Contents: Introduction: Measuring and Explaining Productivity Gaps Between Developed Countries Gilbert Cette, Michel Fouquin, Hans-Werner Sinn 1. Comparing Growth in GDP and Labour Productivity: Measurement Issues Nadim Ahmad, François Lequiller, Pascal Marianna, Dirk Pilat, Paul Schreyer and Anita Wölfl 2. The Breaks in Per Capita Productivity: Trends in a Number of Industrial Countries Tristan-Pierre Maury and Bertrand Pluyaud 3. ICT Diffusion and Potential Output Growth Gilbert Cette, Jacques Mairesse, Yusuf Kocogluc 4. Yeasty Investment and Mushroom Productivity Growth: An Industry Perspective on European and American Performance, 1987–2003 Robert Inklaar and Marcel P. Timmer 5. Growth Patterns in the OECD Area: Evidence from the Aggregate, Industry and Firm Level Dirk Pilat 6. Information Technology and the G7 Economies Dale W. Jorgenson 7. Productivity, Innovation and ICT in Old andNew Europe Bart van Ark and Marcin Piatkowski 8. Information Technology and the Japanese Economy Dale W. Jorgenson and Kazuyuki Motohashi 9. Outsourcing and Productivity Growth: Sectoral Evidence from Germany Theo Eicher, Thomas Fuchs and Hans-Günther Vieweg 10. Determinants of Productivity per Employee: An Empirical Estimation using Panel Data Nicolas Belorgey, Rémy Lecat and Tristan-Pierre Maury 11. Labour Quality and Skill Biased Technological Change in France Johanna Melka and Laurence Nayman

analyzing Ger man sectoral prOO~c1ivily grOl>'1h , The purp",. is not on ly to contrast the lwo meastJres, btlt to highlight too effocI of oUlSourci ng on Ger· man producti,'i ty growth

Our two measures of produc ti~i ty growth IIJe based OIl valu: added and on groos output. n.e formc, e"dudea thoc contr ibution of inte.mediatf inputs (materials, energy and services in sectoral prod<><;tion) .... hi l~ the latter in· d U<ks all inpUl S into prod<><;tion, The IWQ measW'es thus differ r(MIghly by !be degree 10 .... hic h changes in inlermediates are undeneportal . 00 counted as productiv ity dUllljcS in the val ue added measure .

Higher le vds of seclora l auregation diminish the differenres bet"",en the tWQ Outpul measures; ho .... ever, even al the nalional level the tneasures d iffer 10 lhe exten t that itllemlCd ialC inpolS are suurced from imports (Schreyer 2001. p. 42). we can thu.<; utiliu the d ifferential between v~lue· addct.l pro­doctivily and gro •• OUlptJl productivity to provide i n ference~ aboUl Ille chang­ing outsou rc ing sM.r"" acfl)!;.S German """,tors.

Gro!lS outpot o r the "alue-added measures ha,"e been used extensively to characterize tmal faClor prodoc tivity (TF P) growth. Valuc·added productiv­ity re ports cap ita l-labour TFP .... hile gmss-output producli"ity reporU capi ­ta l-labour-intermediate TFP is reflected by the gross output measure (sec DECO 200 1. p. \0), The ne w [fo Productivity database allow. u~ to bre8k aggregate sectoral performance into its cOlllflOnenlS of capital , labour 300 in temJediale ~ood, and hence e."'ablish the fi rsl gross OUlp'Jt prodUCtivity measures for Germany during the informa tion lechnology boom. In addi. lio n, the Ifo Productivi ty Oatabase contain~ assembll.'<l cap;taJ sloc ks a!l(l capi lal services flows data Ihat more lICcuratdy reflect the e~ntribution' of (aclOr inpuL' in se<;toral and aggregate growth accounling nercises than most other s tudi es. Our resulting productivity measures are important belK'hmarks for comparisons with long e~U1bl i shed value-added productiv­ity measures in the li terature.

The significance of intcrmediates in production varie s aeros< countries and industries_ For ttle case of Ge rmany. tile med ian ratio o f imermMi8te:'\ over gro:;.s outp ut across sect= is 60.5 per cenl in 2001. This natillie empl1asiT-CS oow drama tica lly clla n!:,, ' in Ihe inlCmled iatc inte nsity in p, od<><;(ion will af­fect Output and pror:!oc tivity: an effect that i, documented eXlens ively below. The large ~M.re of inte rmediates in output at the German seaoral leve l also hil!hl i&l". the importanc e of conoi<lerinl! intermediate input< in prod""tiYity groMh accounting.

All producli"ity measure. are inhcrc ntly sensitive It) subStitution bet .... een factor inpulS (ioctuding intermediate inputs) and outsourc ing (of intermediate services): the queniOil is OOW this sensitivity is reflected in toe productivity <tatiSlic. The "al ue-added productivity meosure will be show. 10 exaggerate prodoct;,-ity s;~ it implicitly include!! changes in the S\nI(;twc or ; ~tmnedi-

ates inputs, There . re seve ral theoretic.1 rea""", to prefer the gron OUtput productivity measure, part icularly a! the 5CctoraJ leve l. First, the rl1CasUre does OOt ig rl<:lfe i mprove~nt s i~ the price -efficiency IlItio of intermed iate i~pUlS,

second, it allows for an adequate account of inle rsecloral productioo S{l illo"cr effect., and fi nally the groM output productiv ity mea~u.., doc~ nm ncglt<!t in­te"""diates-i nput-embodied technological change_ TIH: disadvantage of the gross output based measure is_ OOv."C\'er, th.t it i ~ not a re liable repre.'lenllu ion of a sector's cootribUlion lowards agzregate T FP Irends. To calc~ lalc this co nuibu tion (:orn:Ctty, Ihe sectoral producti vilies must be combined wilh Domar wcighiS.

Wh ile (he German Statistical om"" does not pro" ide gross OOtput produc­tivity measure!>, these figures are repont:<! by the US Bureau of Labor Stat is­tics for manufacturi ng indu<lfies_ Previous aUlhors Ihal bave uamined gross output TFP at tn.: seclOral levei 81"e Jorgensen et a1. (1 981), Gullickso n and Hwper (1 99%) and (2002 ). Baldwin ct a1. (2001 ) for the US. QuitO" and O'Mallooy (1 994) for the UK, Gu and Ho (2001) and Gu, Lee and Ta ng (200 I) for Canada. Aside from CllIftS and Mills (2001 ), who OOIllflMC the UK and West Germany until 1996, to our l nowledge there nists no gross outpu1 productivity measure for Unified Germany_

Valt>e-added prodoclivity melIS~re5 are popular in oomparatiV( anal~s,

they lower the data requirements and thui facilitate cross eountry compari­sons . f orcxample Van Ark and Pilat ( 1993) and Bernard and Jones ( 1996a..~) use Ille measure in an OECO interOOllnrry a~proach. Harri s aad Trainor (1991 ) for the regional UK. In Spain. Hernando and Nul'kz (2002) use a value-added approach for regional oomparisons o f productivity.

llIcre are also a number of Studks that compare sectoral pmductivily growth rates accordi n8 to vah",-addcd and gloss output-based estimates. Ou l­too and O'Mahony (1994) provided estimates of aggregate maJufacturing TFP for the period 1953- 1986 ""sed on both metbods. Van der Wil'l ( 1999) has estimated labour productivity and TFP according to both approaches for Dutch manufacturing and. service industrie._ iLIId Si(:lIel (2001) ~o"ided a comrariw n For the US communication sector. Harch300i el aL (2001) pro­vide ",(imales of industry TFP in Canada II/Xording to sc:vCTlll diQ'en:nt OUI­put me a..ures. Ou11on (2000) o btained gross out pm _based eSlin.,nes of TFP gro\\l h for UK indUSlry sectors by using the <alio of value added to gron out­P'" to <:onve n v.lue_adde<i ""sed TR' ~S!;"", ..... Goerlich 300 On. (1 9094, 1996) underline d ifferences in sectoral TFP eslimalCS at me aggregate level for the Spanish ecooomy. In all cases the empiricli confirm the analytical !"C­

suits wc revicw below in that the value_added prod uctivity measure consis­tently exceeds the gross outfJOl l e<limales TFP growth by a facIO. equal 10 the rEliio of gross outpu t to vatue-added.

By ronsuucling!he grMll 00lpul productivily measure ~ find !hat. for LIte case of GerrNLny. output productivity growth ;~ significantly unOetst.H~d by 11K: val .... -added measure. wflil~ T FP cro .... ·th is signifICantly overs tated if gross oo.rIp\It is not uken into account These reMJlu ani: similar 10 those 00-wined by Ouhoo and O'Mahony ( 1994) for ~ UK. We liso provide prel imi­nary evidence that many occtcn do no! $ali$fy tlte assumptions to al low value added 10 pro~y for gross outpu t !TlCIISUres.

In terms of ~ soort:eS of economic lfflWlh. we find that, ~ to the ~-01)'. Olllsourcing i~ a key Vananl betwun the two measures. and that appar­ently moSt of Ihi~ oulSOUrl:i ng !las it! origin in iocl"elsed German itllpOt't$. MO$I striking is tMt LIte factor wh ich drove producti vity «,OWIh in Oer1llllly betlll'Cen 1993 and 200] i~ inlumediates and here hig h intumediate growth also translated not on ly into high o~tput growth but alllO high sectoral TFP gro ..... th. T!Ie~ re.sult! are reminiscent of Jorgenson et at (1 987) and Jorge nson and Stiroh (200t ) for II!c case of the US.

"The stnlClun: of Ihe d laptcr is as fOllo .... s. f illit ~ O\Itl ine the di fferences. advaIltagcs and d is.advantages of the grou and val"".added productivity measuru. The ... "'e describe the novc l datasel that has been assembled att!le lFO Institutc. Finally _ compare !he p"rfonnance of the val ue-ad<,k(! and &fOSS OUlput meas."es .... ilh ~cial emphasis 00 the contribulions of interme­diate$ and outsourd ... g to sectoral productiYity. Section 9.6 cooclldes.

9.3 METHODOLOGY: VALUE·ADDED VS. GROSS OUTPUT MEAS URES

There exi" tWO standard concept! in the economics lilerature 10 ~.r.tabli""

·outpUl ·. Gross output i. the total ~a! .... of SIII~ and opentin, receipr~, whi le val"" added wbtracu from gross Output ~ va/ae or good5 and servicu pur. c~ ascd from other sector.; 115 intenncdiate inpuU. "The difference~ bctwee/ltl!c measures are both practiQl and concc:plual .

Value added is attrKti vc becau~ of iu relative simplicity; doll ... income. or pIIyrne~u to ,,"-pila l and laboor "'" readi ly available fro m LU data. Time $Cries dtu.~ .ISO requires, ho~~er. tlte USC of lieparatc dcflaron for sales and inputs. GrolS OUtpUl is closer to Ihe ordiMry notion of productivity M it represe nts sales per cmploy~ . As gross OUtput dea ls explicitly with intermc­~iale toed! it ;s not subjecl to distonions .... hen lhe primary a nd intermediate input mix cllanges.

The value-lidded pnXh.lCt i\'ity COfICC'pt is .... idely ....w for international compari", ... bcca"", the data are m<"" readily a~ai lab lc from countries ' na· tional KCOU"'U ..... !>creM sectoral Voss 00lpu l requires men: dcta.ilo:d and di$-­auregated dala a nd is theref"", Ius easily available. Hence. "'hile ~ US

Bun:1IIJ of Labor stati~tics hn determined thal. gross output i. the comet basis for US measu= of prodUl':livity. il ac\:n.owledges mal there ate other con~id­erations Ihat rna) make value added a better concept for intemationJ,l com· pillUom such as differences among countries in the extent of vcnical integra· tion of indllstrits. We COmmtnu ,",,'ilh a brief ~"iew of !he issucs Killed to gl'QS!i and value·1Jddcd OI'!p<It produclivity measUK. aoo then eumioe their performance in Germany.

Gros/I Oulpu l

The most natural mellSu~ of productivity is d~ctly Kilt«! to the "Qlume cbaJlge~ in o utput. TIle $fOSS ootput prodoclivity ........ lIfe i. a nalUral analog to the model of Hicks ne\llral technical ch.angc. Here technological change could be R&D or learni n&-by-<loing. The productivity increase is disembod­ied tedm ic al change because il is lhe ~idu31 output change that is not physi· cally tied to any faclor of pnxiuctioo.

The basic growth accou nting eq ... tion " ,ing tho: gross output concept i~:

-L - K -II 6 1 I1Y~ : ~.6Inl.." +~~ 6InK. +1'/0 6 1nM. +AlnTFPyi , (9.1)

with 1' .. as gross OUtJ)UC in indwll'y i duri nl period ,. l. i. the flow of labour inpu ts.. K are capital service flows. ~ intermediate inputs.. and TFP. is gross oulpllt lotll fXlor productivity. I'~ reprosenl'l the input slw'es in gross output. under constant ,et\lf1lS they fulfill thecondilion: ; ! • ;: . ;: _I .

In a perfect world. the gross output conCl:pl is. 11$ highl i&hted by Jor&e nson and Stiroh (2001). !he appropriate concept for induwy-aevel growth &<XOOnt­ing £tuditll. IIo_ve,. the mc.asln has t;Cveral .hortromings. Sichel (2001. p.7) ~vie~ the literatun: and highlights that in practice. gross ""ll"'t TFP rroeasure.'l also ~nect a numher of addition.1 innuenccs includinl ctwlges in efficiency. economies (If scale. capncity uti li zation. martel. structln and meMUremenl error. Und~r the assumption of competitive markers and con· stant re\ums to sc.I~, the.., facton are assumed to De constant and to have no effect on TFP.

The val...,_added producti"ity measure is, ho"''IlVCf. iiUbjec\ 10 the WIle

criticism, The advantage of lhe gross oo\l"'t· based productivity mca.\ul't' ;.1 that ;1 ac<:o ~nts for intermediate inpul'l lIS a iWUroo of iooustr)' growlh lIJId ho:nce appropriately considers differences in the ;nput m; ~ as c ~planat iOflS for outpUt and TFP p-owlh. Some "1"" tNIl this rcnde .. a more complele reprc· s.entation of the proo\ICtion process (Sichel 2001, p_ 1 ), Jor~nson and Stiroh (2001) have provided I .. uample of tile impon....:e of intermedi :ue iapu gt"Q\\otb in lhe analy~is of TFP. when they showed thaI ploduct;vity improve· ments red\ICed s.emi·condoctor prices and inc,cucd !heir now tIS intermed.illl

"puts 10 ether industriel. A full account of !he centribu tion of tN:!IC' input5 lJlPrtlpriately redll(:e$ b iases e f the attribute(! proo:lllCliyj ty eonuibution to )Jimary inpu ts (eapitnlllabour ) and IT-P across sr:c:t()(i.

By correctly IICcnunling for the quantity and qual ity of intermediate inputs. he gross outpu t concept .lIow5 aggregate TFP gains to be beucr alloca ted ,mong indu~tries (l orgen~n and Stiroh 2001, p. S3). The melhod ofaggrega­j on is cruci.l (Gullickson 1995. p. IS). Aggregate outputs and inputs an: IIOt ;imple !iUms of their sect(ll"l.l counterparu and ioconsistcocies can arise be . ""·Un the productivity e.<IimaWi. At,gregllte ou!pUt and input measures u _ :iude a ll intermediate rransactions bet",-een domestic indusoY$ \Q 1\00 dou.. lIIe counting a nd 10 capture mo~ements in inputs ,nd outputs resulting from redlllologkal change and other effICiency change5. Im_vu. industry inputs incl!lde purchases from other iool/"Sl ries IIIld induruy outputs include SllIes te ..:her industries a, well u s.ales tn final demand. As a resu lt. ag~~ate pre>­~ur:t i vity growtk cannot be nt>Lai ..... d as an avenge using I set of ",eights thm IORI tn nne (Gullicbon wwJ Harper 1999a, p. S I).

ConsiSte ncy between a~gate and S«tOl"llI estimates ef TFP bawd en !VOSS outpu t is enhanced by the exc lusion of ' nln _industry inputs and by the adoption e f a sectoral productivity weigl1ting system In deri~ aggregate es· timates. As 1M seCtor size increases. the proportien e f a ll rransactiens that are inw-sector tends to ri.., and the ratio of interrroediate inputs to , ·.11Ie added tends to fall. Equi~alen tly. as the level ef sectonl a~galien iocreascs. the d ifference be tween vo~s output·based estimates e f TFP growth and value­added based eStimatU tends to dc<.;rease. In the case of a desed eceoomy. sectoral out(lllt at the ITIOS{ aggregate level is identical \0 tnUlI val"" added (OECD 2001. p. 9t).

Domar (1961 , proposed that TFP gro",th at the aggregate le,·el shO\lld be mea<;ure<\ as a wcilNr:d sum n f indu<lr)"_level TFP growth rates (sec: o..l ton and O·Malln ny 1994. PI'. 13-1 4 and pp. 118-2 1). The ind\lStry prodtJC1ivity growth rates are estimair:d usinl I""'" output and ;lICOrpOrnte imcm"l«liate inputs from ether sectors. The '[)Qmar· weight is the rltio of !he value e f Voss OUt(lllt ef an indUStry/sector te the sum of value added in a\l indus· trie&lsec.lotS. Thi s "'·eigh ting scheme can he odapted to diffcn:nt aggregates. ,,"'hc!lIer I scctQl1\I anregate or the market ecenomy.

The effect of weighting industry gro ..... tII rites is tn scale the ind ustry 1FI' estimates by thei r reluti'·e importance to perm'l . ec<;ttoeiliatien w;lh aggregate esti mates. These weigl1ts renect the d irect contribution of sectoral productiv­ity change: to economic ,roWlh through delive ries te final demand and the in · direct contribu tion tll ...... gl1 dcli,..,rie. te intermediate deJ11llnd (J orgenson et . 1. 1981. p. 1). This weight ing methodnlogy implies that economy-wide TFP growth can VOW faster titan productivity in Uly .;n,le industry. since prodnc·

!i"i!), pi"" 'Il' ma&ni flCd as they .... ork m.,ir .... ay !hroul!h !Ile prudOCI;on Jl'fllI%Ss (J<>rgell§Of1 and Sci.oh 2001, p. '3).

Tobit 9.1 Indulry OUlP " ! growlh alld a" tl'llge CO/tlribUliO/t~ I99J- 1OO I

C"""""nkotioM lUI T_ Eqoipmcn< .I. ... Fi........,-.....d,1boo ,71

ROll E>I ... A.cIi • • .t 0. •. s.n.;"", I ," ~. J.U T........- .1.)0

El<c , Md El<co ..... Eqaip. IN... .\.,. ~_, '" PIa;':. ~ ,.l

Elec:tn<i'y.o.. ...d ...... " S. pply lAl Wood It. Pro. of WoodO<1dC..... I'" h,le ~ctaI. A Fob,M,"I I'1<>ol",," I ,.. A&n '.Ir"~ """''''1 ond ~,"'i~ , 4' Non·Mod<oI Sotvi<a I .... Othot s....i<a 0.,., Food.~,t T_.. o.n ~·M<t.oI~c l.Ii_ Plod.... 0,. R.".;n or>d .. -.. .. ,""'" 0.41 R .... I... Oll Me<f\oftical ~I 014 C<IMINOtion .all PIolp. It. P.opo:o- I'rod.; F'>in' . ,t Publ. -0.00 ...... , .... ~1, 1<. M ... r.; =)'cli... ·U2 Hoo.h '" CMtrinl . 2.68 tn' .. ..... ~. Foot .... It.CIorhi"l ·1 .. , 1.1 ....... Oil Ref .. C ... It N"de..- .0,00 Min'"l ..... ~"II .9,1\2

0,1.l 0./10 ·.UII .om 0.1lI .0.112 Il.IlI 0.1l

0 .19 0... "-" UN : ,02 0.'0 lUO 0.'1(1 0.41 (1.0} .(U) (1.(11

o.a. IlIQ .0.11 0.09 0.11'1 -Gin ..Q,44 (1.0.

OJ) OJ)< ..Q.G' (1. 11 IJO GDI -.Q.09 Q~

o.n 0.00 -.Q." 0, 11 0." 0.00 ..Q':1 -G,OI OUO (lID -(l,.' 0,"

0." 0.115 0(11;1 QW

0.61 Il l) 0911 0..11

\I." 4 ". -iUIO 0,011 0... .{UI, ..". , 01' o.n -G.Ol: ..,,(lI IlXl 0.J6 Ilo< o.D"2 QXI

0-" .0.'" -.Q.J6 0.0) .000 .0.0' -G.t) 0.'9 (1..14 Il(\l .(l.U QW

Il.l. G.O' -(l,1,6 0,1*

.... :10 0.1l> 0.." Q.J' OJ)< .0.0) _1':2 -Go1l

0-11 0.0' -<),08 000 0.00 0.00 ·1..16 ..w.

709 6-< 1 • . , ) .0.1'/ 1.16 .(1404

0.11 ·L(U ,. ~ I " "." ,.'" -(l.IJ t il O-Ift

,~ -,.10 (1,0' ,,,. 0.,.,

0"'"' 0 .'6 ~.., .ft.5,

0 .0 1 .1,11

CU1 0." O-a. -lU2 0.1I .0.112 .(11' -Go,.

CU7 -Go'l

~ -IU ' -OJ, -0.10 -I XI .~ -I." _1,72 .0.42

"7! G.!9 ·'.11 ·:1.506

N"" , I'ttMU/ll< r""",. dr : o..'P" (i_'do; <II':': "",,·10' <dflj-.( COOI'ri",,' ..... JIi!": tcT rnpiMI ~rib.,;"", <lUi: """,";""11"" "'''''''' Jo". .. ""'*<d. JW, """'rljo ..... "',.", _""~ •• <J1oI: _,ri __ "';01#_ ;0,...',-"""""""" TFI': _"""'" P"'_ doocll",,,

The Domar 8Urcg31ion of &ross OOlpul·based TFP meaSure! across indus· tries provides Hn acc urruc pictun: of the coniribu!iOfl ~ of ioo uslriu 10 ~8ire· BOle TFP c ... ngc . Howe"cr, I.IIere an:: significant data problems anociatoo willi inpu!-OUlptl! tab les "nd tho ir ,onsi 'le~~ with national acoounts. This issue i$ di~u • .ed below. 0 "" COS! of lhi' approach is lnat ",""Ioral producti'" il)' VO"'th nteHannot be compan.d willi the aagngn/! because the aagnaate is built up 11$ wdJIIlcd stnns. but IIQ( averages, from ;t, components. In con­trast, a$ JIQIed earlier. value·added b~ productivity measures of aggr",ates

are weighted averag.es. of their components and can be c:ornparcd across levels of awe,Mion beelus(" tile _igilts IIIdd to uniey.

Table 9. 1 represents me, average gross ootp'" gnw.·th between 1993 and 2001 Ind the contributions of various inputs to the average OUtput groWlh . Table 9 .1 indica(($that. for uample in the elM' ohlle comm un ications sector 7.49 per cent of thoc 13.31 per ccont pn:xluclivity growth is derived from in[(r. mediate$. TfP growth contributed 6.41 per ccom to the lverage growth rate in bctwec~ 1993 and 2001.

TIle lOp S sectors in tttm.. of average growlh nues nre Communications. Transportatloo Equipment Fituutcial Intermediation. Rell Enate Ind Chemi­cal s. MOSt of the gro ... 1h in these seeton; is derived via growth in ;ntamedi­ale' (e~cepl in Ihe c!ISC of real estate where t:apital ac.:umulntion dominated) . The same Is true ( II" the 5 biggeSt losi ng sectors. Minioa. Minera ls . Teuiles. HOiel. and Furniture. where most of the contraction is derived from reduc­tions in inccrmedlates. Hence their gross output measure t.i&hliahts the impor­tarlCe of Germ31l growtli in the 19'1Os being driven by c hanges ;n outsO\l rcing or in((rmed iates.

Note thal TFP C011tributes positi"ely in only II of the' 26 seeton. Mos t ;n_ terest ing is lhal the correlation bctwun the TFP cooui bulion to outpul gro ... ·lh and Ihe intermediates contribution to ootput UO .. 1h i§ 0 .66. Juueslios that IhQ!;e sectors that had large expansions (contractions) in intermediate growlh also nperieoced large TW C(lnlribmions 10 output growth. Only the correla­tion tJo,l",,,,,n TFP and outpm YOWlh is lIisl>er wilh 0 .77. clearly iooic.ting the: important rtlatiomhip belween total faclOf productivity and output per­formance.

Value Addfll

The niue-added ProdUCliyity measure pos.u:uc s 1"'0 wt)ng advantages. First. it can tJo, derived wilh . ellltively low data requirement>. as il abstrac •• frtH11 ;nter-sectoro l 110001 Mod service flows. Seccndly. the approach PI""" vides. simple link between sectoral ~nd aggregate TFP gro .... lh (see OECD 2001. p. 30). Sectoral val ue added CUI al.-.o be simply aggregaled using wcilllllS tha. repre~ent each sector's CUffenl ~ice s hare in tOtal value;added . This implies that value_added productivi ty can be ea.~ily compi\f"td acrrn; s • ...:too or industr;e., and that .ectflfal productivity ,rowth can be wm~~.,j 10 Ihe nationa l a\·erage.

Tbc ba!.ic vatue· added gro ... ·th 1ICC0untin~ ntuatioro is:

(9.2)

wherc VA iI .. Iuc Mldcd. and TFP •• i! value-added loul faclor prodU(:livily. v. ~nl!/lle input shares of labour &lid capital for wah .. fI<lIkd : lhey M ­fill !he conditio n: V~L + ii./{ '" I .

Thil cOII venieno:ly impklMnloXlle and c~ IMlhodoIogy romcs II I pri..c . Vllue-added prodllC1ivily il M limes criticized as pnn'jdina II be$!. an 1IItlbi.- pocwrc oIlhc ICWlII p!"odue!i viry, due to in at.wxtlo. I'rt>m iot-1ermcdiMes, l lId ..... 10 Ihc flet 100 rul --'d anal", to value added illCtu­ally prodllCN by plllllJ (Slldit and Finan- 1981. p. 14; (lullon IIfId O ' Mahooy 199-1. p. )3: Hullen 2000. p. 58).

Since value ao;kIed i$ the diffen:nc:c bet .... «n $epatllCly denalcd S!0$lI 0lI1-put, y. and inlermedlate: input., M. lhe f.XIofI(cpI rrqui'e5 an additivcly ilepano­bk productiof! function o f the type Y . VA + M. Sucll • function imposes 5tron, ruulclions 011 ~nerality 100 the 10k of ICChoolocical chat1&e (see Gallop 1919. p. 320: Bruno 1980: Die_rt 1980). EmpiricallCSlin, IUUCSlll, for example. thl! !/lieu u.iSti 00 Kpanbilily bclween lhe •. a1 .... !Idded func­lion and Inlel1J'lCO.lillle inputs. Jor~nSOft CIIII. ( 1987) find that the concl ilioll$ ...., ... ..,.. and sumd" nt for the txis~l'I«; o f I il<ctoral . ·Ilue·ado;kd functlOll d id not U !SI in ffJny 0\It offony·nvc US industrin lnal)'Z"'d.

An ado./itioMI ' impli roc.uio n Mtlle val ... ·addcd measure il that it ablHractl; from ,uhstiMion between capi l. III.lM.tr . nd inlo:rmed;a~ inpucs. This as­sumption •• ~pecillly problemalie in I;OIIntric:swi th ",idl.bour mwuu; where oo..nsili", limits introduce fr iction and slack into s.ect0fll1 employ­menL One might v, .... mat in this m:pea, ClemwIy i$ a good candidate: for C$lbl1Ihi"l &IV" outpul prfJduaiYilies.

f'riec clll .. ,., in i,oI(;, um." lllpUu.. pDlmI .. Uy ';gnifocant dn~ers of 1«­

toni prod...:tmly vowm. are I5SUmed MlIIl to 'ff«t.thc relau.e IAIIc of other inputs (IfJrJl'nson ClIl. 1987. p. 9: Dean Ind Harper 2000. p. 48). TIK yalue­atIded COlICept thus assuflll'5 IIIaI prices of inl(rmcdialcs a!WJ,YS fl~ II the s.ame ralC All tho!e or , II other raclors. F"iMl1 y, lechlKllogicll clwtgt is as· IoUmW to .ffecl only capital Bnd Ilhour. Jil>ce intermedi~le us.q;e Is. by IS· SUmpt ;OII. no! affe<:led by the 'mprO~cm..,na in productivity (Oollop 1919. p. 322.00 G\iUk kMln 1995. p. 17). Hence I cO\lilll)' th.Il UpuicDCCS rapid pro. dual"ily gflWIlI throo.ogh oul$(lUn:i na and I ~lIlft of sect""" prfJd"':IH:In 10-... "Vdlt hi"," prodUC'tlvity intermed ille goods will find .wch Ctw.JI'1 aurib­uted to capital Ind I.boa in !he \/al .... ·1ddcd measure.

Val"".addcd prvdudlV;ty utiml~ can thu~ be up«ted to be lria he. llIan gross ou!pUI·bucd e,tilNlle5. sinc:c value-lIOkIcd pfO,hx:th'ity ,c"",,1t impl icit cbanp in InIC rm..:liaie inpuu. Follow' n, DieW1:f1 (2001. p. IS ). ¥fO'S OlIIpUi prodoctivlly cln be wriuen as Y I(M +1.. + K) and vah ... ·added productivity is then ddirlCd as VA1 t..+ K) .. (Y - M )I(L+ K ). [n Il'al te~, .. prOOUCI;V­;1)' impro:l"" .... nt ... r AY tlw eannot be Ittribuled to ellanJl'. in M. L. " tide .. .. """ oulpul productivil)' cr ... wth f1ItC or

T;P~ _ r -l-~Y / r = 1+4Y M +L+ K "' + 1. +1( Y

(9.)

.hid! .,leu.llaft (he real value.added producliv i.y growl/t rale gi.e~ by

TF~v .. . Y+ ... r-M / Y- M _.+ toy L+K Lot- I( Y-M

(9.4)

The smaUer denominalor in \he value-added 1l'P measun: translates into arger TFP growth mcasun for "atue added. Comparing the h.U growth niles. LI (0110""11 m...

• • n 'PVA =(Y IVA)TFP. (9 .S)

Which clearly highlighu thai value-added productivity is uptCled 10 elceed TFP prodUCtiv ity_

Comparisol1'l of grGSS OOlplll a~ d value-addeo:l measures of -'toral l H ' growth in Germany. confirm !his oOOervalion:

Table 9.2 highlights ""vernl imporUln( factors. Imercslingly In 18 oul 0(26 §.C:Clors the: value-added concept D"er l'epll'5CnlS flei llal OOlplll growth. In addi · lion. vah",· addcd TFP growth indelld significatllly exceeds Ihe.!l1'OSS output productivity musure (in n of 26 cases). however. mit uniformly so. The last two columns of Tablc 9.2 hodicale Ihe uiMCnce of signifK;am di iT_ncr! be­tween It.. e~pected rat ios of TFPvJ['I"P.,. and !he ""male<! QIOCS IIsing the &rOSS QI,ltp\l\ alld value-lidded concept. 'There u i5t large diffc,enoe5 between the two .... 1;05 in jU51 about all sectors. r R 17 OUI of 26 indlls.trir~ lhe ratio IS smaller than o~ and in nine out of Ihcsc lrumen illduslrie~ the ratin be~omc:s e.en negativ<:. which clearly conlrlldicl$ lhe ~\I I"opose<i by the ll\eoQ in C<jualion (9.S). ll\C '''-0 columM provNk lhe first evidence that the aswmp­tiOfL'i necessary 10 derive qualitati vely (&rid somelimes c'~n quantitatively) ~alid 5tatemenlS of IndusU')l-levd TFP u..ing lhe value-added concept arC too reslriclive for the else of Germany. Belo .... ~ prciiiCnt some anal)'SCs for a Simple iiiCCtor 10 indicate 1m .... Illc omission o f inlennr:di"te inputs can Cluse lhe abovc described dislOftions in me value·added concept .

A ca~cal that must be added to Table 9_2 is. ""~v<:r, thaI intersectoral cOffip¥ison:; of prodocI;.ily growth .... y be distorted becauiiiC IWO seelOrs mIIy share !he same prodoctiv;Iy gr()\to1h On a gro •• 0lI1put basis. 001 diffe-n:nt r~ of prodoclivlly growth on, value-Iddcd buis. As di!CU$Sed above, this may ocCllr because the proportion of inlennedi,\e inpou in total cost. differs in thcse two sectors, This p1»\si bility is dcmonstratro for several Dutch illdus-. tries by ~an dcr Wiel 1999 ,nd in our data this i~ highli &hted by. for cumplc: the German Eloe<:trical and Electrooic sector and Mechanical Engineerin,. Both sec tors sh8ll: illcntical &rO" OlllpUl prodoct j"ilies, bUI differ in their value_added productivi ty.

0",,,,,,,,,111, aNI PnxllLfi.iry Gro~.Ir: S«rom/ £wdma ""'" (d"..,."y '" T .. bl,9.1 C"""""iso" of t: ' OSS ()~Ip~I ""d .... /",,-addm m~"swus.

Gt,."w,y /99J-1001

.... _1993_2000

n t\.!ff ~,.

n .", ~, ~" ' NO '. ~ .. (VIVA)

< ,~, '" ... IG .• ' ~ , .. ~,

T .. _F~ '" D.ll '" .W '" .w = _~11 •• ", , H' ,. ~ ,. ~ . \I9.!(i . .. IInI f.oIaor 11«1 .... ___ , .. .. , . , !I ' .. ,~ ~ ~,

~-. .. ,~ •• ,n W l . II '" '- , .. 1.11 ." '" '. H, '" ....,...., .. __ ..... " ..... 1_. ,. . " ." '" , . . ' 1.11 .," _.io- U, .. , ." ! .~' ,.' 1~ .1e ,U9

F-',rui.,;.',Ga. ....... _~, U' '.n •• , . ,. " 'l '" ... """ .. _of""..,. ... C .... ,n ... ." ", 1.'1 ,'7,00 ,w • _ _ .Io ,oto· ___

on •• ' .n ,w w 7.' 1 '" Apic-. . ......., .... "...... ,.W ,~ ... , .. = 11.11 , . --... ~ .... -.. ,. I.Jl {I.l l ... l-Jl •• •• ~Scr>"" on 1.01 • ) Jj ., . ,. •• 0.') -..v.ooo.T_ • n .. , '" '" HI '" , . __ ....... ,- .. ." ." , . , .. -J" .'.1It R<poir>'" _ .. I< ""'" ... . ~ •• , . 2,~1 ." .W -... •• "a, .. '" ,,, an ." __ f p '. '" = ." ,~ U , _II.U .,"" -- .0.11 . , .. , •• '" 2,61 , .. '" ........... "- "' ...... Pri •. .to_. •• · I .IT ' .n . " , .. a~ ." -.. ... , ... _ ...... ...,.~ .1.U . " .,~ .", ~ , .. "' _.c-. .1.61 .,~ _I." " .JI ,." '" ,. T . .. ;o", ~. Fooo_" C"""",,, . l.41 '" ... '" ,n • \.'!6 • • M_ .. Oil Rdo., Co/..o .. II",,,,,,, "'" <.W .~ .~ . " .. •• .. Mioiotc_~ •• ·11." .~ ." ,n , . ..

Soo,,"~· A.,,,,,,,,' ",'<"OIJz""". "J/~ "'" 1ft> Pmd.<~';" I~.

[)Ift"erences in Interttmporai Produtllvlry Trends

n.e .... Iio of gross OUlpullD Yalue added may Y1rJ dnomatkally arccoroin! to the ch¥acteristics of a IiCClDr. This implies tllat not only • ...,rag~ TfP growth ralel may differ. hut al~ tlla! in!enernporal contjWisons of productivity .. e distoned (Gullickson and Harper 1~. p. 19). n.e mealJ.lfCS diIT~r in­tenemponolly bec:a\l5e the valuc:_added ~ in gross output chu,1:!< as the colflJlO'>'itioo b."""'~n im~rmccliales and cal'italilabour is a lltrM. This compo-­sition is largely a.tfccttd by out5OOl«'ing. as denlQMlntr:d by Oullkk$on ....J Harper (l099 b. p. 18) who cmphuizrd the role o f oulSCll1f<:Cd production

abroad by ,c"Tiling!hl: n: lalion belween Tfl'v~ ,'010'111 and TWG gro ... 1h in :qual ion (9.5) abo,'c all

• • (9.5')

If sectors increasingly o utsource pans of the produclion process, and in_ termediale inpuls accounl for Kn ever incrca'5; nl ~ .... e of value !IlldOO in Ihal <;eC1Or, val",,-added pr<)duc:tivity win grow ever faYer than gross output pro­:luctivily as I oonscqucnce o f \he outsourcing. Hence on an aggregab: 1e~1 ,

the dirren:nti~1 bel~n va l ..... >dde.;J productivily and g os:! otllpUl producli v­il)' c.a n be ~n a.s ~ mug .. measure or ouWl..-cing acli vit)'. This oUl<Oun:ing mip.! be intemllliooal or domeslie;, H""",,'fa", in Gefnlarty it i~ lempIing 10 !IearCh for effects of out$oun:ing thai are driv-en wle!y by in ~maliorlal factors.

,~

.~ • e '1101

! '. • ! '. '. , .. . \1

,, --

Fiswrt! 9.1

-~---.

V.1~Add.

Gn:>" Ou\l:H.JI f'roduc\iViIy

, - ~ , I. I'll IIIiII ... lJlI

Cotnp{l,,'S01l impom "nd TFl'v,,;rFP~ J99J- ]OQI

The ,",pM in FilW"C 9 .1 niccly hiabtiam the diffm nces between !he vallie_added and grQSs output productivity measures, Fi rst. the notio of the lwo !' "oea5llre5 r¢f1ects \he upward trendi in imporu belw«n 1993 utd 2000 in Jermany . A~ 0lI1!;OOrdng falls after 2000 (value_lidded TFP lICtual\y d,ops llelow a&a,egate 11'f' in 200 1) poss ibly d"" to the \.argc contraCtion of the :onstruction indusHy, which conSCitule5 a large share in German "" tplll , im· ;>oris followe<l uil.

Figure 9. 1 a lso Uiows thaI (he irllertcmporal rate at ... 'hi ell value-.added pm­juctivity overestimates lfOOs output producti vity i. lar&Cly . table ; around one ;>er cent on the _"",gate Icvd. Thi$ ~ility coilK'ides wi th a lie of relat ively <table irnpon gJ"O"111 bel_en 1994 a nd 2000. However. varialions in imports

lin' ill\ll1fdiately . eflected .1., ,,,.he amooA! by ""'""" "a!_~ "I"P ower· stales Jf'O$S output prodUC: li~l ly . 15 seen in 1994 Ind 200 l.

Finally il 'I Ilw: contribulion of any given f~Of 10 gro "" l~ thai il mosI se· ~Iy dmonro by u.., ... hoe.....sded mellSUres. If WI: absIrII:t from illlelTll«li. alel In lhe sludy of produc:tivily l1O'".h • • 11 Impro..:menl$ in produc:.i",.y growth iIIal actulily arise Ihrough OII~inl or i"l~rmedille produc:.ivity gtoIIo'th are instead •• tributed to capital , nd laboIrr. Se~erat 1eC!OfS, such as fin;m::e and compuler snvm. rely heavily on Ilw: 1fO'Oo'1h of ,nterrl'latialle in· puts' efficiency. I~ the entire RAD based li~ture l1li pnel'lled . tOt·

IllIc illdliWy of model, mat rely on JIffldUC:livily incnasel gcnet1l1ed by the ;ntermedi'!f: inpull. In addi.ion, any impmemc:nll in producti .. i.y in inll:r' medi.llc aupplying industries ..... y connibute 10 IInprovt:l!lIInfS in producthilY in fhe downstream leCtor in !Ievcral ways. Suppliers mi,1lt incrc:allC OUtput qual!!y without changing inpuu used .nd oowm;tream firrm might benefil from lUCh qual i.y impl'O~el!lllnt in the form of illl:rcased productivity (pr ... ~umin, the chang<d quality is not fully compen~ed fQf by prit c changc.).

For the US, JQf..,n"", et . , . ( 1987) IJId Jorlcnwn .nd Stiroh (2001 ) show thai IR\trrnedilft inpuu In' the pudominllll 5OW1:<: of OUtpUt gm .. lh II Ilw: sectoral 1e\'C1. c.\CllWin! boOt produc:tivily aro""'th ,nd the conlribuliom of capilli tod labour in the IlrJc majority of illdusme.. n.e estimates rorepartd by CNlIOn r.nd O'Mahony (19904 ) fQf the UK show tlta! inplll &fO"1h upt.ins a much hiper ,N""""11oft 01 productivity gowth 8CCordi~, 10 1M ,1'055 out· put estimale!l IlIan in the case of the vlh.w:·lldded nli .... ,t'I. The same is U\IC

in me Gennan dtll3llere.

9.4 THE IFO PRODUCTIVITY DATABASE

Ow dill. is deriwed from !he German Producti~"y Dallbluc thai is cllfTtndy beinlaacmbLcd .t the Ua In!.l itute . AI this pilii'll the dalaime coven'2 sec· 101'5 belween 1991 and 200 I. the Il'isocialed capital $lock,. U5e1 types, prien. depreciation niCS and labod'. in 199' and cumnl pricu. Cum:ntIy thll diu. bue i. beill! upanded 10 ~ 1960-200 1. "The lfo ProducriyilY o.taba.oc: .110 ..... (Qf JeCIOf· and ~'lJIeC i flC price def1l1on in • !fOWllt I/;Count,nl study rOO' Ge. lIWty.

Primary dati 001 .... 1 and nomiroal OUtpUt. v.loo 8ddc:d and intmnedillCs inpul data IS obIained from the Nation.l N:CQUnrs lUIf.Qticl of the German Statiuic.t office. The o.:.m.n Stat;llitll Offi~ classifies Ihis dltll accordi n, \0 • 'Syuematik de. WirUch.f~_lp· (WZ !iJ)) I)'mm. wIlictt ~ ba5ed on !lie NAC'E Indu!.lry ClassifICation.' The JCIIII'Ce of !he CIfIi .... ltoc!o; daI.I is the Ifo f~I'f"'D"'I1'K""""1l DDI<JIHI:w . .. hielt IUllloenl'i Gennan SlailSIicll Office !"VUI_nt·series with Ifo $lIf~y dati for 52 ICCton and ticve" ;n ~est_m

assetiilKO'ices.' 1be ,reM advan!a,c of the capital Mock dall in tile Ifo Pro­duct' ''it y Ollabase il thus il5 rd ianee 00 iCdoral capital WICk and sectoral in>lOWlIenl dil;tribuuQftll d ... Depreciation .. ell are 'lUrYC)' ~ Go:rmany­sp:cifK, induslI)'-IpC('ific .00 UlCt~flC a!Id 1ft, ~in' 10 001' k(l(lwl­edge, unique and for!he rtrst lime emplO)'f:d In a gN"1h Krot.Inling sludy (Of

Germany. Tile capital SIO\!k is tlw:n cak:III;llcd usi'li the: perpet ..... invemory """In!:

(9.6)

wt>crc S i~ the capil,1 $lOCk, I Is inYCSlmenl, oJ is. C(lC$WIt Ilcpm:ialioR ra!e, aIIII tile .ut.c:ript), rde~ to lhe ISXIlypc! TIl!: Ifo InwMOretnd!R .... Da­tabase permits the aklilitioa of capi .... Mock dlta for eK"Il investmcllilype in nch seclor will>loul fUU"iclinlll$JUmptions re,ardillj; in¥C$l.l11Cnl diwibulionli (e.g .• II nifClm1 d illlribution ~1 industries and in"Hlment tyres) that are ~sually nec .. sary ,n the lilCrature.

[""clIment data II also laken (rom the Ifo InvClitorcarechnung Database. " 'hiclt recalculates: d~ from the CienIWI SllIiilical Off..:.: lI$inllfo survey dlta. and tile Ifo Investitionste<l. Tbc Iro InveMMenrcd,nun, Oala~~ lit ... providei i!lduSlrY-lpCciflC inv«lInen! distribulions tltat cnablim • "niqUl! li nk bet.....,cn!he economic U5e o f lUI invulmcnlll$5el. and;1:II SllIlistical c'-ui­flClOllOll.' ~ 1(0 Productivity D'"'''''''' employs the JMP'nwn and Sliroh (2001)

method of caleulalinl tlte capital ie/vires Row over. liven period. rather titan Ute capital StOCk In tltis period. Capital services nows an: aSlumed 10 be proportional to the a"entlC' c-pila l $lOC k .uilable at !he beaiooin, anti end of • a;i~1 period:

(9.7)

where q dcOOle5 I oonst.aot of proportionality. TIw: Igre,.e capital no .... data is ba5ed 01 diNU"'PICd teCtoni dau to. uch..-. C~tly the ;nve:Sllnr:n1 price dcRalOf$ ute(I to value It. capital "ot:~ fIR:

obta;~ed from tile German Stalistical Offtce:. Ilowe_c:.-, the Iro Prod",!!vh, intabasc i5 cummtly C1pt1~ to include quality c!tanses ~sed on the US 8_,", Ilf Labor Slitistic hedonic price indiCC$. Preliminary calculalions "'. ins the US hedonic price de:n ... un f(lf Icr-investmeno: indicate t!wll !he: d iffer­en<:e!O resultin. from the use ofhedonic price denlt<nare small.

The qu.ntity of I"" C8p;taJ ".nice~ now ove r . ,,_en period i. gi.en by. ren .... price formull

(9.81

,,'hictl a~",e~ investors "'" indifferent bet...,...., a ",nul r~ c. and a retutn on the capital market with nominal interest r~e,/ In order to decrease the im­pact of c~clical flLIClUations in the cak:ulatioos of the capi tal _vices fiow. we employ a j -year nlO'-;ng-a"erage smooching method .

The: ",nud arbittage eq ~alion yields the familiar co.t of capilal equation :

C orr -fro, '"" +J- -F., •. ,j , '. J "-, . J '.j./J

whe", tbe industry and ~t_speo;-ific cap illli gai ns are &i""n b~

(9.9 )

(9 _10 )

The: cost of capital equation inc ludes expected capital gains of assets. v.tlich may differ lICl'06S industrie" n.e speciflCJ.t;on allows for both industry and 111<set ~ific price differe nce~. The adva ntage of ind ustry and asset spe­cifIC investment prio:ing i~ t"at thi s metllod "'strict. investment-price-deflators to a lesser degn:e and hence delivtfs I1lOI'e real istic and appropriate "'nll$ compared to other method~.

llIe tfo ProdLICbvity dau. a l50 C()nla;ns data on labour inpuu deri ved using the Jorgenson and Sliroh (2001 ) methodology (wllich is simi lar to Ihe clPiw Slod; methodology ). Labour input dala are obtaiaed from the German Slali . ­tical OffICe dati . which conLl.ins information on "'ages and toooI ho::>un worted! For now the dataSet incllJdes labour quality d ala from the Oronincen Growth IU1d ~""Iopmenl Data_'

Llboor ~rvices fIowJ are ."umed \0 be proportional 10 toWl boors worked:

(9.11 )

where H" are toW] 1looT5 worUd in indu$Uy i in period I and q" is a COOSUnt of P'Op<>i tiona]ity_ This methodology is s imilar 10 the one implied by SlCindel and Stirch (200 l) and llS!itrn'1i cO!llJml1'bility with OIher growlh accounting studieJ for Germany.

9.5 EXPLANATION OF KEY INDUSTRIES

A prime eKarttple of the differences bt:t"<:en lhe ~allie·added and Jf(I5I; 1"0-ducti~ity ITC8'lUres is the IThInufac:ture of vehicle, (the nanspon "luipment _tor). ,,·IIicll IIa& ..... ayA ~n un;q""ly dqlelldenl on complex $Ub<:on!rac:'l­

Ing networks_ Since the intn)duclion of lean-management and j USl -jn-lime production the mid-198Of., the industry "- creawl .srruc:rurcs of Icplly indc_

pcndenc compa~ie~ Ih~t are dO$Cly ~Iated . 'oog clle va!",,_~ chlon, fiNt· Ii« 5)'S1eItl4U(l(lI~rs are know·how driven cOIl1pIInin "'ith M difttl lCCell 10 the major allm maltllflldure<S. TIocn follow ICVfflI! ")'eTII of p<IfI n od c0mpo­

nent $Upp1~1'l! thar ~y or may not reside in the _ tor' s classificarion. (h.ing the sec:and IIaIf oflhe 199Ot. lhe d~l\aIion oflCT IcchnolOSin

provided an additional stimu!UIi co spec1.! ilC pnxIuttioll among subool'ltracl­in& compuics. E-broslllns 11K btooi.oe ..,iclllll"ly importarll in dlil 5eC1OI".

allhoush mostly for commoditiel .nd not for key car parU and componcnu. T1Ie impac:1 of lhe IUloonoti ve jlfOOuclIOO nelwod:s on growdi and Of! ptodut· Ilvily cln he di!ICUMed on 1"'0 le,'c!\. We can cumine fim the TiK of em· cieney as • result of special iMtion .nd S'''I~","n( 'M\lCtion of IrInslIIO'lion OO$U in add, tion 10 the reduclioo of labour CO!01:i dlle 10 OIl!Jour'l:i ng : and lee­

oOOly. the scope of R&D in tile , rowth and proouc,ivity of the automotive iooUlitry. The 1.lIer poIlIl is key 10 the under"SWll1in& of Ihe dlff~nces itt tilt TFP conaPIi.

VII"" added ,n Ihe 1IlIOmOli~ indllStry lit defined by Ihe Europeu Il1O­

_nc lllU~ . roup 34 (NACE Rev. I) ~ the manufllCtllft of clllira. car bodocs. dri~ trains erc. Ho~ver. I nwnber ofparu WI are dc5tined 10 be _mbIed (&l.IC1I IS .. ind-screen", pl",1 Cit.) ha. e Dever ~n pilll 0( tIM: aulOmOllve induW)' production. T1IeIe products are ;lIICrmedi.lel u..c IIfC

ukcn into account only in Tfl'y bul not in the 11'J>VA concept. As long IS theR arc no chanJljeS In IIIe autorooc ive industry struc:ttue. ICChnoloty and ils upstRlm linb,&cs "'ilhin the prod..cliOll ~Iwork. both conc~pl!I ....... ld pr0-

vide inlcl1cmporlUy comparabk: n::sulll mal differ oaly in \c.vels. Two rrlIIJOr developments lutve. ho~w:r. Iffected uf'W"um lillbJljeS: i.·

no'·alion Ind OII[J()Ur<:ln,. Both aeti~ ll ie, induced I reduclion of the man,,­faelUTin, 1X~lfllion. A,5 measu.ed by the share o f vllue added on loul Jl"OSS-OIllput . manuflCIUrin.tl peMlnu ion to/llrac:tcd d ramalically between 1993 I nd 200 1.

TlII'hnoologica l Pf"OglCSi IW to Ihe dc~eIOp~nt of a broad IlInJe of ~w compoMnu rh.1 m ind ispeo ... bIe in tile rnode.n car. M".t of lhesc elc: ments are nOl ~llflCturcd by tho: lutomot ive induury. ,,-& . ""Ii_blockin, ')'SIems and tlllL"lioo control ( .... as. ASR ). 1ir bap. '11~itionin • . n- ill1tnnedi. ~ iapuu m:ei,'c 11\ ever- hilher share i. 1he I0Il1 val"" 0/ 1 car IW1d haw: be­_ imporlant 10 .Ior! OOflllnuro growth in !he demuId for . cllielcs. G __ output lull been lIimullttd by thil dudopmc nt. "'''''IUS lhe v.lue added ClllIII)!: .x_I for rhi5 increase ova" Ihe period under invc,,;,.ioo. Ou~inl lo wboonlfaclOfl; has lOll, been • priority in the ",1OmOI;ve

industry. Th i5 ~nd has been drive n by cc:onomics-of·scopc and price differ· entials in f.,tor input •. Beyond the opportuniliei provided by !he idv.nced ICT. tM elutOon of the Single M",ket and lhe JltteSSion <;>I CEE COlI_trieS ha~e I I,.., cuotn buwd to c,.....bo!'deT OUIlQUfCin&. GencnlLy speak;",. rhr:K

acTi~ilie.~ induce em! reductions based on lIulorootivc companiu' cleciJions 10 mau·or-bIIy . In ll'aJ lerms, the u.lo(:a!i()n h.:I~. u'~ds p(";b~$. the June im­pact on !he 11-"Py and TFPv .. "'MAe"n rompanie$ shift ~broad their capaci­ties withoul fric: lioII . If fador nwUlllfC ItOI ptrfeetly flexible. ho_vn. 1VP ",in be; a roe",!;VC durin, ~ IdjuSU1>Cnt pniod. This m:g"'iv~ dip ""ill be ..-" proIIOUlICed fuf-~ TFPv~ C'OfIC%pI.

9.6 CONCLUSIONS

The OOCP Productivity Manua l (2001. p. 10) 5IUtu tllat the (hoi~ between produclivity mcuura depuds in part on the purpo5e of the productivity measure. In princ iple. the value-added and &ross output·ba$cd measul'Cl an: me&iIIru of two diffnUI concepti. We ronSlfUCI ami employ . no'"el 1(0 Productivity D"abaw COIIMnlC\ 10 allow the analysis of JI'OSII OIItput prodllC_ livily ;Q Oennany. we Iho:- compIfc SttIOnI productiville!; with uirin, ,"alur·addcd mc;mares in m. l ilefallll~

The OOCD (2001. p. 21 ) coocludCl ttw nch lhta5lJfc l1li$ its ptac.c , de:­pend in,oo the l tafldpoint 1doptNl. 0... re ... lu suggest that von OOlpul and value·addlld hued TFP nu:asures arc ...seful romplemclKS. 8 0th folio", roughly the lame tu.nd on the aun:g.ale .• 1though they do divngc signlfi· cantly on the secloral leve l. When techn ical pr08R'lS aff~1S all faclOrS of produclioo propor1ionally. the value·added measure i9 the silJlpler mI'.,..u. of Itthnicll chanae. Generally. the ~ OOlpul-tNosed TfP mcuurc if leu., ... sili", It! " 'ualions of OtIIIOUrCInJ. , .~. \0 cllanaes In the degree 01 verticar in_ ICgrarion iM:IIWCn KCIOn- v.rpe...-dded ba$cd 11-;> measures .no corumi_ naled by clllll1,e:I in OIruOVrc: ing .nd provide an indica1ion of the;~ of the producTh it)' impro~errcnl for lite economy as • wIIole.

NOTES

I . ""' . ...... '_ oT ... _,.._ ..... " ............. _._a..- __ 131a" 110 1'I\doo<I'' 'y rw+w ........... .. ""- ............ ....,.. ..... _ tIIo a .... t.hIo "'"" OIl "'" " ... , ~ _v'~. n". " tIIo ... for __ ..... .., .. _ """'" _ n.. _~ __ .. __ . ", ... -.t ....t~,.... 1100: .. ;_ .. _....,... ............... ''"''''0l0i& .... _ .... ---l. llo< 11'0 1'\ , t " h it) ...... _""" " ..... 0( , II< 0..- _ .. 0fI\<& .. ........ rOf 0.-...... tIeotion.

l . n.. .. I""" .. _,.... ..... "' ...... ~,,_and "'" '" .. ,_ ... ~ ...... 4. n.. 111_ _ ... -. io ... _ '" -.....r_;.. OIl .... '''''"'''l' ..... duo ks _ '"

." .".. .......... __ .. boot < ..... r ....... __ ,.".. _ .. """_ "_-..-.pI. Tho ...... dO __ ""_~ ,_ ~ .. ,,.. ...... ""'" 1. __ ot: .....

..... __ ............. _ io_ in .... ...-,_,..,.-. ..,-..Ii<4 ...... iIIoINIrt_........... ,"' _ cf .... . "'" S. floe _ .,( ,"" ...,..... , ...... ' ,..., , ...... ,I0000 , II< ... .,... , .. 0( m ....... hott.d ... So.;n<I<l .. 511 .... (10)1).

6. Totool -. -ud om ...... _ d,,,,",",,,,",,, <lola "'.....,..., .......... for 29, 1 ,its. DIlly _ au 'I. <lola .. ....... , hi&kr i.....,. .,>d ...... -. ........... izoj ") ..... if>< ............ _ --.t _ ~ .. -. _ _ ... "'" _ t ... load ,100_ <"- iIo ............. _ .... __ ." .. "" o ... ' ilk ..... .

7 no. W"G l"Iodoo<:!i>il) ""_;. o ....... ,~ bei,. •• 1*Idt<! '0 i",,_ 1 ... ",_ ond Soi ..... (2DJ11 _'''''''''11-' Pk_ .. fer '" ........ .zzd<._1ar """" _"",00 .... . . .....

loo ... o( ...... qoMI~y

BtBLiOGRAPHY

ArIc . 8 . wan _ O. Pi lll ( •• b . 199]). Erp4w1"'1 u-lMk G __ I.: £s-p '" lIo.r­"", of An .... MaddJ.o.r . Sooh·HoIbnd. "1I'I!oImIIm.

Dally, M.N., C. Hil i,." and O. Campbell (1 9<)2). "The <liStribu""" 0( prto<Iu<"1 i';,y ,II -..tlttUrin. pl .. ,,' , BfO<OiMl' l't>(Kn ... ~ Aaiviry; Mic~.

PI'- 187-2.9. a.ldwin. I.R., T.M. IUn:tII<Iui -' 1._P Ma)Nfd (2(1U ). 'ProdIIcl>"" Y .,. .... h in Canod."'" ,ho UnilOd 51 __ ' . in I .R. 8aldwi~ . D. 8oc:klleOd. N. 0Il,1i_1. R. 0.. .

.--I, V. aa""rc .... l!. T.M. Harehaooi. J. HOIein. M. KICi. and f.·P. Maynard, Pro· dMt;.iwry 0 ,., .. ,11 '" C ... ,... SIMOOco Canada. 011 ..... "". jI-6(l.

I ........ M. MCI J. l-WUI (lOOO), 'PrndllClivicy in tile 199Ot: •• "idao;:e rrom lri.W1 pI_', Qo....... "h.y. Un!\CtsIlY of loDdon, hllj>:II_'W.qm ... lOC.II kI .... goel '31.

Ihrtel. rnan. E.I . -' M. DoIN (1000), ' UndcnLM>di •• pn><lll<l •• i.y: t.es.-u r,om tonai,"di,.., mief'QdaLo' , 'o..mal of &<>ot-x U'm .. ~IY. 13 (3). PII. ,69--'1(.

Bernard, " . and C.1. Jo_ (11196&). ·!'ru<Iu<Ii."y ond ec."U,.hCC Aator.I US S,_ ond 11IIdIutries', c.p;ril-al c..-...I<s. 11. 1'1'. 113-1 n .

Bcmonl. A.. .... CJ. Icnc:I ( 1996b). ·Produah'.y _ , Indusui(, and COUIllnao: nme ScriQ Theory and Ev,dmoe·. R,wwc{u....otIIIa....J SrtJIl .. ia. 7a (I) , pp. 135--146.

8cnxII. E.R. ond 0 .0 . Wood (1 911), "TecMoIo!y. pricu ond the dcnocd dmWId I'uo­meru'. Rr.w oj u-Ia aM ~. 5'7 . pp. lS'I-6S.

&n,o"", M. (1980). 'DuoI~y. intermediat. ''''''''. ond .-.l..e-oddlld' , in f .- M. ooId MeFoddao. D. (<<Il). Pm4..c<W1I &o!tomb: A Dwll App"""'" Ii} Throry alld Ap. pl.ca'""", Vol. 2. Nonh.llollond, " ........ dam, I'll , 3- 16.

Cnfu.. N. I<tId T. Mills (2001 ). 1W OrowIb In BtI.W! and 0 ....... M ..... t."mriRI 19»-- I 996'. CfJ'R dl.:.wl"" flaP" J01S.

0-.. B.R. tnd M. J. Ibtpor (2000). ~ ilLS PT-odue!i";'Y MeawocmeM P'<>8' .. ,,'. US 8 wuu of Laboo- S .... 't!.CI . ..... uhinl" ...

Dean. E.R .• MJ. Ilwpct ond M.S. S~ (1991\). · ProWtti"~y _KmcnI .. ~h -~~t:hI indcan of ""'puIS ond i!!pUll·. In OECD. l..dowry ProtlMalvlry: I_~~_ Mftl. • ........ 1 .. -... 0EC0, Pwts. pop. 183-21'. Die_. B.W. ( 1980). ' Hido' ~ .heon:m _ the ni_ 0( a real valuo­oddod "' ... ioII'. in M. "" .. and 0 MeFoddtn (odll. "rod, .. "io~ F_ "' /N: A o.."J ~r. Ii} ~ aM AppliC<>Jiotr.£ .... 01. 1. NorttI ·HoIlond. AmI",rOam. pp. ll_J I.

Die .... ". E, W, (2001). ' ProduClivil) m:oo. and dotermi""nu in Cao>dli·. DilallSKln Poper no. 01 _ 1 ~. o..""n"", ... nf 1lc<KJOmics. Uni.eml, of !kiliw Columbi a. hUp:llwww.ec<III. ulx: .<:I/di "'P"P<'r'$/dpO I 1 , .pdf

Dic:wcrt. E. W. """ A.O. Nabnuro (1998 1. 'A "''''e)' nf empirical mr:Ol~ 0( productiv. it Y """",,,_,' • _<p:lll i]y.iP<' .ucltica&o.cdul-k1,*""iIIon<IonI!'ap<diewnak.pdf

Doniiii'. E.D. 1961 . '00 ,he moasurunem 0( ""'hnological </lanSC· . &o.rlJlrOC J""T' 1101. 71 . !'P. 709-29.

fl";&. 0. on<! V. S..,iDel' (1993). 'Sean:hI Rg fo< .... product.,,!) _ : ..,.,.,. .... . prilil\i findi",1 from We!' G<:rman M",u flOC.'lurinl·. RM~'" of &(IJt_k. ""d S'd/i,,;"" 15. pp. 57-65

(Joeoli<h. f'J . • 1Id V. Om (1 994). ·~I ar .. n pru:i'H~'" marJinal "nd economl .. de c\I<al. "" I. induillria ~Ia: I %4-191!~' . Rrn<la rh E",,_I(j !tplkwM, (; (2).

pp. 29-'3. Cloo'rlich. F.J. and V. On l 11996). 'Economfas de Meola. e.tcmal idade. and IIlesc.

...,..,ien,o de tW>.jo en I. indumia ospallola. 1%04-1989' . R~.'il'" J~ ENMo",{a !tpl;cudtl. I I (. ). pp. 151 _ 11>6.

QoIlop. F.M, ( 1979~ ·Acoounlin. for imamt<iilll. input: The liM: bet_" I<'(lorai and oa;rcp lO IJ'II:2SUlCS of prod"",;.i" growth'. In N""onal ~""h C""""il. M_~"' and Inurprr" .. I0/l of Prodw:,M,y. National A<:odemy of S<:~. WlSIt inllOl'l. pp. 318- 33.

001101'. F. M. at>d IdJ. RoI>eru (198 1). '1"'1'0"«1 i","""""i • .., in ..... : Its I.".,... Oft

oecto<" prooJ"'1iv~y ;n US m"lUf"",.rin!·. in Doyarna.;:i. A. Ind Adom, K it. (tdli. !tUrt~"'~ and IIId~s,,,..vwl Prod.eIM" """'yro. Man;nu. N,jt.ofl' Pub­lishing. Bo.lOfl. 1'1'. 149-g(),

Gn:~on. P.K. . 1. Gali ..,d D. Parham (2002). 'Uptab; and iml""'l ' of Icr. in ,he Au ..... lian f!WOomy. E~ from aggregate, S«\oolI -' firm Ic vd , ·.1'ap!r ~ ro< the Wooiuh<>p on leT ond 0.. , ;,,"" 1'I=r1'0Im0I'W:<'. OIlCO. I>ari .. 'I Dew'iba,

Qu. W. and M , S. Ho (2001 ). 'A C(II"I'IIIUon of pIO<i"""vi,y a", .... h ;0 lIWIuflClUrillj beI .. 'ceot Canada and .he Uniltd Stal<:$. 1 % 1_9'. in lotgel'6Oll. D.W. """ I.cc. FT. (eds). 1.od"'I?'I~wl ProdMc.iVIIJ ".od IltUO.flllitHlill C""'IH'iliwMSS B""'~~ C""",," "l1li 1M Ullil~d S''''''$. Ind ... ,,, Co_, Oua...... !'P. 121 - 1 Sol. hllp:l{OI/'Oo1e&i., ic .I!C .~_«nmyfmtral"'RdOCll)l.btml.

0.. . w .. P.e. l.cc """ •. TIInJ (2001 J. '~ """ prnd..:l,"';ry ",,>loth in Canadi.., in .... """"·, in JCIf'JC'I"OO. D.W. and lce, F.e. (tdt). /~tw/ Pn;>dwlWity IIild /"",,,,,,,/onal C~ili""~aJ! &_ CD1Iado a1td 1M UIIi"'" S_s. IndiQUy c.., . .:Ii, DIu ..... 1'1'. n-I20. b~p~/"""" i" ic. ... c",,,,,_ecnmy/muWengd0002..htmi.

Gullickson , W. ( 1995). ' Mea:iIlll:menl o f prod\lCli,ity growth in US manufactu';",·. Men/hly u.bo, RM~",.!'P' 1 )_28.

OullicUoo. W, """ M, Hall'"' (19991,). ' ''''''ibl~ InCL\Uremctll bias in ",~,31C pro. d"",;,i" "" .... h .. MIMIMy LtJI>o, R .... /r.;. 1'1'. 47-67.

GuliidJon. W. and M. Hupcr (1mb). 'Prodoctioo Function$. [nput.().t.pII' Tllbles. ond tile Rclalion;hip bel"""", I ........ " and AU Il:p'e PnxI"",i,i" M_",,', au · ~ 01' Lobo< bi.ua., Washington . teb"*J_

o..lIid.s.on. W. ond M. Haoptr (2002). ' Oill$ in au"' ... " prndllC1i"';ty ,/'efIds rev; .. itrd· . Mon,/ofJ I...aOO, R~~. 1'1'. 3~.

Horc_i. T.M .. M. K..,; ..... ' .·P. Maynard (20011. 'The SLalistics CaMda pro<Iuc. ';vi' ~ ""'..,...,: COlI«pO' and _thod,·. in J.R Baldwin . D. Boc:1tJ_. N, Dhah· .... , R. Durand, V. <Jao<Irnauh n T.M. ~i . J. lI_in., ),f, Ka:i and 1.· P. May· ....u. P""'_";'Y Gro ... 1t it! Co"",*" S .. i$lie< C..-, Ottawa. Pl'. 143_76.

Ilatris. R. ..,d M. TrajOOf (1991). 'Pnxl""'i.i'~ m o.ov.th in UI( Itc,ion!;. 1%11-91' . O,rf<wd 8ull"';" of~' """s",IUI;a. 59 (. ~ [>1', ' 85-SIIl

lI otnMklo, I. ond S. !,,~fIc, (lOO2), "ille C"" ,ribluion of leT .~ &ooornic Ae,i,i,y: A Oro"ltb A«o\In.ina E>.erci!l<': Wi,h SporoWl Firm L",,01 n.a'. a .... o doo E!.p.I\oo. 1I.-rd1l)q)onmtnl. MA.

Hul.tn. C.R. 2000. 'T ..... F.., ... I'todu<t,";." A Shoo 8ibliotnflhy· . NUI""",I 8~ · ""'~ ofUonOlftk Rn""",hW",l ,"MII P"P"r. 7~71. NIlER. CAmbn. MA.

10fJC'II0ft. D. W .• F .M. 0011"1' ond 8 .M. F ..... ".,...; (1981). Prod..::';. ;" "'" US &1>. _G ... ~ .... 1Ian&l\l Uni-.ity Pn::os.,CM"boidt<. M it..

lor .......... D.W and K. S,ir"oIo (lOO')' ·Railin, ,he ........ ,,"',,: US cconQnlil: IJOWOh in ,he information .,..'. in D.W. l<q<n'lOll ItId P.C. L.- (odl ). 1,..I"",'y .. U>.M1 P>DilIlCrMl1l11W1lrur"""nOftlJl C~ht"'n ..... Ik_ C .......... .-.I ,Iv U""N Sf.Hu. lndlntryCanlda. au ....... pp. 2~75 ..

lu",. S .. K.S ... 8 .C. MO)'C' ond R. E. y...uv. (2000), ·hnpm ¥Cd ... imola 0( "",.

",00"'" by iI,dun/}, for 1 941~98". S~rwJ "fC.u",II 8 "" ......... June. pp. lJ....54. OECD (OrJ ...... ion ro. F ............. ~ion and Dc 'dopooOUll.. 2(01 ), f'ruJw.. 1M" M_..aI· A c.-. ... rIv M_,-", II{ INlfUl"..kwllittll AU"'''''' 1' ....

dw#Yiry 0", ,,'11 .. DECO. Pari ... Ouh"" . N. (1991). ~oq><ti,ion and .he di'9U'ion or ,..... product •• ;')'-.101

UK ~I". OKlo'" """" ...... ,.. ... 51 (I), pp .. U-1I. Ouno.. N .. (2000). "MUS! On:>wIh II .... Dodi....? a OPnoo!' IU_""""CI Oto""" Rc-

."iled· . B_~ of FA,"'''' / W~/"M Paprr. 107. J"""",y . ......... banko(tnjlland.o;:u. ~i ... popcrsI\to'p ' 01.pctf

0..11OIl, N. ond M. O ' Mahonl ( 1994). PMJ..mwiry ..... 0 ..... 110. c.mt.rio;IF Uni,,,,,· lilY PttU, CombridJl!.

P. {200II. 'TIle OECD produot;" ly manpai : A i Pi(lc ,y , ... ""'lIluremm, of ., .J't'Ptc prodto;:ll.i, y·, /M/r"",1i1!ffllJ Prodw:ri .. " MOIIito<,

In 'he eonwrunic.,ioo • • • '",r. '" o,~iew· , Pop'" ;, !kwk.