Direct Effect and Spillover Effect of ICT on Electricity

Research ArticleDirect Effect and Spillover Effect of ICT on ElectricityConsumption in China Evidence from a Spatial Panel Analysis

Jianjun Xu 12 and Xiaowei Yang 3

1College of Science amp Technology Ningbo University Ningbo 315300 China2School of Economics and Management Dalian University of Technology Dalian 116024 China3Business School Ningbo City College of Vocational Technology Ningbo 315000 China

Correspondence should be addressed to Jianjun Xu xujianjunmaildluteducn

Received 8 December 2020 Revised 30 January 2021 Accepted 22 February 2021 Published 9 March 2021

Academic Editor Chiara Boccaletti

Copyright copy 2021 Jianjun Xu and Xiaowei Yang +is is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in anymedium provided the original work isproperly cited

+e extensive use of information and communication technologies (ICT) has facilitated peoplersquos lives and promoted the im-provement of productivity In the meantime ICT has a profound effect on the efficiency of electricity utilization and the demandfor electricity +e existing studies consider the direct effect of ICTon electricity consumption (EC) but neglect the spillover effectof ICT on EC and their action channels Under the assumption of cross-section dependence this paper introduces spatialmodeling techniques to confirm the positive direct effect and negative spillover effect of ICTon EC+e positive direct effects andnegative spillover effects of information technology on EC are similar to those of the communication technology and the absolutevalue of the former is also greater than the latter Additionally the results of meditation effect modeling also confirm that thereexists an incomplete mediating effect in the process of the ICT affecting EC through the channels of economic growth and theadjustment of the industrial structure +is study provides freshly empirical evidence for people to better understand the role ofICT in EC and opens fresh insights for policymakers to make corresponding policy adjustments

1 Introduction

+e pervasive use of information and communicationtechnologies (ICT) such as the Internet and mobile phoneshas had a wide impact on peoplersquos lives [1] Many studieshave shown that the extensive application of ICT has playedan important role in global economic growth productivityimprovement and energy consumption [2] Althoughscholars basically agree that the use of ICT can improveenergy efficiency there are different views on whether ICTcan promote energy consumption or energy-saving Somestudies have noted that the application of ICT is beneficial toenergy conservation and emission reduction [3 4]According to Greenpeace International [5] the use of ICTreduced global CO2 emissions by 2 +e other studiesbelieve that the application of ICT will increase moreelectricity demand [6 7] Faucheux and Nicolaı [8] foundthat a large part of the electricity consumption (EC) in

Europe is related to ICTproducts and services InternationalEnergy Agency [9] pointed out that in the past 20 years theuse of ICT has led to an annual increase of about 7 in EC

As far as China is concerned remarkable achievementshave been made in the development and application of ICTAccording to ldquothe 45th China Statistical Report on InternetDevelopmentrdquo issued by China Internet Network InformationCentre (CNNIC) as of March 2020 Chinarsquos Internet usersreached 904 million the Internet penetration rate (Internetusersaggregate population) reached 645 mobile Internetusers reached 897 million and the proportion of Internetusers using mobile phones to access the Internet reached993 Compared with five years ago the average downloadrate of mobile broadband increased by about six times andthe level of mobile Internet traffic dropped bymore than 90ldquoSpeed up and fee reductionrdquo project in China promoted asubstantial increase in mobile Internet traffic with an averagemonthly mobile traffic of 72GB 12 times the global average

HindawiMathematical Problems in EngineeringVolume 2021 Article ID 6688800 13 pageshttpsdoiorg10115520216688800

mobile Internet access traffic consumption reached 5539billion GB a year-on-year increase of 1073 +ere is nodoubt that the use of ICT will play a more significant role inChinarsquos economic development Meanwhile the use of ICThas a profound impact on the demand and supply of elec-tricity According to the data released by the China Federationof Electricity Enterprises on January 29 2019 Chinarsquos total ECin 2018 was 684 trillion kWh an increase of 85 year onyear Among them in 2018 the EC of Chinarsquos secondaryindustry was 472 trillion kWh with a year-on-year growth of72 and that of the tertiary industry was 108 trillion kWhwith a year-on-year growth of 127 especially that of in-formation transmission software and information technologyservices with a year-on-year growth of 235 In order toalleviate the increasing pressure of electricity demand broughtby the development of ICT we should correctly coordinatethe relationship between the development of ICT and ECdemand and the premise of coordinating the relationship isto correctly understand the complex relationship betweenthem

At present some achievements have confirmed thedemand growth effect of ITC on EC [10ndash16] Unfortu-nately the existing empirical literature on the relationshipbetween ICT and EC does not consider the spatial de-pendence of EC Because these methods being used inexisting studies assume that the cross-section elements areindependent [14ndash16] the existing studies only considerthe direct impact of ICT on EC in decision-making unitsbut neglect the indirect effect of EC caused by the spilloverof ICT in spatial neighboring units which may under-estimate or overestimate the comprehensive impact ofICT on EC In addition the existing studies such asShahbaz et al [12] Saidi et al [14] and Afzal et al [16]regarded ICT and economic growth as independentvariables affecting EC and did not reveal the mediationrole of economic growth in the process of ICT promotingEC In this study we take China for example consider thecross-section dependence of EC variables and introducethe spatial modeling techniques and the mediation effectmodel to conduct an empirical test Our focus is con-centrated on the following aspects what role does ICTplay in Chinarsquos EC Is there a spatial spillover effect of theICTon EC Is there a significant heterogeneity effect of theinformation technology and the communication tech-nology on EC Are economic growth and the adjustmentof industrial structure the action channels of ICTaffectingelectricity consumption

+is paper provides freshly empirical evidence for peopleto better understand the direct effect and indirect effect ofICT affecting EC and their action channels and opens freshinsights for policymakers to make corresponding policyadjustments +e contribution of this paper lies in (1) theprevailing spatial panel modeling techniques are adopted toanalyze the relationship between ICT and EC in China +isstudy not only confirms the direct effect of ITC on EC butalso take into account the role of the spillover effect of ITCon EC in China (2) the analysis of mediation effect modeldemonstrates that there exists an incomplete mediatingeffect when ICT effect on EC through the channels of

economic growth and the adjustment of the industrystructure

+e rest of the study is organized as follows Section 2provides a literature review of ICT and EC Section 3presents the model and method Section 4 elaborates on thevariables and data Section 5 elaborates on the empiricalresults Section 6 presents conclusions and discussion

2 Literature Review

21 EmpiricalAnalyses on theDirectRelationshipbetween ICTandEC Some researchers pay attention to the effect of ITC onEC considering not only the contribution of ITC for thegrowth of the EC but also the electricity-saving capacity of ICTFor example Cho et al [17] utilized the logistic model to studythe effect of ICT investment electricity price and oil price onindustrial EC in South Korea +eir empirical analysis showsthat ICTinvestment inmanufacturing industries with EC couldpromote the transformation of input factor substitution fromlabor-intensive to electricity-intensive In addition the resultsshow that investment of ICT in certain manufacturing sectorshelps to reduce EC while investment of ICT in services andmost manufacturing sectors increases EC Sadorsky [6] in-vestigated the effect of ICTon EC in emerging economies Hisempirical results of the dynamic panelmodel showed that therewas a positive and statistically significant relationship betweenICT and EC

Schlomann et al [10] analyzed the net effect of ICT onfuture EC in Germany +ey found that there was greatpotential for technological progress in reducing the EC ofICT equipment Afzal and Gow [7] studied the effect of ICTon EC in the next 11 emerging economies during 1990ndash2014and their results showed that there was a significant positivecorrelation between ICT and EC Kishita et al [18] con-structed an electricity demand model for the telecommu-nications industry and used scenario planning methods toanalyze changes in EC in the telecommunications industryand the possibility of electricity-saving +eir results showedthat improving energy efficiency equipment and mobilecommunication equipment for Internet Protocol (IP)communications were the most effective measures to reduceelectricity demand With quarterly data from 1993q1 to2014q4 Faisal et al [15] used the ARDL approach to confirmthat there was a U-shaped relationship between Internet useand EC

However the existing literature only takes into accountthe direct effect of ICT on EC but neglects the spatial thespillover effect of ICT on EC In this study we consider thecross-section dependence of variables and use the spatialeconometric method to calculate the direct effect and thespillover effect of ITC on EC

22 -e Channel of ICT Influencing EC ICT not only has adirect effect on EC but also indirectly affects the final ECthrough different path of the mediation channels One of theimportant paths may be economic development includingeconomic growth and the adjustment of the industrialstructure Existing empirical literature whether based on

2 Mathematical Problems in Engineering

time series research [11 18] or based on panel data modeling[19ndash25] had proved the positive role of ICT in economicdevelopment Moreover Cardona et al [26] reviewed 150studies and found that most of them confirmed the sig-nificant positive effect of ICT on productivity growth Vuet al [27] integrated 208 papers and most of the researchconclusions indicated that ICT had a positive impact oneconomic growth

Meanwhile a large number of documents also confirmedthat economic development has increased the demand forelectricity energy Whether based on time series analysis[28ndash34] or panel data analysis [35ndash39] most of the empiricalresults showed that there was a cointegration relationshipbetween economic growth and EC and there was a Grangercausality relationship between economic growth and EC

Furthermore some literature also discussed the con-nection between ICT economic growth and EC in a unifiedframework For example Salahuddin and Alam [11] withthe annual data from 1985 to 2012 used the ARDL boundarycointegration test method to confirm that the use of theInternet and economic growth stimulated the EC in Aus-tralia Based on the data of the United Arab Emirates from1975 to 2011 Shahbaz et al [12] applied the BayerndashHanckcointegration method to confirm that ICT and revenuegrowth increased electricity demand but electricity pricereduced electricity demand and the relationship betweenICT and EC was inverted U-shaped With panel data ofOECD from 1985 to 2012 Salahuddin and Alam [19] used apanel cointegration test and mixed mean group regressiontechnology to analyze the relationship among ICT eco-nomic development and EC+e results show that the use ofICT and economic growth stimulates EC Combined withthe dynamic panel data model of 67 countries from 1990 to2012 Saidi et al [14] investigated the impact of ICT andeconomic growth on EC+eir results demonstrated that theimpact of ICT on EC was significantly positive and eco-nomic growth had a positive statistically significant impacton EC Based on the analysis of the dynamic panel datamodel from 1990 to 2014 Afzal et al [16] analyzed theimpact of ICT and economic growth on EC in BRICs andthe results showed that ICT and economic growth both hadsignificantly positive effects on EC

Regrettably the above literature regarded ICT andeconomic growth as independent variables affecting EC anddid not reveal the mediation role of economic growth in theprocess of ICT promoting EC In this study the mediatingeffect model is introduced to investigate the mediating roleof economic growth and industrial structure adjustmentvariables in the analysis of the impact of ICT on EC

3 Theories and Methods

31 -e Effect Mechanism of ICT on Electricity Consumptionfrom a Spatial Perspective According to Schlomann et al[10] the impact of ICTon electricity demandmay be double

On the one hand the impact of local ICT developmenton EC is mainly reflected in the EC of ICT equipment andthe EC is required for the industrial development using ICT+e growth of ICTequipment will undoubtedly increase theconsumption of electric energy and promote the growth ofelectric energy In addition the development of local ICTcanalso affect EC through mediating variables eg the use ofICT will promote economic growth which will lead to anincrease in the demand for EC resulting in the so-calledrebound effect [40 41]

On the other hand the use of ICT may improve theefficiency of electric utilization and reduce the EC per unitproduct [42 43] and the use of ICT may promote thetransformation of high electricity consuming industry to lowelectricity consuming industry resulting in the decrease ofEC [44] In addition the development of local ICT can alsoaffect EC through mediating variables for example it willpromote the optimization and adjustment of economicstructure especially the transformation from high EC in-dustry to low EC industry which may reduce EC moreoverFinally whether ICT promotes or hinders EC in differentcountries or regions remains to be tested

+e forgoing analysis only considers the impact of localICT on EC but neglects the spillover effect of the ICT inneighboring regions If different countries or regions areadjacent to each other not only the development of ICT inlocal region can affect EC through the abovementioned waysbut also the development of ICT in neighboring regions canalso affect the local EC by the direct path and mediationpath and this is so-called spillover effect of ICT Figure 1shows the effect mechanism of ICT on EC

32 Model and Method

321 Point Estimation of Direct Impact and Spatial Spillover+e purpose of this study is tantamount to analyze the directand spillover effects of ICTon EC and their action channelsConsidering that spatial econometric modeling technologyis a very effective tool to analyze the direct effect andspillover effect [45 46] and the mediation effect model is avery effective tool to analyze the action channels [47ndash49]this paper introduces these two methods for analysis asfollows

(1) Test Method for Cross-Section Dependence Before de-ciding whether it is necessary to introduce a spatialeconometric model it is necessary to judge whether thevariables are cross-section dependency +e cross-sectiondependence (CD) test proposed by Pesaran [50] and thebias-corrected scaled Lagrange multiplier (bias-correctedLM) test proposed by Baltgi et al [51] are two effectivemethods to judge whether the cross section is dependentAmong them the CD test constructs test statistics based onthe correlation coefficient of sample regression residual andthe bias-corrected LMmethod constructs test statistics based

Mathematical Problems in Engineering 3

on the square of correlation coefficient of sample regressionresidual +e formulas are as follows

CD

2T

N(N minus 1)

1113971

1113944

Nminus 1

iminus 11113944

N

ji+1ρij

1113868111386811138681113868111386811138681113868111386811138681113868

1113868111386811138681113868111386811138681113868111386811138681113868 (1)

bias minus corrected LM

1

N(N minus 1)

1113971

1113944

Nminus 1

i11113944

N

ji+1Tρ2ij minus 11113872 1113873

minusN

2(T minus 1)

(2)

where ρij is the correlation coefficient of residuals series i andj and its calculation formula is

ρij 1113936

Tt1 1113954μit1113954μjt

1113936Tt1 1113954μ2it

1113868111386811138681113868111386811138681113868111386812

1113936Tt1 1113954μ2jt

11138681113868111386811138681113868

1113868111386811138681113868111386812 (3)

where 1113954μit is an estimated value of μit based on the least square(LS) regression

(2) Expression of the Spatial Panel Model According to thedifferent expression methods of ldquospacerdquo in the setting ofspatial panel model the spatial panel model is mainly dividedinto two types [45 46] one is the spatial auto-regression panelmodel (SAR) which is mainly used to analyze the influence ofEC behavior of neighboring provinces on EC behavior of thelocal province the other is spatial error panel model (SEM) inwhich the relationship between provinces is reflected by errorterms +e spatial Durbin model (SDM) contains the en-dogenous and exogenous variables with spatial lag which canbe regarded as a more general form than SAR and SEM

[45 46]+e expressions of SAR SEM and SDMare shown inmodel (4) model (5) and model (6) respectively

Lneleit λ0 + ρWlowastLneleit + λ1Lnictit

+ ζj 1113944 Lnctrjit + μit(4)

Lneleit λ0 + λ1Lnictit + ζj 1113944 Lnctrjit + μit

μit ψ 1113944N

j1wijμit + ]it

(5)

Lneleit λ0 + ρ3WlowastLneleit + λ1Lnictit + ζj 1113944 Lnctrjit

+λlowast1WlowastLnictit + ζlowast

j1113944 WlowastLnctrjit + μit

(6)

where ele represents EC ict represents the development ofICT ctr represents the control variable μit is the randomdisturbance term i represents the number of the i-th unit trepresents the t-th period and Ln represents logarithmicprocessing WlowastLnXit 1113936

Njne i wijLnXjt If the weighted

spatial interaction items in the model are not significantthen there is no so-called ldquospillover effectrdquo otherwise if oneor more spatial weighted interaction items are significantthe ldquospillover effectrdquo exists

(3) Selection Method of the Different Spatial Panel Model Inorder to judge whether the setting form of the SDMmodel isappropriate we applied Wald statistics and likelihood ratio(LR) statistics to judgment [45 52] According to Elhorst[45 52] the Wald statistics and LR statistics are subject tothe χ2 distribution of degree of the freedom k If theWald_spatial_lag statistics and Wald_spatial_error statisticsboth reject null hypotheses (H0) it indicates that SDM is

Loca

l ICT

Elec

tric

ity co

nsum

ptio

n

Mediatingeffect

Industry structure adjustment

Direct effect

Including spatial effectsNo spatial effects

Economic growth

Electric energy of ICT equipment

Industrial EC using ICT technology

Nei

ghbo

urin

g IC

T

Mediatingeffect

Direct effect



Economic growth


Figure 1 Effect mechanism of ICT on EC


more reasonable Likewise If LR_spatial_lag statistics andLR_spatial_error statistics both reject null hypotheses (H0)it also indicates that SDM is more reasonable

(4) Parameter Estimation Strategy of the Spatial Panel Model+e parameters of LS estimation are not only biased but alsoinconsistent if the model contains the spatial lag explanatoryvariables +e parameters of LS estimation meet the unbi-ased and consistent but no longer meet the validity as-sumption if the model contains the spatial error term+erefore it is necessary to find an effective alternativemethod +e maximum likelihood estimation (MLE)method can get unbiased and consistent estimation results[53 54] and Elhorst [54] further proves that this method caneffectively solve the endogenous problem

322 Partial Differential Method of Direct Impact andSpillover Effect LeSage and Pace [55] pointed out that theresults obtained by the point estimation method may bebiased and proposed the partial differential method tocalculate the direct effect and spatial spillover effect If theexplained variable is represented by matrix Y and theexplaining variable is represented by matrix X the SDM canbe rewritten into the following form

Y (I minus δW)minus 1αlN +(I minus δW)

minus 1(Xβ + WXθ)

middot+ (I minus δW)minus 1ε

(7)

Among them the partial differential of the variablenamed Y ofN regions with respect to the 1st to k-th variablesnamed X is shown as follows

zY

zx1k

middot middot middotzY

zxNk

1113890 1113891

zY1

zx1k

middot middot middotzY1

zxNk

⋮ middot middot middot ⋮

zYN

zx1k

middot middot middotzYN

zxNk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(I minus δW)minus 1

βk w12θk middot middot middot w1Nθk

w21θk βk middot middot middot w2Nθk

⋮ ⋮ ⋮ ⋮

wN1θk wN2θk middot middot middot βk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(8)

Among them LeSage and Pace [55] defined the meanvalue of diagonal elements in the rightmost matrix of for-mula (7) as a direct effect and the mean value of the sum ofnondiagonal elements in each row or column was defined asthe spillover effect In SDM the direct effect of the k-th

variable namedX onY is equal to βk and the average value ofthe spillover effect on the adjacent area is 1113936

Ni1 1113936

Nj1 wijθkN

323 Analysis Method of Action Mechanism In order toreveal whether ICT has significantly affected Chinarsquos ECthrough the channels of economic growth and the adjust-ment of the industrial structure according to the practice ofLi et al [48] and Xu et al [49] the following linear models(9) (10) and (11) are introduced to analyze the mediationrole of ICT on EC

Lneleit α0 + α1Lnictit + α2Lnctrit + ε1it (9)

Lnmedit β0 + β1Lnictit + β2Lnctrit + ε2it (10)

Lneleit c0 + c1Lnictit + c2Lnmedit + c3Lnctrit + ε3it

(11)

where ele represents EC ict represents the development ofICT med represents the mediation variables ctr representsthe control variable and εit is the random disturbance termSubstituting model (10) into model (11) we can get

Lneleit c0 + c2β0( 1113857 + c1 + c2β1( 1113857Lnictit + c2β2 + c3( 1113857

Lnctrit + ε3it

(12)

In model (12) c1 is the direct effect of ICT on EC andc2β1 is the mediation effect of ICT on EC

Mediation effect estimation strategy as follows the firststep is to test the significance of the coefficient in the model(9) If the coefficient α1 is not significant it means that ICT isnot related to EC and the mediation effect analysis isstopped if the coefficient α1 is significant the further test iscarried out +e second step is to test the significance of theestimation coefficient β1 in model (10) If the estimationcoefficient β1 is not significant the test is stopped otherwisethe third step is to test +e third step is to test the sig-nificance of the estimated coefficients c1 and c2 in model(11) If the coefficient c2 of mediation variables is significantit indicates that there exists a mediation effect At this timeconsider the significance of the estimated coefficient c1 If c1is significant there exists an incomplete mediation effect ifc1 is not significant there exists a complete mediation effect

4 Variables and Data

41 Variables EC is an explanatory variable +ere are twokinds of approaches tomeasure EC One way is the aggregateindex [56 57] the other way is the average index [58ndash60]Per capita EC (avele) is obtained from aggregate EC (toele)divided by the aggregate population +is paper choosesboth of them as the explained variable respectively

+e development of ICT is the main explanatory vari-able +ere are two types of measurement indexes one is a


single indicator For example some literature used thenumber of Internet users (net) as an indicator measuringinformation technology the number of landline and mobilephone users (tele) as an indicator measuring the use ofcommunication tools [6 11 14] and the industrial addedvalue of telecommunication services (ICTP) measuring thedevelopment of ICT industry [13] the other is to synthesize acomprehensive indicator Concerning studies combinedseveral indexes such as Internet broadband Internet userslandline and mobile phone and computer users into oneindex to measure the development of ICT based on theprincipal component analysis (PCA) [12 21 35 61] Re-ferring to the existing literature and improve its practice thisstudy chooses Internet users (net) landline and mobilephones per 100 people (tele) and the industrial added valueof telecommunication services (ICTP) to measure the threespecific performance of the development of ICT respec-tively Because the first two indicators are highly correlatedand the correlation coefficient is more than 083 (see Sup-plementary Materials (available here)) if adding both ofthem in the same model may cause multicollinearity we usethe PCA method to reduce the dimensions and extractcommon factors to measure the development of ICT +ecomposition index is represented by ICTCI

+e other variables include economic growth the ad-justment of the industrial structure patent innovationdomestic credit and import trade Among them economicgrowth is mostly measured by real gross domestic products(RGDP) We use 1999 as the base period and the real GDPequals the nominal GDP divided by the GDP deflator +eadjustment of the industrial structure is measured by theproportion of the tertiary industry in GDP which isexpressed by str Patent innovation is measured by the actualnumber of authorized patents divided by the population atthe end of the year which is represented by pat Domesticcredit is obtained by the year-end credit divided by nominalGDP which is represented by crd import trade is calculatedby the import amount divided by GDP which is representedby imp

It should be pointed out that there are numerous optionsfor the spatial weight matrix [62 63] In this paper refer tothe practice of Yu et al [63] economic distance weight (W)is adopted in which the definition of element wij in W is asfollows If ine j wij 1|yi minus yj| otherwise i j wij 0where yi and yj are the average of the real GDP of the i-thprovince and the j-th province during the sample period

42 Data +e sample regions include 31 provinces mu-nicipalities and autonomous regions in the Chinesemainland with a sample period of 1999ndash2018 years +ebasic data come from China Statistical Yearbook during2000ndash2019 China Energy Statistical Yearbook during2000ndash2019 China Financial Statistical Yearbook during2000ndash2019 China Regional Financial Operation Report

during 1999ndash2018 and National Economic and Social De-velopment Operation Report of each province municipali-ties and autonomous regions during 1999ndash2018 +estatistical description of each variable is shown in Table A1(see Supplementary Materials) and the correlation coeffi-cients are shown in Table A2 (see Supplementary Materials)

5 Empirical Results and Interpretation

51 Test of Cross-SectionDependence It is necessary to judgewhether the variables are cross-section dependent before wedecided to introduce a spatial econometric model Weadopted CD statistics and bias-corrected LM statistics to testwhether the variables are cross-section dependent +e re-sults of the cross-section dependence test with CD statisticsand bias-corrected LM statistics are shown in Table 1 +eCD statistics and bias-corrected LM statistic of Lntoele rejectthe null hypothesis (H0) of cross-section independence at thesignificance level of 1 which means that the aggregate ECof different provinces is not independent Likewise +e CDstatistics and Bias-corrected LM statistics of Lnavele LnictciLnnet Lntele Lnpat Lncrd Lnimp Lnrgdp and Lnstr alsorejectH0 at the level of 1 significance whichmeans that theper capita EC the development of ICT the number ofInternet users telephone penetration rate patent innova-tion domestic credit import trade economic growth andthe adjustment of the industrial structure are also cross-sectional dependent respectively

52 Direct Effect and Spillover Effect Are Calculated by PointEstimation Method Table 2 shows the estimated results ofthe model (6) M(1a) M(2a) and M(3a) are the estimationresults with Lntoele as the explained variable and M(1b)M(2b) and M(3b) are the estimation results with Lnavele asthe explained variable According to the Hausman test [64]all of Hausman statistics in M(1a) M(2a) M(3a) M(1b)M(2b) and M(3b) reject the null hypothesis H0 at thesignificance level of 1 which indicates that the fixed effectform is superior to the random effect formWald_spatial_lagstatistics Wald_spatial_error statistics LR_spatial_lag sta-tistics and LR_spatial_error statistics in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) all reject the null hy-pothesis H0 at the significance level of 1 which indicatesthat SDM is suitable and it is unnecessary to simplify SDMto SAR or SEM Next we analyze the meaning of eachparameter as follows

Firstly the effect of comprehensive development of ICTon EC is analyzed According to Table 2 the estimatedcoefficients of the main explanatory variables named Lnictciin M(1a) and M(1b) are 02282 and 03927 and meets the t-test at the significance level of 1 which means the de-velopment of ICTpromotes on the aggregate (per capita) ECin China+is further validates the conclusions of Afzal et al[16] Afzal and Gow [7] Salahuddin and Alam [19] and


Table 2 +e effect of ICT on EC

VariableLntoele Lnavele Lntoele Lnavele Lntoele LnaveleLnictci Lnictci Lnnet Lnnet Lntele LnteleM(1a) M(1b) M(2a) M(2b) M(3a) M(3b)

Lnictci 02282lowastlowastlowast 03927lowastlowastlowast[56286] [97714]

Lnnet 01488lowastlowastlowast 02370lowastlowastlowast[58044] [96590]

Lntele 00854lowastlowastlowast 01867lowastlowastlowast[32977] [74404]

Lnpat minus 00868lowastlowastlowast minus 00545lowastlowastlowast minus 00871lowastlowastlowast minus 00531lowastlowastlowast minus 00873lowastlowastlowast minus 00539lowastlowast[minus 47190] [minus 29899] [minus 47450] [minus 30224] [minus 46512] [minus 29612]

Lncrd 03219lowastlowastlowast 02500lowastlowastlowast 03136lowastlowastlowast 02413lowastlowastlowast 03356lowastlowastlowast 02754lowastlowastlowast[85908] [67297] [83649] [67225] [87918] [74463]

Lnimp minus 00752lowastlowastlowast minus 00439lowastlowast minus 00748lowastlowastlowast minus 00424lowastlowast minus 00857lowastlowastlowast minus 00592lowastlowastlowast[minus 48513] [minus 28566] [minus 48572] [minus 28744] [minus 54499] [minus 38850]

Lnrgdp 05132lowastlowastlowast 03930lowastlowastlowast 05146lowastlowastlowast 04127lowastlowastlowast 05401lowastlowastlowast 04209lowastlowastlowast[55988] [43242] [56485] [47308] [57920] [46581]

Lnstr 03679lowastlowastlowast 04266lowastlowastlowast 03784lowastlowastlowast 04585lowastlowastlowast 04355lowastlowastlowast 05049lowastlowastlowast[60978] [71294] [63774] [80664] [73203] [87526]

W lowast lnictci minus 03499lowastlowast minus 02657lowast[minus 23465] [minus 17711]

W lowast Lnnet minus 01880lowast minus 01434[minus 19484] [minus 15282]

W lowast Lntele minus 01325lowast minus 00718[minus 17172] [minus 09447]

W lowast Lnpat minus 01434lowastlowast minus 01471lowastlowast minus 01288lowastlowastlowast minus 01283lowastlowast minus 01597lowastlowast minus 01718lowastlowastlowast[minus 23939] [minus 24807] [minus 21521] [minus 22418] [minus 25996] [minus 28882]

W lowast Lncrd 02062 03609lowastlowast 01773 03765lowastlowast 01379 03298lowastlowast[11986] [21264] [10406] [23152] [08167] [20224]

W lowast Lnimp minus 00678lowast minus 00122 minus 00722lowast minus 00146 minus 00762lowastlowast minus 00312[minus 17167] [minus 03128] [minus 18548] [minus 03932] [minus 18697] [minus 07914]

W lowast Lnrgdp 22355lowastlowastlowast 25934lowastlowastlowast 20418lowastlowastlowast 25087lowastlowastlowast 23397lowastlowastlowast 27896lowastlowastlowast[57515] [67389] [55031] [70674] [58332] [71864]

W lowast Lnstr minus 09695lowastlowastlowast minus 07151lowastlowastlowast minus 09306lowastlowastlowast minus 05646lowastlowastlowast minus 10367lowastlowastlowast minus 07083lowastlowastlowast[minus 52921] [minus 38962] [minus 50907] [minus 31624] [minus 54160] [minus 37419]

W lowast Depvar 01196 minus 00594 01568lowastlowast minus 01510lowastlowast 01841lowastlowast minus 00970[16171] [minus 07907] [21600] [minus 19299] [25446] [minus 12541]

R-squared 09898 09775 09899 09773 09893 09756Corr-squared 05038 05545 05057 05507 04788 05175Log likelihood 4738777 47885102 4752999 4760878 45961325 minus 2178122Wald_spatial_lag 731734lowastlowastlowast 637459lowastlowastlowast 67975lowastlowastlowast 630889lowastlowastlowast 740587lowastlowastlowast 75855lowastlowastlowastLR_spatial_lag 733308lowastlowastlowast 646918lowastlowastlowast 677502lowastlowastlowast 594992lowastlowastlowast 74424lowastlowastlowast 710421lowastlowastlowastWald_spatial_error 733840lowastlowastlowast 625727lowastlowastlowast 683616lowastlowastlowast 606834lowastlowastlowast 755035lowastlowastlowast 725566lowastlowastlowastLR_spatial_error 726930lowastlowastlowast 623797lowastlowastlowast 672241lowastlowastlowast 577388lowastlowastlowast 739207lowastlowastlowast 686425lowastlowastlowastHausman test 7543281lowastlowastlowast 906921lowastlowastlowast 1032146lowastlowastlowast 516559lowastlowastlowast 600381lowastlowastlowast 1389592lowastlowastlowastMethod MLE MLE MLE MLE MLE MLEObs 620 620 620 620 620 620Note lowast lowastlowast and lowastlowastlowast denote significance levels of 10 5 and 1 the values in square brackets are t statistics

Table 1 Cross-section dependence test results of variables

Variable CD test Bias-corrected LMLntoele 9514387lowastlowastlowast 2798158lowastlowastlowastLnavele 9474905lowastlowastlowast 2773882lowastlowastlowastLnictci 9583513lowastlowastlowast 2840940lowastlowastlowastLnnet 9574304lowastlowastlowast 2835191lowastlowastlowastLntele 9456901lowastlowastlowast 2762868lowastlowastlowastLnpat 9268280lowastlowastlowast 2650118lowastlowastlowastLncrd 6154365lowastlowastlowast 1392010lowastlowastlowastLnimp 6730324lowastlowastlowast 2585710lowastlowastlowastLnrgdp 9072525lowastlowastlowast 2545939lowastlowastlowastLnstr 3771847lowastlowastlowast 8900097lowastlowastlowast

Note lowastlowastlowastdenote significance levels of 1


Solarin et al [61] and reconfirm that the use of ICTpromoteselectricity demand +e estimated parameter ofW lowast Lnictciin M(1a) and M(1b) are minus 01434 and minus 02657 and meets thet-test at the significance level of 5 and 10 respectivelywhich indicates the spillover effect of ICT is negativelycorrelated with the aggregate (per capita) EC +is meansthat the development of ICT in neighboring provinces is notconducive to aggregate (per capita) EC in local regions +ismay be related to the existence of competitive relationshipbetween neighboring regions

Secondly the differential effect of ICT on EC is com-pared According to Table 2 the estimated coefficients ofLnnet in M(2a) and M(2b) are 01488 and 02370 the es-timated coefficients of Lntele inM(3a) andM(3b) are 00854and 01867 and all of them meet the t-test at the significanceof 1 which confirms that the use of the Internet andcommunication tools both directly contribute to aggregate(per capita) EC and the direct role of Internet use is slightlygreater than that of information and communication toolsIt is consistent with the results of Afzal et al [16] and Saidiet al [14] and those studies also prove that the use of theInternet and communication tools stimulate EC +e esti-mated coefficients of W lowast Lnnet in M(2a) and M(2b) areminus 01880 and 01434 and only the former meets the t-test at10 significance level which only confirms that the use ofthe Internet has a significant negative spillover effect onaggregate EC the estimated coefficients of W lowast Lntele inM(3a) and M(3b) are minus 01325 and minus 00718 and the firstestimated coefficient meets the t-test at a significance level of10 which only confirms that the popularization of com-munication tools has a significant negative spillover effect onaggregate EC

Finally we further analyze the parameter estimationresults of different control variable According to Table 2 theestimated coefficients of Lnpat in M(1a) M(2a) M(3a)M(1b) M(2b) and M(3b) are significantly negative whichindicates that patent innovation is conducive to aggregate(per capita) electricity-saving +is supports the conclusionof An et al [44]+e estimated coefficients of W lowast Lnpat inM(1a) M(2a) M(3a) M(1b) M(2b) and M(3b) are signif-icantly negative which indicate that the spillover effect ofpatent innovation is not conducive to aggregate (per capita)EC

+e estimated coefficients of Lncrd in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) are significantly positivewhich indicates that the direct effect of domestic creditpromotes aggregate (per capita) EC and this supports theconclusion of Lin and Zhao [41] and Rafindadi and Ozturk[65] who deemed there existed a significant promotioneffect of financial development on renewable EC +e esti-mated coefficients ofW lowast Lncrd inM(1a)M(2a) andM(3a)are insignificantly positive and the estimated coefficients ofW lowast Lncrd in M(1b) M(2b) and M(3b) are significantlypositive which indicate that the spillover effect of domesticcredit promotes per capita EC

+e estimated coefficients of Lnimp in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) are significantly negative

which indicates that the direct effect of import trade is notconducive to aggregate (per capita) EC and it only partiallysupports the conclusion of Lin and Zhao [41] they deemedthat trade open had a significant negative effect on renewableEC in the long-run +e estimated coefficient ofW lowast LnimpinM(1a)M(2a) andM(3a) is significantly negative and theestimated coefficient of W lowast Lnimp in M(1b) M(2b) andM(3b) is insignificantly negative which only confirms thatthe direct effect of import trade is not conducive to aggregateEC

+e estimated coefficients of Lnrgdp in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) are significantly positivewhich means economic growth promotes to aggregate (percapita) EC +is is consistent with the conclusion of Hamdiet al [28] Osman et al [36] and Saidi et al [14] All thesestudies confirmed that there existed a positive and statis-tically significant effect of economic growth on EC +eestimated coefficients of W lowast Lnrgdp in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) are significantly positiveand it means that the spillover effect of economic growthsimilarly promotes to aggregate (per capita) EC

+e estimated coefficients of Lnstr in M(1a) M(2a)M(3a) M(1b) M(2b) and M(3b) are significantly positivewhich indicates the direct effect of the adjustment of theindustry structure promotes to aggregate (per capita) EC Itis consistent with the conclusion of Liu et al [66] +eestimated coefficients ofW lowast Lnstr inM(1a)M(2a)M(3a)M(1b) M(2b) and M(3b) are significantly negative whichindicates that the spillover effect of the adjustment of theindustry structure does not contribute to aggregate (percapita) EC

53 Direct Effect and Spillover Effect Are Calculated by thePartial Differential Method In order to get the robustnessresults we applied the partial differential method to calculatethe direct effect and spillover effect Table 3 is the decom-position results of the effect of on EC

According to Table 3 the direct effect of Lnictci onLntoele (Lnavele) is 02273 (03945) and it meets the t-test at1 significance level the spillover effect of Lnictci onLntoele (Lnavele) is minus 03559 (minus 02764) and it meets thesignificant t-test which indicates that the development ofICT has a positive direct effect and a negative spillover effecton aggregate (per capita) EC Comparison of the direct andspillover effect of the usage of Internet and communicationtools on aggregate (per capita) EC the direct effect of Lnnetand Lntele on Lntoele (Lnavele) are 01491 (02370) and00833 (01870) respectively and all of them are significantwhich confirms the direct promotion role of the informationtechnology and communication technology on aggregate(per capita) EC +is is consistent with the correspondingresults based on point estimation in Table 2 +e spillovereffect of W lowast Lnnet on Lntoele (Lnavele) isminus 01704(minus 01750) and it meets the t-test at 10 significancelevel which confirms the spillover effect of the usage ofInternet is not conducive to aggregate (per capita) EC +e


spillover effect of W lowast Lntele on Lntoele (Lnavele) isminus 01423(minus 00934) However neither of them passed thesignificance t-test

In addition to that according to Table 3 the direct effectsof Lnpat and Lnimp on aggregate (per capita) EC are sig-nificantly negative the direct effects of Lncrd Lnrgdp andLnstr on aggregate (per capita) EC are significantly positive+ey are consistent with the results of the point estimationmethod

+e spillover effects of Lnpat Lnimp and Lnstr onaggregate (per capita) EC are significantly negative thespillover effects of Lncrd and Lnrgdp on aggregate (percapita) EC are significantly positive+ey are consistent withthe corresponding results of the point estimation method

54 Empirical Test of Action Channels We introduce themediation effect model to analyze the action channels of ICTaffecting EC According to the test steps of the mediation

effect model the first step is to fit model (9) We take Lntoeleand Lnavele as the explained variables respectively and takeLnictci as the explanatory variables to obtain the estimatedcoefficients α1 in M(4a) and M(4b) which are 04710 and06499 at the significance level of 1 respectively (seeTable 4) and confirm the direct effect of ICTon aggregate ECand per capita EC the second step is to estimate model (10)in which the mediation variables named Lnrgdp and Lnstrare taken as the explanatory variables and the estimatedcoefficients β1 of Lnictic inM(5a) andM(5b) are 01287 and03581 respectively and both meet the significant t-testindicating that the development of ICT is conducive toeconomic growth and adjustment of the industry structurethe third step to estimate model (11) the estimated coeffi-cients c1 of Lnictci in M(6a) and M(6b) are 02461 and04174 which meet the requirement of 1 significant t-testIn addition the estimated coefficients of Lnrgdp and Lnstr inM(6a) are 05329 and 04366 respectively and both of themmeet the significant t-test +e results show that the

Table 3 Decomposition results of the effect of on EC

Lntoele Lnavele Lntoele Lnavele Lntoele LnaveleDirect effect




Lnpat minus 00875lowastlowastlowast minus 00547lowastlowast minus 00862lowastlowastlowast minus 00539lowastlowast minus 00884lowastlowastlowast minus 00548lowastlowast[minus 47275] [minus 29275] [minus 47214] [minus 29948] [minus 47087] [minus 8354]


Lnimp minus 00752lowastlowastlowast minus 00435lowastlowast minus 00748lowastlowastlowast minus 00423lowastlowast minus 00864lowastlowastlowast minus 00590lowastlowast[minus 48878] [minus 28359] [minus 49918] [minus 28015] [minus 55962] [minus 36103]

Lnregdp 05205lowastlowastlowast 03929lowastlowastlowast 05124lowastlowastlowast 04166lowastlowastlowast 05590lowastlowastlowast 04299lowastlowastlowast[55756] [43803] [56917] [45843] [60933] [45505]


Spillover effect

Lnictci minus 03559lowastlowast minus 02764lowast[minus 20978] [minus 19075]

Lnnet minus 01704lowast minus 01750lowast[minus 18278] [minus 18133]

Lntele minus 01423 minus 00934lowast[minus 14288] [minus 11479]

Lnpat minus 01758lowastlowast minus 01380lowastlowastlowast minus 01457lowastlowast minus 01193lowastlowastlowast minus 02162lowastlowast minus 01769lowastlowastlowast[minus 25027] [minus 23421] [minus 22855] [minus 20264] [minus 26449] [minus 26518]

Lncrd 02717 03243lowastlowast 02395 03311lowastlowast 02462 03044lowastlowast[13830] [19747] [14360] [19996] [11580] [17445]

Lnimp minus 00905lowast minus 00114 minus 00820lowastlowast minus 00095 minus 01129lowastlowast minus 00292[minus 19240] [minus 03147] [minus 20128] [minus 2561] [minus 21466] [minus 06985]


Lnstr minus 10508lowastlowastlowast minus 07016lowastlowastlowast minus 08890lowastlowastlowast minus 06459lowastlowastlowast minus 11980lowastlowastlowast minus 08088lowastlowastlowast[minus 45564] [minus 39559] [minus 44858] [minus 37180] [minus 45885] [minus 38529]

Notes lowast lowastlowast and lowastlowastlowast denote significance levels of 10 5 and 1 the values in square brackets are t statistics


incomplete mediation effect exists Among them the me-diation effect of ICTon aggregate EC through the channel ofeconomic growth and the adjustment of the industrialstructure is 02249 +e estimated coefficients of Lnrgdp andLnstr inM(6b) are 03953 and 05072 respectively and bothof them meet the t-test at the significance of 1 +e resultsshow that the incomplete mediation effect exists Amongthem the mediation effect of ICT on per capita EC throughthe channels of economic growth and the adjustment of theindustrial structure is 02325

6 Conclusions and Discussion

+e widespread use of ICT has a significant effect on EC+eoretical analysis shows that local and neighboring ICTcan lead to the growth of local EC because of the EC of ICTequipment and the development of related industries At thesame time it can also reduce the EC because of the appli-cation of local and adjacent ICT In order to better un-derstand the complex relationship between ICT and ECbased on the panel data of 31 provinces in China this paperintroduces spatial modeling techniques and the mediationeffect model to investigate the effects of ICT on Chinarsquos ECand their action channels +e conclusions are as follows

Firstly the point estimation results of SDM are basicallyconsistent with those of partial differential method whichshow that the direct effect of ICT significantly promotes ECwhile the spillover effect is opposite +is means that ig-noring the spatial spillover effect will overestimate the totalimpact of ICT on EC +is provides new empirical evidencefor us to fully understand the impact of ICT on EC +econclusion also has some enlightenment for the predictionof EC In the process of electricity demand forecasting it isnecessary to integrate ICT factors into prediction model Ifthe ICT factor is not included as an explanatory variable itmay miscalculate the actual electricity demand which inturn will cause a shortage or surplus of electricity demand

Secondly there exists an incomplete mediating effectwhen the ICT influences EC through the action channels ofeconomic growth and the adjustment of the industrialstructure +is means that ignoring the indirect effect ofincomplete mediating effect may overestimate the directeffect of ICTon EC+e policy implication of this conclusionis that controlling the speed of economic growth and op-timizing the industrial structuremay serve to bring down thepressure of the gap between electricity supply and electricitydemand Specifically speaking related means such aseliminating the high electricity consuming industries ortransforming the high electricity consuming industries intolow electricity consuming industries through technologicaltransformation or promoting the development of emerginglow electricity consuming industries are all effective ways toreduce electricity consuming

+irdly the direct and spillover effects of domestic crediton EC are significantly positive while innovation and im-port are just the opposite +e result means that it is difficultto effectively control the excessive growth of electricityconsumption by controlling the domestic credit scale whileinnovation and import may be an effective means forelectricity-saving in China +erefore related policy shouldsupport the development of clean power generation tech-nology encourage enterprises to adopt clean productiontechnology and encourage development and application ofelectricity-saving technologies and products in terms ofimport the government should actively cultivate all kinds ofimporters and give priority to the introduction of electricity-saving technologies and products

Admittedly there are still some limitations and inade-quacies in this study First of all the stability of the results ofspatial panel regression is affected by the selection of spatialweight In addition to the economic distance alreadyadopted in this study there are many alternatives such as thespatial weight matrix such as binary contiguity matrixgeographic distance matrix and institutional distance

Table 4 Estimation results of mediation effect model with the full sample

VariableStep 1 estimation model (9) Step 2 estimation model (10) Step 3 estimation model (11)Lntoele Lnavele Lnrgdp Lnstr Lntoele LnaveleM(4a) M(4b) M(5a) M(5b) M(6a) M(6b)

Lnictci 04710lowastlowastlowast 06499lowastlowastlowast 01287lowastlowastlowast 03581lowastlowastlowast 02461lowastlowastlowast 04174lowastlowastlowast[127310] [177647] [83708] [147677] [61439] [105767]

Lnpat minus 00272 00000 00296lowastlowastlowast 00229lowast minus 00530lowastlowastlowast minus 00233[minus 14474] [00003] [37804] [18566] [minus 30331] [minus 13534]

Lncrd 03522lowastlowastlowast 02723lowastlowastlowast minus 00716lowastlowastlowast 00340 03755lowastlowastlowast 02834lowastlowastlowast[90351] [70657] [minus 44222] [13311] [102921] [78840]

Lnimp minus 00848lowastlowastlowast minus 00428lowastlowastlowast minus 00190lowastlowastlowast 00140 minus 00808lowastlowastlowast minus 00424lowastlowastlowast[minus 51961] [minus 26546] [minus 27973] [13099] [minus 53469] [minus 28507]

Lnrgdp 05329lowastlowastlowast 03953lowastlowastlowast[58152] [43783]

Lnstr 04366lowastlowastlowast 05072lowastlowastlowast[75119] [88581]

R2 03350 04068 01545 02831 04425 05063Log likelihood 3824923 3894710 9268840 6444934 4371531 4463790Method LS LS LS LS LS LSObs 620 620 620 620 620 620Notes lowast lowastlowast and lowastlowastlowast denote significance levels of 10 5 and 1 the values in square brackets are t statistics


matrix How to select appropriate spatial weights is also anunsolved problem Secondly when we analyze the role ofmediation variables this paper only considers the role ofeconomic growth and the adjustment of the industrialstructure Perhaps ICT can also affect EC by the channel ofimproving productivity promoting population migrationand so on +ese channels are not taken into account in thispaper +irdly the influence of ICT on EC in each industryperhaps is also heterogeneous which should be paid moreattention in the future

Data Availability

+e data used to support the findings of this study are in-cluded in Section 32 All data can also be obtained from theEPS website httpolapepsnetcomcn

Conflicts of Interest

+e authors declare no conflicts of interest

Acknowledgments

+is work was sponsored by the National Natural ScienceFoundation of China (no 71773057) Humanities and SocialSciences Research Youth Foundation Project of Ministry ofEducation in China (no 16YJCZH124 17YJCZH218) andKey Funding Projects of Humanities and Social Sciences ofUniversities in Zhejiang Province (no 2016QN005)

Supplementary Materials

Table A1 description statistics of variables Table A2 cor-relation matrix (Supplementary Materials)

References

[1] J D Moyer and B B Hughes ldquoICTs do they contribute toincreased carbon emissionsrdquo Technological Forecasting andSocial Change vol 79 no 5 pp 919ndash931 2012

[2] M Shahiduzzaman and K Alam ldquoInformation technologyand its changing roles to economic growth and productivity inAustraliardquo Telecommunications Policy vol 38 no 2pp 125ndash135 2014

[3] W Walker ldquoInformation technology and the use of energyrdquoEnergy Policy vol 13 no 5 pp 458ndash476 1985

[4] R Bernstein and R Madlener ldquoImpact of disaggregated ICTcapital on electricity intensity in European manufacturingrdquoApplied Economics Letters vol 17 no 17 pp 1691ndash1695 2010

[5] Greenpeace International Clicking Clean How CompaniesAre Creating the Green Internet April 2014 SB BusinessWeekly Greenpeace International Amsterdam Netherlands2014

[6] P Sadorsky ldquoInformation communication technology andelectricity consumption in emerging economiesrdquo EnergyPolicy vol 48 pp 130ndash136 2012

[7] M Afzal and J Gow ldquoElectricity consumption and infor-mation and communication technology in the next elevenemerging economiesrdquo International Journal of Energy Eco-nomics and Policy vol 6 no 3 pp 381ndash388 2016

[8] S Faucheux and I Nicolaı ldquoIT for green and green IT aproposed typology of eco-innovationrdquo Ecological Economicsvol 70 no 11 pp 2020ndash2027 2011

[9] International Energy Agency (IEA) Gadgets and GigawattsPolicies for Energy Efficient Electronics OECD PublishingParis France 2009

[10] B Schlomann W Eichhammer and L Stobbe ldquoEnergysaving potential of information and communication tech-nologyrdquo International Journal of Decision Support Systemsvol 1 no 2 pp 152ndash163 2015

[11] M Salahuddin and K Alam ldquoInternet usage electricityconsumption and economic growth in Australia a time seriesevidencerdquo Telematics and Informatics vol 32 no 4pp 862ndash878 2015

[12] M Shahbaz I U Rehman R Sbia andH Hamdi ldquo+e role ofinformation communication technology and economicgrowth in recent electricity demand fresh evidence fromcombine cointegration approach in UAErdquo Journal of theKnowledge Economy vol 7 no 3 pp 797ndash818 2016

[13] B Han DWangW Ding and L Han ldquoEffect of informationand communication technology on energy consumption inChinardquo Natural Hazards vol 84 no 1 pp 297ndash315 2016

[14] K Saidi H Toumi and S Zaidi ldquoImpact of informationcommunication technology and economic growth on theelectricity consumption empirical evidence from 67 coun-triesrdquo Journal of the Knowledge Economy vol 8 pp 1ndash152017

[15] F Faisal T Tursoy and N Berk ldquoLinear and non-linearimpact of internet usage and financial deepening on electricityconsumption for Turkey empirical evidence from asymmetriccausalityrdquo Environmental Science and Pollution ResearchInternational vol 4 pp 1ndash20 2018

[16] M N I Afzal J Gow and A Rahman ldquoEconomic and in-ternet growth effect on electricity consumption in the BRICScountriesrdquo International Advances in Economic Researchvol 25 no 3 pp 339ndash346 2019

[17] Y Cho J Lee and T-Y Kim ldquo+e impact of ICT investmentand energy price on industrial electricity demand dynamicgrowth model approachrdquo Energy Policy vol 35 no 9pp 4730ndash4738 2007

[18] Y Kishita Y Yamaguchi Y Umeda et al ldquoDescribing long-term electricity demand scenarios in the telecommunicationsindustry a case study of Japanrdquo Sustainability vol 8 no 1pp 52ndash16 2016

[19] M Salahuddin and K Alam ldquoInformation and communi-cation technology electricity consumption and economicgrowth in OECD countries a panel data analysisrdquo Interna-tional Journal of Electrical Power amp Energy Systems vol 76pp 185ndash193 2016

[20] S Sassi andM Goaied ldquoFinancial development ICTdiffusionand economic growth lessons from MENA regionrdquo Tele-communications Policy vol 37 no 4-5 pp 252ndash261 2013

[21] Z Latif Y mengke L Danish et al ldquo+e dynamics of ICTforeign direct investment globalization and economicgrowth panel estimation robust to heterogeneity and cross-sectional dependencerdquo Telematics and Informatics vol 35no 2 pp 318ndash328 2018

[22] A Das M Chowdhury and S Seaborn ldquoICT diffusion fi-nancial development and economic growth new evidencefrom low and lower middle-income countriesrdquo Journal of theKnowledge Economy vol 9 no 3 pp 928ndash947 2018

[23] G G Haftu ldquoInformation communications technology andeconomic growth in Sub-Saharan Africa a panel data


approachrdquo Telecommunications Policy vol 43 no 1pp 88ndash99 2019

[24] C Cheng M Chien and C Lee ldquoICT diffusion financialdevelopment and economic growth an international crosscountry analysisrdquo Economic Modelling vol 94 pp 662ndash6712020

[25] S Asongu and N M Odhiambo ldquoForeign direct investmentinformation technology and economic growth dynamics inSub-Saharan Africardquo Telecommunications Policy vol 44Article ID 101838 2020

[26] M Cardona T Kretschmer and T Strobel ldquoICT and pro-ductivity conclusions from the empirical literaturerdquo Infor-mation Economics and Policy vol 25 no 3 pp 109ndash125 2013

[27] K Vu P Hanafizadeh and E Bohlin ldquoICT as a driver ofeconomic growth a survey of the literature and directions forfuture researchrdquo Telecommunications Policy vol 44 no 2Article ID 101922 2020

[28] H Hamdi R Sbia and M Shahbaz ldquo+e nexus betweenelectricity consumption and economic growth in BahrainrdquoEconomic Modelling vol 38 pp 227ndash237 2014

[29] J E Payne ldquoA survey of the electricity consumption-growthliteraturerdquo Applied Energy vol 87 no 3 pp 723ndash731 2010

[30] I M Ouedraogo ldquoElectricity consumption and economicgrowth in Burkina Faso a cointegration analysisrdquo EnergyEconomics vol 32 no 3 pp 524ndash531 2010

[31] S-H Yoo and S-Y Kwak ldquoElectricity consumption andeconomic growth in seven south American countriesrdquo EnergyPolicy vol 38 no 1 pp 181ndash188 2010

[32] M Shahbaz C F Tang and M Shahbaz Shabbir ldquoElectricityconsumption and economic growth nexus in Portugal usingcointegration and causality approachesrdquo Energy Policyvol 39 no 6 pp 3529ndash3536 2011

[33] R Sbia M Shahbaz and I Ozturk ldquoEconomic growth fi-nancial development urbanisation and electricity consump-tion nexus in UAErdquo Economic Research-EkonomskaIstrazivanja vol 30 no 1 pp 527ndash549 2017

[34] J Wang J Zhao and H Li ldquo+e electricity consumption andeconomic growth nexus in China a bootstrap seeminglyunrelated regression estimator approachrdquo ComputationalEconomics vol 52 no 4 pp 1195ndash1211 2018

[35] R P Pradhan M B Arvin and N R Norman ldquo+e dynamicsof information and communications technologies infra-structure economic growth and financial development ev-idence from Asian countriesrdquo Technology in Society vol 42pp 135ndash149 2015

[36] M Osman G Gachino and A Hoque ldquoElectricity con-sumption and economic growth in the GCC countries paneldata analysisrdquo Energy Policy vol 98 pp 318ndash327 2016

[37] N Apergis and J E Payne ldquo+e electricity consumption-growth nexus renewable versus non-renewable electricity inCentral Americardquo Energy Sources vol 7 no 1ndash4 pp 423ndash4312012

[38] Y Wolde-Rufael ldquoElectricity consumption and economicgrowth in transition countries a revisit using bootstrap panelgranger causality analysisrdquo Energy Economics vol 44pp 325ndash330 2014

[39] W-C Lu ldquoElectricity consumption and economic growthevidence from 17 Taiwanese industriesrdquo Sustainability vol 9no 1 p 50 2017

[40] L G Brookes ldquoEnergy efficiency and economic fallacies areplyrdquo Energy Policy vol 20 no 5 pp 390ndash392 1992

[41] B Lin and H Zhao ldquoTechnological progress and energyrebound effect in Chinarsquos textile industry evidence and policy

implicationsrdquo Renewable and Sustainable Energy Reviewsvol 60 pp 173ndash181 2016

[42] Y Li L Sun T Feng and C Zhu ldquoHow to reduce energyintensity in China a regional comparison perspectiverdquo EnergyPolicy vol 61 pp 513ndash522 2013

[43] J Yin S Wang and L Gong ldquo+e effects of factor marketdistortion and technical innovation on Chinarsquos electricityconsumptionrdquo Journal of Cleaner Production vol 188pp 195ndash202 2018

[44] H An J Xu and X Ma ldquoDoes technological progress andindustrial structure reduce electricity consumption Evidencefrom spatial and heterogeneity analysisrdquo Structural Changeand Economic Dynamics vol 52 pp 206ndash220 2020

[45] J P Elhorst Spatial Econometrics From Cross-Sectional Datato Spatial Panels Springer Heidelberg Germany 2014

[46] L Anselin Spatial Econometrics Methods andModels KluwerAcademic Publishers Dordrecht NL USA 1998

[47] R M Baron and D A Kenny ldquo+e moderator-mediatorvariable distinction in social psychological research con-ceptual strategic and statistical considerationsrdquo Journal ofPersonality and Social Psychology vol 51 no 6 pp 1173ndash1182 1986

[48] B Li S Shi and Y Zeng ldquo+e impact of haze pollution onfirm-level TFP in China test of a mediation model of laborproductivityrdquo Sustainability vol 12 no 20 pp 1ndash21 2020

[49] Y Xu X Fan Z Zhang and R Zhang ldquoTrade liberalizationand haze pollution evidence from Chinardquo Ecological Indi-cators vol 109 no 2 pp 1ndash13 2020

[50] M H Pesaran ldquoGeneral diagnostic tests for cross sectiondependence in panelsrdquo Cambridge Working Papers in Eco-nomics No 0435 2004 httpwwweconcamacukresearch-filesrepeccampdfcwpe0435pdf

[51] B H Baltagi Q Feng and C Kao ldquoA Lagrange multiplier testfor cross-sectional dependence in a fixed effects panel datamodelrdquo Journal of Econometrics vol 170 no 1 pp 164ndash1772012

[52] J P Elhorst ldquoMatlab software for spatial panelsrdquo Interna-tional Regional Science Review vol 37 no 3 pp 389ndash4052014

[53] L-F Lee and J Yu ldquoEstimation of spatial autoregressive paneldata models with fixed effectsrdquo Journal of Econometricsvol 154 no 2 pp 165ndash185 2010

[54] J P Elhorst ldquoSpecification and estimation of spatial paneldata modelsrdquo International Regional Science Review vol 26no 3 pp 244ndash268 2016

[55] J P LeSage and R K Pace Introduction to Spatial Econo-metrics CRC Press New York NY USA 2009

[56] C Bao and R Liu ldquoElectricity consumption changes acrossChinarsquos provinces using a spatial shift-share decompositionmodelrdquo Sustainability vol 11 no 9 pp 2494ndash2515 2019

[57] S Liu and H Li ldquoDoes financial development increase urbanelectricity consumption Evidence from spatial and hetero-geneity analysisrdquo Sustainability vol 12 no 17 pp 7011ndash70172020

[58] M Bouznit M Pablo-Romero and A Sanchez-BrazaldquoResidential electricity consumption and economic growth inAlgeriardquo Energies vol 11 no 7 pp 1656ndash1717 2018

[59] Y Han J Shi Y Yang and Y Wang ldquoDirect rebound effectfor electricity consumption of urban residents in China basedon the spatial spillover effectrdquo Energies vol 12 no 11pp 1ndash16 2019

[60] P C Bosah S Li G K M Ampofo D A Asante andZ Wang ldquo+e nexus between electricity consumption eco-nomic growth and CO2 emission an asymmetric analysis


using nonlinear ARDL and nonparametric causality ap-proachrdquo Energies vol 13 no 5 pp 1ndash24 2020

[61] S A Solarin M Shahbaz H N Khan and R B Razali ldquoICTfinancial development economic growth and electricityconsumption new evidence from Malaysiardquo Global BusinessReview vol 17 no 6 pp 1480ndash1496 2019

[62] X Liu L Yu Y Zhang and Y Chao ldquoSpatial spillover effectof financial flexibility on investment in Chinarsquos conventionand exhibition listed companiesrdquo Mathematical Problems inEngineering vol 2020 Article ID 3926747 10 pages 2020

[63] X Yu M Shen W Shen and X Zhang ldquoEffects of landurbanization on smog pollution in China estimation ofspatial autoregressive panel data modelsrdquo Land vol 9 no 9pp 1ndash16 2020

[64] J A Hausman ldquoSpecification tests in econometricsrdquo Econ-ometrica vol 46 no 6 pp 1251ndash1271 1978

[65] A A Rafindadi and I Ozturk ldquoEffects of financial devel-opment economic growth and trade on electricity con-sumption evidence from Post-Fukushima Japanrdquo Renewableand Sustainable Energy Reviews vol 54 pp 1073ndash1084 2016

[66] X Liu T Sun Q Feng and D Zhang ldquoDynamic nonlinearinfluence of urbanization on Chinarsquos electricity consumptionevidence from dynamic economic growth threshold effectrdquoEnergy vol 196 no 4 pp 1ndash11 2020


mobile Internet access traffic consumption reached 5539billion GB a year-on-year increase of 1073 +ere is nodoubt that the use of ICT will play a more significant role inChinarsquos economic development Meanwhile the use of ICThas a profound impact on the demand and supply of elec-tricity According to the data released by the China Federationof Electricity Enterprises on January 29 2019 Chinarsquos total ECin 2018 was 684 trillion kWh an increase of 85 year onyear Among them in 2018 the EC of Chinarsquos secondaryindustry was 472 trillion kWh with a year-on-year growth of72 and that of the tertiary industry was 108 trillion kWhwith a year-on-year growth of 127 especially that of in-formation transmission software and information technologyservices with a year-on-year growth of 235 In order toalleviate the increasing pressure of electricity demand broughtby the development of ICT we should correctly coordinatethe relationship between the development of ICT and ECdemand and the premise of coordinating the relationship isto correctly understand the complex relationship betweenthem

At present some achievements have confirmed thedemand growth effect of ITC on EC [10ndash16] Unfortu-nately the existing empirical literature on the relationshipbetween ICT and EC does not consider the spatial de-pendence of EC Because these methods being used inexisting studies assume that the cross-section elements areindependent [14ndash16] the existing studies only considerthe direct impact of ICT on EC in decision-making unitsbut neglect the indirect effect of EC caused by the spilloverof ICT in spatial neighboring units which may under-estimate or overestimate the comprehensive impact ofICT on EC In addition the existing studies such asShahbaz et al [12] Saidi et al [14] and Afzal et al [16]regarded ICT and economic growth as independentvariables affecting EC and did not reveal the mediationrole of economic growth in the process of ICT promotingEC In this study we take China for example consider thecross-section dependence of EC variables and introducethe spatial modeling techniques and the mediation effectmodel to conduct an empirical test Our focus is con-centrated on the following aspects what role does ICTplay in Chinarsquos EC Is there a spatial spillover effect of theICTon EC Is there a significant heterogeneity effect of theinformation technology and the communication tech-nology on EC Are economic growth and the adjustmentof industrial structure the action channels of ICTaffectingelectricity consumption

+is paper provides freshly empirical evidence for peopleto better understand the direct effect and indirect effect ofICT affecting EC and their action channels and opens freshinsights for policymakers to make corresponding policyadjustments +e contribution of this paper lies in (1) theprevailing spatial panel modeling techniques are adopted toanalyze the relationship between ICT and EC in China +isstudy not only confirms the direct effect of ITC on EC butalso take into account the role of the spillover effect of ITCon EC in China (2) the analysis of mediation effect modeldemonstrates that there exists an incomplete mediatingeffect when ICT effect on EC through the channels of

economic growth and the adjustment of the industrystructure

+e rest of the study is organized as follows Section 2provides a literature review of ICT and EC Section 3presents the model and method Section 4 elaborates on thevariables and data Section 5 elaborates on the empiricalresults Section 6 presents conclusions and discussion

2 Literature Review

21 EmpiricalAnalyses on theDirectRelationshipbetween ICTandEC Some researchers pay attention to the effect of ITC onEC considering not only the contribution of ITC for thegrowth of the EC but also the electricity-saving capacity of ICTFor example Cho et al [17] utilized the logistic model to studythe effect of ICT investment electricity price and oil price onindustrial EC in South Korea +eir empirical analysis showsthat ICTinvestment inmanufacturing industries with EC couldpromote the transformation of input factor substitution fromlabor-intensive to electricity-intensive In addition the resultsshow that investment of ICT in certain manufacturing sectorshelps to reduce EC while investment of ICT in services andmost manufacturing sectors increases EC Sadorsky [6] in-vestigated the effect of ICTon EC in emerging economies Hisempirical results of the dynamic panelmodel showed that therewas a positive and statistically significant relationship betweenICT and EC

Schlomann et al [10] analyzed the net effect of ICT onfuture EC in Germany +ey found that there was greatpotential for technological progress in reducing the EC ofICT equipment Afzal and Gow [7] studied the effect of ICTon EC in the next 11 emerging economies during 1990ndash2014and their results showed that there was a significant positivecorrelation between ICT and EC Kishita et al [18] con-structed an electricity demand model for the telecommu-nications industry and used scenario planning methods toanalyze changes in EC in the telecommunications industryand the possibility of electricity-saving +eir results showedthat improving energy efficiency equipment and mobilecommunication equipment for Internet Protocol (IP)communications were the most effective measures to reduceelectricity demand With quarterly data from 1993q1 to2014q4 Faisal et al [15] used the ARDL approach to confirmthat there was a U-shaped relationship between Internet useand EC

However the existing literature only takes into accountthe direct effect of ICT on EC but neglects the spatial thespillover effect of ICT on EC In this study we consider thecross-section dependence of variables and use the spatialeconometric method to calculate the direct effect and thespillover effect of ITC on EC

22 -e Channel of ICT Influencing EC ICT not only has adirect effect on EC but also indirectly affects the final ECthrough different path of the mediation channels One of theimportant paths may be economic development includingeconomic growth and the adjustment of the industrialstructure Existing empirical literature whether based on











32 Model and Method





CD

2T

N(N minus 1)

1113971

1113944

Nminus 1

iminus 11113944

N

ji+1ρij

1113868111386811138681113868111386811138681113868111386811138681113868

1113868111386811138681113868111386811138681113868111386811138681113868 (1)


1

N(N minus 1)

1113971

1113944

Nminus 1

i11113944

N

ji+1Tρ2ij minus 11113872 1113873

minusN

2(T minus 1)

(2)


ρij 1113936

Tt1 1113954μit1113954μjt

1113936Tt1 1113954μ2it

1113868111386811138681113868111386811138681113868111386812

1113936Tt1 1113954μ2jt

11138681113868111386811138681113868

1113868111386811138681113868111386812 (3)







μit ψ 1113944N

j1wijμit + ]it

(5)




(6)





Loca

l ICT

Elec

tric

ity co

nsum

ptio

n

Mediatingeffect


Direct effect


Economic growth



Nei

ghbo

urin

g IC

T

Mediatingeffect

Direct effect



Economic growth








minus 1(Xβ + WXθ)


(7)


zY

zx1k


zxNk

1113890 1113891

zY1

zx1k


zxNk


zYN

zx1k


zxNk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦




⋮ ⋮ ⋮ ⋮


⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(8)



Ni1 1113936

Nj1 wijθkN





(11)



Lnctrit + ε3it

(12)



































































Spillover effect





























Data Availability




Acknowledgments




References


















































































32 Model and Method





CD

2T

N(N minus 1)

1113971

1113944

Nminus 1

iminus 11113944

N

ji+1ρij

1113868111386811138681113868111386811138681113868111386811138681113868

1113868111386811138681113868111386811138681113868111386811138681113868 (1)


1

N(N minus 1)

1113971

1113944

Nminus 1

i11113944

N

ji+1Tρ2ij minus 11113872 1113873

minusN

2(T minus 1)

(2)


ρij 1113936

Tt1 1113954μit1113954μjt

1113936Tt1 1113954μ2it

1113868111386811138681113868111386811138681113868111386812

1113936Tt1 1113954μ2jt

11138681113868111386811138681113868

1113868111386811138681113868111386812 (3)







μit ψ 1113944N

j1wijμit + ]it

(5)




(6)





Loca

l ICT

Elec

tric

ity co

nsum

ptio

n

Mediatingeffect


Direct effect


Economic growth



Nei

ghbo

urin

g IC

T

Mediatingeffect

Direct effect



Economic growth








minus 1(Xβ + WXθ)


(7)


zY

zx1k


zxNk

1113890 1113891

zY1

zx1k


zxNk


zYN

zx1k


zxNk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦




⋮ ⋮ ⋮ ⋮


⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(8)



Ni1 1113936

Nj1 wijθkN





(11)



Lnctrit + ε3it

(12)



































































Spillover effect





























Data Availability




Acknowledgments




References










































































CD

2T

N(N minus 1)

1113971

1113944

Nminus 1

iminus 11113944

N

ji+1ρij

1113868111386811138681113868111386811138681113868111386811138681113868

1113868111386811138681113868111386811138681113868111386811138681113868 (1)


1

N(N minus 1)

1113971

1113944

Nminus 1

i11113944

N

ji+1Tρ2ij minus 11113872 1113873

minusN

2(T minus 1)

(2)


ρij 1113936

Tt1 1113954μit1113954μjt

1113936Tt1 1113954μ2it

1113868111386811138681113868111386811138681113868111386812

1113936Tt1 1113954μ2jt

11138681113868111386811138681113868

1113868111386811138681113868111386812 (3)







μit ψ 1113944N

j1wijμit + ]it

(5)




(6)





Loca

l ICT

Elec

tric

ity co

nsum

ptio

n

Mediatingeffect


Direct effect


Economic growth



Nei

ghbo

urin

g IC

T

Mediatingeffect

Direct effect



Economic growth








minus 1(Xβ + WXθ)


(7)


zY

zx1k


zxNk

1113890 1113891

zY1

zx1k


zxNk


zYN

zx1k


zxNk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦




⋮ ⋮ ⋮ ⋮


⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(8)



Ni1 1113936

Nj1 wijθkN





(11)



Lnctrit + ε3it

(12)



































































Spillover effect





























Data Availability




Acknowledgments




References













































































minus 1(Xβ + WXθ)


(7)


zY

zx1k


zxNk

1113890 1113891

zY1

zx1k


zxNk


zYN

zx1k


zxNk

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦




⋮ ⋮ ⋮ ⋮


⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦

(8)



Ni1 1113936

Nj1 wijθkN





(11)



Lnctrit + ε3it

(12)



































































Spillover effect





























Data Availability




Acknowledgments




References





































































































































Spillover effect





























Data Availability




Acknowledgments




References



























































































































Spillover effect





























Data Availability




Acknowledgments




References



































































































Spillover effect





























Data Availability




Acknowledgments




References
























































































Spillover effect





























Data Availability




Acknowledgments




References



























































































Data Availability




Acknowledgments




References










































































Data Availability




Acknowledgments




References

































































































































Documents

Direct Effect and Spillover Effect of ICT on Electricity