Do Intangibles Contribute to Productivity Growth in East Asian Countries? Evidence from Japan and Korea

Transcription

1 RIETI Dscusson Paper Seres 15-E-055 Do Inangbles Conrbue o Producvy Growh n Eas Asan Counres? Evdence from Japan and Korea Hyunbae CHUN Sogang Unversy MIYAGAWA Tsuomu RIETI Hak Kl PYO Seoul Naonal Unversy TONOGI Konom Kanagawa Unversy The Research Insue of Economy Trade and Indusry hp://

2 RIETI Dscusson Paper Seres 15-E-055 May 2015 Do Inangbles Conrbue o Producvy Growh n Eas Asan Counres? Evdence from Japan and Korea * Hyunbae CHUN (Sogang Unversy) MIYAGAWA Tsuomu (Gakushun Unversy and RIETI) Hak Kl PYO (Seoul Naonal Unversy) TONOGI Konom (Kanagawa Unversy) Absrac Usng he Japan Indusral Producvy (JIP) and he Korea Indusral Producvy (KIP) daabases and oher prmary sascs n Japan and Korea we esmae nangble nvesmen n Japan and Korea a he ndusry-level. Comparng our esmaes from wo-counry daa we fnd ha he growh n nangble nvesmen n Korea has exceeded ha n Japan n he pas 30 years. Inangble nvesmen/gross value added (GVA) raos n he machnery ndusres n Japan are hgher han n Korea because Japanese machnery ndusres are research and developmen (R&D) nensve. On he oher hand raos n some servce ndusres n Korea are hgher han n Japan because Korean servce ndusres are nformaon and communcaons echnology (ICT)-nensve. When we conduc growh accounng analyss wh nangbles we fnd ha he conrbuon of nangble nvesmen o economc growh afer 1995 n Japan decreased sgnfcanly. In addon he conrbuon of nangbles o producvy growh n Japan afer 1995 s lower han no only Korea bu also he European Unon (EU) counres and he Uned Saes. The lack of synergy effecs beween ICT and nangbles n Japan may be he cause of low producvy growh n he 2000s. Keywords: ICT Inangble nvesmen Growh accounng Complemenary JEL classfcaon numbers: E01 E22 O31 O47 O53 RIETI Dscusson Papers Seres ams a wdely dssemnang research resuls n he form of professonal papers hereby smulang lvely dscusson. The vews expressed n he papers are solely hose of he auhor(s) and neher represen hose of he organzaon o whch he auhor(s) belong(s) nor he Research Insue of Economy Trade and Indusry. * Ths paper s wren beased on our presenaon a he 3rd World KLEMS Conference. We hank Professor Jorgenson (Harvard Unversy) Professor Tmmer (Unversy of Gronngen) and Professor Fukao (Hosubash Unversy) and he parcpans a he 3rd World KLEMS Conference a Tokyo on May for her helpful commens and dscussons. We also hank Dr. Fuja Presden of RIEYI and Dr. Morkawa Vce-Presden of RIETI for her helpful suggesons Excellen research asssance by Professor Hsa (Kanagawa Unversy) and Mr. Makno (Hosubash Unversy) s also apprecaed. Ths sudy s conduced as a par of he Projec Sudy on Inangble Asses n Japan underaken a Research Insue of Economy Trade and Indusry(RIETI) and suppored n par by a Gran-n-Ad for Scenfc Research from he Mnsry of Educaon Culure Spors Scence and Technology (No ) of Japan. The sudy on Korea was conduced hrough he KIP projec a Korea Producvy Cener suppored by Korean Mnsry of Trade Indusry and Energy. 1

3 1. Inroducon The ICT revoluon n he 1990s and he producvy growh caused n he US have nspred new neres amongs economss n explorng sources of growh. Table 1 shows he sandard growh accounng n Korea and Japan. In he manufacurng secor n boh counres TFP growh acceleraed afer 1995 whle he conrbuon of capal o economc growh declned. Ths has led economss o look for new sources of economc growh. Hall ( ) Bresnahan Brynjolfsson and H (2002) and Basu e al. (2003) emphaszed nangble asses -- ha are complemenary o ICT asses -- and play a crucal role n producvy mprovemen. However hey had o ndrecly esmae he role of nangble asses due o he challenges n measurng nangbles. 1 (Place Table 1 around here) Corrado Hulen and Schel (2009) (hereafer referred o as CHS) overcame hs challenge and measured nangble nvesmen a he aggregae level n he US for he frs me. Based on her esmaon hey found ha he rao of nangble nvesmen o GDP exceeded he rao of angble nvesmen o GDP n he early 2000s. Afer her success n measurng nangble asses many economss followed her mehod and esmaed nangble nvesmen n her own counres. 2 One of he major conrbuons of CHS s work was o show he conrbuon of nangbles (whch have been hdden n he conrbuons of capal asses and TFP) o economc growh. CHS argued ha one hrd of he producvy growh n he lae 1990s and he early 2000s s arbuable o he growh n nangble asses and he OECD (2013) emphaszed ha he effecs of nangbles on producvy growh are greaer han hose of angbles. The framework consruced by CHS has been developed furher manly n wo drecons. One s o measure nangble nvesmen by ndusry. Aggregaed daa does no provde enough dealed nformaon o conduc a producvy analyss. As Jorgenson Ho and Sroh (2005) Inklaar O Mahony and Tmmer (2005) and Fukao e al. (2011) suggesed here s a sgnfcan producvy gap beween ICT ndusres and non-ict ndusres. In addon even n ICT-nensve servce ndusres here s a producvy gap beween he US and Japan. To undersand he above gaps we requre nangble nvesmen daa a he ndusry level. Moreover he aggregae seres also consrans our analyss. The measured me seres nangble nvesmen daa are avalable for a mos 30 years. Ths sze of daa s no suffcen for several economerc analyses. As a resul economss sared measurng nangble nvesmen a he ndusry level n a few counres. Chun e al. (2012) esmaed nangble nvesmen by ndusry n Japan and Korea usng he Japan Indusral Producvy (JIP) daabase and he Korea Indusral Producvy (KIP) daabase and followng he framework developed by CHS (2009). In her paper hey found ha alhough he rao of nangble nvesmen o GDP n Japan s hgher han n Korea he gap n he raos beween wo counres has 1 Myagawa and Km (2008) also consdered he role of nangble asses on producvy mprovemen hrough he ndrec measuremen n nangble asses by usng frm-level daa. 2 Burnes and McClure (2009) for Ausrala Corrado e al. (2013) for he EU counres Fukao e al. (2009) for Japan and Pyo Chun and Rhee (2010) for Korea. 1

4 conraced n many ndusres. Usng he same daa n Korea Chun and Nadr (2013) conduced growh accounng ncludng nangbles. They dvded 27 ndusres no nangble-nensve ndusres and non-nangble-nensve ndusres and found ha he producvy growh n he former ndusres was hgher han n he laer. They also showed ha nangbles are key drvers n producvy growh n he nangble-nensve ndusres. Myagawa and Hsa (2013) examned he effecs of nangble nvesmen on producvy growh by usng daa on nangbles n Japan. They found posve and sgnfcan effecs of nangble nvesmen on producvy n ICT nensve ndusres. Ther resuls mply here are complemenares beween ICT echnology and nangbles. In Europe Nebel e al. (2013) esmaed nangble nvesmen by 11 secors n 10 EU counres by usng n INTAN-Inves daabase. They found ha he conrbuons of nangbles o producvy growh are hghes n he manufacurng and fnancal secors. Followng he CHS approach Crass e al. (2015) also esmaed nangble nvesmen by sx secors n Germany. They also suppored he vew ha he manufacurng and fnancal and busness servce secors are nangble-nensve secors. In hese secors nnovave capal conrbues o labor producvy growh. Comparng hs wh nangble nvesmen n he UK developed by Gl and Haskel (2008) hey found ha he UK nvess more n sofware frm-specfc human capal and organzaonal srucure han Germany whle Germany nvess more n R&D and adversng han he U.K. The second drecon hs framework was developed owards s o reformulae CHS s componens of nangbles. When CHS sared o measure nangbles only sofware nvesmen was ncluded as capal formaon n he SNA based on he 93 SNA manual. However he 08 SNA manual recommends ncludng R&D nvesmen as capal formaon and has been accouned for as capal formaon n many advanced counres. Therefore Corrado e al. (2014) measure nangble nvesmen n he EU counres reclassfyng angbles and nangbles: non-ict angbles ICT R&D and non-r&d nangbles. Based on hese new classfcaons hey examne complemenares beween capal asses n parcular ICT asses and R&D and ICT asses and non-r&d nangbles. Followng he wo developmens of CHS s sudy we examne nangble nvesmen by ndusry n Japan and Korea and conrbuons of nangbles o producvy growh. Alhough Japan and Korea are he mos advanced counres n Eas Asa he growh pahs n he recen pas of he wo are dfferen. The Japanese economy has suffered from he long-erm sagnaon of he economy snce he collapse of he bubble economy n On he oher hand he Korean economy has recorded hgher economc growh han oher advanced economes -- even hough experenced a serous downurn due o he fnancal crss n Therefore we expec ha we wll fnd he dfferen effecs of nangbles o economc growh n he wo counres. 3 In he nex secon we nroduce new asse classfcaons developed by Corrado e al. (2014) and explan how o measure hese asses n Japan and Korea. In he hrd secon we show some feaures of nangble nvesmen a he ndusry level n Japan wh some comparsons wh esmaes n Korea. As our ndusry classfcaon s more sophscaed han hose n prevous leraure we are able o provde useful nformaon 3 Usng he framework developed by McGraan and Presco ( ) Myagawa and Takzawa (2011) showed ha he conrbuon of nangbles o economc growh n Korea s lareger han ha n Japan. 2

5 for he composon of capal asses by ndusry. In he fourh secon usng growh accounng wh nangbles we examne he conrbuons of nangble asses o producvy mprovemen. In he ffh secon we examne he correlaons beween ICT asses and nangbles. In he las secon we summarze our resuls. 2. Classfcaon and measuremen of Capal Asses Economss have also developed classfcaons of nangbles. In Korea Pyo (2002) defned nangbles as compuer sofware mneral exploraon culural producs such as eneranmen leraure orgnal fne ars and unproduced nangble asses such as paens and lcensng of moble communcaon ec. van Ark (2004) argued ha human capal knowledge-based capal organzaonal capal markeng of new producs and socal capal n addon o ICT capal should be aken no accoun n he knowledge economy. The caegory of nangbles developed by CHS ha we use o measure nangbles s broader han Pyo (2002) and narrower han van Ark (2004). Inangble asses n CHS conss of compuerzed nformaon nnovave propery and economc compeences. However as Corrado e al. (2014) negrae angbles and nangbles and reclassfy no four ypes of asses: non-ict angbles ICT R&D and non-r&d nangbles we wll explan how we measure hese asses by ndusry. Our measuremen depends manly on he JIP daabase for Japan and he KIP daabase for Korea and we adjus ndusry classfcaons n boh counres o 27 ndusry classfcaons because boh ndusry classfcaons are slghly dfferen beween Japan and Korea Measuremen of ICT nvesmen and non-ict angbles nvesmen As Fukao e al. (2011) showed ICT asses conss of compung equpmen communcaon equpmen and sofware. We oban he daa for he frs wo asses from he JIP daabase and he KIP daabase. The las asse corresponds o compuerzed nformaon defned n CHS Compuerzed nformaon consss of cusom and packaged sofware and own accoun sofware. In Chun e al. (2012) he daa n cusom and packaged sofware n Japan were obaned from he Japanese SNA. However snce 2012 he Japanese governmen has publshed own accoun sofware nvesmen as well as cusom and packaged sofware nvesmen. Therefore for he aggregae sofware nvesmen no only Korea bu also Japan follows he SNA daa. To measure sofware nvesmen by ndusry we allocae he oal sofware nvesmen no each ndusry by usng he Fxed Capal Formaon Marx (FCFM) n each counry. Non-ICT angbles are all angbles excep compung equpmens and communcaon equpmen. We also oban hs daa by ndusry from he JIP daabase and he KIP daabase Measuremen of R&D nvesmen Innovave propery defned n CHS consss of scence and engneerng R&D mneral exploaon copyrgh and lcense coss and oher produc developmen desgn and research expenses. Corrado e al. (2014) defned he frs wo componens of nnovave propery as R&D nvesmen whch has been already couned n SNA n advanced 4 In he JIP and KIP daabases we measure capal sock whch frms own and do no coun capal sock whch frms ren. In he measuremen of nangble nvesmen by ndusry we ake he same approach as he capal accoun n he JIP and KIP daabases. 3

6 counres. In Japan we esmae scence and engneerng R&D coss by usng he Survey of Research and Developmen publshed by he Sascal Bureau of he Mnsry of Inernal Affars and Communcaons. In Korea we esmae scence and engneerng R&D coss by usng he Survey of Research and Developmen publshed by Mnsry of Educaon Scence and Technology. In boh counres we focus only on R&D expendures by he prvae secor. We oban he daa for nvesmen n mneral exploaon from he SNA n Japan and Korea Measuremen of non-r&d nangbles Inangbles excep sofware and R&D are consdered o be non-r&d nangbles n Corrado e al. (2014). Non-R&D nangbles conss of copyrgh and lcense coss oher produc developmen desgn and research expenses brand equy frm specfc human capal and organzaonal srucure. Copyrgh and lcense coss are consruced from he IO able n each counry. They conss of he nermedae npus from he publshng ndusry and he vdeo pcure sound nformaon characer nformaon producon and dsrbuon ndusry. In he new esmaes by Corrado e al. (2013) expendures n eneranmen and arsc orgnals are measured nsead of copyrgh and lcense coss. However we use only nermedae npus from he vdeo pcure sound nformaon characer nformaon producon and dsrbuon ndusry. As for he esmaon of produc developmen n fnancal servces he esmaon mehod by CHS was very conroversal because hey assumed ha 20 percen of nermedae npus produced by fnancal servces can be assumed o be expendures n nangble asses. Recenly Corrado e al. (2013) suggesed ha he cos of new produc developmen n fnancal servces was almos 8 percen of he compensaon of hgh sklled workers n he fnancal ndusry o harmonze her esmae wh esmaes n EU counres by COINVEST and INNODRIVE projecs. Thus followng Corrado e al. (2013) s suggesons we also assume ha 8 percen of he compensaon of college graduaes n he fnancal and he nsurance ndusres can be regarded as expendures n nangble asses n hese ndusres. Brand equy frm specfc human capal and organzaonal srucures caegorzed as economc compeences n CHS. To measure brand equy we oban he oupu daa of he adversng ndusry and allocae no 27 ndusres by usng he IO able n he JIP and KIP daabases. In esmang frm specfc human capal we focus on off-he-job-ranng coss. In Japan we esmae he rao of off-he-job ranng coss o oal labor coss from he General Survey on Workng Condons by ndusry publshed by he Mnsry of Healh Labor and Welfare. We esmae off-he-job ranng coss by frms by ndusry by mulplyng hs rao by oal labor coss n he JIP daabase. For he opporuny cos of off-he-job ranng n erms of workng hours los we use he resuls obaned by Ook (2003). Ook calculaed he average rao of he opporuny cos of off-he-job ranng o drec frm expenses for ranng n 1998 for he enre busness secor usng mcro-daa of he Survey on Personnel Resrucurng and Vocaonal Educaon/Tranng Invesmen n he Age of Performance-based Wage Sysems (Gyosek-shug Jda no Jnj Ser o Kyoku/Kunren Tosh n Kansuru Chosa) conduced by he Japan Insue for Labor Polcy and Tranng. The value was 1.51 and we use hs same value o esmae he opporuny cos. 4

7 In Korea employer-provded ranng coss are obaned from he Repor on Labor Cos of Enerprse Survey publshed by he Mnsry of Labor. The survey ncludes ranng coss only for esablshmens wh 30 or more employees. We assume ha esablshmens wh fewer han 30 employees spend 50% (relave o he oal labor coss) less han hose wh 30 or more employees and we esmae he employer-provded ranng coss for all esablshmens n each ndusry. Toal ranng coss are defned as he sum of he drec coss and he opporuny coss of ranng. As he opporuny coss of ranng are no avalable we assume ha he drec coss of ranng are equal o he opporuny coss. To esmae expendures no organzaonal srucure CHS assumed ha 20% of he remuneraon of execuves can be consdered nangble asses for an organzaonal srucure. However n Japan we beleve 9% s a more accurae number because only 9% of he oal workng me of execuves s spen on organzaonal reform and he resrucurng of organzaon accordng o Robnson and Shmzu (2006). We calculae he rao of he remuneraon of execuves o value added usng he Fnancal Saemens Sascs of Corporaons by Indusry publshed by he Mnsry of Fnance. Then we fnd he expendure for he organzaonal srucure by ndusry by mulplyng hs rao o value added n he JIP daabase. In Korea consulng coss are consdered o be frm-specfc nvesmens n organzaonal resources. To oban expendures on consulng a he aggregae level we use he gross oupu of consulng ndusry. The ndusry-level consulng coss are also esmaed from he IO ables Inangble Invesmen n Japan and Korea Our measure of nangble nvesmen (excludng hardware ICT nvesmen) n Japan and Korea s shown n Fgure 1. 6 In Japan he amoun of nangble nvesmen was abou 40 rllon yen n I peaked n 2007 afer whch declned due o he depresson caused by he Global Fnancal Crss ha occurred n On he oher hand nangble nvesmen n Korea has no been affeced by busness cycles excep he Asan Fnancal Crss n Growng rapdly reached abou 80 rllon won n Whle growh n ICT nvesmen s he hghes (5% per annum from 1995 o 2010) among he hree componens n Japan growh n R&D nvesmen s he hghes (11% per annum from 1995 o 2010) n Korea. R&D nvesmen n Japan has no ncreased n he pas 15 years. We also fnd a gap n he growh rae n nvesmen n non-r&d nangbles beween Japan and Korea. In he pas 15 years s annual growh rae n Korea was 6.5% whle s growh rae n Japan was only 1.1%. However he rao of nangble nvesmen o GDP n Japan (7.5%) s sll hgher han n Korea (7.4%) n he 2000s. (Place Fgure 1 around here) 5 The measuremen of each componen n nangble asses are explaned n Chun e al. (2012) and Chun and Nadr (2013). 6 The ndusry-level Inangble nvesmen daa n Japan s uploaded a he followng webse: hp:// 5

8 In Fgure 2 we compare nangble nvesmen by ndusry and by asse beween Japan and Korea n In he manufacurng secor he composon of he asses n Japan s smlar o Korea n he sense ha he share of R&D nvesmen s he larges. However n he servce secor sofware nvesmen n Korea s greaer han n Japan n some ndusres such as nformaon and communcaon and busness servce ndusres. In he non-marke secor such as educaon and healh and socal work ndusres sofware nvesmen n Korea s also larger han n Japan. In addon n busness servce educaon and culure and eneranmen servces ndusres nvesmen n economc compeences n Korea s greaer han n Japan. (Place Fgure 2 around here) We also compare nangble nvesmen n Japan and Korea wh Germany and he UK n Table 2. 7 Table 1 shows he nangble nvesmen/gross oupu rao by secor n four counres. The nangble nvesmen/oupu raos n European counres are hgher han hose n Japan and Korea. In parcular nangble nvesmen/oupu raos n he servce secor are much hgher han hose n oher counres. 8 (Place Table 2 around here) We esmae capal sock n nangble asses based on he measuremen of expendures n nangble asses. The capal formaon seres n nangble nvesmen s measured n nomnal erms. Usng he deflaor by asses n he JIP and KIP daabases we consruc a real capal formaon seres n nangble asses. Then we generae he capal formaon seres by usng he perpeual nvenory mehod o fnd real capal sock n nangble asses. To measure capal sock n nangbles we use he deprecaon raes used n Corrado e al. (2013). To fnd he composon of capal asses by ndusry we have he share of each asse by ndusry n Fgure 3. The share of non-ict angbles for mos ndusres s he larges among he four ypes of asses n boh counres bu we fnd some dfferences n he shares of oher asses beween Japan and Korea. In he manufacurng secor Japanese machnery ndusres such as machnery equpmen elecrcal and elecrc equpmen precson nsrumens and ransporaon equpmen are more R&D-nensve han Korean machnery ndusres. On he oher hand mos Korean servce ndusres such as nformaon and communcaon busness servce educaon and culure and eneranmen are more ICT-nensve han Japanese servce ndusres. (Place Fgure 3 around here) 4. The Role of Inangble Asses on Producvy Improvemen 7 The daa n European counres are obaned from Crass e al. (2015). 8 The nangble nvesmen/oupu raos n he manufacurng secor n Korea seem o be very low. These low raos n Korea are caused by low value added/gross oupu rao n he manufacurng secor. When we look a he nangble nvesmen/gross value added rao n he manufacurng secor n Fgure 2 he raos n machnery ndusres are very hgh. 6

9 7 Usng our esmaes on nangbles we conduc a growh accounng ncludng nangble asses o examne her mpacs on producvy growh. A feaure of he new growh accounng s ha we can break down he conrbuon of capal asses no four ypes of asses as shown n Fgure 3. Based on he asse classfcaon n Fgure 3 he producon funcon n ndusry for value-added V can be expressed as (1) ) ( Z R C K L F A V = where L s labor npu K C R Z are non-angble asses ICT asses R&D asses and non-r&d nangble asses respecvely. A s TFP. Value-added (V ) s adjused for nangbles as H Y V H P Y P V P + Y s value added before consderng nangble asses H s nangble nvesmen ha s no ncluded n he SNA Then we may express a growh accounng relaon from equaon (1) as (2) a z V Z Z F r V R R F c V C C F k V K K F l V L L F v = where j j N j N j j Xj j Xj X X X w X w X X x 1 1 = = = = (X=LKC R or Z) j X represens an npu facor ha has he characer j n ndusry and j w represens he prce of an npu facor whch has he characer j n ndusry. As for he esmaon sraegy for growh accounng we can esmae he producon funcon n equaon (1) usng he growh raes of he value added and npus whle nerpreng coeffcens on each npu as a share. Alernavely we can calculae he share of s npu usng he daa on labor ncome and capal spendng and use he growh raes of value added and npu o conduc he above growh accounng exercse. In our growh accounng we ake he alernave approach. To measure he share of capal ncome n he alernave approach and capal servce we calculae he cos of capal for asse j as follows (3) ) ) /(1 ( j j j j u p CC + = p δ where p j s he prce of nvesmen goods for capal j s he nomnal neres rae (calculaed as he weghed sum of he bond rae and long-erm prme rae) 9. δ j s he deprecaon rae of capal j p j s he expeced rae of ncrease n he prce of nvesmen goods. We assume perfec foresgh for he expeced rae of ncrease n he 9 The weghs used here are he raos of deb o oal asses by ndusry.

10 prce of nvesmen goods and use he growh rae of he prce beween hs perod and he nex perod. We can calculae he cos of capal (CC) for each ype of angble and nangble asses. 10 Table 3 shows he resuls of growh accounng based on equaon (2) n Japan and Korea. In Table 3 we fnd ha he growh rae n he marke economy slowed down drascally afer 1995 n boh counres due o he fnancal crses n he lae 1990s. Whle all producon facors ncludng nangbles nduced he slowdown n he growh rae afer 1995 n Japan he conrbuon rae of R&D asses o value added growh n Korea has kep a consan rae from 1985 o (Place Table 3 around here) When we compare he manufacurng and he servce secors n Japan he growh rae of value added n he servce secor has slowed more han n he manufacurng secor. Whle boh he non-ict angbles and R&D asses and TFP growh are major conrbuors o value-added growh n he manufacurng secor we fnd a low conrbuon of non-r&d asses and a negave conrbuon of TFP growh o value-added growh n he servce secor afer In Korea value added growh n boh secors has declned dramacally afer However he conrbuon rae of R&D asses o value added growh has kep he same pace snce Whle TFP growh rae n he manufacurng secor acceleraed afer 1995 TFP growh rae n he servce secor urned o be negave afer 1995 as shown n he growh accounng n Japan. Nex we compare a growh accounng wh nangbles o a radonal growh accounng shown n Table 1 n boh counres. The value added growh n a radonal growh accounng s dfferen from n he growh accounng wh nangbles as spendngs n nangbles are reaed as nvesmens raher han nermedae expenses and hus are par of value-added. Due o he ncluson of nangbles no capal npu he conrbuon of capal npu n he growh accounng wh nangbles s greaer han n he radonal growh accounng. In conras TFP growh n he growh accounng n Table 3 s smaller han ha n Table 1 excep for he case of he servce secor afer 1995 n Korea. In parcular TFP growh subsanally decreased n he Japanese manufacurng durng he perod when nangbles were rapdly accumulaed. Lasly we compare our growh accounng resuls wh hose for he US and Europe n Corrado e al. (2013) ha also ncorporaes nangble asses. Corrado e al. (2013) broke down he labor producvy growh rae whch was measured usng hours worked no angble asses capal deepenng rae of nangble asses change n labor composon and TFP growh rae. We follow her mehod n breakng down he labor producvy growh rae by usng our daa. Table 4 shows he nernaonal comparson of growh accounng afer akng nangble asses no accoun. We fnd ha Japanese capal deepenng rae of nangble asses s 0.2% and lower han he average nernaonal sandard. On he oher hand he 10 We ake he effecs of capal ncome ax on cos of capal no accoun only n he case of compuerzed nformaon because oher nangbles are reaed as nermedae npus n he curren fnancal saemen. 8

11 capal deepenng rae of nangbles n Korea s 0.6% and s slghly hgher han he average rae n European counres. The share of capal deepenng of nangble asses n he labor producvy growh rae s 9.5% n Japan and 13% n Korea compared o 33.7% n he US and 19.9% n he EU respecvely. Afer 1995 he accumulaon of nangble asses played a key role n producvy growh n developed counres. On he oher hand he labor producvy growh n Japan afer 1995 has been arbued o a composonal shf n he labor marke -- an ncrease n hgher qualy labor due o a hgher demand for hgher educaon. However here s a lm o hs rend and once he number of people pursung hgher educaon hs he celng hs composonal effec would be mued. In ha sense s necessary o accumulae nangble asses up o he level comparable o oher developed counres. In he case of Korea as angble capal deepenng and TFP growh are wo man conrbuors o labor producvy growh. The source of labor producovy growh n he Korean economy has changed from massve angble capal nvesmens n he rapd growh perod of 1970s and 1980s oward nangble nvesmen and TFP growh afer he 1990s. However he conrbuon of nangbles o labor producvy growh n Korea s sll relavely small compared o hose n he US and European counres. (Place Table 4 around here) 5. Correlaons beween ICT and R&D and ICT and non-r&d nangbles Bresnahan Brynjolfsson and H (2002) and Basu e al. (2003) argued ha nangbles play a complemenary role on he effecs of ICT asses on producvy growh. Our growh accounng n Table 3 also sugges ha ICT asses and non-r&d asses moved ogeher. However he growh accounng exercse only capures he ndependen conrbuon of each asse. Then we examne he correlaons beween ICT and R&D and ICT and non-r&d nangbles n hs secon. We ake he fve-year movng average growh of ICT asses R&D asses and non-r&d nangbles n Japan and Korea from 1990 o 2010 and examne correlaons among hese asses. Fgure 4 shows he correlaons beween growh n ICT and R&D asses n Japan and Korea. Whle we fnd posve correlaons beween he wo asses n Japan and Korea. However as shown n Table 5 hese posve correlaons are affeced by he srong posve correlaons beween he wo asses n he1990s. The correalons n Japan urned o be negave n he 2000s. In Korea alhough he posve correlaons were kep n he 2000s hey werre no sgnfcan. (Place Fgure 4 and Table 5 around here) Fgure 5 shows he correlaons beween ICT and non-r&d nangbles. As shown n Table 4 ICT s posvely and sgnfcanly correlaed wh non-r&d nangbles only n he 1990s n boh counres. In he 2000s he negave correlaons beween ICT and non-r&d nangbles n boh counres were found. Alhough hese fgures do no show any causales among asses low producvy growh n he 2000s n Japan may be parly caused by he lack of synergy effecs among ICT asses R&D asses and non-r&d asses. 9

12 (Place Fgure 5 around here) We also examne he cross-seconal correlaons beween ICT and nangbles. Fgure 6 shows cross-seconal correlaons beween ICT and R&D n Japan and Korea. 11 The correlaon beween wo asses n Japan s very week (r=0.06) whle we fnd he hgh posve correlaon (r=0.77) n Korea. Ths fndng mply ha ICT nensve ndusres nves n R&D aggressvely n Korea. On he oher hand ICT nvesmen s conduced ndependenly from R&D nvesmen. (Place Fgure 6 around here) Fgure 7 whch shows he complemenary beween ICT and non-r&d nangbles s smlar o Fgure 6. In Korea ICT nensve ndusres nves n non-r&d nangbles aggressvely. On he oher hand growh n ICT asses s no assocaed wh growh n non-r&d asses. (Place Fgure 7 around here) 6. Concludng Remarks Based on he framework of Corrado Hulen and Schel ( ) and Corrado e al. (2014) we esmaed nangble nvesmen by ndusry n Japan and Korea usng he JIP and KIP daabases. Comparng nangble nvesmen n Japan wh ha n Korea we fnd ha Japanese growh n nangble nvesmen slowed due o he fnancal crss ha occurred n 1997 and urned o be negave afer he World Fnancal Crss whle Korean nangble nvesmen has grown a a hgh pace. Whle only sofware nvesmen grew afer 1995 n Japan boh R&D and sofware nvesmens n Korea have rapdly grown. As a resul he gap n he rao of nangble nvesmen o GVA beween Japan and Korea subsanally reduced n he 2000s. When we examne nangble nvesmen by ndusry he rao of nangble nvesmen o GVA n Japan s hgher han n Korea n many machnery ndusres because R&D nvesmen o GVA raos n hese ndusres n Japan s hgher han n Korea. However n he servce secor he raos n Korea are greaer han hose n Japan because he Korean servce secor s more ICT-nensve han he Japanese servce secor. Usng nangble nvesmen daa we consruc capal sock n Japan and Korea by ndusry usng he perpeual nvenory mehod. Examnng he composon of angble and nangble capal we fnd ha machnery ndusres are R&D-nensve and he servce ndusres are more ICT-nensve alhough non-ict angbles domnae and ake up he larges share n mos ndusres n boh counres. We conduced growh accounng analyss wh nangbles. Our resuls n growh accounng showed ha he conrbuons of nangble asses o economc growh afer 11 In Fgure 6 we examne he cross-seconal correlaon n he manufacurng secor because R&D nvesmen s no conduced n some ndusres n he servce secor. A numbers n Fgures 6 and 7 corresponds o each ndusry. The correspondng able beween number n Fgures 6 and 7 and ndusry classfcaon s shown n Appendx able. 10

13 1995 are lower han hose before 1995 n Japan. In parcular he non-r&d nangbles have no conrbued o economc growh afer 1995 n Japan because expendures on off-he-job ranng has decreased rapdly due o he harsh resrucurng. On he oher hand he conrbuons of R&D asses o economc growh are consan from 1985 o 2010 n Korea. The Korean economc growh rae decreased afer 1995 bu he role of R&D asses o economc growh has ncreased. Comparng growh accounng wh radonal accounng we fnd ha he conrbuon of capal asses o economc growh and TFP growh n he radonal growh accounng ncludes he conrbuons of accumulaon n nangble asses. We also fnd ha he conrbuon of nangble capal accumulaon o producvy growh s weaker n Japan han n oher developed counres. On he oher hand he conrbuon of nangbles o producvy growh n Korea s smlar o hose n European counres. Fnally we examne he correlaon beween ICT asses and nangbles. Whle we fnd posve correlaons beween ICT asses and nangbles n he 1990s ICT s no posvely correlaed wh nangbles n Japan n he 2000s. Low producvy growh n Japan n he 2000s may be due o he lack of he synergy effecs of ICT asses and nangbles. Ths mplcaon s confrmed by cross-seconal correlaons beween ICT and nangbles. However o examne causales among ICT asses and nangbles more sophscaed analyss s requred. Our sudy suggess wo mporan polcy mplcaons for he long-erm producvy growh of he Japanese and Korean economes. Frs snce he conrbuons of nangbles o economc growh n boh counres are no hgher han hose n oher advanced counres here s room o mprove labor producvy hrough nangble nvesmen. Second our resuls shed lgh on complemenares among ICT asses and nangbles suggesed by Basu e al. (2003) Corrado e al. (2013) Myagawa and Hsa (2013) and Chun and Nadr (2013). In he case of Japan he lack of hese complemenares may be a facor n he low producvy growh n he 2000s. Therefore polces promong nangble nvesmen should be accompaned wh polces promong ICT nvesmen. 11

14 Appendx able: Indusry classfcaon of he paper Indusry classfcaon 1 Agrculure foresry and fshng 2 Mnng and quarryng 3 Food beverages and obacco 4 Texles and leaher 5 Wood paper and prnng 6 Peroleum coal and chemcals 7 Non-meallc mneral producs excep peroleum and coal 8 Meal Fabrcaed meal producs 9 Machnery equpmen 10 Elecrcal and elecronc equpmen 11 Precson nsrumens 12 Transpor equpmen 13 Furnure and oher manufacurng ndusres 14 Elecrcy gas and waer supply 15 Consrucon 16 Wholesale and real rade 17 Resaurans and hoels 18 Transpor and sorage 19 Fnancal nermedaon 20 Real esae and renng 21 Informaon and communcaon 22 Busness servces 23 Publc admnsraon and defense 24 Educaon 25 Healh and socal work 26 Culure and eneranmen servces 27 Oher servce acves 12

15 References Barnes Paula and Andrew McClure (2009) Invesmens n Inangble Asses and Ausrala s Producvy Growh Producvy Commsson Saff Workng Paper. Basu Susano John G. Fernald Ncholas Oulon and Sylaja Srnvasan (2003) The Case of he Mssng Producvy Growh: Or Does Informaon Technology Explan Why Producvy Acceleraed n he Uned Saes bu no n he Uned Kngdom? n Mark Gerler and Kenneh Rogoff (eds.) NBER Macroeconomcs Annual Cambrdge MA: MIT Press Bresnahan Tmohy Erk Brynjolfsson and Lorn H (2002) Informaon Technology Workplace Organzaon and he Demand for Sklled Labor: Frm-Level Evdence Quarerly Journal of Economcs 117(1) Chun Hyunbae Kyoj Fukao Shoch Hsa and Tsuomu Myagawa (2012) The Measuremen of Inangble Invesmen by Indusry and Is Role n Producvy Improvemens wh Some Comparave Sudes beween Japan and Korea RIETI Dscusson Paper Seres 12-e-037. Chun Hyunbae and M. Ishaq Nadr (2013) Inangble Invesmen and Changng Sources of Growh n Korea mmeo. Corrado Carol Charles Hulen and Danel Schel (2005) Measurng Capal and Technology: An Exended Framework n Carol Corrado John C. Halwanger and Danel E. Schel (eds.) Measurng Capal n he New Economy Chcago: Unversy of Chcago Press Corrado Carol Charles Hulen and Danel Schel (2009) Inangble Capal and U.S. Economc Growh Revew of Income and Wealh 55(3) Corrado Carol Jonahan Haskel Cecla Jona-Lasno and Massmlano Iomm (2013) Innovaon and Inangble Invesmen n Europe Japan and he US Dscusson Paper 2013/1 Imperal College Busness School. Corrado Carol Jonahan Haskel Cecla Jona-Lasno and Massmlano Iomm (2014) Inangbles and Indusry Producvy Growh: Evdence from EU mmeo. Crass Drk Georg Lch and Bena Peers (2015) Inangble Asses and Invesmens a he Secor level Emprcal Evdence for Germany n Ahmed Bounfour and Tsuomu Myagawa (eds.) Inangbles Marke Falure and Innovaon Performance Sprnger. Fukao Kyoj Tsuomu Myagawa Kenaro Muka Yuko Shnoda and Konom Tonog (2009) Inangble Invesmen n Japan: Measuremen and Conrbuon o Economc Growh Revew of Income and Wealh 55(3) Fukao Kyoj Tsuomu Myagawa Hak K. Pyo and Keun Hee Rhee (2011) Esmaes of Mul-Facor Producvy ICT Conrbuons and Resource Reallocaon Effecs n Japan and Korea n Malde Mas and Rober Sehrer (eds.) Indusral Producvy n Europe: Growh and Crss Chelenham UK: Edward Elgar Publshng. Gl Valenna and Jonahan Haskel (2008) Indusry-level Expendure on Inangble Asses n he UK hp:// DTI_Nov08-1.pdf. Hall Rober (2000) E-Capal: The Lnk beween he Sock Marke and he Labor Marke n he 1990s Brookngs Papers on Economc Acvy

16 Hall Rober (2001) The Sock Marke and Capal Accumulaon Amercan Economc Revew 91(5) Inklaar Rober Mary O Mahony and Marcel Tmmer (2005) ICT and Europe s Producvy Performance: Indusry-Level Growh Accoun Comparsons wh he Uned Saes Revew of Income and Wealh 51(3) Jorgenson Dale Mun Ho and Kevn Sroh (2005) Informaon Technology and he Amercan Growh Resurgence Cambrdge MA: MIT Press. McGraan Ellen and Edward C. Presco (2005) Expensed and Swea Equy Federal Reserve Bank of Mnneapols Workng Paper no McGraan Ellen and Edward C. Presco (2010) Unmeasured Invesmen and he Puzzlng U.S. Boom n he 1990s Amercan Economc Journal: Macroeconomcs 2(1) Myagawa Tsuomu and Shoch Hsa (2013) Esmaes of Inangble Invesmen by Indusry and Producvy Growh n Japan Japanese Economc Revew 64(1) Myagawa Tsuomu and YoungGak Km (2008) Measurng Organzaonal Capal n Japan: An Emprcal Assessmen Usng Frm-Level Daa Seoul Journal of Economcs 21(1) Myagawa Tsuomu and Mho Takzawa (2011) Producvy Dfferences beween Japan and Korea and he Role of Inangble Asses Kazum Asako Nobuo Iduka and Tsuomu Myagawa (eds.) Grea Recessons n he Global Economy and Busness Cycle Analyss Tokyo: The Unversy of Tokyo Press (n Japanese). Nebel Thomas Mary O Mahony and Maranne Saam (2013) The Conrbuon of Inangble Asses o Secoral Producvy Growh n he EU ZEW Dscusson Paper No OECD (2013) New Sources of Economc Growh. Ook Ech (2003) Performance-based Wage Sysems and Vocaonal Educaon/Tranng Invesmen (Gyosek-shug o Kyoku Kunren Tosh) Implcaons of Performance-based Wage Sysems for Indvduals and Organzaons (Ko o Soshk no Sekashug) n K. Konno (ed.) Chuo Keza Sha Tokyo (n Japanese). Pyo Hak K. (2002) A Sudy on Indrec Esmaon of Inangble Asses Korea Naonal Sascal Offce Repor (n Korean). Pyo Hak K. Hyunbae Chun and Keun Hee Rhee (2010) The Producvy Performance n Korean Indusres ( ): Esmaes from KIP Daabase presened a RIETI/COE H-Sa Inernaonal Workshop on Esablshng Indusral Producvy Daabase for Chna (CIP) Inda (IIP) Japan (JIP) and Korea (KIP) held on Ocober Robnson Parca and Norhko Shmzu (2006) Japanese Corporae Resrucurng: CEO Prores as a Wndow on Envronmenal and Organzaonal Change The Academy of Managemen Perspecves 20(3) Van Ark Bar (2004) The Measuremen of Producvy: Wha Do he Number Mean? n George Gelauff Luuk Klomp Sephan Raes and Theo Roeland (eds.) Foserng Producvy: Paerns Deermnans and Polcy Implcaons Amserdam: Elsever. 14

17 Table 1: Tradonal Growh Accounng n Japan and Korea Japan Korea Marke economy GDP growh 3.13% 0.43% 9.29% 4.23% Labor npu 0.44% -0.39% 2.13% 0.70% Capal npu 1.64% 0.44% 5.14% 1.84% TFP growh 1.05% 0.38% 2.02% 1.68% Manufacurng GDP growh 2.80% 1.30% 10.90% 6.40% Labor npu -0.28% -0.94% 2.12% 0.05% Capal npu 1.56% 0.45% 5.67% 2.27% TFP growh 1.52% 1.80% 3.11% 4.07% Servce GDP growh 3.57% 0.12% 10.04% 3.36% Labor npu 0.86% -0.18% 3.42% 1.65% Capal npu 1.70% 0.45% 5.56% 1.76% TFP growh 1.00% -0.16% 1.05% -0.05% Table 2: Comparson of Inangble Invesmen/Gross Oupu Rao (%) Japan Korea Germany UK Japan Korea Agrculure and mnng Manufacurng Uly Consrucon Real Hoel and Transporaon Fnancal and busness servces

18 Table 3: Growh Accounng wh Inangbles Japan Korea Marke economy GDP growh 3.03% 0.62% 9.46% 4.32% Labor npu 0.38% -0.37% 2.00% 0.60% Capal npu 2.09% 0.61% 5.61% 2.11% Non-ICT angbles 1.00% 0.22% 3.88% 1.33% ICT 0.54% 0.29% 1.04% 0.36% R&D 0.29% 0.08% 0.30% 0.29% Non-R&D nangbles 0.26% 0.02% 0.38% 0.12% TFP growh 0.56% 0.38% 1.86% 1.61% Manufacurng GDP growh 2.51% 1.53% 11.14% 6.55% Labor npu -0.29% -0.77% 1.87% -0.18% Capal npu 2.49% 0.70% 6.40% 2.84% Non-ICT angbles 0.90% 0.28% 3.61% 1.74% ICT 0.46% 0.17% 1.60% 0.25% R&D 0.85% 0.23% 0.74% 0.72% Non-R&D nangbles 0.28% 0.02% 0.45% 0.13% TFP growh 0.31% 1.59% 2.87% 3.88% Servce GDP growh 3.57% 0.25% 10.14% 3.38% Labor npu 0.80% -0.18% 3.27% 1.54% Capal npu 2.02% 0.67% 5.92% 1.84% Non-ICT angbles 1.03% 0.19% 4.58% 1.19% ICT 0.62% 0.37% 0.84% 0.47% R&D 0.04% 0.02% 0.08% 0.06% Non-R&D nangbles 0.34% 0.09% 0.42% 0.12% TFP growh 0.74% -0.23% 0.94% -0.01% 16

19 Table 4: Inernaonal Comparson of Labor Producvy Growh ( ) Labor producvy growh Capal deepenng Labor composon (%) TFP growh Tangble asses Inangble asses Japan Korea Ausra Belgum Chech Republc Denmark Fnland France Germany Ireland Ialy Neherlands Slovena Span Sweden UK US * In Japan and Korea labor producvy growh from 1995 o 2010 s decomposed and he decomposon of Labor producvy growh conduced by Corrad e al. (2013). Table 5: Correlaons beween ICT and Inangbles ICT and R&D ICT and Non-R&D Japan Korea Japan Korea (0.065) (0.030) (0.047) (0.075) (0.009) (0.003) (0.022) (0.098) (0.226) (0.286) (0.287) (0.150) *Sandard devaons are shown n parenhess. 17

20 Fgure 1-1: Inangble Invesmen by Indusry and Componen n Japan rllons of JPY Fgure 1-1 Inangble Invesmen by Indusry and Componen n Japan (2010) Non-R&D nangble R&D ICT (Sofware)

21 Fgure 1-2: Inangble Invesmen by Indusry and Componen n Korea Fgure 1-2. Inangble Invesmen by Indusry and rllons of Korean Won Componen n Korea (2010) Non-R&D nangble R&D ICT(sofware)

22 Fgure 2-1: Rao of nangble nvesmen o GVA by ndusry n 2010 (Japan) Agrculure foresry and fshng Mnng and quarryng Food beverages and obacco Texles and leaher Wood paper and prnng Peroleum coal and chemcals Non-meallc mneral producs excep peroleum and coal Meal Fabrcaed meal producs Machnery equpmen Elecrcal and elecronc equpmen Precson nsrumens Transpor equpmen Furnure and oher manufacurng ndusres Elecrcy gas and waer supply Consrucon Wholesale and real rade Resaurans and hoels Transpor and sorage Fnancal nermedaon Real esae and renng Informaon and communcaon Busness servces Publc admnsraon and defense Educaon Healh and socal work Culure and eneranmen servces Oher servce acves 0% 10% 20% 30% 40% ICT(Sofware) R&D Non-R&D nangbles 20

23 Fgure 2-2: Rao of nangble nvesmen o GVA by ndusry n 2010 (Korea) Agrculure foresry and fshng Mnng and quarryng Food beverages and obacco Texles and leaher Wood paper and prnng Peroleum coal and chemcals Non-meallc mneral producs Meal Fabrcaed meal producs Machnery equpmen Elecrcal and elecronc equpmen Precson nsrumens Transpor equpmen Furnure and oher Elecrcy gas and waer supply Consrucon Wholesale and real rade Resaurans and hoels Transpor and sorage Fnancal nermedaon Real esae and renng Informaon and communcaon Busness servces Publc admnsraon and defense Educaon Healh and socal work Culure and eneranmen servces Oher servce acves 0% 5% 10% 15% 20% 25% ICT(Sofware) R&D Non-R&D nangbles 21

24 Fgure 3-1: The composon of capal asses by ndusry n 2010 (Japan) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Agrculure foresry and fshng Mnng and quarryng Food beverages and obacco Texles and leaher Wood paper and prnng Peroleum coal and chemcals Non-meallc mneral producs excep peroleum and coal Meal Fabrcaed meal producs Machnery equpmen Elecrcal and elecronc equpmen Precson nsrumens Transpor equpmen Furnure and oher manufacurng ndusres Elecrcy gas and waer supply Consrucon Wholesale and real rade Resaurans and hoels Transpor and sorage Fnancal nermedaon Real esae and renng Informaon and communcaon Busness servces Publc admnsraon and defense Educaon Healh and socal work Culure and eneranmen servces Oher servce acves Non-ICT angbles ICT R&D Non-R&D nangbles 22

25 Fgure 3-2: The composon of capal asses by ndusry n 2010 (Korea) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Agrculure foresry and fshng Mnng and quarryng Food beverages and obacco Texles and leaher Wood paper and prnng Peroleum coal and chemcals Non-meallc mneral producs excep peroleum and coal Meal Fabrcaed meal producs Machnery equpmen Elecrcal and elecronc equpmen Precson nsrumens Transpor equpmen Furnure and oher manufacurng ndusres Elecrcy gas and waer supply Consrucon Wholesale and real rade Resaurans and hoels Transpor and sorage Fnancal nermedaon Real esae and renng Informaon and communcaon Busness servces Publc admnsraon and defense Educaon Healh and socal work Culure and eneranmen servces Oher servce acves Non-ICT angbles ICT R&D Non-R&D nangbles 23

26 Fgure 4-1: The Correlaon beween ICT and R&D n Japan R&D R² = ICT Fgure 4-2: The Correlaon beween ICT and R&D n Korea R&D R² = ICT 24

27 Fgure 5-1: The Correlaon beween ICT and non-r&d n Japan non-r&d R² = ICT Fgure 5-2: The Correlaon beween ICT and non-r&d n Korea non-r&d R² = ICT 25

28 Fgure 6-1: The Cross-seconal correlaons beween ICT and R&D n Japan 5.00% 4.00% % R&D 2.00% 1.00% % % 2.00% 4.00% 6.00% 8.00% 10.00% % -2.00% ICT *The number corresponds o he ndusry shown n he Appendx able. Fgure 6-2: The Cross-seconal correlaons beween ICT and R&D n Korea 16.00% R&D 14.00% 12.00% 10.00% 8.00% % 4.00% % % -2.00% 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% ICT *The number corresponds o he ndusry shown n he Appendx able. 26

29 Fgure 7-1: The Cross-seconal correlaons beween ICT and non-r&d n Japan non-r&d 8.00% 6.00% % % % % 2.00% 4.00% 6.00% % 10.00% % -2.00% -4.00% % ICT *The number corresponds o he ndusry shown n he Appendx able. Fgure 7-2: The Cross-seconal correlaons beween ICT and non-r&d n Korea 15.00% 10.00% % % % 0.00% 5.00% 10.00% 15.00% 20.00% % non-r&d % ICT *The number corresponds o he ndusry shown n he Appendx able. 27