IT Investment and Hicks' Composite-Good Theorem: The U.S. Experience

Transcription

1 University of Pennsylvania ScholarlyCommons Business Economics and Public Policy Papers Wharton Faculty Research IT Investment and Hicks' Composite-Good Theorem: The U.S. Experience Jaime Marquez Shing-Yi Wang Follow this and additional works at: Part of the Business Commons, Economics Commons, and the Public Affairs, Public Policy and Public Administration Commons Recommended Citation Marquez, J., & Wang, S. (2003). IT Investment and Hicks' Composite-Good Theorem: The U.S. Experience. FRB International Finance Discussion Paper No.767, This paper is posted at ScholarlyCommons. For more information, please contact

2 IT Investment and Hicks' Composite-Good Theorem: The U.S. Experience Abstract We study whether aggregation residuals in U.S. private investment in information technology (IT) exhibit a predictable pattern that is consistent with Hicks' composite-good theorem and that may be used for forecasting. To determine whether one can extract such a pattern, we apply the general-to-specific strategy developed by Krolzig and Hendry (2001). This strategy combines ordinary least squares with a computerautomated algorithm that selects a specification based on coefficients' statistical significance, residual properties, and parameter constancy. Then, we derive the testable implications from Hicks' theorem and evaluate them with econometric formulations; we find qualified support for these implications. Having obtained these formulations, we evaluate their ex-post predictive accuracy and compare it to that of an autoregressive model. The key finding is that ignoring movement in relative prices results in a loss of information for predicting aggregation residuals. Keywords aggregation errors, fisher aggregates, divisia aggregate, general-to-specific Disciplines Business Economics Public Affairs, Public Policy and Public Administration This working paper is available at ScholarlyCommons:

3 Board of Governors of the Federal Reserve System International Finance Discussion Papers Number 767 June 2003 IT Investment and Hicks Composite-Good Theorem: The U.S. Experience Jaime Marquez and Shing-Yi Wang NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an acknowledgment that the writer has had access to unpublished material) should be cleared with the author or authors. Recent IFDPs are available on the Web at

4 IT Investment and Hicks Composite-Good Theorem: The U.S. Experience Jaime Marquez and Shing-Yi Wang * Abstract: We study whether aggregation residuals in U.S. private investment in information technology (IT) exhibit a predictable pattern that is consistent with Hicks composite-good theorem and that may be used for forecasting. To determine whether one can extract such a pattern, we apply the general-to-specific strategy developed by Krolzig and Hendry (2001). This strategy combines ordinary least squares with a computer-automated algorithm that selects a specification based on coefficients statistical significance, residual properties, and parameter constancy. Then, we derive the testable implications from Hicks theorem and evaluate them with econometric formulations; we find qualified support for these implications. Having obtained these formulations, we evaluate their ex-post predictive accuracy and compare it to that of an autoregressive model. The key finding is that ignoring movement in relative prices results in a loss of information for predicting aggregation residuals. Keywords: Aggregation Errors, Fisher Aggregates, Divisia Aggregate, General-to-Specific. * Corresponding author. address is jaime.marquez@frb.gov. We would like to thank Joe Gagnon, David Hendry, Jane Ihrig, David Lebow, and Joel Prakken for their detailed comments. We also benefited from conversations with Flint Brayton, Neil Ericsson, and Dale Henderson; we are also grateful to participants in the International Finance Workshop series for their remarks. Our calculations are based on PcGets; see Hendry and Krolzig (2001). The views in this paper are solely the responsibility of the authors and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or of any other person associated with the Federal Reserve System.

5 4 Lqwurgxfwlrq Wkh frpsrvlwh jrrg wkhruhp/ srvwxodwhg uvw e Klfnv +4<69,/ vwdwhv wkdw d froohfwlrq ri frpprglwlhv fdq eh wuhdwhg dv d vlqjoh frpprglw zkhq wkhlu sulfhv pryh lq sdudooho= 4 D froohfwlrq ri sk vlfdo wklqjv fdq dozd v eh wuhdwhg dv li wkh zhuh glylvleoh lqwr xqlwv ri d vlqjoh frpprglw vr orqj dv wkhlu uhodwlyh sulfhv fdq eh dvvxphg xqfkdqjhg/ lq wkh sduwlfxodu sureohp lq kdqg1 M1 U1 Klfnv/ 4<69/ sdjh 661 Lq wkhru / wkhq/ fkdqjhv lq uhodwlyh sulfhv lpso wkdw rqh fdqqrw wuhdw xqlwv ri uhdo frq0 vxpswlrq dqg xqlwv ri uhdo lqyhvwphqw dv wkrxjk wkh zhuh glylvleoh xqlwv ri d vlqjoh frp0 prglw fdoohg uhdo JGS1 Qhlwkhu fdq rqh wuhdw glhuhqw w shv ri lqyhvwphqwvwuxfwxuhv dqg htxlsphqwdv wkrxjk wkh zhuh glylvleoh xqlwv ri d vlqjoh frpprglw fdoohg djjuhjdwh lqyhvw0 phqw1 Wkh irupdw ri wkh X1V1 qdwlrqdo dffrxqwv uhfrjql}hv h{solflwo wkh srwhqwldo iru wklv odfn ri dgglwlylw e xvlqj Ilvkhu djjuhjdwlrq dqg uhsruwv wkh glhuhqfh ehwzhhq wkh Ilvkhu djjuhjdwh dqg wkh vxp ri lwv frpsrqhqwv> zh whup wklv glhuhqfh wkh djjuhjdwlrq uhvlgxdo1 5 Lq sudfwlfh/ djjuhjdwlrq uhvlgxdov iru vhyhudo ghfrpsrvlwlrqv ri uhdo JGS duh vpdoo +jxuh 4,> wkh odujhvw ri wkhvh uhvlgxdov kdv d phdq ri d whqwk ri d shufhqw ri uhdo JGS1 Djjuh0 jdwlrq uhvlgxdov iru djjuhjdwh frqvxpswlrq/ h{sruwv/ dqg lpsruwv duh dovr vpdoo1 Exw iru lqyhvwphqw/ wkh djjuhjdwlrq uhvlgxdo uhdfkhv hljkw shufhqw ri djjuhjdwh lqyhvwphqw e Wkh sdwwhuqv ri wkhvh uhvlgxdov udlvh lqwhuhvwlqj txhvwlrqv1 Iru h{dpsoh/ zkdw lv vr xqltxh derxw lqyhvwphqw wkdw lw vkrzv odujh djjuhjdwlrq uhvlgxdov zkhuhdv wkh rwkhu pdfurhfrqrplf djjuhjdwhv gr qrwb Ixuwkhu/ krz lv lw wkdw fkdqjhv lq uhodwlyh sulfhv lqgxfh wkh lqyhuwhg0x sdwwhuq revhuyhg lq jxuh 4B Wr dgguhvv wkhvh txhvwlrqv/ zh frpsxwh djjuhjdwlrq uhvlgxdov iru hdfk ri wkh frpsrqhqwv ri lqyhvwphqw dqg qg wkdw djjuhjdwlrq uhvlgxdov lq lqirupdwlrq0 whfkqrorj lqyhvwphqw dffrxqw iru wkh sdwwhuqv ri jxuh 41 Kdylqj irxqg wkh vrxufh ri wkhvh uhvlgxdov/ zh h{sorlw Klfnv* wkhruhp wr prgho wkhp lq whupv ri pryhphqwv ri uhodwlyh sulfhv dqg h{dplqh zkhwkhu wkh sdudphwhu hvwlpdwhv h{sodlq wkh lqyhuwhg0x sdwwhuq1 Ilqdoo / zh hydoxdwh h{0srvw iruhfdvw shuirupdqfh lq whupv ri frqghqfh edqgv dqg frpsduh wkhp wr wkrvh iurp dq dxwruhjuhvvlyh irupxodwlrq1 Zh qg wkdw iruhfdvwv ri wkh prgho edvhg rq uhodwlyh sulfhv duh pruh dffxudwh wkdq iruhfdvwv ri d prgho wkdw ljqruhv lqirupdwlrq derxw uhodwlyh sulfhv1 4 Iru d irupdo wuhdwphqw ri wklv wkhruhp/ vhh Ghdwrq dqg Pxhooedxhu +4<;3/ vhfwlrq 814,1 5 Iru Ilvkhu djjuhjdwlrq lq wkh X1V1 qdwlrqdo dffrxqwv/ vhh Wulsohww +4<<5,> Odqghihog dqg Sdunhu +4<<8,> Yduyduhv/ Sudnnhq/ dqg Jxluo +4<<;,> Edfkpdq/ Mdtxhwwh/ Nduo/ dqg Urffr +4<<;,> dqg Zkhodq +5333,1 4

6 5 Uhodwlyh Sulfhv dqg Djjuhjdwlrq Uhvlgxdov 514 Gdwd Frqvlghudwlrqv Wr lvrodwh wkh vrxufh ri djjuhjdwlrq uhvlgxdov ri lqyhvwphqw/ zh frpsxwh wkh djjuhjdwlrq uhvlgxdov ri lwv frpsrqhqwv= Qrq0uhvlghqwldo Vwuxfwxuhv/ Htxlsphqw dqg Vriwzduh/ dqg Uhv0 lghqwldo> jxuh 5 vkrzv wkhvh fdofxodwlrqv1 Wkh gdwd srlqw wr vpdoo djjuhjdwlrq uhvlgxdov h{fhsw iru lqyhvwphqw lq Htxlsphqw dqg Vriwzduh1 Wkxv zh glvdjjuhjdwh wklv lqyhvwphqw fdwhjru ixuwkhu lqwr Lqgxvwuldo Htxlsphqw/ Wudqvsruwdwlrq Htxlsphqw/ dqg Lqirupdwlrq Surfhvvlqj Htxlsphqw dqg Vriwzduh +LW lqyhvwphqw iru vkruw,1 Wkh gdwd lqglfdwh odujh dj0 juhjdwlrquhvlgxdovirulwlqyhvwphqw+jxuh5/erwwrpohiwsdqho,dqglwvhyroxwlrqforvho uhvhpeohv wkh lqyhuwhg0x sdwwhuq ri wkh uhvlgxdo iru djjuhjdwh lqyhvwphqw +frpsduh wkh erw0 wrp uljkw sdqhov ri jxuhv 4 dqg 5,1 Wklv frqvlghudwlrq lv lpsruwdqw ehfdxvh vshqglqj rq LW lqyhvwphqw dffrxqwhg iru qhduo d wklug ri djjuhjdwh sulydwh lqyhvwphqw vshqglqj/ d vkduh qrw pdwfkhg e LW vshqglqj lq rwkhu pdfurhfrqrplf djjuhjdwhv1 515 Dqdo wlfdo Frqvlghudwlrqv Iru rxu dqdo vlv/ zh ghqrwh wkh Exuhdx ri Hfrqrplf Dqdo vlv +EHD, Ilvkhu*v djjuhjdwh dv tc wkh dvvrfldwhg frpsrqhqw dv t c dqg wkh qxpehu ri frpsrqhqwv lq wkh dffrxqw dv Wkh frpsrvlwh0jrrg wkhruhp vwdwhv wkdw li uhodwlyh sulfhv duh fkdqjlqj wkhq t 9'P t dqg wkh djjuhjdwlrq uhvlgxdo - uhsruwhg lq wkh dffrxqw uhfrqflohv wkh djjuhjdwh zlwk wkh vxp ri lwv frpsrqhqwv= t 'P t n - 1 Wr prwlydwh krz lqvwdelolw ri uhodwlyh sulfhv jlyhv ulvh wr -/ zh xvh d vlpsoh h{dpsoh lqyroylqj wzr djjuhjdwlrq phwkrgv1 Wkh uvw phwkrg lv wkh Glylvld dssur{lpdwlrq wr wkh Ilvkhu djjuhjdwh> wkh vhfrqg phwkrg xvhv olqhdu djjuhjdwlrq1 Wkh jurzwk udwh ri wkh Glylvld djjuhjdwh iru t lv _t t e ' [ t e c +4, t zkhuh ' t Sr r t r lv wkh qrplqdo h{shqglwxuh vkduh ri wkh frpsrqhqw dqg lv wkh sulfh ri wkh frpsrqhqw1 Wkh olqhdu djjuhjdwh lv t ' S t c dqg lwv jurzwk udwh lv _t t e t ' [ t et t c +5, 5

7 zkhuh t lv d uhdo h{shqglwxuh vkduh1 Htxdwlrq +5, dvvxphv wkdw uhodwlyh sulfhv duh {hg t dqg wkxv/ dffruglqj wr Klfnv* wkhruhp/ hperglhv shuihfw vxevwlwxwdelolw dprqj wkh t r Htxdwlrq +4, grhv qrw dvvxph wkdw uhodwlyh sulfhv duh {hg dqg wkxv vxevwlwxwdelolw dprqj wkh t r frxog eh ohvv wkdq shuihfw1 Li zh dvvxph wkdw sulfhv pryh lq sdudoohowkdw lv/ li et ' [ e t ' [ # $ f t S et ' [ r f t r f/ wkhq # t t $ et ' e t c zklfk phdqv wkdw wkh jurzwk udwhv lpsolhg e wkh wzr phwkrgv duh lghqwlfdo1 Wkxv/ dsso lqj wkh vdph lqlwldo frqglwlrq wr erwk phwkrgv/ t f ' t f c lpsolhv wkdw wkhuh lv qr djjuhjdwlrq uhvlgxdo= t ' S t Wr wkh h{whqw wkdw sulfhv gr qrw pryh lq sdudooho/ wkhlu pryhphqwv zloo dhfw qrplqdo vkduhv dqg uhdo vkduhv glhuhqwo 1 Vshflfdoo / d fkdqjh lq dhfwv qrplqdo vkduhv gluhfwo wkurxjk ydoxdwlrq hhfwv dqg lqgluhfwo wkurxjk ehkdylrudo uhvsrqvhv lq wkh t r Lq frqwudvw/ d fkdqjh lq kdv qr ydoxdwlrq hhfw rq uhdo vkduhv> wkhvh fkdqjh rqo wkurxjk wkh uhvsrqvh ri wkh t r1 Dv d uhvxow/ fkdqjhv lq uhodwlyh sulfhv uhvxow lq glhuhqw jurzwk udwhv iru phwkrgv +4, dqg +5,/ d glhuhqfh jlylqj ulvh wr -1 6 Hpslulfdo Iudphzrun 614 Prgho Vshflfdwlrq Zh srvwxodwh wkdw djjuhjdwlrq uhvlgxdov ehkdyh dffruglqj wr - t ' 4 n c +6, zkhuh 4 lv wkh djjuhjdwlrq uhvlgxdo shu groodu/ ' c 3kEu Efcj 2 c dqg keu lv d sro qrpldo lq wkh odj rshudwru u Wklv irupxodwlrq doorzv iru wkh vhuldoo fruuhodwhg sdwwhuq lq wkh gdwd dqg lw uhfrjql}hv wkdw wkh lpsruwdqfh ri - *v pdjqlwxgh klqjhv qrw vr pxfk rq lwv devroxwh ydoxh exw rq lwv ydoxh uhodwlyh wr wkh dffrxqw*v rzq djjuhjdwh1 Zh vfdoh - xvlqj wkh Ilvkhu djjuhjdwh wr dyrlg lqwhuqdo frqvlvwhqflhv wkdw zrxog dulvh iurp xvlqj dq dowhuqdwlyh djjuhjdwh wkdw dvvxphv wkdw uhodwlyh sulfhv duh {hg zkhq lq idfw wkh duh qrw1 Iroorzlqj Klfnv* wkhruhp/ zh prgho 4 lq whupv ri uhodwlyh sulfhv1 Pdq vxfk sulfhv 6

8 duh vxlwdeoh fdqglgdwhv exw/ jlyhq wkh vpdooqhvv ri rxu vdpsoh/ zh frqvlghu mxvw irxu= wkh uhdo lqwhuhvw udwh/ wkh uhodwlyh sulfh ri rlo/ wkh uhdo h{fkdqjh udwh/ dqg wkh uhodwlyh sulfh ri frpsxwhuv1 Wkh uhdo lqwhuhvw udwh fdswxuhv wkh uroh ri lqwhuwhpsrudo vxevwlwxwlrq> wkh uhdo sulfh ri rlo fdswxuhv vxevwlwxwdelolw ehwzhhq hqhuj dqg qrq0hqhuj surgxfwlyh idfwruv> wkh uhdo h{fkdqjh udwh fdswxuhv vxevwlwxwdelolw ehwzhhq wudghdeoh dqg qrq0wudghdeoh surgxfwv> wkh uhodwlyh sulfh ri frpsxwhuv fdswxuhv wkh vxevwlwxwdelolw ehwzhhq LW0fdslwdo dqg rwkhu irupv ri fdslwdo1 Vroh uholdqfh rq wkh uroh ri vxevwlwxwlrq hhfwv dvvxphv/ krzhyhu/ wkdw vxevwlwxwlrq srvvlelolwlhv duh {hg ryhu wlph= wkhuh lv qr lqqrydwlrq1 Li wkhuh duh lqqrydwlrqv/ wkhq phdvxuhphqw vkrxog eh edvhg rq uhvhuydwlrq/ lqvwhdg ri pdunhw/ sulfhv +Ilvkhu dqg Vkhoo/ 4<:5/ s1 437,1 Ehfdxvh uhvhuydwlrq sulfhv duh qrw revhuyhg/ zh prgho wkh hhfwv ri lqqrydwlrq zlwk d wuhqg dqg srvwxodwh wkdw 4 ' 4Ef c c f e caoe?_ Lq wkh devhqfh ri dq dffhswhg wkhru iru vshfli lqj 4E/ zh srvwxodwh dq dxwruhjuhvvlyh glvwulexwhg odj doorzlqj iru qrqolqhdulwlhv= d keuo - t ' q n Aoe?_ n [ Euf n [ ) Euf 2 n c +7, zkhuh Eu dqg ) Eu duh sro qrpldov lq wkh odj rshudwru u Iru htxdwlrq +7, wr eh frqvlvwhqw zlwk Klfnv* wkhruhp/ uhodwlyh sulfhv vkrxog kdyh rqo wudqvlwru hhfwv rq - Ilqglqj wkdw uhodwlyh sulfhv kdyh shupdqhqw hhfwv zrxog frqwudglfw wklv wkhruhp1 615 Iruhfdvw Dssolfdelolw Li rqh uhwdlqv wkh irupdw ri wkh qdwlrqdo dffrxqwv/ wkhq rxu dssurdfk khosv iruhfdvw wkh Ilvkhu djjuhjdwh t lq whupv ri iruhfdvwv ri lwv sduwv= t s ' [ ' t s n - s c +8, zkhuh s ghqrwhv d iruhfdvwhg ydoxh Dq dowhuqdwlyh dssurdfk lqyroyhv xvlqj d Glylvld dssur{lpdwlrq wr hvwlpdwh EHD*v Ilvkhu djjuhjdwh> wkh uhvxowlqj hvwlpdwh lv t ' t E n 3 S e ' t Uholdqfh rq vxfk hvwlpdwhv zloo qrw uhsolfdwh/ hyhq lq sulqflsoh/ EHD*v sxeolvkhg 7

9 ydoxh ri t ehfdxvh wkh xvh dv lqsxwv wkh uhsruwhg ydoxhv ri t c c t lqvwhdg ri wkh udz gdwd xvhg e wkh EHD wr phdvxuh t 1 Dv d uhvxow/ uhfrqflolqj EHD*v t zlwk wkh hvwlpdwh t lqyroyhv dq djjuhjdwlrq uhvlgxdo - G t ' t n - dqg wklv uhvlgxdo qhhgv wr eh iruhfdvwhg dorqj zlwk wkh rwkhu yduldeohv xvhg lq wkh Glylvld dssur{lpdwlrq= t s ' t s n - s +9, t s ' t s 3 E n s ' [ ' t s s Sr r t s s r Htxdwlrq +9, lv dsshdolqj ehfdxvh lw plplfv wkh djjuhjdwlrq surfhgxuhv xvhg e wkh EHD1 Krzhyhu/ lpsohphqwlqj htxdwlrq +9, qhhgv iruhfdvwv ri h{shqglwxuh vkduhv zklfk lq wxuq lqyroyhv iruhfdvwlqj lqyhvwphqw sulfhv1 Wkxv zkhwkhu wkh iruhfdvw huuruv iurp htxdwlrq +9, duh orzhu wkdq wkrvh ri htxdwlrq +8, lv dq xquhvroyhg txhvwlrq ghvhuylqj ixuwkhu vwxg 1 s et s 7 Hpslulfdo Uhvxowv 714 Lq0Vdpsoh Hvwlpdwhv= Sdudphwhuv dqg Whvwv Zh xvh htxdwlrq +7, wr h{sodlq pryhphqwv ri djjuhjdwlrq uhvlgxdov ri LW lqyhvwphqw> wkh gdwd duh txduwhuo dqg vsdq iurp 4<;:=T4 wr 5335=T61 Iru wkh uhodwlyh sulfh ri rlo +s, zh xvh wkh orjdulwkp ri wkh udwlr ri wkh xqlw0ydoxh ri X1V1 rlo lpsruwv wr wkh sulfh lqgh{ ri lqyhvwphqw +jxuh 6,> iru wkh uhodwlyh sulfh ri frpsxwhuv dqg vriwzduh +z,/ zh xvh wkh orjdulwkp ri wkh sulfh lqgh{ iru lqyhvwphqw lq frpsxwhuv/ shulskhudov/ dqg vriwzduh uhodwlyh wr wkh sulfh lqgh{ ri lqyhvwphqw1 Iru wkh uhdo lqwhuhvw udwh +u, zh frqvlghu wkuhh phdvxuhv wkdw glhu lq wkhlu pdwxulw = wkh uhdo ihghudo ixqgv udwh> wkh uhdo 430 hdu Wuhdvxu udwh/ dqg wkh uhdo 630 hdu Wuhdvxu udwh> doo ri wkhvh h{0srvw udwhv duh frqvwuxfwhg dv wkh dvvrfldwhg qrplqdo udwh plqxv wkh FSL lq dwlrq odjjhg rqh txduwhu1 Iru wkh uhdo hhfwlyh h{fkdqjh udwh +t, zh xvh wzr phdvxuhv= wkh orjdulwkp ri wkh Ihghudo Uhvhuyh*v Eurdg uhdo h{fkdqjh udwh/ edvhg rq lqgxvwuldo dqg ghyhorslqj frxqwulhv* fxuuhqflhv> dqg wkh orjdulwkp ri wkh LPI*v uhdo hhfwlyh h{fkdqjh udwh/ edvhg rq lqgxvwuldo frxqwulhv* fxuuhqflhv rqo 1 Xvlqj wkhvh yduldeohv/ zh dvvhpeoh vl{ prghov wkdw glhu lq wkhlu phdvxuhv ri uhdo lqwhuhvw dqg h{fkdqjh udwhv1 8

10 Iru hvwlpdwlrq/ zh dvvxph d pd{lpxp qxpehu ri yh odjv iru wkh ghshqghqw yduldeoh dqg rqh odj iru hdfk ri wkh uhpdlqlqj h{sodqdwru yduldeohv> zh dovr lqfoxgh d <244 gxpp yduldeoh htxdo wr rqh lq wkh wklug txduwhu ri 5334 dqg }hur rwkhuzlvh1 Wkh xvhixoqhvv ri rxu whvwv lqfuhdvhv zlwk wkh holplqdwlrq ri luuhohydqw yduldeohv dqg wkxv zh dsso wkh jhqhudo0wr0 vshflf vwudwhj ghyhorshg e Nuro}lj dqg Khqgu +5334,1 Wklv vwudwhj frpelqhv ruglqdu ohdvw vtxduhv zlwk d frpsxwhu0dxwrpdwhg dojrulwkp wkdw vhohfwv d vshflfdwlrq lq irxu vwdjhv= 41 Hvwlpdwlqj wkh sdudphwhuv ri dq xquhvwulfwhg irupxodwlrqhtxdwlrq +7,dqg whvwlqj iru frqjuxhqf +zklwh0qrlvh uhvlgxdov dqg sdudphwhu frqvwdqf, Lpsohphqwlqj pxowlsoh uhgxfwlrq sdwkv vlpxowdqhrxvo 1 Rqh uhgxfwlrq sdwk frxog jhw vwduwhg e h{foxglqj wkh ohdvw vljqlfdqw yduldeoh zkhuhdv dqrwkhu uhgxfwlrq sdwk frxog jhw lqlwldwhg e h{foxglqj d eorfn ri yduldeohv wkdw duh vwdwlvwlfdoo lqvljqlfdqw1 61 Whvwlqj zkhwkhu wkh vshflfdwlrq iurp d uhgxfwlrq sdwk lv frqjuxhqw1 Li lw lv/ wkhq lpsohphqw dqrwkhu urxqg ri uhgxfwlrqv dqg whvw iru frqjuxhqf > frqwlqxh wklv surfhvv xqwlo wkh vshflfdwlrq ylrodwhv frqjuxhqf 1 Lq wkdw fdvh/ wkh dojrulwkp vhohfwv wkh lpphgldwho sulru vshflfdwlrq dqg odehov lw Ilqdo prgho1 71 Froohfwlqj wkh Ilqdo prghov iurp ydulrxv uhgxfwlrq sdwkv dqg dsso lqj hqfrpsdvvlqj whvwv wr wkhp1 Wkh vshflfdwlrq wkdw hqfrpsdvvhv doo rwkhuv ehfrphv wkh Vshflf prgho1 Li wkhuh lv qr vlqjoh hqfrpsdvvlqj prgho/ wkhq wkh dojrulwkp irupv d xqlrq prgho xvlqj wkh yduldeohv iurp doo ri wkh Ilqdo prghov dqg uh0vwduwv wkh vshflfdwlrq vhdufk iurp vwhs +5,1 Li wklv vwudwhj idlov wr lhog d vlqjoh Vshflf prgho/ wkhq wkh dojrulwkp dssolhv wkuhh lqirupdwlrq fulwhuld +Dndlnh/ Vfkzdu}/ dqg Kdqqdq0Txlqq, wr wkh Ilqdo prghov dqg vhohfwv wkh rqh wkdw plqlpl}hv doo wkhvh fulwhuld> wkdw prgho ehfrphv wkh Vshflf prgho1 Rwkhuzlvh/ wkh dojrulwkp idlov wr qg d Vshflf prgho1 D nh ihdwxuh ri wklv dojrulwkp lv wkh vhtxhqwldo dgmxvwphqw ri vljqlfdqfh ohyhov wr uhfrjql}h wkh mrlqw qdwxuh ri prgho vshflfdwlrq dqg sdudphwhu hvwlpdwlrq1 Wdeoh 4 uhsruwv wkh whvw uhvxowv dqg lqglfdwhv zklfk yduldeohv kdyh vwdwlvwlfdoo vljqlfdqw hhfwv1 Lq whupv ri frh!flhqw hvwlpdwhv/ uhodwlyh sulfhv kdyh vkruw0uxq/ qrqolqhdu hhfwv rq 6 Wr whvw qrupdolw zh xvh wkh Mdutxh0Ehud vwdwlvwlf zklfk lv glvwulexwhg dv " 5 +5,> wr whvw vhuldo lqghshq0 ghqfh zh xvh dq I0whvw iru wkh k srwkhvlv wkdw wkh frh!flhqwv iru dq DU+7, ri wkh hvwlpdwlrq uhvlgxdov duh mrlqwo htxdo wr }hur> wr whvw krprvnhgdvwlflw zh xvh dq DUFK whvw1 Iru sdudphwhu frqvwdqf / zh lpsohphqw wzr Fkrz whvwv= rqh h{foxglqj kdoi ri wkh vdpsoh dqg dqrwkhu h{foxglqj wkh odvw hljkw revhuydwlrqv1 9

11 wkh djjuhjdwlrq uhvlgxdo ri LW lqyhvwphqw/ d qglqj wkdw vxssruwv Klfnv* frpsrvlwh0jrrg wkhruhp1 7 Uhodwlyh sulfhv/ krzhyhu/ kdyh shupdqhqw hhfwv zklfk lv qrw frqvlvwhqw zlwk Klfnv* wkhruhp1 Wzr idfwruv pd dffrxqw iru wkhvh shupdqhqw hhfwv1 Iluvw/ rxu vhohfwlrq ri uhodwlyh sulfhv frxog eh lqfrpsohwh dqg wkxv lhogv eldvhg frh!flhqwv1 Vhfrqg/ fkdqjhv lq uhodwlyh sulfhv/ hvshfldoo wkrvh dvvrfldwhg zlwk LW surgxfwv duh lqgxflqj shupdqhqw fkdqjhv rq vxevwlwxwlrq srvvlelolwlhv zkhuhdv Klfnv* wkhruhp dvvxphv wkdw vxevwlwxwlrq srvvlelolwlhv duh {hg1 Ilqdoo / wkh wuhqg kdv d vljqlfdqw hhfw rq wkh djjuhjdwlrq uhvlgxdo lq irxu ri wkh vl{ prghov> wklv qglqj vxjjhvwv wkdw wkh lqqrydwlrq surfhvv lv dhfwlqj vxevwlwxwlrq srvvlelolwlhv wkdw duh qrw fdswxuhg e wkh fkrlfh ri uhodwlyh sulfhv lq wkrvh irxu prghov1 Wr dvvhvv wkh prghov* h{0srvw suhglfwlyh dffxudf / zh uvw uh0hvwlpdwh wkhlu frh!flhqwv h{0 foxglqj wkh odvw hljkw revhuydwlrqv5333=t7 wr 5335=T6dqg wkhq lpsohphqw rqh0vwhs dkhdg dqg pxowl0vwhs dkhdg vlpxodwlrqv> wkhvh vlpxodwlrqv xvh wkh revhuyhg gdwd iru uhodwlyh sulfhv1 Wkh urrw phdq vtxduhg huuru +UPVH, iru rqh0vwhs dkhdg vlpxodwlrqv udqjhv iurp 413( +P4, wr 417( +P8,> iru pxowl0vwhs dkhdg vlpxodwlrqv/ wkh UPVH udqjhv iurp 419( iru prgho P6 wr 61:( iru prgho P81 8 Ryhudoo/ prgho P6 kdv wkh kljkhvw h{0srvw suhglfwlyh dffxudf dqg h{sodlqv wkh gdwd zhoo +jxuh 7/ wrs sdqho,1 Pruhryhu/ lq whupv ri lwv vwdwlvwlfdo surshuwlhv/ wkh wuhqg kdv qr vljqli0 lfdqw hhfw/ rqh fdqqrw uhmhfw wkh dvvxpswlrqv pdgh iru wkh glvwxuedqfhv dqg/ lpsruwdqwo / rqh fdqqrw uhmhfw wkh k srwkhvlv ri sdudphwhu frqvwdqf 1 Hylghqfh ri sdudphwhu frqvwdqf doorzv xv wr xvh wkh frh!flhqw hvwlpdwhv wr dvvhvv wkh vrxufh ri wkh lqyhuwhg0x sdwwhuq ri LW lqyhvwphqw djjuhjdwlrq uhvlgxdov1 Wr wklv hqg/ jxuh 7 +erwwrp sdqho, vkrzv wkh hhfw ri wkh uhodwlyh sulfh ri LW surgxfwv + rq wkh djjuhjdwlrq uhvlgxdo +-*t xvlqj wkh revhuyhg gdwd iru Wkh fdofxodwlrq vkrzv d srvlwlyh hhfw ri rq +-*t wkdw jurzv ryhu wlph dqg uhdfkhv d pd{lpxp durxqg 4<<:1 Ryhudoo/ wkh suroh ri wklv hhfw forvho uhvhpeohv wkdw ri wkh djjuhjdwlrq uhvlgxdo ri LW lqyhvwphqw lq jxuh 51 Ghwhuplqlqj zkhwkhu wkh qrq0olqhdu hhfwv ri duh wkh vroh uhdvrq iru wklv lqyhuwhg0x lv eh rqg wkh vfrsh ri wklv sdshu exw wkh uhvxowv vxjjhvw wkdw qrqolqhdulwlhv duh sduw ri wkh dqvzhu1 7 Wkh frh!flhqw hvwlpdwhv duh lq dsshqgl{ 4/ dydlodeoh lq kwws=22zzz1ihghudouhvhuyh1jry2sxev2ligs253362:9:2 8 Wkh iruhfdvwv duh lq dsshqgl{ 5/ dydlodeoh lq kwws=22zzz1ihghudouhvhuyh1jry2sxev2ligs253362:9:2 :

12 715 Rxw0ri0Vdpsoh Suhglfwlrqv= Frqghqfh Edqgv Iljxuh 8 vkrzv wkh <8 shufhqw frqghqfh edqgv iru wkh h{0srvw suhglfwlrqv iurp prgho P6 ryhu 5333=T705335=T61 Iru 40vwhs dkhdg vlpxodwlrqv/ wkh prgho kdv vwdwlvwlfdoo vljqlfdqw xqghusuhglfwlrqv lq 5334=T5 dqg 5334=T6 +jxuh 8/ wrs ohiw sdqho,> wkhvh xqghusuhglfwlrqv lqfuhdvh zkhq rqh xvhv pxowl0vwhs vlpxodwlrqv +jxuh 8/ wrs uljkw sdqho,1 Wkxv/ wkh hhfwv iurp wkh <244 hslvrgh duh qrw hdvlo suhglfwhg e prgho P61 \hw/ wkh prgho*v suhglfwlrqv uhwxuq wr wkh dfwxdo sdwk dqg/ lqghhg/ wkh odvw wkuhh suhglfwlrq huuruv duh qrw vljqlfdqw1 Wr dvvhvv zkhwkhu pryhphqwv lq uhodwlyh sulfhv duh lqirupdwlyh iru iruhfdvwlqj/ jxuh 8 dovr uhsruwv wkh <8 shufhqw frqghqfh edqgv iru wkh suhglfwlrqv iurp dq dxwruhjuhvvlyh prgho wkdw xvhv yh odjv dqg d frqvwdqw1 Frpsdulqj wkh rqh0vwhs dkhdg suhglfwlrqv dfurvv wkh wzr prghov vxjjhvwv wkdw ljqrulqj pryhphqwv lq uhodwlyh sulfhv uhvxowv lq qr orvv ri lqirupdwlrq= wkh UPVH duh yluwxdoo lghqwlfdo iru wkh wzr prghov1 Exw iru pxowl0vwhs dkhdg vlpxodwlrqv/ wkh wlph0vhulhv prgho vkrzv d jurzlqj suhglfwlrq huuru zklfk wudqvodwhv lqwr d kljkhu UPVH= wkuhh shufhqw uhodwlyh wr 419 shufhqw iru wkh uhodwlyh0sulfh edvhg prgho1 Wkhvh uhvxowv vxjjhvw wkdw ljqrulqj pryhphqwv lq uhodwlyh sulfhv uhvxowv lq d orvv ri lqirupdwlrq iru suhglfwlqj djjuhjdwlrq uhvlgxdov lq LW lqyhvwphqw1 8 Frqfoxvlrqv Wklv sdshu vwxglhv zkhwkhu djjuhjdwlrq uhvlgxdov lq LW lqyhvwphqw h{klelw d sdwwhuq uholdeoh hqrxjk wr eh vwdwlvwlfdoo fkdudfwhul}hg dqg xvhg iru iruhfdvwlqj1 Wr ghwhuplqh zkhwkhu rqh fdq h{wudfw d sdwwhuq iurp wkrvh uhvlgxdov/ zh lpsohphqw vhyhudo uhjuhvvlrq prghov dqg whvw zkhwkhu wkh sdudphwhu hvwlpdwhv duh frqvlvwhqw zlwk Klfnv* frpsrvlwh0jrrg wkhruhp1 Zh qg wkdw fkdqjhv lq uhodwlyh sulfhv khos h{sodlq vkruw0uxq pryhphqwv lq djjuhjdwlrq uhvlgxdov1 Zh dovr qg wkdw fkdqjhv lq uhodwlyh sulfhv kdyh shupdqhqw hhfwv rq djjuhjdwlrq uhvlgxdov zklfk frqwudglfwv d nh suhglfwlrq ri Klfnv* frpsrvlwh0jrrg wkhruhp1 Rqh idfwru wkdw pd dffrxqw iru wkhvh shupdqhqw hhfwv lv wkdw rxu vhohfwlrq ri uhodwlyh sulfhv lv lqfrpsohwh1 Dqrwkhu idfwru lv wkdw wkh surgxfw pl{ ri lqirupdwlrq0whfkqrorj surgxfwv lv fkdqjlqj ryhu wlph gxh wr lqqrydwlrq zkhuhdv Klfnv* wkhruhp dvvxphv dq xqfkdqjhg surgxfw pl{1 Zlwk wklv fdyhdw lq plqg/ zh lpsohphqw h{0srvw vlpxodwlrqv dqg qg vwdwlvwlfdoo vljql0 fdqw xqghusuhglfwlrqv iru djjuhjdwlrq uhvlgxdov lq LW lqyhvwphqw gxulqj wkh <244 hslvrgh1 Wkhvh xqghusuhglfwlrqv duh whpsrudu / krzhyhu/ dqg wkh suhglfwlrqv uhwxuq wr wkh dfwxdo ;

13 sdwk1 Pruhryhu/ xvlqj dq dxwruhjuhvvlyh prgho dv dq dowhuqdwlyh lhogv odujhu suhglfwlrq huuruv vxjjhvwlqj wkdw ljqrulqj pryhphqw lq uhodwlyh sulfhv uhvxowv lq d orvv ri lqirupdwlrq1 9 Uhihuhqfhv Edfkpdq/ G1/ S1 Mdtxhwwh/ N1 Nduo/ dqg S1 Urffr/ 4<<;/ Wkh ZHID X1V1 Pdfur Prgho zlwk Fkdlq0Zhljkwhg JGS/ Mrxuqdo ri Hfrqrplf dqg Vrfldo Phdvxuhphqw/ 57/ Ghdwrq/ D1 dqg M1 Pxhooedxhu/ 4<;3/ Hfrqrplfv dqg Frqvxphu Ehkdylru/ Fdpeulgjh= Fdp0 eulgjh Xqlyhuvlw Suhvv1 Ilvkhu/ I1 dqg N1 Vkhoo/ 4<:5/ Wkh Hfrqrplf Wkhru ri Sulfh Lqglfhv/ Dfdghplf Suhvv/ Qhz \run1 Khqgu / G1 I1 dqg K1 Nuro}lj/ 5334/ Dxwrpdwlf Hfrqrphwulf Prgho Vhohfwlrq Xvlqj SfJhwv/ Orqgrq= Wlpehuodnh1 Klfnv/ M1U1/ 4<6</ Ydoxh dqg Fdslwdo/ R{irug= R{irug Xqlyhuvlw Suhvv +4<:; sulqwlqj/ vhfrqg hglwlrq,1 Nuro}lj/ K1 dqg G1I1 Khqgu / 5334/ Frpsxwhu Dxwrpdwlrq ri Jhqhudo0wr0Vshflf Prgho Vhohfwlrq Surfhgxuhv/ Mrxuqdo ri Hfrqrplf G qdplfv ) Frqwuro/ 58/ ;640;991 Odqghihog/ V1 dqg U1 Sdunhu/ 4<<8/ Suhylhz ri wkh Frpsuhkhqvlyh Uhylvlrq ri wkh Qdwlrqdo Lqfrph dqg Surgxfw Dffrxqwv EHD*v Qhz Ihdwxuhg Phdvxuhv ri Rxwsxw dqg Sulfhv/ Vxuyh ri Fxuuhqw Exvlqhvv/ Mxo / 6406;1 Wulsohww/ M1/ 4<<5/ Hfrqrplf Wkhru dqg EHD*v Dowhuqdwlyh Txdqwlw dqg Sulfh Lqgh{hv/ Vxuyh ri Fxuuhqw Exvlqhvv/ Dsulo/ 7<0851 Yduyduhv/ F1/ M1 Sudnnhq/ dqg O1 Jxluo/ 4<<;/ Pdfur Prgholqj zlwk Fkdlq0w sh JGS/ Mrxuqdo ri Hfrqrplf dqg Vrfldo Phdvxuhphqw/ 57/ Zkhodq/ N1 5333/ D Jxlgh wr wkh Xvh ri Fkdlq Djjuhjdwhg QLSD Gdwd/ Ilqdqfh dqg Hfrqrplfv Glvfxvvlrq vhulhv > Zdvklqjwrq/ G1F1 = Erdug ri Jryhuqruv ri wkh Ihghudo Uhvhuyh V vwhp1 <

14 0 Billion of 1996 Dollars GDP by Expenditure (a) GDP by Sector (b) GDP by Product (c) Percent of GDP GDP by Expenditure GDP by Sector GDP by Product Expenditures Sector Product Mean Std. Dev ADF* Expenditures Sector Product Mean Std. Dev ADF* Billion of 1996 Dollars 0.0 Percent of Own Aggregate 0-50 Consumption (d) Imports (e) Exports (f) Investment (g) -2.5 Consumption Imports Exports Investment Consumption Investment Exports Imports -5.0 Consumption Investment Exports Imports -100 Mean Std. Dev ADF* Mean Std. Dev ADF* * ADF test uses three lags and a constant; the 5% critical value is Iljxuh 4= Djjuhjdwlrq Uhvlgxdov lq X1V1 Qdwlrqdo Lqfrph dqg Surgxfw Dffrxqwv1 Vrxufh= Vxuyh ri Fxuuhqw Djjuhjdwlrq uhvlgxdo iru JGS e w sh ri H{shqglwxuh +QLSD wdeoh 415,1 K Djjuhjdwlrq uhvlgxdo iru JGS e w sh ri Vhfwru +QLSD wdeoh 41:,1 S Djjuhjdwlrq uhvlgxdo iru JGS e w sh ri Surgxfw +QLSD wdeoh 417,1 _ Djjuhjdwlrq uhvlgxdo iru Frqvxpswlrq +QLSD wdeoh 516,1 e Djjuhjdwlrq uhvlgxdo iru Lpsruwv ri Jrrgv dqg Vhuylfhv +QLSD wdeoh 717,1 s Djjuhjdwlrq uhvlgxdo iru H{sruwv ri Jrrgv dqg Vhuylfhv +QLSD wdeoh 717,1 } Djjuhjdwlrq uhvlgxdo iru Sulydwh Il{hg Lqyhvwphqw +QLSD wdeoh 818,1 43

15 0 Billion of 1996 Dollars Billion of 1996 Dollars Published Residual for Aggregate Investment (a) Computed Residual for Investment in Non-residential Structures (b) Billion of 1996 Dollars 0 Billion of 1996 Dollars Computed Residual for Investment in Equipment and Software (c) -100 Computed Residual for Residential Investment (d) 0 Billion of 1996 Dollars 0.0 Percent Computed Residual for Investment in Information Processing Equipment (e) -5.0 Computed Residual for Investment in Information Processing Equipment and Software Relative to Own Fisher Aggregate Iljxuh 5= Djjuhjdwlrq Uhvlgxdov lq Djjuhjdwh Sulydwh Lqyhvwphqw1 Vrxufh= Vxuyh ri Fxuuhqw Djjuhjdwlrq uhvlgxdo iru Sulydwh Lqyhvwphqw +odvw olqh ri QLSD wdeoh 818,1 K Frpsxwhg e dxwkruv dv wkh glhuhqfh ehwzhhq Sulydwh Lqyhvwphqw lq Vwuxfwxuhv dqg wkh vxp ri h{shqglwxuhv rq Qrq0uhvlghqwldo Exloglqjv/ Xwlolwlhv/ Plqlqj H{sorudwlrq/ Vkdiwv dqg Zhoov/ Rwkhu Qrq0uhvlghqwldo Vwuxfwxuhv1 S Frpsxwhge dxwkruvdvwkhglhuhqfhehwzhhqwkhilvkhudjjuhjdwhirusulydwhlqyhvwphqwri Htxlsphqw dqg Vriwzduh dqg wkh vxp ri Lqirupdwlrq Surfhvvlqj Htxlsphqw dqg Vriwzduh/ Lqgxvwuldo Htxlsphqw/ Wudqvsruwdwlrq Htxlsphqw/ Rwkhu +Htxlsphqw dqg Vriwzduh,1 _ Frpsxwhg e dxwkruv dv wkh glhuhqfh ehwzhhq wkh Ilvkhu djjuhjdwh iru Sulydwh Lqyhvwphqw lq Uhvlghqwldo dqg wkh vxp ri Vlqjoh0idplo Vwuxfwxuhv/ Pxowlidplo Vwuxfwxuhv/ Rwkhu Uhvlghqwldo Vwuxfwxuhv/ dqg Uhvlghqwldo Htxlsphqw1 e Frpsxwhg e dxwkruv dv wkh glhuhqfh ehwzhhq Ilvkhu djjuhjdwh iru Sulydwh Il{hg Lqyhvwphqw lq Lqirupdwlrq Surfhvvlqj Htxlsphqw dqg Vriwzduh dqg wkh vxp ri Lqyhvwphqw lq Frpsxwhuv dqg Shulskhudo Htxlsphqw/ Vriwzduh/ dqg Rwkhu Lqirupdwlrq Surfhvvlqj Htxlsphqw dqg Vriwzduh1 44

16 300 Aggregate Price Index for Investment Price Index for Investment in Information Technology 150 Real Effective Exchange Rates IMF (USD/FX) 1990 = FRB(FX/USD) January 1997 = US Dollars per barrel Unit value of Oil Imports Percent fed. funds rate 30-year 10-year CPI Inflation Iljxuh 6= Hyroxwlrq ri H{sodqdwru Yduldeohv1 Vrxufh= Vxuyh ri Fxuuhqw Exvlqhvv/ Exuhdx ri Oderu Vwdwlvwlfv/ Lqwhqudwlrqdo Prqhwdu Ixqg/ dqg Ihghudo Uhvhuyh Erdug1 45

17 0 Percent of IT Investment -5 Actual Fitted for Model M3 Estimated Parameters (standard errors) R/Y = *R/Y(t-1) *w *w.sqrd (0.0825) (11.8) (4.87) (0.566) Effect of "9/11" Dummy *p.sqrd *q *q.sqrd *r (0.059) (13.5) (19.9) (0.306) *r(t-1) *r.sqrd *Trend *d911 (0.0609) (0.0417) (0.0525) (0.49) SER: R.SQRD: ( ) mean(r/y): var(r/y): Contribution of Relative Price of IT Investment (1987=100) 110 Estimated Effect of w= (25.75*w *w.sqrd)*100/value in 1987Q1 Estimated Standard Deviation: 4: Iljxuh 7= Hpslulfdo Uhvxowv iru Prgho P60Ilwwhg Ydoxhv dqg Hhfw ri Uhodwlyh Sulfh ri LW Lqyhvwphqw 46

18 Percent of Aggregate Percent of Aggregate 0 1-step Forecasts Actual 0 Multi-step Forecasts Actual Model M3 RMSE 1.1% -10 Model M3 RMSE 1.6% step Forecasts Actual 0 Multi-step Forecasts Actual Model AR(5) RMSE 1.3% -10 Model AR(5) RMSE 3.0% Iljxuh 8= <8( Frqghqfh Lqwhuydov iru H{0srvw Iruhfdvwv ri Uhvlgxdov lq LW Lqyhvwphqw 47

19 Table 1: Parsimonious Model of Aggregation Residuals Private Investment in Information Equipment a Ordinary Least Square Estimation Results :Q1-2002:Q3 Sensitivity to Model Specification b M1 M2 M3 M4 M5 M6 Significant short-run price effects Linear w, p, q, r w, p, q w, q, r w, q w, q w, q Nonlinear w w, p, q w, p, q, r w, p, q w, p, q w, p, q Significant long-run price effects Linear w, r w, p, q w, q, r w, q w w, q Nonlinear w w, p, q w, p, q w, q w, q w, q Is Trend Significant? no yes no yes yes yes Adj. R SER (percent) Number of Parameters /DJJHG'HS9DULDEOH (std. error) (0.09) (0.10) (0.08) (0.09) (0.09) (0.09) Hypothesis Testing c Properties of Disturbances Normality Serial Correlation Homoskedasticity Parameter Stability Exclude Half Sample Exclude Last Eight Obs * * Forecasts Error (RMSE) d Multi-step Ahead One-step Ahead a Aggregation residual for Private Investment in Information Equipment (NIPA table 5.5): difference between the BEA aggregate for nonresidential, private investment in information equipment and the sum of investment in computers and peripheral equipment, software, and other information processing equipment and software. q: Real effective exchange rate, r: Real interest rate, p: Price of oil relative to aggregate s own price w: Price of computer equipment purchases relative to aggregate s own price b Models differ in their choice of exchange-rate and interest-rate measures: Real Interest Rate Real Effective Federal Funds 10-year 30-year Exchange Rate IMF-Narrow M1 M2 M3 FRB-Broad M4 M5 M6 c Significance level needed to reject the associated hypothesis; an entry less than 0.05 means that one can reject the null hypothesis at the five percent significance level-denoted with an asterisk. d Root mean squared error over 2000:Q4 to 2002:Q3. 15