VAR approach. Telecommun Policy 2014; 38(7):634e49. [102] Pradhan RP, Bele S, Pandey S. Internet-growth nexus: evidence from cross country panel data. Appl Econ Lett 2013;20(16): 1511e5. [103] Pradhan RP, Bele S, Pandey S. The link between telecommunication infrastructure and economic growth in 34 OECD countries. Int J Technol Policy Manag 2013;13(3):278e93. [104] Pradhan RP, Mukhopadhyay B, Gunashekar A, Bele S, Pandey S. Financial development, social development, and economic growth: the causal nexus in Asia. Decision 2013;40(1e2):69e83. [105] Pyka A, Andersen ES. Introduction: long term economic development: demand, finance, organization, policy and innovation in a schumpeterian perspective. J Evol Econ 2012;22(4):621e5. [106] Rajabiun R, Middleton CA. Multilevel governance and broadband infrastructure development: evidence from Canada. Telecommun Policy 2013;37(9):702e14. [107] Ramlan J, Ahmed EM. Information and communication technology (ICT) and human capital management trend in malaysia's economic development. Appl Econ Lett 2009;16(18):1881e6. [108] Roller L, Waverman L. Telecommunications infrastructure and economic development: a simultaneous approach. Am Econ Rev 2001;91(4):909e23.

149

[109] Saci K, Holden K. Evidence on growth and financial development using principal components. Appl Financ Econ 2008;18(19): 1549e60. [110] Saunders RJ, Warford JJ, Wellenius B. Telecommunications and economic development. Washington DC: The John Hopkins University Press; 1983. [111] Schreyer P. The contribution of information and communication technology to output growth: a study of the G 7 countries. Working Paper, No. 2000/2. Geneva: Organization of Economic Cooperation and Development; 2000. [112] Schumpeter JA. The theory of economic development. Cambridge: Harvard University Press; 1934. [113] Shahiduzzaman M, Alam K. The long-run impact of information and communication technology on economic output: the case of Australia. Telecommun Policy 2014;38(7):623e33. [114] Sharma S. Applied multivariate techniques. New York, NY: Wiley; 1996. [115] Shiu A, Lam PL. Causal relationship between telecommunications and economic growth in China and its regions. Reg Stud 2008; 42(5):705e18. [116] Shiu A, Lam PL. Causal relationship between telecommunications and economic growth: a study of 105 countries. In: Paper presented at the 17th biennial conference of the international telecommunications society. Montreal, June 24e27; 2008. [117] Siegel D, Griliches Z. Purchased services, outsourcing, computers, and productivity in manufacturing. NBER Working Paper, No. 3678. Cambridge: National Bureau of Economic Research; 1991. [118] Swierczek FW, Shrestha PK. Information technology and productivity: a comparison of Japanese and Asia-Pacific banks. J High Technol Manag Res 2003;14(2):269e88. [119] Veeramacheneni B, Ekanayake EM, Vogel R. Information technology and economic growth: a causal analysis. Southwest Econ Rev 2007;34(1):75e88. [120] Wolde-Rufael Y. Another look at the relationship between telecommunication investment and economic activity in the United States. Int Econ J 2007;21(2):199e205. [121] Wolde-Rufael Y. Re-examining the financial development and economic growth Nexus in Kenya. Econ Model 2009;26(6): 1140e6. [122] World Economic Forum. The financial development report, 2011. New York, NY: World Economic Forum USA, Inc.; 2011. [123] Wu J, Hou H, Cheng S. The dynamic impacts of financial institutions on economic growth: evidence from the European Union. J Macroecon 2010;32(3):879e91. [124] Yoo SH, Kwak SJ. Information technology and economic development in Korea: a causality study. Int J Technol Manag 2004;27(1): 57e67. [125] Zahra K, Azim P, Mahmood A. Telecommunications infrastructure development and economic growth: a panel data approach. Pak Dev Rev 2008;47(4):711e26. [126] Zaman K, Izhar Z, Khan MM, Ahmad M. The relationship between financial indicators and human development in Pakistan. Econ Model 2012;29(No. 5):1515e23. [127] Prieger JE. The broadband digital divide and the economic benefits of mobile broadband for rural areas. Telecommun Policy 2013; 37(Nos. 6e7):483e502. [128] Jorgenson DW, Stiroh KJ. Information technology and growth. Am Econ Rev 1999;89(No. 2):109e15. [129] Gillen D, Lall A. The economics of the internet, the new economy and opportunities for airports. J Air Transp Manag 2002;8(1):49e62. [130] Hardy A. The role of the telephone in economic growth. Telecommun Policy 1980;4(No. 4):278e86. [131] Madden G, Savage SJ. CEE telecommunications investment and economic growth. Inf Econ Policy 1998;10(No. 2):173e95. [132] Lee J. Empirical evidence of causality between information communications technology and economic growth in China, Japan and South Korea. In: Paper presented in 11th International DSI and the 16th APDSI Joint Meeting, Taipei, July 12e16; 2011. [133] Gries T, Kraft M, Meierrieks D. Linkages between financial deepening, trade openness and economic development: causality evidence from Sub-Saharan Africa. Center for International Economics, working paper, No. 15. Germany: University of Paderborn; 2009.