WORKING PAPERS IN ECONOMICS AND ECONOMETRICS Servce Olgopoles and Australa s Economy-Wde Performance* Rod Tyers College of Busness and Economcs Australan Natonal Unversty Lucy Rees** Ban and Company rst revson, May 2008 Key words: Regulaton, olgopoly, servces, prce caps, prvatsaton JEL codes: C68, D43, D58, L3, L43, L5, L80 Workng Papers n Economcs and Econometrcs No.490 College of Busness and Economcs Australan Natonal Unversty * undng for the research descrbed n ths paper s from Australan Research Councl Dscovery Grant No. DP0557885. Thanks are due to lavo Menezes for valuable nput at the formaton stage of ths project, Chrs Jones for useful dscussons, Marcn Pracz for assstance wth ths research durng 2006 and to Pngkun Hsu and Ian Ban for research assstance snce then. lavo Menezes also offered valuable comments on an earler draft as dd partcpants at a March 2008 semnar n the ANU s College of Busness and Economcs, ncludng Phllpa Dee, Ben Smth, Chrs Jones, Martn Rchardson, Wllam Coleman and Rchard Cornes. ** Lucy Rees contrbuted to the early development of ths research durng her studes at the ANU. The results presented and opnons offered n no way represent the postons of Ban and Company on regulaton polcy.
Servce Olgopoles and Australa s Economy-Wde Performance Abstract The retreat from publc ownershp of servce frms and ndustres has left behnd numerous prvate monopoles and olgopoles supervsed by regulatory agences. Servces ndustres n government and prvate ownershp generate two-thrds of Australa s value added, whle the newly prvatsed ones, utltes, telecommuncatons, fnance and transport, supply a ffth. Ths study offers an economy-wde approach that represents monopoly and olgopoly behavour explctly. It examnes the mplcatons of olgopoly rents for factor markets and the real exchange rate, the extent of sectoral nteractons and the potental economy wde gans from tghter prce cap regulaton, wth the results confrmng the mert of an economywde approach. External shocks, lke the present Chna boom, are also smulated. Such postve shocks are shown to expand the potental for olgopoly rents and therefore to rase the bar for regulatory agences. Moreover, less than tght prce caps are shown to exacerbate entry-ext hysteress n boom and bust cycles.. Introducton ollowng mcroeconomc reforms of the 980s and 990s, excludng government agences, corporatons subject to regulaton now provde almost half of Australan s employment and of ts GDP. The analyss of regulatory regmes must, therefore, account for economy-wde mplcatons, such as effects on the real exchange rate and factor markets as well as on other ndustres through the cost of ntermedate servces. Ths s best acheved by modellng the whole Australan economy n a way that allows explctly for the monopoly and olgopoly behavour requrng regulaton n the frst place. Such a model s offered n ths paper. It s desgned to help clarfy the mplcatons of changes n the regulaton of olgopoly prcng for the structure and performance of Australa s servce ndustres whle at the same tme examnng nter-sectoral effects and assocated changes n the performance of labour markets and the economy as a whole. The retreat from the government s drect provson of nfrastructure servces has left prvate frms and publcly-owned corporatsed enttes (government bodes subject to corporatons law) n ndustres that are lttered wth olgopoly structures and component monopoles. These frms and enttes are therefore supervsed by regulatory agences. They nclude: transport, electrcty, water supply, gas dstrbuton, telecommuncatons, fnance and nsurance, educaton and health. 2 Whle regulatory polces cover both product prcng and See ABS 2003: Tables 2.2 and 2.3, wth publc servce output and employment subtracted. 2 The health sector s not addressed ndependently n ths study, prmarly because ts actvty s dffcult to dstngush n the avalable economy-wde database we use but also because t s rfe wth nformaton asymmetres that make ts regulaton more complex than the sectors consdered. 2
qualty, we focus entrely on prce regulaton, ncludng prce survellance as well as prce caps, and hence the control of economc costs assocated wth dstortons due to mperfect competton. Some exstng studes suggest an ndrect lnk between the prvatsaton and regulatory reforms and productvty 3, whle other studes follow a long tradton n regulatory economcs of applyng ndustry-specfc (partal equlbrum) comparatve statcs. 4 In recent years economy-wde mplcatons have been examned usng models underpnned by perfectly compettve behavour, wth olgopoly rents mpled by the choce of parameters, closure or productvty shocks. 5 Industry-specfc fx-ups n such models stll requre the assumpton of perfect competton n all other ndustres to generate the economy-wde effects. Whle ths approach has been very useful durng the mcroeconomc reform transton, t tends to gnore the fact that most other economc actvty s also mperfectly compettve and subject to regulaton, and that regulatory changes to one ndustry are unlkely to occur wthout mplcatons for the regulaton of others or for the performance of the whole economy. 6 Central to ths paper s the mathematcal model of the Australan economy t ntroduces and the economy wde database that serves t. Ths model represents monopoly and olgopoly prcng behavour and the regulatory envronments facng major frms. Its behavoural structure s based on early work by Harrs (984), Horrdge (987), Gunasekera and Tyers (990) and Tyers (2005), whch emphassed homogeneous product olgopoles wth frms nteractng on producton n an almost small home economy. Subsequent extensons ncluded dfferentated product olgopoles nteractng on prces. The focus n these pror applcatons has been trade reform and manufacturng olgopoly, wheren consderable attenton has been gven to the pro-compettve effects of trade lberalsaton (Hertel 994; Ianchovchna et al. 2000; Tyers 2005). In ts new guse the model s structured to focus on the more commonly regulated servces ndustres. To ths end, a more complete representaton of the Australan tax system s ncorporated. Ths s needed for two reasons. rst, the response of frms to regulaton 3 See, for example, Productvty Commsson (999), Madden et al. (2002). 4 Classc partal equlbrum studes nclude those by Averch and Johnson (962), Courvlle (974) and Wellsz (963). 5 These nclude the global model by Dee (2003), whch s nnovatve n that t recognses the mportance of domestc locaton for overseas-based servce frms. requently appled s hghly detaled MONASH model of the Australan economy (Dxon and Rmmer 2002). Ours s not an attempt to compete wth ether of these approaches. Rather, we seek to construct a more drect means to evaluate regulatory polces by embeddng more realstc monopoly and olgopoly behavour n the economy-wde context. 6 Ths s notwthstandng the fact that regulaton can be appled dfferently across dfferent jursdctons n Australa and across dfferent ndustres. See, for example, NECG (2003). 3
depends on the rates of tax to whch they are subject, and second, taxes and transfers offer an alternatve approach to the achevement of regulatory objectves. urther, the model ncludes generc foregn ownershp, thereby allowng for proper representaton of net factor ncome flows and ther effects on the real exchange rate, the composton of the current account and GNP. Usng ths machnery, ths paper begns wth an assessment of the scale of costs the Australan economy would bear were ts olgopolstc servce ndustres to be allowed to cartelse. Ths hypothetcal experment s used to llustrate the extent of economy wde nteractons assocated wth olgopoly behavour. It then assesses the potental for further gans from tghter cappng of prces, once agan gvng emphass to economy wde nteractons. nally, t offers a stylsed representaton of the Australan economy s response to the recent Chna drven global commodty boom, whch drectly affects ts agrcultural and mnng sectors but whch has rased the relatve prces of Australan servces and so t has mportant mplcatons for olgopoly rents n servces and ther regulaton. Indeed, the results show that the mantenance of tght prce caps through the boom would measurably ncrease the beneft from t n the short run and avod excessve entry n the long run should the effects of the boom be transtory. The secton to follow brefly revews developments n the structure and regulaton of Australa s servces sector. Secton 3 revews the behavoural structure of the model used, the detaled descrpton of whch s consgned to appendces, and the economc structure emboded n ts database. The potental economc losses from olgopoly behavour are assessed n Secton 4 whle the gans to be derved from tghter prce cappng are consdered n Secton 5. The response of olgopolstc servces to external shocks s examned n Secton 6. Conclusons are offered n Secton 7. 2. Regulaton Research and Australa s Servces Australa s servce ndustres have experenced rapd growth over the last ffty years. Ther regulaton s seen as redress for market falures that nclude lack of nformaton, monopoly power, externaltes or socal objectves (ncome dstrbuton or servce qualty). 7 Many servce ndustres requre networks wth hubs that consttute substantal recurrent fxed 7 See ndlay (2000: 0). 4
profts. The reach of regulatory polces n Australa has rsen snce these Acts were wrtten, cost and the control of whch creates barrers to entry, conferrng monopoly power. 8 The economc ratonale for servce regulaton s therefore strong. Stgler (97) argued that the analyss of regulaton should concentrate on three mportant questons 9 : who wll receve the benefts and who wll bear the burden of regulaton, the form and nature of the regulatory nterventon and the effect of regulaton on resource allocaton. There are very substantal subsequent lteratures on the generc benefts from regulaton and, more partcularly, the specfc regulatons mposed on Australa s servces. 0 Devces range between prce controls, ownershp restrctons, lmts on foregn drect nvestment and capacty constrants. ederal responsblty for competton regulaton and montorng rests wth the Australan Competton and Consumer Commsson (ACCC), whch admnsters the Trade Practces Act 974 ( TPA ) and the Prces Survellance Act 983 ( PSA ). The TPA promotes competton and far tradng whle also provdng consumer protecton. The PSA crcumscrbes the ACCC s montorng of prces, costs and due to the extensve prvatsatons assocated wth the mcroeconomc reform era and the pace of technologcal change. The latter, partcularly n telecommuncatons, electrcty and gas, has made t possble to unbundle ndustry segments, leavng some as natural monopoles or olgopoles but wth others organsed around supervsed new markets that foster compettve behavour (such electrcty producton and the retalng of gas, electrcty and telephone servces). The ntroducton of prce caps as ncentve regulaton was amed at the monopoly and olgopoly elements, where compettve prcng could not be otherwse nduced. These consequences of prvatsaton dd, however, dstort behavour as nvestment sought to escape the prce caps. 2 In telecommuncatons, ar transport and the producton and dstrbuton of natural gas and electrcty these changes have been stark. 3 Inevtable dstortons notwthstandng, reforms have been shown to contrbute to mprovements n both economc and government performance, accordng to the OECD (997) to the tune of fve percent of GDP. 8 Ibd. In the Australan context a recent example of ant-compettve behavour attrbuted to monopoly power over a network s Telstra s prcng for ts broadband servce. See ACCC (2003). 9 See Stgler (97). 0 See Trade Practces Act 974 (Cth), revewed n Productvty Commsson (2003). See ACCC (2003). 2 We are grateful to lavo Menezes for ths pont. or related dscusson, see Menezes et al. (2006). 3 See Doove et al. (200: 43). 5
In the late 970s there was sgnfcant ant-regulatory sentment n developed countres. The practce of rate-of-return regulaton was found to be ncompatble wth ncreased competton. Lttlechld (983) changed ths negatve percepton of regulaton wth hs report on the Brtsh telecommuncatons ndustry n whch he suggested prce caps as a regulatory polcy tool. Ths sgnalled a movement towards a more ncentve-based and less heavy handed approach to regulaton. The result has been very wdespread applcaton of prce-caps n servces, whch are charactersed by: product-specfc prce celngs, basket celngs that offer frms greater flexblty, and perodc adjustments of celngs to ensure that consumers share n the gans from techncal change and market formaton. 4 Theoretcal studes have been hghly stylsed and sector-specfc but they demonstrate that prce-caps, even as second best measures, can protect consumers aganst monopoly power, promote competton, mprove productve effcency and nnovaton and reduce the admnstratve burden of regulaton. 5 Emprcal follow-up by Xaver (995) assesses prcecap schemes n the UK, the USA and Australa. Hs Australan focus s on the (then) Telecom basket prce cap between 989 and 992. He fnds that the scheme reduced the average prce of Telecom s domestc servces n real terms by 3 percent. Internatonal call prces fell n real terms by 25 percent n ths perod, however, suggestng that the scheme fell short of delverng a far share of technologcal gans to the Australan consumer. He takes a sceptcal vew of some prce-cap mechansms, preferrng the fosterng of compettve forces where ths s possble. Turnng to economy-wde approaches wth explct representaton of mperfect competton, Blanchard and Gavazz (2003) offer an elemental general equlbrum model to nvestgate the combned effects of product market and labour market regulaton. Ther closed economy model ncorporates monopolstc competton n the goods market and barganng n the sngle factor (labour) market. In seekng competton, a government mght try to rase the elastcty of substtuton. In the short-run they fnd that the ncreased competton s benefcal because t forces frms to lower ther mark-up, leadng n turn to reduced captal returns but a hgher real wage. In the long run, however, there s ext by frms and reduced product varety. The assumpton of monopolstc competton leaves no pure profts to erode and nvarant recurrent fxed costs must see the mark-up return to ts orgnal value, so there are no long run benefts. If, nstead, the government attacks barrers to entry (recurrent fxed costs), the effects are unambguously welfare mprovng n the short and long 4 See Vogelsang and Acton (989). 5 The key works n ths area are: Cabral and Rordan (989), Bradley and Prce (988), and Brennan (989). 6
runs. There s an ncrease n the number of frms, a hgher elastcty of demand, a lower mark-up and thus lower unemployment and a hgher real wage. Whle t s not made clear how a government mght alter the elastcty of substtuton or entry costs, ths research sgnals an mprovement over pror studes of regulaton through ts charactersaton of market structure n an economy-wde context and t s n ths sprt that the research presented n ths paper has been undertaken. The precse extent of mperfect competton n Australa s servce ndustres s dffcult to quantfy. We offer short qualtatve summares for the key sectors n whch prvatsaton and regulaton have brought most change. Telecommuncatons The ACCC s analyss ndcates that ths sector s slowly becomng more compettve, wth most mprovement at the retal level, as opposed to nfrastructure provson. In lne wth global changes n telecommuncatons technology, there have also been consderable mprovements n productvty, as ndcated by Madden et al. (2002). 6 Telstra contnues to be the domnant frm wth about two thrds of the sector s lsted market captalsaton and between a thrd and three quarters of the markets for the dfferent telecommuncatons products (Telstra 2003). Whle these facts suggest a hgh level of concentraton, n areas such as moble and long dstance telephony and data transfer, competton s ntense. Telstra s explotable market power s n fxed telephony and network access. 7 Electrcty New market mechansms as well as regulatory reform have been ntroduced n ths sector. It has nonetheless been found that generators, whose numbers reman small, have often been able to ncrease prces substantally above compettve levels for sustaned perods. Indeed, Short et al. (200) ndcate that the electrcty market s subject to sgnfcant departures from compettve outcomes. Hgh Lerner ndces (prce-margnal cost margns) suggest the collecton of substantal rents. Yet, gven ths ndustry s hgh fxed costs, a better measure mght have been the mark-up over average cost. Gas 6 The controversal lnk between the market power of frms and nvestments n R&D, along wth assocated productvty performance, s not the focus of ths paper. Our approach offers a capacty to nvestgate ths ssue, however, and so t wll be the subject of further research. 7 or ths pont we are grateful to lavo Menezes. 7
Whle offcal barrers to the free flow of natural gas across state borders have been removed, the market remans hghly concentrated on the supply sde and t carres many legacy agreements that lmt competton. The resultng lack of lqudty n Australan gas markets has mpeded the development of transparent spot markets. 8 There are three supplers n the eastern Australan gas markets that account for more than 95 percent of the supply gas. The two ncumbents BHP Bllton and ExxonMobl account for 38 percent and 4 percent respectvely. As to nfrastructure, the largest ppelne owner n Australa s Australan Ppelne Trust (APT) whch owns a thrd of the total transmsson ppelne system. 9 Australa s second largest ppelne owner s Epc Energy, wth about half the capacty of APT. In 997 the Australan Government ntroduced a Gas Code The Natonal Thrd Party Access Code for Natural Gas Ppelnes whch s admnstered by the ACCC and the Natonal Competton Councl (NCC). 20 The Code ensures that gas can be transmtted through the ppelne network on reasonable terms and condtons, though n practce these have attracted controversy. Ar Transport The Australan arlne domestc market has long had a duopoly structure, changes of players n the 990s notwthstandng. Because of volatlty assocated wth these changes, the market share of the only remanng ncumbent, Qantas, has been measured at and above 70 percent. 2 Nonetheless t remans n the nterest of both the major carrers to mantan an ndustry structure whch allows both to generate sustanable proftablty wthout encouragng further entry. Agan, the ACCC montors prces and frequent flyer schemes for antcompettve elements. As to avaton nfrastructure, pror to 2002, arports n Australa were subject to prce-cap regulaton. However, the Productvty Commsson concluded that whle the major metropoltan arports have substantal market power, t s not n ther nterests to abuse ths power n such a way that would confer large costs onto the economy. 22 Hence, the government has largely deregulated arports, replacng prce caps wth prce montorng. 23 Debate contnues, however, as new ar servce entrants seek access to arport servces. 8 See Short, C. et al. (2003). 9 See Australa Daly (2004). 20 See Moran (2002). 2 See reed (2004). 22 See Productvty Commsson (2002). 23 Rather than collude to rase carrer costs, owners of prvatsed arports have sought and found proftablty through the development of arport property by explotng relatvely relaxed federal regulatons governng the use of arport land. 8
Ths very bref revew makes t clear that the regulaton of olgopoly servce ndustres n Australa s made more complex by the trend toward the subdvson of each ndustry nto more and less compettve components. In ths paper, however, our purpose s to take a broad brush to the estmaton of economy wde effects of servce olgopoly behavour. We therefore work at the level of 0 sectors, necessarly averagng out sectoral detal. In nterpretng the research that follows t should be borne n mnd that the task of sectoral regulators s not only made dffcult by the non-transparency of the costs we model but also because product lnes and the degree of dfferentaton between frms are not stable through tme n the way we model them. 3. An Olgopoly Model of the Australan Economy The model s a development of that used to examne pro-compettveness effects of trade lberalsaton n manufacturng by Tyers (2004, 2005). 24 The verson descrbed here dffers n fve key respects. rst, behavoural equatons have been added to represent the effects of regulatory polcy, ncludng Ramsey prce-caps; second, a government sector has been ncluded to dstngush the government s expendtures from those of the collectve household. Prevously, net revenue from border taxes was transferred to the collectve household n lump-sum. Lke the model s sngle prvate household, the government now has Cobb-Douglas preferences over goods and a constant elastcty of substtuton (CES) subaggregaton of home goods wth mports. It s not, however, treated as an optmsng agent maxmsng a utlty functon lke the representatve household. 25 A balanced budget s assumed so that government spendng changes n lne wth tax revenue. Thrd, the border tax system n the earler verson of the model has been extended to nclude a more detaled representaton of Australa s tax system wth both drect (ncome) taxes leved separately on labour and captal ncome and ndrect taxes ncludng those on consumpton, mports and exports. 26 ourth, a new database s constructed that emphasses Australa s servce ndustres. Ths database ncorporates government consumpton and the complete tax system. 27 fth and fnally, stablty problems encountered n the tradable goods sectors due to a unform structure suted to the more closed servces sectors necesstated a 24 It s a dstant descendent of the models of Harrs (984) and Gunasekera and Tyers (990). 25 One approach, not adopted n ths paper, would be to ncorporate the government sector so that ts spendng and taxaton decsons are endogenous and determned by the maxmsaton of an assumed socal welfare functon. The dffculty wth such an approach s that t does not capture the poltcal economy of government polcy. Thanks are due to Chrs Jones for useful dscussons on ths topc. 26 Income taxes are approxmated by flat rates deduced as the quotent of revenue and the tax base n each case. 27 These frst four advances are due to Rees (2004). 9
restructurng of the model s treatment of foregn goods. These are now consdered homogeneous n each broad sector but dfferentated from correspondng home products, whch themselves are dfferentated by varety. 28 Model structure The scope of the model s defned n Table. It dvdes the Australan economy nto ten sectors, of whch seven offer servces, and four prmary factors. The ten sectors are chosen to reflect the domnance of servces n domestc demand and the mportance of regulaton to mperfectly compettve servces sectors. rms n all ten sectors are olgopolstc n ther product prcng behavour wth the degree of prce-settng colluson between frms represented by conjectural varatons parameters. The magntudes of these parameters are consdered to represent the flexblty allowed the frms prcng behavour by the survellance of regulatory agences. Each frm bears fxed captal and sklled labour costs, enablng the representaton of unrealsed economes of scale. Home products n each sector are dfferentated by varety 29 and output s Cobb-Douglas n varable factors and ntermedate nputs. Intermedate nput demands are CES subaggregates of home and mported products. Despte ther olgopoly power n product markets, frms have no olgopsony power n the markets for prmary factors or ntermedate nputs. The sophstcaton wth whch home product markets are represented notwthstandng, the modellng of a sngle economy necesstates crudeness n the representaton of foregn frms. Thus, mports are seen as homogeneous, dfferentated from home products as a group, so that mport varetal dversty never changes. 30 In the long run, physcal captal s homogeneous and fully moble between sectors and nternatonally, whle the domestc endowments of other factors are fxed. A short run closure s also constructed, wheren captal use n each ndustry s fxed and rates of return on captal can vary between sectors. In ths verson, the closure can be adjusted correspondngly, so that the real wage of unsklled labour s fxed and unsklled employment s endogenous. The quantty of domestcally-owned captal s fxed both n the short and long runs, so that 28 The earler verson requred very large elastctes of substtuton between varetes, leadng to unrealstc behavour n response to changes n border tarffs. Moreover, n a sngle economy model there s no satsfactory way to endogense the number of foregn product varetes. 29 Product dfferentaton s assumed to be of the Spence-Dxt-Stgltz type. Ths means that each ndvdual derves utlty from consumng a number of varetes of a gven product. 30 Snce all home varetes are exported there s no movement on the extensve margn of the type that s evdent n the models of non-homogeneous export sectors by Meltz(2003) and Balstrer et al. (2007). We see ths as acceptable n ths study because our focus s on largely non-traded servces. 0
changes n the total captal stock affect the foregn ownershp share and hence the level of ncome repatrated abroad. The economc profts or losses earned by frms are dependent on the closure, under whch ether the number of domestc frms (varetes) can be fxed whle profts are endogenous, or flexble whle economc profts are fxed. The economy modelled s almost small, mplyng that t has no power to nfluence border prces of ts mports but ts exports are dfferentated from competng products abroad and hence face fnte-elastc demand. 3 The effectve numerare s the mport product bundle, snce mport prces are exogenous n all experments. Its current account defct (of about a tenth of ts ntal GDP) s fxed n terms of foregn prces n all experments. 32 or presentatonal purposes, prces and values can be dvded through by the consumer prce or the GDP prce ndex so that the ntal consumpton or producton bundle becomes the numerare. The consumer prce ndex s constructed as a composte Cobb-Douglas-CES ndex of home product and after-tarff mport prces, derved from the sngle household s expendture functon and measured after consumpton taxes are appled. Ths formulaton of the CPI ads n the analyss of welfare mpacts. Because collectve utlty s also defned as a Cobb-Douglas combnaton of the volumes of consumpton by product aggregate, proportonal changes n overall economc welfare correspond wth those n real GNP. 33 rms n any sector supply dfferentated products and nteract on prce. Cobb- Douglas producton drves varable costs so that average varable costs are constant f factor and ntermedate product prces do not change. Consequently, whle ever factor and ntermedate product prces are constant, average total cost declnes wth output. The magntudes of recurrent fxed costs are calbrated from data on ndustry proftablty, gross 3 Ths follows the practce n natonal modellng snce the frst sgnfcant economy-wde model by Dxon et al. (982) and the frst publshed economy-wde olgopoly model by Harrs (984). 32 Ths mples that the acquston of Australan assets by foregners remans constant rrespectve of shocks mposed on the model and ther consequences. 33 When the utlty functon s Cobb-Douglas n consumpton volumes, the expendture functon s Cobb-Douglas C n prces. If the consumer prce level, P, s defned as a Cobb-Douglas ndex of prces, the equvalent varaton n ncome can be expressed n terms of the proportonal change n ths ndex. Thus, followng any shock, the ncome equvalent of the resultng changes to ncome and prces s: C C C ΔP Δ W = Y Y0 + EV ( P0, P, Y) = Y Y0 Y, P C whch can be expressed n proportonal change form as: C ΔP Y C Y0 C ΔW P ΔY ΔP =. C W Y0 Y0 P Ths s, approxmately, the proportonal change n real GNP.
value of output and value added. 34 rms charge a mark-up over average varable cost so that t s at least possble for them to cover ther average fxed cost n a zero-pure-proft monopolstc competton equlbrum. They choose ths mark-up strategcally, however, so that ther capacty to push ther prce beyond ther average varable costs wthout beng undercut by exstng compettors then determnes the level of any pure profts and the potental for entry by new frms. As ntuton would suggest, under free entry pure profts are eroded and the mark-up just covers average total costs. Each frm n ndustry s regarded as producng a unque varety of ts product and t faces a downward-slopng demand curve wth elastcty ε (< 0). The optmal mark-up s then: () where p m = =, v + ε p s the frm s product prce, ν s ts average varable cost and ε s the elastcty of demand t faces. rms choose ther optmal prce by takng account of the prce-settng behavour of other frms. A conjectural varatons parameter n ndustry s then defned as the nfluence of any ndvdual frm k, on the prce of frm j: (2) μ p j =. pk These parameters are consdered to ndcate the power of prce survellance by such nsttutons as the ACCC. The Nash equlbrum case s a non-collusve dfferentated Bertrand olgopoly n whch each frm chooses ts prce, takng the prces of all other frms as gven. In ths case the conjectural varatons parameter (2) s zero. When frms behave as a perfect cartel, t has the value unty. Ths parameter enters the analyss through the varetal demand elastcty, whch s formulated n Appendx 2. To study the effects of prce-caps a regulated Ramsey mark-up, R m s formulated as: (3) m afc + ν =. R ν rms are permtted to choose compromse mark-ups by alterng the parameter ϕ n the followng: ( ) ( ) C R (4) m = ϕ m + 2 ϕ m. 34 In the startng equlbrum t s assumed that each ndustry has pure profts equal to one percent of gross earnngs. Consstent secondary data are not readly avalable to determne the share of pure profts n captal returns n all ndustres. 2
C C R Thus, when ϕ =, m = m, and when ϕ = 2, m = m. Crtcal to the mplcatons of mperfect competton n the model s that the product of each ndustry has exposure to four dfferent markets. It can be consumed by prvate households or by government, used as an ntermedate nput n another ndustry or t can be exported. The elastcty of demand faced by frms n ndustry, ε, s therefore dependent on the elastctes of demand n each of these four markets, as well as the shares of the home product n each. More precsely, the four sources of demand for home produced products are fnal demand (), ntermedate demand (I), export demand (X) and government demand (G). or sector, the elastcty sought s a composte of the elastctes of all four sources of demand. (5) ε I I X X G G = s ε + s ε + s ε + s ε, where j s denotes the volume share of the home product n market for each source of demand j. These share parameters are fully endogenous n the model. Because the dfferent sources of demand are dfferently elastc, wth export demand most elastc and ntermedate demand least, any shock that reapportons demand between them necessarly changes the compettve behavour of the frms. 35 Almost all concevable shocks do ths to some degree. Thus, the strategc behavour of frms, and hence the economc cost of servce olgopoles, s affected by conjectural varatons parameters as they represent collusve capacty on the one hand and regulatory prce survellance on the other, and by the composton of demand as t nfluences the elastctes of demand faced by each frm. Of course, the capacty frms have to reduce ther prces also depends on ther productvty performance, whch we do not examne n ths paper, and on ther numbers, hence the sectoral fxed cost burden. The database and ts representaton of broad economc structure The model database s constructed from the GTAP Verson 5 global database for 997 (Dmaranan and McDougall 2002). 36 It combnes detaled blateral trade, transport and protecton data characterzng economc lnkages among regons, together wth ndvdual country natonal accounts, government accounts, balance of payments data and nput-output 35 Export demand s found to be more elastc because of the larger number of substtutable product varetes avalable abroad whle ntermedate demand s relatvely nelastc because of frms reluctance to alter arrangements for ntermedate nput supply whch may depend on locaton or just n tme relatonshps. These ssues are addressed by Harrs and Cox (983). 36 Documentaton on the GTAP 5 Data Package may be vewed at: <http://www.gtap.agecon.purdue.edu/databases/v5/v5_doco.asp>. 3
tables whch enable the quantfcaton of nter-sectoral flows wthn regons. rom the database key elements of our representaton of the Australan economy emerge. rom Table 2, t s evdent that the prvatsed servces, electrcty, water, gas, telecommuncatons, fnance and transport, supply about a ffth of the economy s GDP, yet ther partcpaton n nternatonal trade s tny compared wth agrculture, manufacturng and mnng. Moreover, the prvatsed servces are shown n Table 3 to be more ntensve n skll and physcal captal than are the tradable sectors so that ther comparatve performance has partcular mplcatons for the sklled wage premum and total captal use. The flows represented n the database do not reveal detals of ndustral structure. In partcular, addtonal nformaton s requred on frm numbers, pure profts, fxed costs and mnmum effcent scale for each sector. Whle some detals are avalable on these varables for some ndustres, there s no readly avalable source that s consstent and comparable across sectors. Wth the support of the few ndustry studes already mentoned and the Mornngstar nancal Analyss Database, 37 these varables are calbrated n the followng manner. rst, pure profts are requred as a share of total captal ncome (operatng surplus) n each ndustry. Ths s needed to fnalse the flow database but also to calbrate ndustry compettve structure. or ths we have resorted to data on the proftablty of lsted frms from the Mornngstar Database. 38 Addtonal detal as to our approach s offered n Appendx 3. Second, rough estmates of strategcally nteractng frm numbers n each ndustry and ther correspondng conjectural varatons parameters are requred. It s not suffcent smply to record the number of establshments n each ndustry, however. Unless ndustres are subdvded fnely, consderable dversty of frm sze and product s emboded n each. Indeed, wthn a partcular ndustry classfcaton, many frms supply ntermedate nputs to other frms n the same classfcaton. Prces of the products that emerge from a partcular ndustry are very lkely determned by a small proporton of the frms wthn t. Agan, we resort to the Mornngstar database for measures of ndustry concentraton. rom ths we assgn the crude ndex of frm numbers ndcated n Table 4 and we also post the correspondng conjectural varatons parameters shown n the same table. Agan, addtonal detal as to our approach s provded n Appendx 3. 37 The database s formally the Aspect nancal Analyss Database. It s suppled by Aspect-Huntley, and the copyrght s held by Huntleys' Investment Informaton Pty Ltd (HII) (a wholly owned subsdary of Mornngstar, Inc). 38 After tax profts rates are compared wth the prme borrowng rate n the perod 997-2007 to obtan measures of pure profts. rm statstcs were drawn from http://www.aspectfnancal.com.au/af/fnhome?xtmlcensee=fnanalyss.for and the data on ndustral borrowng rates was from www.rba.gov.au. 4
Thrd, to complete the formulaton of ndustry demand elastctes, elastctes of substtuton between home product varetes and between generc home and foregn products are requred for each sector. These are drawn from the estmaton lterature. 39 Intal ndustry demand elastctes are then calculated for each source of demand (fnal, ntermedate, government and export), va the equatons n the appendces, and the results are also lsted n Table 4. Intal shares of the demand facng each ndustry are then drawn from the database to enable the calculaton of weghted average demand elastctes for each ndustry. Mark-up ratos are then deduced from these, fxng average varable cost n each sector, va equaton (). The ntal equlbrum ndustry shares, average elastctes and mark-up ratos for each sector are gven n Table 5. Note that the elastctes appear large n magntude at frst glance. Ths s because they do not represent the slopes of ndustry demand curves for generc goods. Rather, they are the elastctes faced by supplers of ndvdual varetes and are made larger by nter-varetal substtuton. Ths completes the demand sde calbraton. It enables us to turn to the calbraton of the supply sde, where we begn by usng the mark-up ratos to deduce the ntal level of average varable cost n each sector. Next, we turn to pure profts. The proporton these make up of total turnover s deducted from the mark-up to arrve at fxed cost shares of total turnover. 40 Total recurrent fxed cost n each sector then follows. The results of ths calbraton are summarsed n the frst three columns of Table 6. It s now possble to obtan a sense of the scale of producton. 4 Under our assumpton of Cobb-Douglas technology n varable factor use, combned wth recurrent fxed costs, f ndustres could expand ndefntely wthout changng unt factor rewards (the partal equlbrum assumpton that s relaxed here), average fxed cost would approach average varable cost asymptotcally from above. ollowng Harrs and Cox (983) we choose an arbtrary mnmum effcent scale (MES) product volume at the pont where average fxed cost would declne to a twenteth of average varable cost. The mpled scale parameters are dsplayed n the fnal column of Table 6. They confrm expectatons that fxed costs are most promnent n electrcty, gas, water, telecommuncatons and transport servces, due to fxed physcal nfrastructure and 39 Summares of ths lterature are offered by Dmaranan and McDougall (2002) and at http://www.gtap. purdue.edu/databases/.. 40 xed costs take the form of both physcal and human captal costs usng the rule of thumb (based on estmates by Harrs and Cox, 983) that physcal captal has a fxed cost share of 5/6. 4 The actual calbraton process s more complex than ths because the elastctes of ntermedate demand depend on ntermedate cost shares, whch depend on the varable cost share. It s therefore necessary to calbrate teratvely for consstency of elastctes and shares. 5
network mantenance costs. The results also suggest, plausbly, that the sectors closest to ther mnmum effcent scale are agrculture, mnng, fnance and other servces. 4. Sectoral Interactons wth Olgopoly To explore the nterdependence of the prvatsed servce sectors and the potental mpacts of ther non-compettve behavour on the economy as a whole we begn by consderng the effects of complete explotaton of market power n all sectors. In partcular, on the presumpton that olgopoly frms fal to collude and form cartels (or consoldate nto monopoles) manly because of government prce survellance and (the threat of) ant-trust actons, 42 we magne what the Australan economy would have looked lke had these government actvtes never occurred. A long run closure s selected n whch physcal captal s nternatonally and ntersectorally moble and labour markets clear at flexble wages. The entre economy s frst allowed to cartelse, by rasng all conjectural varatons parameters to unty. 43 Then, ndvdual sectors are cartelsed one by one n a bd to dentfy non-lneartes that mght mply the necessty of economy-wde analyss. The results of ths exercse are summarsed n Table 7. 44 Clearly the economy would have been substantally smaller f all sectors had been cartelsed. Real GDP would have been smaller by a thrd and real wages smaller by more than half. In general, cartel rents mply hgher home product prces and hence an apprecated real exchange rate and reduced trade wth the rest of the world. The agrcultural and mnng sectors are exceptons n that, there, cartelsaton reduces home producton wth less mpact on product prces due to foregn competton. Moble factors are shed, however, reducng ther rewards and ths tends to reduce servces costs and hence to deprecate the real exchange rate. Manufacturng s specal because, as Table 8 shows, t uses and supples manly ntermedate nputs. Snce elastctes of substtuton between ntermedate nputs are low, home cartelsaton sees servce ndustry costs rse due to lmted substtuton to competng mports. If the cartelsaton had only occurred n the recently prvatsed servces, electrcty, 42 Ths gnores the roles of contestablty and the free rder problem n the mantenance of cartels. 43 The number of frms s held constant n ths closure but t s, n any case, mmateral so far as prcng s concerned when ndustres are cartelsed. Consoldaton to a monopoly would reduce fxed costs and thereby ncrease monopoly profts, however, a development not explored here. 44 It stretches credblty to magne that sectors wth large numbers of small frms, such as agrculture, could overcome free rder and communcaton costs n ths way. Takng agrculture as an example, however, the Australan sector s rfe wth organsed boards desgned to extract rents for farmers (Seper, 982). Even the other servces sector s full of state and local government regulatons drected at reducng competton, such as zonng rules for such specalst retal outlets as pharmaces and news agents. All ths sad, our purpose here s not to suggest that full cartelsaton s possble or lkely but merely to use ths carcature of olgopolstc behavour to explore economy-wde effects. 6
water, gas, telecommuncatons, fnance, transport and other servces, GDP would have been smaller by just over a tenth and real wages by almost a thrd. The central block n the table ndcates how cartelsaton by each ndvdual sector affects overall economc performance. It s clear that sectors lke manufacturng and other servces, whch have large ntal shares of GDP, also have the largest mpacts on the economy followng cartelsaton. 45 Cartelsaton creates rents that accrue to captal owners. Yet t reduces output and therefore varable factor and nput demand. Unt factor rewards to moble factors therefore fall n all cases. 46 The only cartelsaton that reduces the average gross rate of return on captal s that of the manufacturng sector. Ths s, agan, because manufactured nputs are extensvely used n other sectors, as ndcated n Table 8, the performance of whch are retarded by hgh manufactured product prces. The bottom rows of Table 7 allow an assessment of the model s lnearty n proportonal changes followng cartelsaton shocks. Where collectve cartelsaton yelds results dfferent from the sum of the proportonal changes due to sectoral cartelsaton, the case for economy-wde analyss s made clearer. Whle ths non-lnearty s evdent when the cartelsng sectors nclude the tradable ones and the government-ntensve other servces, t s not strong when only the prvatsed servce sectors are ncluded. Opposng sectoral nteractons mght be expected to cancel when cartelsaton occurs n all sectors and so t follows that the gross effects are smaller n ths case than when sectoral cartelsatons are summed. When the traded sectors are ncluded, however, the elastc supply of competng products from abroad appears to further damp the collectve, relatve to the sectoral, mpacts of cartelsaton. Even though non-lnearty s not always strong, the case for economy-wde analyss s further supported by the substantal mpacts on GDP and real wages of the ndvdual sectoral cartelsatons shown n the table. The extent to whch sectors nteract s further clarfed from Table 9, whch shows the effects of cartelsaton by the column sectors on gross rates of return n row sectors. The frst row reproduces the sxth column of Table 7. The frst column gves the effects on all sectors of cartelsaton throughout the economy. rom ths t s evdent that nteracton between sectors causes gross returns n some to fall n spte of cartelsaton. Manufacturng s one of these, for the reasons ndcated above. Electrcty, water, fnance and transport all yeld net 45 The water sector also has a comparatvely large mpact. Ths s because the frms nvolved are few and manly state-owned and because they do not presently explot ther market power. The water prce would ncrease by 600 per cent f they dd! 46 The possblty that rents mght be shared wth sectoral workforces s real n Australa, though t s not modelled here. See Dowrck (993) and Mumford and Dowrck (994). 7
rses n rates of return n spte of hgher nput costs due to correspondng changes n other sectors. The second column shows that the effects of market power n the prvatsed servces s large at the natonal level and as t affects returns n the tradable sectors. Scannng the other columns, the non-dagonal elements ndcate the extent of sectoral nteracton. Ths s largest for manufacturng, for reasons dscussed above, but t s also sgnfcant for servces lke electrcty, telecommuncatons, fnance and transport. 5. Sectoral Interactons under Prce Cap Regulaton Here the ntal equlbrum, wth the pure profts generated n all sectors as ndcated n Table 6, s subjected to tght prce cap regulaton, whereby product prces are forced to equal average costs. No such regulaton s represented n the calbrated ntal equlbrum, even though t was probably nfluental n 997. Instead, the conjectural varatons parameters are set to ndcate consderable constrants to colluson n the olgopoles of the tme, due to regulatory polces. In our experment, tght prce caps are frst mposed, va equaton (4), smultaneously on all sectors. Because sx of the 0 sectors do not earn pure profts ths mples that mark-ups are regulated to decrease n only four and to ncrease n the remanng sx an unrealstc prospect ncluded only for completeness. The next experment mposes tght prce caps only on the four ndustres earnng pure profts (agrculture and food, mnng and energy, telecommuncatons and fnance). Then tght prce caps are mposed on the proftable prvatsed servces only. Ths s followed by prce caps on each of the proftable sectors ndvdually. The effects on overall economc performance are gven n Table 0. The very frst row of the table shows the effects of average cost prcng n all ndustres and, snce sx ntally earn less than market rates of return, the results are net ncreases n mark-ups and prces, and declnes n real wages. The more realstc mposton of tght prce caps on the four proftable sectors, however, shows that consderable addtonal economc actvty mght be obtaned n ths way. GDP and real wages are consderably boosted, at the expense of gross returns on assets. Interestngly, although the two proftable servce ndustres are not sgnfcantly larger contrbutors to GDP than ether agrculture and food on the one hand or mnng and metals on the other, ther prce caps are the more sgnfcant because ther ntal proftablty s hgher (Table 6). The results for ndvdual ndustres bear ths out. Indeed, t s the fnancal sector where reduced pure profts and hence lower product prces would have the most natonal 8
mpact. As n the cartelsaton experments of the prevous secton there s agan a contrast between the real exchange rate effects of prce caps n the tradable and the servces sectors. In the tradable sectors reduced home product prces cause ncreased demand and expanded output and hence ncreased wages and resource rents. The servces sectors face hgher wage costs and hence rase ther prces. But they also redrect ther demand to meetng the ntermedate requrements of the expanded tradable sectors and so the elastctes they face fall and ther mark-ups rse. Ths further contrbutes to hgher servce prces and hence to net ncreases n the real exchange rate. The prce caps on prevously proftable servces, on the other hand, merely reduce non-traded prces and the results are straght-forward Balassa- Samuelson real deprecatons. As n the case of cartelsaton, the non-lnearty of economy-wde responses to sectoral prce caps s tested n the last two rows of the table. By contrast wth the case of cartelsaton, prce caps on the four proftable sectors exhbt neglgble non-lnearty n proportonal changes. These do not nclude shocks to the large and dosyncratc manufacturng and other servces sectors, and the magntudes of the shocks are smaller than those due to cartelsaton. So t would seem that the non-lneartes are assocated wth the scale of shocks to prces and the partcular behavour of manufacturng and other servces. Ths offers weak support for economy-wde analyss. The drect nter-sectoral effects summarsed n Table offer further weak support n the sense that prce caps n the servces sectors have measurable economywde effects and that, especally n the case of fnance, nteractons are strong wth the tradable goods sectors. 6. Olgopoly and External Shocks: the Chna Boom Shocks to Australa s external terms of trade, to nflows on ts captal account and n ts trade polcy regme all have obvous effects on ts relatvely small agrcultural and ndustral sectors. They also change the real exchange rate and hence they ndrectly affect the state of ts largely non-traded servces sector. We show n ths secton that these ndrect effects have mplcatons for compettve behavour n both the tradable and servces sectors of the economy. They occur through the reapportonment of demand for olgopolstcally suppled goods and servces toward more elastc exports or less elastc ntermedate demand, va equaton (5). Both the terms of trade and the current account balance have been affected by the recent Chna boom. The extraordnary nature of the assocated commodty prce shocks s 9
clear from gure. Money prces of wheat and ron ore, both major Australan exports, have rsen n the last year or so by several hundreds of per cent. And the shocks go beyond those two commodtes. Australa s overall terms of trade rose n ths perod by 50 per cent, as shown n gure 2. At the same tme, foregn acquston of Australan assets has also rsen, almost doublng the current account defct (captal account surplus) between the turn of the century and 2007. 47 To examne the effects of these shocks on prvatsed servce performance, we subject the model to 50 per cent ncreases n the foregn prces of agrcultural products and mnng and energy products, combned wth an expanson of the current account defct by 50 per cent. Ths tme we use both a short run closure, n whch physcal captal s fxed and specfc to each sector and producton (unsklled) employment s flexble at a fxed real wage, and a long run closure n whch captal s nternatonally and ntersectorally moble and employment s fxed. The number of frms n each sector s held constant n the short run, allowng pure profts to vary. Two versons of the short run experment are carred out. In the frst, no prce caps are mposed and frms are permtted to choose ther mark-ups. 48 In the second, two new ntal equlbra are frst calculated, by mposng prce caps ether n all proftable sectors or n the proftable prvatsed servces only. These prce caps are then retaned whle the economy s subjected to the Chna boom shocks n each case. or the long run analyss the boom shock s frst mposed wth fxed numbers of frms. Then a new ntal equlbrum s calculated n whch entry and ext are allowed and pure profts reduced to zero. To ths ntal equlbrum the Chna boom shocks are appled n such a way as to retan zero pure profts but to allow entry of new frms. Taken collectvely, n both the short and long runs, these shocks are very postve for Australa. Ther prncpal short run effects are ndcated n the frst column of Table 2. Real GNP, whch serves as a preparedness to pay measure gven the formulaton we use (as dscussed n Secton 3), rses sgnfcantly, captal returns ncrease and ether real wages or employment levels rse. There are Dutch dsease elements, however (Corden and Neary 982). The natural resource based sectors expand at the expense of manufacturng, whch suffers from hgher prced factors and nputs and a real apprecaton mpars ts competton 47 It must be noted that these extraordnary shocks stem not only from the surge n the growth of Chna and other economes n transton. The US-ntated fnancal crss that began n 2007 saw a retreat to commodtes, further boostng prces and captal flght from the US. 48 Of course, ther strategc nteracton s assumed to be constraned by survellance, whch prevents the enlargement of ther conjectural varatons parameters. 20
aganst foregn products. 49 The servce sectors expand, however, suffcently to rase overall employment n the short run and real wages n the long run. The only sgnfcant exceptons to ths are electrcty and gas, whch are large supplers of nputs to manufacturng. Ther gross output levels are lower wth the boom. In the case of the gas sector, so also s ts gross rate of return on captal. urther nsght as to the role of prcng behavour n ths short run smulaton s avalable from the frst column of Table 3. Mark-ups fall n the natural resource based sectors as the share of elastc exports n the demand they face ncreases. or the electrcty sector, the mark-up falls because manufacturng s relatvely nelastc ntermedate demand contracts and so the elastcty t faces rses. Wth the excepton of electrcty the prvatsed servce sectors all experence less elastc demand as the share of ntermedate use by natural resource based sectors (and by each other) expands. Ths s partcularly true of the telecommuncatons sector, whch rases ts mark-up, reduces ts output and ncreases ts pure profts by half. The boom therefore causes these sectors to exhbt less compettve behavour that mght be expected to challenge regulatory agences. Returnng to Table 2, the remanng four columns detal the effects of the Chna boom shocks on the economy f prce caps are mantaned tght enough to elmnate all pure profts before and after the shocks. The mposton of the caps prevents ndustres from rasng rents assocated wth the boom and, as a consequence, the short run gans from the boom are larger, by amounts that depend on the number of sectors subjected to the prce caps. If all four proftable sectors are so regulated, the addtonal gan s nearly a per cent of real GNP. The real sklled wage would rse by a further per cent and producton employment would rse by more. The ncrease n telecommuncatons output due to the boom would be larger by.5 per cent. Indeed, output volumes would be larger for all the servce sectors. If the prce caps were restrcted to the proftable prvatsed servces the constranng effects are smaller and the addtonal boost derved from the Chna boom s smaller accordngly. The caps do, nonetheless, yeld measurable gans n economc performance that are evdent n the labour markets and n the supply of water, gas, telecommuncatons fnance and transport servces. Two apparently anomalous results emerge from Table 3. rst, the manufacturng sector, whch contracts n both the short and long runs (Table 4) and whose captal suffers 49 These changes are observed n the Australan economy. Also, the real apprecaton necesstates ether nflaton or a nomnal apprecaton. Australa s central bank prefers the latter, but to brng t about some monetary tghtenng s requred, placng low-margn mortgage holders under pressure. Ths lnk s explored emprcally by Bloch et al. (2006, 2007). 2
declnes n ts gross rates of return n both, appears to transton from negatve to postve pure profts as a consequence of the boom. Recall that, n the short run, sectoral captal use s fxed. The overall gross return on manufacturng captal falls due to ts reduced output. But the pure proft share of these returns rses because the real apprecaton caused by the boom swtches manufacturng demand away from elastc exports toward nelastc ntermedate markets n the domestc economy. Its overall elastcty of demand falls and ts mark-up therefore rses. The second anomaly s that the boom appears to reduce pure profts n mnng n the short run. In ths case the explanaton s the reverse of that for the manufacturng anomaly. A substantal rse n the export share of mnng output ncreases the elastcty of demand t faces and reduces the mnng mark-up. Other thngs equal, ths reduces the pure proft margn, and ths s the domnant short run force. The long run smulatons, detaled n Tables 4 and 5, yeld generally the same effects, wth substantal overall gans to the economy tempered by the Dutch dsease contracton of manufacturng and assocated contractons n utltes, partcularly the gas sector. The telecommuncatons sector also contracts, but all the other servce sectors expand strongly. The overall gans are conspcuously smaller when free entry and ext are allowed. Ths s because the hgh proftablty the boom brngs to most sectors does rase output n the long run but nduces new entry (Table 5) to the pont that output per frm actually declnes. 50 Ths rases the overall burden of fxed costs. Returnng to the apparently anomalous behavour of manufacturng profts, although the manufacturng mark-up remans larger n the long run due to the boom, pure profts declne substantally. Ths occurs because captal use s flexble at a fxed rate of return n the long run, so lower returns n manufacturng cause ts captal use to declne along wth ts output. In the frst column of Table 5, however, frm exts are not allowed so that fxed captal use remans constant. On reduced output, average fxed costs therefore absorb the entre mark-up, causng the declne n pure profts. In the second column, where entry and ext are free, frm numbers (and therefore fxed costs) contract by a ffth, helpng mantan the zero pure proft equlbrum. In the mnng sector, where pure profts declne n the short run f no prce cap s appled (because the mark-up declnes), n the long run varable captal use expands consderably. When entry s prohbted output also expands and the average fxed cost margn declnes. Pure profts therefore ncrease consderably (Table 5, frst column). 50 Ths s clear from a comparson of the changes n gross sectoral output n Table 4 wth the correspondng changes n frm numbers n Table 5. 22
When entry s allowed, however, output and frm numbers expand by smlar proportons and so average fxed costs change lttle, consstent wth the absence of pure profts. Manufacturng redress? The negatve consequences of the Chna boom for manufacturng have suggested to some that ts protecton should be renstated, or at least that scheduled declnes should be arrested. 5 To address ths, we expermented wth a rse n the power of the manufacturng tarff. Whle dfferences n the short and long run responses of the whole economy occur due to the complextes of mark-up choce and fxed cost margns dscussed above, the domnant outcome from more manufacturng protecton s an economy-wde contracton. Surprsngly, even the manufacturng sector tself fals to beneft from the tarff ncrease, both n the short and long runs. The answer to ths further anomaly les n the manufacturng sector s pattern of ntermedate use, whch s compared wth that of other sectors n Table 8. Of all the sectors manufacturng carres the hghest share of ntermedate nput cost n total turnover and by far the largest share of manufactured ntermedates, a thrd of whch are mported. So there are two mpacts of a tarff rse across the whole manufacturng sector. rst, and strongest, s the effect on ntermedate nput costs. Second, and weaker, s the effect of the tarff n rasng the prce of competng foregn manufactured products. Ths effect s weaker because home manufactures are dfferentated from foregn ones. Consumer substtuton between them s therefore constraned. The results therefore fal to offer support for a return to protecton to redress nequaltes from the Chna boom. Moreover, because a tarff rse n so mportant a tradable goods sector tends to turn the whole economy nward reducng elastctes of demand, mark-ups rse n all sectors except electrcty, transport and other servces. And, even though gross rates of return on captal fall across the board n both the short and long runs, ndustry scale also falls n all sectors, confrmng that such a polcy would doubly mpar Australa s economc effcency. 52 Boom-bust hysteress Lke all booms, that due to the present surge n Chnese growth s lkely to be transtory. Sooner or later there wll be a down-cycle. Yet, because n the up-cycle the 5 Or, at least, that scheduled declnes n protecton should be halted. See, for example, The Age (2008), VACC (2008). 52 Were such a polcy to be consdered n response to the Chna boom, for the manufacturng sector to be a clear benefcary t would need to be gven relef from tarffs on ts manufactured ntermedate nputs. Whle ths would drect the benefts approprately, the economc cost of the protecton would reman large and be borne n other sectors and n labour markets. 23
expandng tradable goods sectors bolster demand for servces as ntermedate nputs and ths demand s less elastc than that for fnal consumpton, the servce ndustres tend to prce less compettvely. Boom condtons therefore rase the bar for regulatory nsttutons. As seen earler, f tght prce caps are retaned, even f only n the presently proftable prvatsed servces, the benefts to the economy from the boom are shown be measurably ncreased. Moreover, the boom encourages exts from manufacturng and entres nto mnng, agrculture and servces, and presumably, the opposte n the followng down-cycle. Some emprcal evdence n support of excessve entry s offered n Table 6, whch shows the numbers of frms lsted on the ASX by Mornngstar sector. Snce 2002 the greatest expansons have been n the energy, materals (mnng products) and servces sectors. Snce ext costs are non-zero, boom-bust cycles must accompany debltatng hysteress. 53 Ths suggests that, at least for the servces sectors, tght prce caps serve two key purposes. rst, they enlarge and better dstrbute the gans durng up-cycles and, second, they prevent excessve entry and hence down-cycle ext costs. 7. Conclusons An economy-wde model wth olgopoly behavour facltates the analyss of ntersectoral and economy-wde effects of olgopoly rents, suggestng that these are potentally very large. Takng the extreme of cartelsaton n each sector as a benchmark, the complete explotaton of market power n all sectors s shown to leave the economy smaller by a thrd. Even f the cartelsaton had taken place only n the newly prvatsed servces the model suggests that Australa s GDP would have been smaller by almost a quarter. More partcularly, sectoral nteractons due to the explotaton of olgopoly power are shown to be large enough to justfy an economy-wde approach. Ths s not only true for olgopoly behavour n the major sectors of the economy but also n the recently prvatsed servces, supplyng as they do only a ffth of GDP. Moreover, the prce caps that would have elmnated over-market proftablty n the food, mnng and metals, telecommuncatons and fnance sectors also cause measurable changes n factor rewards and the real exchange rate. Tghter prce caps are shown to have sgnfcant effects on the performance of other sectors, partcularly when appled to fnance. A fnal set of experments subjects the model to a stylsed representaton of the recent Chna boom. The dea s to explore mplcatons of boom condtons for compettve 53 Ths s akn to the problem explored by Caballero and Lorenzon (2007). 24
behavour n the prvatsed servces sectors and hence ther regulaton. The sheer scale of the boom s made clear, along wth ts net postve mpacts on the economy as a whole. Because one of ts key consequences s an apprecaton of the real exchange rate, however, there s a relatve rse n servces prces. The servce sectors therefore expand. Yet, because the expandng tradable goods sectors bolster demand for servces as ntermedate nputs and ths demand s less elastc than that for fnal consumpton, the servce ndustres tend to prce less compettvely. Boom condtons are therefore lkely to ncrease stress on regulatory nsttutons. If tght prce caps could be retaned across the economy, however, even f only n the prvatsed servces, the benefts to the economy from the boom are shown be measurably larger. Moreover, because strong Dutch dsease consequences are unavodable, the boom encourages exts from manufacturng and entres nto mnng, agrculture and servces. Snce booms are nvarably transtory and ext costs are non-zero, boom-bust cycles nevtably accompany debltatng hysteress. Ths suggests some form of assstance to manufacturng durng the boom to prevent excessve ext and tght prce caps n servces to prevent excessve entry. References ABS (2003), Year Book of Australa 2003, Australan Bureau of Statstcs, Commonwealth of Australa, Canberra. Age, The (2008), Rudd weghs tarff freeze for remanng makers, http://www.theage.com.au/artcles/2008/02/05/20209042833.html, accessed 4 March 2008. Australa Daly (2004), Australa Energy Info, <http://www.australadaly.com/s/australaenergy/> accessed September October. ACCC (2003), ACCC ssues consultaton notce to Telstra over broadband prce-squeeze allegatons, <http://www.accc.gov.au/>. ACCC (2004), What We Do, Australan Competton and Consumer Commsson <http://www.accc.gov.au>. Balstrer, E.J., R.H. Hllberry and T.J. Rutherford (2007), Structural estmaton and soluton of nternatonal trade models wth heterogeneous frms, presented at the 0 Annual Conference on Global Economc Analyss, Purdue Unversty, July. Bloch, H., A.M. Dockery, C. Wyn Morgan and D. Sapsford (2007), Growth, commodty prces, nflaton and the dstrbuton of ncome, Bloch, H., A.M. Dockery and D. Sapsford (2006), "Commodty Prces and the Dynamcs of Inflaton n Commodty-Exportng Natons: Evdence from Australa and Canada," The Economc Record, 82: S97-S09, 09. Bradley, I. and Prce, C. (988) The economc regulaton of prvate ndustres by prce constrants, Journal of Industral Economcs, 37:99-06. Brennan, T. (989), Regulatng by Cappng Prces, Journal of Regulatory Economcs, (2): 33-47. 25
Caballero, R.J. and G. Lorenzon (2007), Persstent apprecatons and overshootng: a normatve analyss, presented at the 0 th Annual Internatonal Economcs Conference, Santa Cruz Center for Internatonal Economcs, 5-6 October, Unversty of Calforna at Santa Cruz. Cabral, L. and M. Rordan (989), "Incentves for Cost Reducton under Prce Cap Regulaton," Journal of Regulatory Economcs, Sprnger, (2): 93-02, June. Corden, M. and Neary, P. (982), Boomng sector and de-ndustralzaton n a small open economy, The Economc Journal, 92: 825-848. Courvlle, L. (974), Regulaton and effcency n the electrc utlty ndustry, The Bell Journal of Economcs 5(): 53-74. Dee, Phlppa S. (2003), Modellng the polcy ssues n servces trade, Econome Internatonale, 94-95: 283-300. Dmaranan, B.V. and McDougall, R.A., 2002. Global Trade, Assstance and Producton: the GTAP 5 data base, May, Center for Global Trade Analyss, Purdue Unversty, Lafayette. Dxon P.B., B.R. Parmenter, J. Sutton and D.P. Vncent (982), ORANI, a Mult-Sectoral Model of the Australan Economy, Amsterdam: North Holland. Dxon, P.B. and M. Rmmer (2002), Dynamc General Equlbrum Modellng for orecastng and Economc Polcy, No. 256 n the Contrbutons for Economc Analyss seres, publshed by Elsever North Holland. Doove, S et al (200), Prce Effects of Regulaton: Internatonal Ar Passenger Transport, Telecommuncatons and Electrcty Supply, Productvty Commsson Staff Research Paper, AusInfo, Canberra, October. Dowrck, S. (993), Enterprse barganng, unon structure and wages, The Economc Record, 69(207): 393-404. ndlay, C. (2000), Introducton to the regulaton of servces, n Achevng Better Regulaton of Servces, Conference Proceedngs, Australan Natonal Unversty, 26-27 June. reed, J. (2004), Qantas takes huge gamble wth Jetstar, Sydney Mornng Herald, May 7. Golley, J.E. (993), Pro-compettve effects of trade reform wth mperfect competton, Honours thess, School of Economcs, Australan Natonal Unversty. Gunasekera, H.D.B. and R. Tyers (990), "Imperfect Competton and Returns to Scale n a Newly Industralsng Economy: A General Equlbrum Analyss of Korean Trade Polcy", Journal of Development Economcs, 34: 223-247. Harrs, R.G. (984), Appled general equlbrum analyss of small open economes wth scale economes and mperfect competton, Amercan Economc Revew 74: 06-032. Harrs, R.G. and D. Cox (983), Trade, Industral Polcy and Canadan Manufacturng, Toronto: Ontaro Economc Councl. Hertel, T.W., (994), The pro-compettve effects of trade polcy reform n a small, open economy, Journal of Internatonal Economcs, 36: 39-4. Horrdge, M. (987), The long term costs of protecton: expermental analyss wth dfferent closures of and Australan computable general equlbrum model, PhD dssertaton, Unversty of Melbourne. Ianchovchna, E., J. Bnkley and T.W. Hertel (2000), Procompettve effects of foregn competton on domestc markups, Revew of Internatonal Economcs, 8(): 34-48. Lttlechld, S.C. (983), Regulaton of Brtsh Telecommuncatons Proftablty, Department of Trade and Industry, Government of Great Brtan, London. 26
Madden, G., H. Bloch and G. Coble-Neal (2002), "Labour and Captal Savng Techncal Change n Telecommuncatons," Appled Economcs, 34(4): 82-28, September. Meltz, Marc J. (2003), "The Impact of Trade on Intra-Industry Reallocatons and Aggregate Industry Productvty," Econometrca, 7(6), 695-725. Menezes,., R. Breung, S. Stacey and J. Hornby (2006), Prce regulaton n Australa: how consstent has t been? The Economc Record 82 (256): 67 76. Moran, A. (2002), Over-regulaton addng fuel to the fre of ndustry resentment, The Age, August 26. Mumford, K.A. and S. Dowrck (994), Wage barganng wth endogenous profts, overtme workng and heterogeneous labor, Revew of Economcs and Statstcs, 76(2): 329-336. NECG (2003) Internatonal Comparson of WACC decsons. Submsson to the Productvty Commsson Revew of the Gas Access Regme. Avalable at www.necg.com.au. OECD (997), The OECD Report on Regulatory Reform: Summary, Organsaton for Economc Cooperaton and Development, Pars <http://www.oecd.org/>. Productvty Commsson (999), Mcroeconomc Reform and Australan Productvty: Explorng the Lnks, Productvty Commsson Research Paper, AusInfo, Canberra. and Australan Natonal Unversty (2000), Achevng Better Regulaton of Servces, Conference Proceedngs, AusInfo, Canberra, November. (2002), Prce Regulaton of Arport Servces, Report no. 9, AusInfo, Canberra. (2003), Regulaton and ts Revew 2002-03, Annual Report Seres, Productvty Commsson, Canberra. Rees, L. (2004), Economy-wde consequences of regulatory reform n an Australan context, Honours Thess, College of Busness and Economcs, Australan Natonal Unversty. Short, C., A. Swan, B. Graham and W. Mackay-Smth (200), Electrcty reform: the benefts and costs of Australa, Outlook 200: Proceedngs of the Natonal Outlook Conference, vol. 3, Mnerals and Energy, ABARE, Canberra. Short, C., A. Heaney and K. Burns (2003), Australan Gas Markets Movng Toward Maturty, Australan Bureau of Agrcultural and Resource Economcs ereport 03.23, prepared for the Australan Gas Assocaton, Canberra, December. Seper, E., (982), Ratonalsng Rustc Regulaton, Centre for Independent Studes, Sydney. Stgler, G. (97), The Theory of Economc Regulaton, The Bell Journal of Economcs and Management Scence, 2(): 3-2, Sprng. Telstra (2003), Annual Report. Tyers, R. (2005), Trade reform and manufacturng prcng behavour n four archetype Asa- Pacfc Economes, Asan Economc Journal 9(2): 8-203, 2005. VACC (2008), Mtsubsh: a message on tarffs, http://www.vacc.com.au/newsadvocacy/medareleases/2008/ebruary2008/5ebrua rymtsubshamessageontarffs/tabd/254/default.aspx, accessed 4 March 2008. Vogelsang, I. and Acton, J. (989), Introducton to Symposum on Prce-Cap Regulaton, RAND Journal of Economcs, 20: 369-72. Wellsz, S.H. (963), Regulaton of natural gas ppelne companes: an economc analyss, Journal of Poltcal Economy 55(): 30-43. Xaver, P. (995), Prce cap regulaton for telecommuncatons: how has t performed n practce? Telecommuncatons Polcy, 9(8): 599-67. 27
gure : Australa s Chna Boom Commodty Prce Shocks 500 500 400 400 300 Wheat Prce Index 300 Iron Ore Prce Index 200 200 00 00 0 3/0/993 28/0/995 24/07/998 9/04/200 4/0/2004 0/0/2006 06/07/2009 0 993 995 997 999 200 2003 2005 2007 2009 Sources: Wheat: Chcago board of trade daly wheat prce n US$/bushel, from the Bloomberg Database. Iron ore: Hamersley fnes, quoted n US cents/dmtu - dry ron unts. If the ore shpped s 62% E (the typcal Hamersley grade) then the prce per tonne of ore s the dmtu prce ( for 2007 that would be US $0.82) X 62 whch means US$50.84/tonne, from the IRL Database. gure 2: Australa s Terms of Trade and Current Account Defct 80 0 60-40 -2 20-3 00 CA/GDP, % Terms of Trade Index -4 80 60-5 40-6 20-7 0 Jan-993 Oct-995 Jul-998 Apr-200 Jan-2004 Oct-2006 Jul-2009-8 Jan-993 Oct-995 Jul-998 Apr-200 Jan-2004 Oct-2006 Jul-2009 Sources: The terms of trade s here the quotent of the ndces of export prces and mport prces, where both ndces are as suppled by the Australan Bureau of Statstcs. The current account balance s n per cent of GDP, compled quarterly, from the Australan Bureau of Statstcs web ste. 28
Table : Model structure Regons Australa Rest of world Prmary factors Natural resources (mneral, energy deposts and land) Sklled (professonal) labour Unsklled (producton) labour Physcal captal Sectors Agrculture and food processng Manufacturng Mnng, petroleum and mnerals Electrcty Water Gas manufacture and dstrbuton Telecommuncatons nance and nsurance Transport Other servces Source: Aggregates of the 57 sector GTAP Verson 5 database from Dmaranan and McDougall (2002). Table 2: Economc Sgnfcance of Prvatsed Servces n the Model Database Value added share of GDP Share of total exports Export share of output Agrculture 6.8 2.7 22.2 Manufacturng 4.0 40.5 9.4 Mnng 5.5 20.3 43.8 Electrcty.9 0.0 0.2 Water.2 0. 0.8 Gas Dstrbuton 0.2 0. 5.5 Telecommuncatons 3..5 6.5 nance 7.4 2.2 4.2 Transport 5.2 2.0 3.8 Other Servces 54.7.6 2.3 Source: Model database (socal accountng matrx), derved from Dmaranan and McDougall (2002). Table 3: actor Intenstes by Industry a Natural resources Sklled labour Unsklled labour Physcal captal Agrculture 2 7 46 34 Manufacturng 0 7 40 42 Mnng 29 5 7 50 Electrcty 0 8 4 78 Water 0 2 2 68 Gas Dstrbuton 0 8 5 77 Telecommuncatons 0 22 23 55 nance 0 23 25 52 Transport 0 2 33 55 Other Servces 0 28 30 42 a These are factor shares of total value added n each ndustry, calculated from the database. Shares sum to 00 per cent horzontally. Source: Model database (socal accountng matrx), derved from Dmaranan and McDougall (2002). 29
Table 4: Conjectural Varatons and Intal Elastcty Values Index of Conjectural Demand elastctes frm varatons numbers a parameter nal Government Intermedate Export Agrculture 50 0. -3.8-3.8-7.0-4.3 Manufacturng 20 0.2-2.7-2.6-5.7-3.6 Mnng 0 0.3 -.2 -.2-6.4-2.7 Electrcty 6 0.4-9.0-9.2-4.0-0.5 Water 6 0.3-9.9-9.9-4.3 -.2 Gas Dstrbuton 2 0.5-5.3-5.6-3. -8.2 Telecommuncatons 4 0.6-6.6-6.6-2.7-8.7 nance 0 0.5-8.7-8.7-4.7-0.0 Transport 0 0.5-8.7-8.7-3.4-0.0 Other Servces 00 0.2-2.7-2.7-4.2-3. a Ths ndex represents the effectve number of strategcally nteractng frms n each sector. Sources: Effectve frm numbers and conjectural varatons parameters are crude estmates, based on ndustry concentraton, from the Mornngstar nancal Analyss Database of lsted Australan frms. Elastctes are calculated va the equatons n the appendces, where elastctes of substtuton are sourced from surveys cted by Harrs and Cox (984) and Dmaranan and McDougall (2002). Table 5: Intal Demand Shares, Average Elastctes and Mark-ups a nal demand share Government demand share Intermedate demand share Export demand share Average demand elastcty Industry mark-ups b Agrculture 43 0 34 22 -.6.09 Manufacturng 7 0 64 9-8.4.4 Mnng 3 52 44-9.4.2 Electrcty 25 0 75 0-5.2.24 Water 6 3 9 0-4.9.26 Gas Dstrbuton 2 0 83 4-3.7.37 Telecommuncatons 28 0 65 6-4.2.3 nance 30 0 66 4-6..20 Transport 2 25 50 4-6..20 Other Servces 45 22 3 2-0.0. a All these varables are endogenous n the model. Intal (base) values are provded here. b Industry mark-ups are the rato of producer prces and average varable costs. Source: Model database (socal accountng matrx), derved from Dmaranan and McDougall (2002). Table 6: Calbrated Pure Proft, Cost Shares and Industry Scale Per cent of ndustry turnover Pure proft a xed cost a Varable cost a Scale b Agrculture 0.4 8.3 9.4 55 Manufacturng -0.2 2.2 88.0 36 Mnng.0 9.6 89.4 46 Electrcty -2.9 22.0 80.9 8 Water -2.9 23.5 79.4 7 Gas dstrbuton -2.9 30.2 72.7 2 Telecommuncatons 4.8 9. 76. 20 nance 9. 7.2 83.6 58 Transport -.0 7.4 83.6 24 Other Servces -0.9 0.8 90.0 42 a The fnal three columns of the table are calbrated. rst, elastctes are estmated, from whch mark-up ratos are calculated. The pure proft shares are then used to deduce the fxed cost resdual. b Scale s defned as the rato (n %) of the gross quantty produced and mnmum effcent scale, whch n turn, s the level of output where unt fxed cost s 5% of unt varable cost. Source: Pure proft proportons are from the Mornngstar nancal Analyss Database of lsted Australan frms. 30
Table 7: Effects of cartelsaton, whole economy and ndvdual sectors a Real sklled wage Real producton wage Real resource rent Average gross rate of return b Real exchange rate c Cartelsaton of: Real GNP Real GDP Whole economy -22.9-32.7-54.7-56.6-48.4 46. 24.8 Prvatsed servces d -2.0-23.9-28.7-29.8-25.2 9.8 9.2 Agrculture -.3-0.3-0.5-3. -4.0 2.5-0.9 Manufacturng -0.5-7.2-8.9-22.8-0.6-5.0.4 Mnng & energy -0.5 -.3 -.0 -. -5.6 0.6-0.2 Electrcty -2.6-4.5-4.7-5.4-6.6 0.7.9 Water -3.2-0.2-0.7 -.0-0.7 0.4 9. Gas -0.2-0.3-0.3-0.4-0.3 0.0 0. Telecommuncatons -2.7-4.9-6.5-6.0-4.4.7 3.3 nance -2.4-4.0-6.7-5.8-3.8 3.6 3.0 Transport -2.3-4.2-5.5-7.2-3.8 3.7 2.0 Other servces -3.3-6.0-39.5-35.8-29. 42.0 2.8 Sum whole econ -39.0-62.9-94.3-98.5-89.0 50.2 4.4 Sum prvatsed srv -3.4-28. -34.3-35.7-29.7 0. 9.4 a The shock here s to rase the conjectural varatons parameter from ts baselne value to unty, for the whole economy, for the prvatsed servce sectors only and, fnally, for each sector n turn. The sums n the bottom rows smply add the sectoral effects to test the non-lnearty of the response. b The gross rate of return ncludes deprecaton. It sums pure profts wth (nternatonal) market captal returns and dvdes the total by the market value of the domestc captal stock, whch s a volume measure tmes the current prce of captal goods (an ndex of manufacturng and servce prces). The changes shown are not percentage pont or bass pont changes. They are proportonal changes n the rates of return, so that a rse n a rate from 5%/year to 6%/year consttutes a 20% change as dsplayed. c The real exchange rate s the home GDP prce dvded by the foregn GDP prce, where both are measured relatve to unchangng mport prces. In these experments t s therefore the equvalent of the change n the home GDP prce. d Prvatsed servces are electrcty, water, gas, telecommuncatons, fnance and transport. Source: Smulatons of the model descrbed n the text, wth fxed frm numbers but otherwse long run closures: full employment and flexble wages, moble captal at a fxed exogenous rate of return. Table 8: Intermedate Cost Shares of Total Turnover All nputs Manufactured Agrculture 65.9 3.6 Manufacturng 66.6 38.4 Mnng 42.0 2.9 Electrcty 49.5 4.3 Water 25.7 0.7 Gas Dstrbuton 4.4.2 Telecommuncatons 34.9 0.9 nance 3.2.3 Transport 5.9 8.2 Other Servces 45.7 3.2 Source: Model database (socal accountng matrx), derved from Dmaranan and McDougall (2002). 3
Table 9: Effects on Gross Rates of Return of Cartelsaton n Each Sector a % change Whole economy Prvatsed servces b Agrculture Manufacturng Mnng energy Electrcty Water Gas Telecoms nance Transport Other servces Average 46. 9.8 2.5-5.0 0.6 0.7 0.4 0.0.7 3.6 3.7 42.0 Agrculture 28. -6.4 48.0-2.9 2.8-0.6-2.4-0. -0.3-0.3-3.0-9.0 Manufacturng -.6-2.6 2.3 6.6-0.7-5.3-2.7-0.4 -.3-0. -3.5-8.9 Mnng & energy -3.2-4.5 3.5-8.9 7.0-3.2 -.0-0. -0.4 0.0 0. -9.6 Electrcty 8.0 85.0 0.7-2.2 -.0 99.5-2.3-0.2-0.8-0.2 -.0-2. Water 74.3 230.4-0.3-6.9-0.3 -.0 252.7-0. -0.5-0.3-0.6-8.6 Gas -25.3 4.8 0. -25.9-0.7-7. -2. 55.0 -.0-0. -.5-2.6 Telecoms 9.4 76.6-0. -0.7-0.5 -.4 -.7-0. 94.8 -. -.9-7.8 nance 35.0 62.5-0.2-4.9-0.2-0.6-0.9 0.0-0.5 69.0-0.6-8.7 Transport 29.6 80.6 -.9-3.3-0.2 -.9 -.8-0.2 -.4-0.3 96.8 -.5 Other servces 78.2-3.8-0. -7.2-0.3-0.9 -.2-0. -0.6-0.4-0.8 04.5 a Here the conjectural varatons parameter s shocked to a level of.0, ndcatng cartelsaton, frst n all sectors, then n only the prvatsed servces and, fnally, n each sector ndvdually. The gross rate of return ncludes deprecaton. It sums pure profts wth (nternatonal) market captal returns and dvdes the total by the market value of the domestc captal stock, whch s a volume measure tmes the current prce of captal goods (an ndex of manufacturng and servce prces). The changes shown are not percentage pont or bass pont changes. They are proportonal changes n the rates of return, so that a rse n a rate from 5%/year to 6%/year consttutes a 20% change as dsplayed. b Prvatsed servces are electrcty, water, gas, telecommuncatons, fnance and transport. Source: Smulatons of the model descrbed n the text. 32
Table 0: Effects of prce caps, whole economy and ndvdual sectors, % a Real sklled wage Real producton wage Real resource rent Average gross rate of return c Real exchange rate d Prce caps n: Real GNP Real GDP Whole economy b -0.34-0.25-0.09-0.84.26-0.27-0.8 Proftable sectors.4.78 3.74 3.8 4.30-2.9-0.86 Prof prv servces e.02.58 3.57 2.86.03-2.62-0.99 Agrculture 0.04-0.02 0.00 0.7 0.26-0.9 0.0 Mnng & energy 0.09 0.23 0.7 0.20 3.40-0.8 0.07 Telecommuncatons 0.26 0.39 0.75 0.62 0.26-0.46-0.22 nance 0.75.7 2.78 2.20 0.76-2.5-0.76 Sum prof sectors.4.76 3.70 3.9 4.67-2.98-0.8 Sum prof prv srv.0.56 3.53 2.83.0-2.6-0.98 a The shock here s to mpose regulated prce caps (P=AC), for the whole economy, for the sectors wth pure profts, for the prvatsed servces among them and, fnally, for each of the proftable sectors n turn. The sums n the bottom rows smply add the sectoral effects to test the non-lnearty of the response. b The frst row dffers n that P=AC s enforced n all sectors, even those prevously makng pure losses markups n these sectors are therefore rased. c The gross rate of return ncludes deprecaton. It sums pure profts wth (nternatonal) market captal returns and dvdes the total by the market value of the domestc captal stock, whch s a volume measure tmes the current prce of captal goods (an ndex of manufacturng and servce prces). The changes shown are not percentage pont or bass pont changes. They are proportonal changes n the rates of return, so that a rse n a rate from 5%/year to 6%/year consttutes a 20% change as dsplayed. d The real exchange rate s the home GDP prce dvded by the foregn GDP prce, where both are measured relatve to unchangng mport prces. e Prvatsed servces are electrcty, water, gas, telecommuncatons, fnance and transport. Proftable amongst these are telecommuncatons and fnance. Source: Smulatons of the model descrbed n the text, wth fxed frm numbers but otherwse long run closures: full employment and flexble wages, moble captal at a fxed exogenous rate of return. 33
Table : Effects on Gross Rates of Return of Prce Caps n Each Sector a % change Whole economy b All proftable sectors All proftable prvatsed servces c Agrculture Mnng & energy Telecoms nance Average -0.27-2.9-2.62-0.9-0.8-0.46-2.5 Agrculture -2.97-2.97-0.35-2.96-0.55-0.08-0.27 Manufacturng.58-0.43-0.42-0.6 0.0 0.04-0.47 Mnng & energy -3.47-3.46-0.9-0.25-3.45-0.0-0.8 Electrcty 7.3 0.06-0.09-0.05 0.9 0.05-0.4 Water 5.69 0.08 0.02 0.02 0.05 0.02 0.00 Gas 6.26-0.07-0.7-0.0 0. 0.07-0.25 Telecoms -3.07-3.05-3.05 0.00 0.0-3.06 0.40 nance -24.94-24.94-24.94 0.0 0.04 0.04-24.94 Transport 3.7 0.5 0.02 0.2 0.04 0.4-0.3 Other servces 3.74 0.6 0. 0.00 0.06 0.05 0.06 a The shock here s to mpose regulated prce caps (P=AC), for the whole economy, for the sectors wth pure profts, for the prvatsed servces among them and, fnally, for each of the proftable sectors n turn. The sums n the bottom rows smply add the sectoral effects to test the non-lnearty of the response. The gross rate of return ncludes deprecaton. It sums pure profts wth (nternatonal) market captal returns and dvdes the total by the market value of the domestc captal stock, whch s a volume measure tmes the current prce of captal goods (an ndex of manufacturng and servce prces). The changes shown are not percentage pont or bass pont changes. They are proportonal changes n the rates of return, so that a rse n a rate from 5%/year to 6%/year consttutes a 20% change as dsplayed. b Here P=AC s enforced n all sectors, even those prevously makng pure losses mark-ups n loss-makng sectors are therefore rased. c Prvatsed servces are electrcty, water, gas, telecommuncatons, fnance and transport. Proftable amongst these are telecommuncatons and fnance. Source: Smulatons of the model descrbed n the text, wth fxed frm numbers but otherwse long run closures: full employment and flexble wages, moble captal at a fxed exogenous rate of return. 34
Table 2: Short Run Economc Effects of the Chna Boom a % changes No prce caps Prce caps n proftable sectors c Dff due to caps Prce caps n proftable prvatsed servces d Dff due to caps Real GNP 9.9 0.7 0.8 0.2 0.2 Real GDP 6. 6.9 0.8 6.3 0.2 Real exchange rate 8.4 8.7 0.3 8. -0.3 Total captal use 0.0 0.0 0.0 0.0 0.0 Real sklled wage.8 2.8.0 2.0 0.2 Real producton wage 0.0 0.0 0.0 0.0 0.0 Prodn employment 5.2 6.7.5 5.5 0.3 Real resource rent 49.7 5.2.5 50.6 0.9 Gross rate of return b Average, all sectors 2.8 3.0 0.2 2.8 0.0 Agrculture 99.0 94. -4.9 00.3.3 Manufacturng -22.5-23.2-0.7-22.2 0.3 Mnng & energy 4.8 40.9-0.9 4.9 0.2 Electrcty 3.3 5..8 3.6 0.3 Water 2.8 4.2.4 3. 0.3 Gas -2.5-0.5 2.0-2. 0.3 Telecoms 6.3 4.5 -.8 3.7-2.6 nance 8.7 7.0 -.7 6.2-2.5 Transport 0.5 2.3.9 0.8 0.3 Other servces.8 3.0.2 2. 0.3 Gross sectoral output Agrculture 67.0 74. 7. 68.. Manufacturng -27.0-25.0 2.0-26.8 0.2 Mnng & energy 7.9 6.3 -.6 8.0 0. Electrcty -3.8-3. 0.7-3.7 0. Water 4.4 4.9 0.5 4.5 0. Gas -7. -6.3 0.8-6.9 0.2 Telecoms -0.5.0.5 0.6. nance.0.0 0. 0.7-0.3 Transport 5.4 6.5. 5.7 0.2 Other servces 5.2 5.8 0.6 5.4 0.2 a Here the shock ncludes 50% rses n the nternatonal prces of agrcultural and mnng and energy products and an ncrease n captal account nflows by 50% (or about 5% of GDP). The short run closure has physcal captal fxed n each sector and mmoble nternatonally and the real wage of producton labour fxed. b The rate of return on physcal captal s here gross of deprecaton and nclusve of pure economc profts. c Ths smulaton commences wth a modfed ntal equlbrum n whch prce caps are mposed on all proftable sectors, agrculture and food, mnng and metals, telecommuncatons and fnance, and pure profts are therefore drven to zero n these sectors. Prce caps are assumed to be tghtly appled on sectors potentally earnng pure profts both before and after the Chna boom. d Ths smulaton commences wth a modfed ntal equlbrum n whch prce caps are mposed only on the proftable prvatsed servces sectors, telecommuncatons and fnance, and pure profts n those sectors alone are therefore zero. Thus, prce caps n these sectors alone are assumed to be tghtly appled before and after the Chna boom. Source: Smulatons of the model descrbed n the text. 35
Table 3: The Role of Imperfectly Compettve Behavour n the Short Run Response to the Boom a % changes No prce caps Prce caps n all proftable sectors b Dff due to caps Prce caps n prvatsed servces c Dff due to caps Mark-ups Agrculture -0.2-0.90-0.69-0.2 0.00 Manufacturng 2.9.53-0.66 2.8-0.0 Mnng & energy -.5-0.56 0.58 -.5 0.00 Electrcty -0.5-0.47 0.04-0.50 0.0 Water 0.20 0.2 0.0 0.20 0.00 Gas.5.56 0.05.48-0.03 Telecoms.52 0.7 -.35 0.8 -.35 nance 0.32 0.0-0.32 0.0-0.3 Transport 0.67 0.74 0.07 0.67 0.00 Other servces 0.03 0.04 0.0 0.03 0.00 Prod prces rel to P GDP Agrculture 5.6 4.9-0.8 5.7 0. Manufacturng -3.5-4. -0.6-3.4 0. Mnng & energy 3.5 3.6 0. 3.7 0.2 Electrcty.8 2.5 0.7 2.0 0.2 Water 0.7.2 0.5 0.8 0. Gas -.4-0.7 0.7 -.2 0.2 Telecoms -.9-2.6-0.8-2.8 -.0 nance 0.3 0.2-0. -0. -0.5 Transport -3.5-3.4 0.2-3.5 0. Other servces -0.8-0.6 0.2-0.8 0.0 Pure profts Agrculture 525.... 532 7 Manufacturng -293-00 d.. d -297-4 Mnng & energy -9.... -9 0 Electrcty 50 49-50 0 Water 2 2 0 2 0 Gas - -5-4 - 0 Telecoms 52........ nance 3........ Transport -57-72 -4-60 -3 Other servces 3 0-3 2 - a Here the shock ncludes 50% rses n the nternatonal prces of agrcultural and mnng and energy products and an ncrease n captal account nflows by 50% (or about 5% of GDP). The short run closure has physcal captal fxed n each sector and mmoble nternatonally and the real wage of producton labour fxed. b Ths smulaton commences wth a modfed ntal equlbrum n whch prce caps are mposed on all proftable sectors, agrculture and food, mnng and metals, telecommuncatons and fnance, and pure profts are therefore drven to zero n these sectors. Prce caps are assumed to be tghtly appled on sectors potentally earnng pure profts both before and after the Chna boom. c Ths smulaton commences wth a modfed ntal equlbrum n whch prce caps are mposed only on the proftable prvatsed servces sectors, telecommuncatons and fnance, and pure profts n those sectors alone are therefore zero. Thus, prce caps n these sectors alone are assumed to be tghtly appled before and after the Chna boom. d The manufacturng sector s ntally unproftable but s rendered proftable by the boom and so a prce cap s also mposed on t n ths smulaton, forcng zero pure profts. Source: Smulatons of the model descrbed n the text. 36
Table 4: Long Run Economc Effects of a Sustaned Chna Boom a % changes No entry/ext c ree entry/ext d Real GNP 0.8 9.5 Real GDP 9.0 7.8 Real exchange rate 9.5 9.6 Total captal use 9.5 2.3 Real sklled wage 2.9 3.2 Real producton wage 8.0 7.9 Producton employment 0.0 0.0 Real resource rent 76.8 76.6 Gross rate of return b Average, all sectors 5.0 0.0 Agrculture 42. 0.0 Manufacturng -8.0 0.0 Mnng & energy 0.7 0.0 Electrcty 0. 0.0 Water 4.4 0.0 Gas -.5 0.0 Telecoms 4.7 0.0 nance 2.4 0.0 Transport 6.3 0.0 Other servces 3.9 0.0 Gross sectoral output Agrculture 8.3 80.0 Manufacturng -33.6-34.6 Mnng & energy 57.2 57.5 Electrcty 0.0 -. Water 9.7 8.8 Gas -8.7-9.7 Telecoms -.3-2.3 nance 2.2.0 Transport 6.2 5. Other servces 6.5 5.2 a Here the shock ncludes 50% rses n the nternatonal prces of agrcultural and mnng and energy products and an ncrease n captal account nflows by 50% (or about 5% of GDP). The long run closure has physcal captal moble nternatonally and ntersectorally and the supply of producton labour fxed. b The rate of return on physcal captal s here gross of deprecaton and nclusve of pure economc profts. c The no entry and ext smulaton begns wth the orgnal ntal equlbrum, whch ncludes frms earnng pure profts n each sector. d The free entry and ext smulaton begns from a zero pure proft development of the ntal equlbrum, generated by a long run smulaton n whch entry and ext are free and pure profts are shocked down to zero. Source: Smulatons of the model descrbed n the text. 37
Table 5: The Role of Imperfectly Compettve Behavour n the Long Run Response to a Sustaned Chna Boom a % changes No entry/ext b ree entry/ext c Mark-ups Agrculture -0.22-0.27 Manufacturng 2.67 2.86 Mnng & energy -.49 -.63 Electrcty -0.45-0.43 Water 0.2-0.9 Gas.63 2.23 Telecoms.75.35 nance 0.38 0.3 Transport 0.77 0.63 Other servces 0.06 0.05 Prod prces rel to P GDP Agrculture 3.2 3. Manufacturng 0.3 0.5 Mnng & energy 7.4 7.2 Electrcty -0.8-0.8 Water -2.7-3. Gas 0.3 0.8 Telecoms -0.4-0.7 nance 0.4 0.3 Transport -2.8-2.9 Other servces -0.9-0.9 Pure profts Agrculture 2532 0.0 Manufacturng 5 0.0 Mnng & energy 675 0.0 Electrcty 22 0.0 Water -70 0.0 Gas 43 0.0 Telecoms 63 0.0 nance 39 0.0 Transport -82 0.0 Other servces -04 0.0 Effectve number of frms Agrculture 0.0 89.0 Manufacturng 0.0-8.4 Mnng & energy 0.0 54.2 Electrcty 0.0-0.7 Water 0.0 6.9 Gas 0.0 -.6 Telecoms 0.0 5.2 nance 0.0 3.0 Transport 0.0 7.8 Other servces 0.0 7. a Here the shock ncludes 50% rses n the nternatonal prces of agrcultural and mnng and energy products and an ncrease n captal account nflows by 50% (or about 5% of GDP). The long run closure has physcal captal moble nternatonally and ntersectorally and the supply of producton labour fxed. b The no entry and ext smulaton begns wth the orgnal ntal equlbrum, whch ncludes frms earnng pure profts n each sector. c The free entry and ext smulaton begns from a zero pure proft development of the ntal equlbrum, generated by a long run smulaton n whch entry and ext are free and pure profts are shocked down to zero. Source: Smulatons of the model descrbed n the text. 38
Table 6: Changes n the Numbers of Lsted rms n Australa 997-2007: Sectors Ranked on Rato of 2007 to 2002 Numbers a Mornngstar sector 997 2002 2007 2007/997 2007/2002 Energy 72 97 50 2..5 Health Care 57 20 63 2.9.4 Materals 252 39 44.6.3 All lsted frms 847 226 428.7.2 Industrals 33 84 208.6. Telecommuncatons 8 32 36 2.0. Utltes 0 24 26 2.6. nancals 35 62 67.9. Consumer Dscretonary 3 85 85.4.0 Consumer Staples 6 72 65. 0.9 Source: Aspect nancal Analyss Database, Mornngstar Inc. 39
Appendx : The Model n Detal We model the real economy and so ncorporate no markets for money or other assets. An exchange rate s defned prmarly as a soluton devce. Its value adjusts to satsfy a balance of payments condton, thereby brngng about changes n relatve domestc prces. Most often, however, the balance of payments condton s elmnated from the model by a closure adjustment, and the artfcal exchange rate fxed, so that all the adjustments to shocks are made by the home prces relatve to the bundle of mported products as the numerare. The balance of payments condton s stll met because t s mpled by the household s and the government s budget constrants. The artfcal exchange rate serves no valuable purpose as a product of model snce what matters for trade and relatve servce sector performance s the real exchange rate the common currency (n ths case common numerare ) average prce of home relatve to foregn products and servces. Demand elastctes or fnal demand the elastcty expresson s: n Pˆ Pˆ ( σ ) H (A.) ε = η + ( σ ) δ + ( η σ )( + ( n ) μ) where η s the elastcty of substtuton of fnal demand home varetes, δ s the home share n fnal demand for product, σ s the elastcty of substtuton of fnal demand for good between domestc and foregn countres, s the number of domestc frms n ndustry, P s the CES composte prces of home varetes, and s the CES ˆH composte of home and foregn fnal product prces n the domestc market, weghted by domestc consumpton shares. Ths expresson s derved n Appendx 2. or exports t s assumed that home frms face such competton n foregn markets that non-collusve prcng behavour s necesstated. The foregn demand elastcty takes the same form as (A.), except that the foregn conjectural varaton parameter, μ, s zero: n Pˆ X 40
n Pˆ Pˆ X ( σ ) X X X X X X (A.2) ε = η + ( σ ) θ + ( η σ )( + ( n ) μ ) where and X ( σ ) ˆ e X X P X = η + σ θ + η σ ˆ X n P X Pˆ ˆ e P X ( ) ( ) e X s the CES composte foregn currency prce of all exported varetes of product s the CES composte of exported and competng foregn fnal product prces n the foregn market, weghted by foregn consumpton shares. oregners dfferentate home exports from correspondng foregn products wth elastcty of substtuton varetes from one another wth elastcty of substtuton η X. or ntermedate demand the expresson s: N I I I I H I I (A.3) ε = s j η + γj + σ φj + ( η σ )( + ( n ) μ ) I j= I σ Pˆ ( ) n Pˆ X σ and home where s j I s the share of ndustry j n the total ntermedate demand for nput and ˆ I P s the CES composte of home and foregn ntermedate product prces n the domestc market, weghted by domestc ntermedate consumpton shares. The correspondng expresson for the elastcty of government demand s: n Pˆ Pˆ G ( σ ) G G G G H G G (A.4) ε = η + ( σ ) δ + ( η σ )( + ( n ) μ) Mark-ups: G We assume constant margnal cost olgopolstc frms n the dfferentated product markets. The assumpton of symmetry wthn each sector mples a common optmal unregulated markup for each frm, as n equaton () of the man text. Domestc prces of mported goods: These are: (A.5) where w M C * p ( + τ )( + τ ) p = e w p s the exogenous foregn currency prce of goods produced n the rest of the world, M C τ s the ad valorem tarff rate and τ s the ndrect consumpton tax n ndustry.
Domestc prces of home products: As n equaton () of the man text, these are marked up over average varable cost. To obtan the latter, recall that producton s Cobb-Douglas n varable factors and nputs, wth output elastctes α for captal, β k for factors k and γ j for nputs j and that the subaggregaton of mported and domestc nputs s CES. Unt varable costs are therefore calculated as: (A.6) K N γ α j βk ˆ I v = br wk P j k= j= where the scale coeffcent b s calbrated from the SAM, as are all the exponents n the equaton, and ˆ I P j s a CES composte of home and mported nput prces weghted by the domestc and mported shares specfc to consumng ndustry : ( ) I ( σ ) I j σ j I (A.7) ˆ I σ j * Pj = φj ( pj ) + ( φj) ( pj ) where φ j s the domestc share of nputs from ndustry j n use by ndustry. Then, domestc producer prces are smply hgher by the mark-up, m. p = mv,. Unt factor and nput demands: These are derved by solvng the frm s cost mnmsaton problem wth Cobb-Douglas producton n varable factors and nputs. It s assumed that frms are prce takers n both factor and nput markets. Therefore, the unt factor demands for captal and other factors are: K αv L βkv (A.8) u =, and uk = k,, r w k where k denotes non-captal factors whch are natural resources, and sklled and unsklled labour. The correspondng unt nput demands are Leontef nput-output coeffcents, except that ther values depend on product and nput prces. or home-produced and mported nputs from ndustry used n the product of ndustry j, respectvely they are: (A.9) A I σ * φjvj p ( φ ) * j vj p j = γj Aj = γj ˆ I ˆ I ˆ I ˆ I Pj Pj Pj Pj,, j Prces of home product exports n foregn markets: These are n foregn currency so they depend on the home producer prce, the exchange rate, the export subsdy rate X s I σ *M and the foregn mport tarff rate, τ : (A.0) p * M e pe ( +τ ) X = ( + s ) 42
Export demand: oregners dfferentate home exports from correspondng foregn products wth elastcty of substtuton σ (>0), and home varetes from one another wth elastcty of substtutonη X. X Ths gves the followng expresson for foregn demand for varety j of home products : (A.) X σ ˆ e θ E P p Hj = n ˆ ˆ ˆ P P P j X X e X X η, where θ s the calbrated reference share of the home export n total consumpton, E s a calbrated constant representng foregn expendture on exports from ndustry, and CES composte of the home export prce, foregn market, weghted by foregn consumpton shares. nal demand: e p, and the foregn product prce, ˆ X P w p, n the Home consumers dfferentate home products from correspondng foregn products wth elastcty of substtutonσ (>0) and home varetes from one another wth elastcty of substtuton η. They have Cobb-Douglas utlty n broad products, wth the result that expendture shares are constant across these groups. nal demand for varety j of home product group s therefore: s a (A.2) D δ ˆ a Y T Y P p H Hj = n ˆ ˆ ˆ P P PH Hj σ η where a s the calbrated reference expendture share of product group, δ s the correspondng share of home goods n fnal demand for product, Y s gross natonal product (GNP), T Y s total drect (ncome) tax, and the composte prce s: ( ) * ( σ ) σ H (A.3) σ = δ + ( δ ) where the home share s (A.4) Pˆ ( p ) ( p ) M Government demand: δ. The expresson for mports s correspondngly gven by: Y T p = * Y ( δ ) a ˆ ˆ P P The formulaton adopted s smlar to that for fnal demand by households. Government demand for home produced products and mports, respectvely, s gven by: σ 43
(A.5) G G G δ ˆ a G P p H Hj = n ˆ ˆ ˆ P P PH Hj G G σ G G η G G p = * * G G ( δ ) a ˆG ˆG P P where the composte prce of government purchases s: (A.6) ˆ G G G ( ) ( ) ( ) ( ) ( ) G σ σ * P = δ p + δ p σ G It s assumed that the government spends all t receves n tax revenues. That s, t mantans a balanced budget. The model s comparatve statc and t does not at present ncorporate prvate or publc savngs and nvestment. 54 Demand for nputs: Ths s derved from the nput-output coeffcents and gross ndustry output, Q. Demands for home-produced and mported varetes of the ntermedate good are: G G σ (A.7) N N * * = j j, = j j j= j= I AQ I AQ Tax revenue: The government rases tax revenue from both drect and ndrect taxaton. The revenue rased from each source s expressed below. Drect ncome tax revenue N ( ) (A.8) T = τ rk + π + τ w L + τ w L S, Y K U U U S S = where K denotes total captal stock n ndustry, π denotes total pure proft n ndustry, U denotes unsklled labour and S denotes sklled labour. Note that the tax rate on captal ncome s not generc. Ths enables the capture of tax polces that dscrmnate between sectors. Consumpton tax revenue (A.9) Import tarff revenue N N C C * C = τ + τ = = T pd p M 54 The mplct assumpton beng that the sectoral composton of nvestment spendng s the same as that of fnal demand. 44
(A.20) Export tax revenue N M * M = τ ( + ) = T M I p e w N = X (A.2) T = ( s ) p X, X where X s denotes the net power of the export subsdy rate. Total tax revenue s then smply a sum of the ndvdual components above. Economc profts or losses: Ths s revenue derved from mark-ups over unt varable costs, less total fxed costs. or sector t s: K L (A.22) π = ( p v ) Q n ( r f + w f ), S where n s the number of frms, f K s the fxed captal requrement per frm and f L s the fxed sklled labour requrement per frm n sector. Net proft n ndustry s therefore: (A.23) π N ( ) ( K L ) ( K S τ ) = p v Q n r f + w f Natonal ncome (GNP): Ths s the sum of payments to domestcally owned factors of producton wth the home share of any net profts or losses made, the net ncome from ndrect taxaton and the net nflow of unrequted transfers from abroad, B. Snce the model has no savng and nvestment, the captal account s formally closed. B could therefore be thought of as an nflow of ad, or, more precsely for Australa, as an exogenous net nflow on the captal account. (A.24) K N N K D B K D * Y = rkd + wklk + π + ( T TY) + + τk r( KT KD) + π k= KT = e K T = where T Y s revenue from drect (ncome) tax. GDP, on the other hand, s a measure of the ncome from producton n the domestc economy, so t excludes factor payments and other flows to and from abroad: (A.25) GDP = rk + w L + π + ( T T ) K T k k k= = N Total factor demands: The model has two captal market closures. In one (the long run closure ) physcal captal s perfectly moble abroad at the exogenous world nterest rate r. In the other (the short run closure ), physcal captal stocks are fxed n each ndustry and ndustry rates of return are Y 45
endogenous. Ether way, physcal captal s fully employed, wth total demand havng varable and fxed components: N K D K (A.26) KT = ( u Q + n f ) = where f K s the total fxed cost outlad by ndustry. Smlarly, the demand for sklled labour also ncludes a varable and fxed component. It s: N L D L (A.27) LS = ( usq + n f ) = nally, demand for all other varable factors (unsklled labour and mneral-energy resources) s: N (A.28) j ( j ) L L = u Q j = 2,..., = In the short run closure, employment of unsklled labour s endogenous, whle the real consumpton wage s exogenous, so that unsklled labour can be unemployed. Appendx 2: nal Demand Elastcty wth Prce Interacton Here the fnal demand elastcty s derved to llustrate the method by whch all the elastcty expressons of Appendx (A. A.4) are arrved at. rom (A.2) the demand equaton for domestc varety j of commodty s: (A2.) d δ ˆ a Y T Y P p H Hj = n ˆ ˆ ˆ P P PH Hj σ η where the composte prces are the average prce of generc product avalable on the home market from both home producton and mports: ( ) * ( σ ) σ H (A2.2) σ = δ + ( δ ) Pˆ ( p ) ( p ) and the average prce of home varetes of product 55 : (A2.3) n ˆ ( η ) PH = = ( phj ) j= n η Substtute (A2.2) and (A2.3) nto (A2.) and the full demand equaton can be re-wrtten as: 55 In equlbrum, because frms have dentcal technologes, these prces are equal, though ths s not perceved by frms n settng ther prces. 46
d Hj σ δ ˆ a Y T Y P p H Hj = ˆ ˆ n ˆ P P PH η δ a = n σ η σ η ( ) ( )( ˆ Y T ) ( ˆ Y P PH) ( phj) η σ ( n σ ) ( ) * η ( ) Y ( H ) ( ) η ( Hj ) j= n δ a σ = ( Y T ) δ p + ( δ ) p p n Dfferentatng wth respect to p Hj gves: d p Hj Hj ( p Hj ) δ a = ˆ n + η j= n n η σ 2 n η η η ( ) ( ) ( ) ( ) ( )( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) η σ * σ η σ Y TY δ ph δ p σ δ ph phj η phj PH ( PHj ) η σ n δ a η σ η η p H η p 2 p H Hj + ( Y TY ) p η p ˆ + η p 2 + + η p P n η j= n n p Hj n p Hj n p Hj δ a ( η ) + ( Y TY ) ( η )( phj) n { }( ) ( ˆ ) σ η σ P ( ˆ PH ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( σ ) Hj H H Hj ( phj ) η η η η Notng that: phj μ j h = phh j = h and notng further that phj = phh j h, because of the assumpton that frms wthn an ndustry behave symmetrcally, the expresson can be wrtten as: 2( ) 2( ) dhj δ 2 a σ η σ ( )( )( ˆ η = Y T ) ( ˆ Y σ P δ PH) ( phj) phj n n ( )( )( ) ( 2 ) δ 2 a η + Y T ˆ ( ) ( ) ˆ Y η σ PH phj + n μ P n n ( )( )( ) ( ) ( ) ( ˆ ) δ a σ η σ η + Y T ( ˆ Y η phj P PH ) n Ths further smplfes to: (A2.4) ( )( ) ( ) η σ σ ( )( ) ( ) ( ) ( ) ( ) ( ) ( )( ) ( ) ( ) ( )( ) ( ) dhj δ a σ η σ ˆ η ˆ σ η σ ˆ η η ˆ η ˆ = Y TY P PH phj σ P δ PH phj + η σ PH ( phj ) ( + ( n ) μ ) + ( η )( phj ) phj n n n 47
So that the elastcty of fnal demand s: d p Hj Hj p d Hj Hj δ a = + + + n n n phj. ( )( ) ( ˆ ) δ a σ η σ η Y T ( ˆ Y P PH ) ( phj ) n ( )( ) ( ) ( ) ( ) ( )( ) ( ) ( ) ( )( ) ( σ ˆ ˆ. ˆ ˆ ˆ ) η σ η σ η σ η η η Y TY P PH phj σ P δ PH phj η σ PH ( phj ) ( ( n ) μ ) ( η )( phj ) ( ) ( ) On the symmetry assumpton ths smplfes to: p Pˆ ˆ ˆ n PH P η ( σ ) Hj H (A2.5) ε = η + ( σ ) δ + ( η σ )( + ( n ) μ ) Appendx 3: Calbratng Olgopoly Parameters No complete set of data on the structure and conduct of Australa s agrcultural, manufacturng and mnng sectors s publcly avalable. Some relevant data s avalable pecemeal, for ndvdual sectors or ndustres, though ths s occasonally at too fne a level of aggregaton for an llustratve economy-wde study such as ths. It has therefore been necessary to extrapolate patterns to some sectors and to make crude assumptons about others. To clarfy our assumptons, ths appendx offers an expanson of the summary gven n Secton 3 of the text. rst, estmates of pure (over-market) profts are requred as shares of revenue n each ndustry. Ths s needed to fnalse the flow database but also to calbrate ndustry compettve structure. or these we have resorted to data on the proftablty of lsted publc frms from the Mornngstar Aspect-Huntley nancal Analyss Database. 56 Accountng proft rates net of deprecaton are compared wth the prme borrowng rate avalable to corporate borrowers n the perod 997-2007 to obtan measures of pure profts. The data on ndustral borrowng rates used n ths comparson s from the RBA (www.rba.gov.au). The resultng paths of pure profts as a proporton of turnover are shown n Table A3.. Ths set of approxmatons s obvously precarous. It consders only lsted frms, thus gnorng most of the farmng communty n agrculture and the small and famly busnesses n the servces sectors, not to menton large prvate frms n all sectors and government-owned 56 The database s formally the Aspect nancal Analyss Database. It s suppled by Aspect-Huntley, and the copyrght s held by Huntleys' Investment Informaton Pty Ltd (HII) (a wholly owned subsdary of Mornngstar, Inc): http://www.aspectfnancal.com.au/af/fnhome?xtm-lcensee=fnanalyss.for. 48
servce frms. Moreover, the concordance wth our sectoral breakdown s necessarly very crude, snce beyond ther ten sector classfcaton, Mornngstar s data gves only the names of lsted frms and not ther actvty. Nonetheless, t offers the only clear ndcaton of frm numbers, szes and performance across the whole economy. The results tend to show a declnng trend n pure proft rates between 997 and 2007. Ths mght not reflect a trend n pure proftablty, however, but merely short term and possbly unsustanable rse n lsted asset values and an assocated declne n P/E ratos. or ths reason, and because we wsh that the numbers used should be of sustaned relevance, we have taken perod averages and appled them to our model database to determne the ntal level of over-market profts n each sector. or estmates of strategcally nteractng frm numbers n each ndustry and ther correspondng conjectural varatons parameters, we examned ndustry structure n each sector, focussng on the numbers of frms wth more than a tenth of market revenue. The results of ths analyss are dsplayed n Table A3.2. In the end the values for the effectve number of frms and the conjectural varatons parameter n each sector are judgemental, takng nto account the numbers of mssng prvate frms and farms and the extent of regulatory survellance lmtng the full explotaton of olgopoly power. 49
Table A3.: Estmated Pure Profts as % of Total Turnover a % 997 998 999 2000 200 2002 2003 2004 2005 2006 2007 Adjusted perod average b Agrculture 0.4 0.4 0.7 0.4 0.2 0.6.0 0.9-0.3 0.3-0.5 0.4 Manufacturng 0..0 0.4 -.0-2. -0.9.9.2.3-3.8-0.7-0.2 Mnng & energy -. -0. -.3 2.9.7 0.4-0.2.4 2.8 5.3-0.6.0 Electrcty -28.9-3. 4.4-3.9.7.2-0.2-6.9-7.5-5.4-9.6-2.9 Water -28.9-3. 4.4-3.9.7.2-0.2-6.9-7.5-5.4-9.6-2.9 Gas -28.9-3. 4.4-3.9.7.2-0.2-6.9-7.5-5.4-9.6-2.9 Telecoms.3 7.0-4.7-2.8 6.8 6.9 8.7 4.3 6.0 7.7.5 4.8 nance 8.3 7.3 66.6-4.8 3.5 9.0 0.4 4.2 4.4-0.8 0. 9. Transport -.0-2.9-3.3-4.3-2.8-0.7 5.2 3.3-0.7-4.7 0.9 -.0 Other servces -.9-2.6-2.8-4.6-3.4 -.3 3.5.4-2.0-5.4 -.4 -.9 a These are pure proft rates derved by subtractng from the net (of deprecaton) rate of return on equty the prme lendng rate (the one year offcal borrowng rate plus 2%) They are then crudely concorded from the Mornngstar classfcaton (consumer staples, ndustrals, nformaton technology, energy, materals, utltes,.telecommuncaton servces, fnancals, consumer dscretonary and health care) to that n the table. b Some outlyng peaks (ncludng for fnancals n 999) and troughs (ncludng for nformaton technology n 2000) are excluded. Source: http://www.aspectfnancal.com.au/af/fnhome?xtm-lcensee=fnanalyss.for. 50
Table A3.2: Estmated Market Structure a No lsted frms Lsted frms >0% Share of frms >0% Effectve no of frms b Conjectural varatons c Agrculture 72 2 70 50 0. Manufacturng 300 3 34 20 0.2 Mnng & energy 378 2 43 0 0.3 Electrcty 9 3 64 6 0.4 Water 6 5 90 6 0.2 Gas 2 00 2 0.5 Telecoms 28 4 97 4 0.6 nance 54 4 64 0 0.5 Transport 40 3 48 0 0.5 Other servces 264 45 00 0.2 a These are crudely concorded from the Mornngstar classfcaton (consumer staples, ndustrals, nformaton technology, energy, materals, utltes,.telecommuncaton servces, fnancals, consumer dscretonary and health care) to that n the table. b These results are judgemental, based on the data n the frst three columns. rm numbers exceedng 00 have neglgble effect on prcng. It s borne n mnd that large numbers of farms and prvate frms are omtted from the data. c the conjectural varatons parameter ranges between zero (non-collusve olgopoly) and unty (cartel). The numbers chosen reflect ndustry concentraton and the extent of exstng regulatory survellance. Source: http://www.aspectfnancal.com.au/af/fnhome?xtm-lcensee=fnanalyss.for. 5