Measuring the Effects of Exchange Rate Changes on Investment in Australian Manufacturing Industry

Transcription

1 Measuring he Effecs of Exchange Rae Changes on Invesmen in Ausralian Manufacuring Indusry Auhor Swif, Robyn Published 2006 Journal Tile The Economic Record DOI hps://doi.org/ /j x Copyrigh Saemen 2006 Blackwell Publishing. This is he auhor-manuscrip version of he paper. Reproduced in accordance wih he copyrigh policy of he publisher. The definiive version is available a Downloaded from hp://hdl.handle.ne/10072/16651 Link o published version hp:// Griffih Research Online hps://research-reposiory.griffih.edu.au

2 Measuring he Effecs of Exchange Rae Changes on Invesmen in Ausralian Manufacuring Indusry Robyn Swif Economics and Business Saisics Deparmen of Accouning, Finance and Economics Griffih Universiy Nahan Campus Brisbane Qld 4111 Ausralia Ph: (+61 7) Fax: (+61 7) This research is par of a projec suppored by a Griffih Universiy Research Developmen Gran (GURD)

3 Measuring he Effecs of Exchange Rae Changes on Invesmen in Ausralian Manufacuring Indusry Absrac This paper uses an empirical framework derived from an opimising adjusmen-cos model of invesmen o examine he relaionship beween exchange raes and invesmen in Ausralian manufacuring indusry beween 1988 and The resuls show ha he response of invesmen o exchange rae changes varies wih he exernal exposure of indusry, posiively wih expor share of sales and negaively wih he share of impored inpus ino producion. Indusry compeiive srucure, represened by markup of price over cos, ineracs wih exernal exposure so ha lower levels of markup increase he responsiveness of invesmen o exchange rae changes. For Ausralian manufacuring indusry, a 10% real appreciaion of he Ausralian dollar (AUD) leads o an average 8.0% decrease in oal invesmen hrough he expor share channel over he period, and an average 3.8% increase in oal invesmen hrough he impored inpu share channel, wih mos of he response occurring hrough invesmen in Equipmen, Plan and Machinery. 1. Inroducion There have been subsanial changes in he value of he Ausralian dollar (AUD) over he las decade bu we sill have lile deailed informaion abou he implicaions of hese movemens for Ausralian indusry. There is now a considerable inernaional lieraure on he facors ha influence he iniial sage of he adjusmen process, he change in relaive prices ha follows an exchange rae change, usually referred o as exchange rae passhrough. 1 However, fewer sudies have exended his analysis o invesigae he longer-erm 1 See Goldberg and Kneer (1997) for a comprehensive review of he lieraure on exchange rae pass-hrough and relaed issues.

4 effecs of exchange raes on invesmen and employmen, and only for a limied number of counries (for example, Campa and Goldberg 1995, 1999; Nucci and Pozzolo 2001). This paper uses quarerly daa from 1988 o 2001 o esimae he size and deerminans of he effecs of exchange rae changes on invesmen in Ausralian manufacuring indusry. The opimal level of invesmen for a firm is an increasing funcion of he curren and expeced value of fuure profis. Exchange rae changes influence profis no only hrough changes in he prices of domesically produced goods sold in expor markes bu also hrough changes in he prices of impored goods used as inpus o producion and in he prices of impors ha compee wih domesically produced goods. A firm wih a larger share of sales in expor markes, for example, would be expeced o reduce invesmen in response o an appreciaion of he domesic currency. Conversely, a firm ha is highly dependen on he use of impored inpus ino producion would face a larger drop in variable coss and would consequenly be expeced o increase invesmen following an appreciaion. The exchange rae pass-hrough lieraure has also highlighed he imporance of he compeiive srucure of indusry in deermining he size of he response o exchange rae changes (see, for example, Dornbusch 1987). Firms in more compeiive indusries wih lower price-over-cos markups have less abiliy o absorb exchange rae changes in heir markups so ha he change in prices for hese firms following an exchange rae change will be larger. Lower levels of indusry markups would herefore be expeced o inerac wih, and amplify, he effecs of exernal exposure on invesmen. The few empirical sudies available have generally confirmed he conclusions of he heoreical models. In a cross-counry survey, Campa and Goldberg (1999) esimaed ha a ne increase of around 1-2% in invesmen in oal manufacuring in he USA would follow a 10% appreciaion of he USD, wih a wide variaion expeced in indusry response, ranging from an increase of 12% o a decrease of 8.4%. The auhors showed ha hese resuls were

5 due o he increasing imporance of impored inpus ino manufacuring in he USA. Japanese indusry generally showed a lower level of response, bu wih an overall increase in invesmen also expeced following appreciaion, and a similarly wide range of indusry responses from an increase of 10% o a decrease of 8.3%. Campa and Goldberg (1999) were unable o find any saisical significance in he exchange rae coefficiens for he UK and Canada, which hey found surprising given he size and exen of he exernal orienaion of Canadian manufacuring. However, Nucci and Pozzolo (2001) repored ha he resuls of heir esimaions of a panel daa se of abou 1000 Ialy manufacuring firms also srongly suppored he implicaions of he heoreical model, wih he overall effecs of swings in he exchange rae on a firm s invesmen depending on which of he opposing expor and impored inpu channels prevailed. The res of he paper is organised as follows. Secion 2 oulines a heoreical framework ha explicily incorporaes boh changes in exernal exposure and indusry markups as deerminans of exchange rae effecs on profis and invesmen. The empirical model described in Secion 3 is direcly derived from his framework, and is esimaed for oal invesmen in Ausralian manufacuring indusry and for wo invesmen sub-classes, Equipmen, Plan and Machinery, and Buildings and Srucures. The resuls presened in Secion 4 confirm he effecs of exernal exposure and markups expeced from he heoreical model. They show ha a 10% real appreciaion of he Ausralian dollar (AUD) leads o a ne decrease of 4.2% in oal invesmen on average over he period, he combined effec of an 8.0% decrease hrough he expor share channel, and a 3.8% increase hrough he impored inpu share channel. Almos all of he invesmen response occurs hrough changes in invesmen in Equipmen, Plan and Machinery. Secion 5 draws some final conclusions.

6 2. Theoreical Framework A firm is assumed o choose he opimal level of invesmen each period (I ) o maximise is marke value (V ), which is equivalen o he discouned sum of expeced fuure cash flows. K, he capial sock a he sar of period, is assumed o be quasi-fixed and subjec o a cos of insalling new capial c(i ) ha is increasing and convex, as in a sandard adjusmen cos model (Chirinko 1993). K is also subjec o a sandard capial accumulaion equaion: K (1 δ ) K I (1) where δ is he rae of capial depreciaion. The profi funcion of he firm exposed o exchange rae changes is represened by Π(K, e ) where e is he real exchange rae in unis of domesic currency per uni of foreign currency. Following he model developed by Campa and Goldberg (1999), he maximised value of he firm a ime is expressed as: τ V ( K, e ) max E β [ Π( K { Iτ } τ τ 0 + τ, e + τ ) c( I + τ ) I + τ ] Ω (2) where he cash flow for every period is given by he profi funcion less he overall cos of invesmen, β is he discoun rae, and E[ Ω ] is he expecaions operaor condiional on all informaion available a ime. The firs order condiion derived by maximising he value of he firm in equaion (2) subjec o equaion (1) represens he opimal decision-making rule for invesmen in period. Tha is, he firm should equae expeced marginal profiabiliy, or he discouned sream of expeced fuure profis from an addiional uni of invesmen, o he coss of ha addiional invesmen: E β Π( K, e (3) c I τ + τ + τ ) + K Ω ( ) [ (1 δ ] 1 τ 1 + τ I

7 The model assumes only a cos of adjusmen funcion ha is increasing and convex. However, for ease of exposiion, a sandard quadraic cos funcion wih parameers γ and µ is used: γ 2 2 c ( I ) ( I µ ) (4) Differeniaing equaion (4), subsiuing ino equaion (3) and rearranging gives an expression for I : I 1 1 Π + ( K +, e + ) ( µ τ τ τ ) E [ β (1 δ ] Ω γ γ τ 1 K (5) + τ The firm is assumed o observe he exchange rae each period (e ) and choose oupu for foreign and domesic markes and quaniies of domesic and impored inpus in order o maximise per-period profi: Π ( K, e ) max p( q, e ) q + e p ( q, e ) q w L e w L (6) q, q, L, LT subjec o q + q f ( K, L, L where q and q are he quaniies supplied by he firm o domesic and foreign markes respecively, p ( q, e ) and p ( q, e ) are he demand funcions faced by he firm, w L and e w L are spending on domesically produced and impored inpus, L and L, respecively, and f ( K, L, L ) is homogenous of degree one. The firs order condiions for equaion (6) are combined and he resuling gross profi funcion differeniaed o give an expression for he marginal profiabiliy of capial: Π K, e ) 1 [ MKUP K K ( 1 1 pq + MKUP e p q ( w L + e w L )] (7) where MKUP and MKUP are he firm s price-over-cos markups in he domesic and foreign markes respecively. These can be represened as:

8 MKUP p MC 1 1 (1 + η ) (8) and MKUP e p MC 1 (1 + η 1 ) (9) where MC is marginal cos and η and and foreign markes respecively. η are price elasiciies of demand in he domesic The model assumes for simpliciy ha he only source of uncerainy is he exchange rae, and ha all exchange rae changes are assumed o be permanen. This implies ha he level of he exchange rae in all fuure periods is expeced o be equal o oday s exchange rae, and consequenly he expeced marginal profiabiliy in equaions (3) and (5) is also equal o oday s marginal profiabiliy as expressed in equaion (7). By using equaion (7) o differeniae equaion (5) wih respec o he exchange rae and rearranging, Campa and Goldberg (1999) derive an expression for he invesmen response o exchange rae changes: I e A AMKUP X [( η p, e ηmkup, e )(1 ) + (1 + η η ) p e MKUP e (1 + η,, w, e X α ) de ] e (10) where β TR A γ [ 1 β (1 δ )] K and η p, e and p, e η are he exchange rae pass-hrough elasiciies o domesic and foreign prices respecively, η MKUP, e and MKUP, e η are he exchange rae elasiciies of markups in domesic and foreign markes respecively, TR is oal revenues, X is he share of TR from sales in foreign markes, α is he share of impored inpus in oal producion coss, η is he exchange rae elasiciy of impored inpu coss and AMKUP is he average markup across all markes. w, e

9 The hree main erms inside he square brackes in Equaion (10) idenify he hree channels hrough which exchange rae changes can influence invesmen. Effecs on he profiabiliy of domesic sales (1-X ) will depend on he exchange rae elasiciies of price and markups in he domesic marke. Effecs on profiabiliy of expor sales (X ) will depend on he exchange rae elasiciies of price and markups in foreign markes plus a direc valuaion effec on revenue from expor sales. These effecs on he revenue side will be offse by opposiely signed changes in oal producion coss, which are dependen on he exchange rae elasiciy of impored inpu coss. The price and markup elasiciies are all dependen on he naure of compeiion in he relevan markes (Dornbusch 1987). For perfecly compeiive firms, domesic prices are equal o marginal coss and markups are accordingly zero, so domesic prices and markups are unaffeced by exchange rae changes ( η η 0 ). Domesic prices of expors p, e MKUP, e are also unchanged so he foreign currency price of expors mus rise fully o reflec currency appreciaion, ( η 1), leading o a corresponding reducion in quaniies sold. In p, e addiion, smaller average markups will amplify he effecs on invesmen in compeiive A indusries from all hree channels, hrough ( AMKUP ). Conversely, firms in imperfecly compeiive markes can decrease markups on expored goods following appreciaion o mainain consan foreign currency prices ( η 0), reducing he impac on quaniies sold. The exen o which foreign currency p, e prices change in response o a currency movemen, and he resuling effecs on invesmen, will be direcly relaed o he degree of compeiion in expor markes and inversely relaed o he size of he markups in he indusry. In all indusries, he larger is X, he share of sales in expor markes, he larger will be he effecs from changes in foreign currency prices. Similarly on he impored inpu side,

10 boh more compeiive markes for inpus (smaller markups and larger η w, e ) and a greaer proporion of impored inpus in producion coss (larger α ) will increase he counervailing effecs of impored inpu coss on profiabiliy and invesmen. 3. Empirical Esimaion and Daa The empirical model o be esimaed is derived from equaion (10) as: I β + β ( AMKUP 0 + β α ( AMKUP ) 1 1 ) 1 1 e + β X 2 1 e + β sales 4 ( AMKUP β r 5 ) 1 e + β I ε (11) where I is invesmen in privae new capial expendiure a consan prices, and e is he real rade-weighed exchange rae in unis of domesic currency per uni of foreign currency. From equaion (10), he coefficien β 1 on he firs exchange rae erm is a funcion of ( p, e ηmkup, e) η ) in he domesic marke. The coefficien β 2 on he exchange rae ineraced wih expor share (X ) is a funcion of he valuaion and marke effecs on he expor side ( 1 η η η p, e MKUP, e p, e + ηmkup, e + ). Similarly, he coefficien β 3 on he final exchange rae erm ineraced wih impored inpu share (α ) is a funcion of impored inpu elasiciy ( 1 η w, e + ). All hree exchange rae erms inerac wih average markup (AMKUP ) as in equaion (10), incorporaing he influence of compeiive srucure on exchange rae effecs. The discoun rae included in he erm in A in equaion (10) is represened by he real ineres rae (r ). The esimaion also includes a variable for he growh rae of real oal sales in he indusry ( sales ) o conrol for oher indusry-specific facors ha may have influenced he growh rae of invesmen over he period. The lagged values of he ineracion erms (expor and impored inpu shares and average markup) and he real ineres rae are used o avoid possible correlaion wih he curren exchange rae, and a lag on oal sales o avoid possible correlaion wih curren invesmen. The final erm in lagged invesmen (I -1 ) is

11 included o represen he adjusmen and implemenaion lags generally expeced in invesmen projecs (Driver and Dowrick 1997, Nucci and Pozzolo 2001). The symbol indicaes firs differences in naural logarihms for all variables excep for he real ineres rae, which is differenced in level form. The inclusion of he lagged dependen variable in equaion (11) raises he possibiliy of serial correlaion and subsequen bias in he esimaes. For his reason, insrumenal variables esimaion (2SLS) was employed, wih addiional lags of invesmen and sales included as insrumens. The esimaes were correced for any remaining serial correlaion using he Newey-Wes echnique (Newey and Wes 1987). Equaion (11) was esimaed for oal invesmen in Ausralian Manufacuring indusry (ANZSIC Division C) and for he wo invesmen sub-classes, Equipmen, Plan and Machinery, and Buildings and Srucures, using quarerly daa for he period 1988:Q1 o 2001:Q4. Deails of daa sources are available in he appendix. Expor share (X ) is calculaed as he raio of he value of expors o he value of oal sales of goods and services. Impored inpu share (α ) is consruced from secoral inpu-oupu daa combined wih impor value shares for each inpu secor o give he oal value of impored inpus used in manufacuring indusry. The share of impored inpus in producion is hen derived as he raio of he value of impored inpus o he oal cos of producion, including boh maerials and labour inpus. 2 Marginal coss for he direc calculaion of price-over-cos markups are no observable. Domowiz e al. (1986) sugges a mehod for compuing a measure of profi margin per uni price: Value of Sales + Invenories Payroll Cos of Maerials MK DHP (12) Value of Sales + Invenories 2 Furher deails on he consrucion of his variable and a discussion of changes in measures of exernal exposure for Ausralian manufacuring indusries over he period are provided in Swif (2005).

12 This formula is used o derive a measure for he average markup (AMKUP) as average priceover-average cos o represen marke power in he indusry: 1 AMKUP 1 MK DHP Value of Sales + Invenories Payroll + Cos of Maerials (13) This mehod is no inended o give a definiive measure of indusry markups, bu i does allow he consisen comparison of changes in profi margins over ime as exchange raes change. 4. Resuls The resuls of he esimaion of equaion (11) for oal invesmen and he wo invesmen sub-classes are given in Table 1. For oal invesmen, he exchange rae variables ineraced wih expor shares and impored inpu shares are boh significan and of he expeced sign, indicaing ha an appreciaion (a decrease in e ) leads o a decrease in oal invesmen hrough he expor price channel and an increase hrough he impored inpu prices channel. The resuls for he ineraced erms in he regression for equipmen, plan and machinery are very similar bu wih slighly smaller coefficiens. Invesmen in equipmen, plan and machinery consiued around 85% of oal invesmen on average over he period, consisen wih he resuls obained. The regression for invesmen in buildings and srucures has markedly poorer fi ( R 0.22) and only he lagged dependen variable is significan. Driver and Dowrick (1997) also found less explanaory power in he equaions for buildings and srucures in heir esimaions of he effecs of invesmen inenions on invesmen in Ausralian manufacuring indusry. Invesmen in buildings and srucures is likely o involve much larger coss, suggesing longer lead imes in planning and implemenaion ha may induce lumpiness in 2

13 invesmen spending. The wedge hus creaed beween he benefis and coss embodied in an opimal invesmen policy may lead o parameer and model insabiliy by weakening he link beween invesmen spending and shifs in economic deerminans such as exchange raes (Chirinko 1993). The negaive and significan coefficiens on he one-period lag of invesmen in all hree esimaions, also found by Driver and Dowrick (1997), suppor he idea ha invesmen spending generally has some degree of lumpiness, as higher levels of invesmen spending las period will be followed by lower levels of spending his period whaever he exernal economic condiions. Table 1: Invesmen in Ausralian Manufacuring Indusry (ANZSIC Div. C) 1988:Q1-2001:Q4 Dependen Variable: Invesmen Equipmen, Plan Buildings and Variables Toal Consan (0.058) and Machinery (0.045) Srucures (0.498) (AMKUP -1 ) -1 e (0.932) X -1 (AMKUP -1 ) -1 e (1.962) α -1 (AMKUP -1 ) -1 e (1.914) sales (0.001) r (0.102) I (0.089) (0.658) (1.257) (1.332) (0.001) (0.080) (0.089) (7.736) (25.351) (19.138) (0.010) (0.856) (0.103) 2 R Noe: Each equaion is esimaed using 2SLS and correced for fourh-order serial correlaion. and indicae significance a 5% and 10% levels respecively. Sandard errors are repored in brackes below parameer esimaes.

14 The coefficien on he real ineres rae for he firs wo regressions is significan, bu small and of he wrong sign. To check ha his did no indicae he need o allow for a longer ime delay on he ineres rae effec, all hree equaions were re-esimaed wih lags up o five quarers on he real ineres rae variable. Insignifican lags were eliminaed using he general-o-specific mehodology, bu only he firs lag remained significan in all hree regressions. Ineresingly, Campa and Goldberg (1995) found similar small bu posiive and significan coefficiens in heir esimaions of he effecs of exchange rae changes on invesmen in manufacuring indusry in he Unied Saes. Campa and Goldberg (1995) offered no explanaion for heir findings, bu he relaive size of he coefficiens for boh Ausralia and he USA suppor he general conclusion of neoclassical models of invesmen ha effecs on oupu or sales, as generaed here by exchange rae changes, have he dominan effec on invesmen wih user cos having a minor role (Chirinko 1993). The implicaions for Ausralian indusry of he invesmen effecs of exchange rae changes can be shown by combining he esimaed coefficiens on he ineraced exchange rae erms in Table 1 wih indusry-specific daa on expor share, impored inpu share and markups. Table 2 gives he responses expeced for oal manufacuring indusry and he 2- digi ANZSIC manufacuring sub-divisions following a 10% real appreciaion of he AUD. Ideally, he exchange rae coefficiens would also be allowed o vary beween sub-divisions, bu he figures calculaed here do show he variaion ha can be expeced in effecs beween indusries as a resul of differences in markup and he degree of exernal exposure. A 10% real appreciaion leads o an overall ne decrease of around 4.2% in invesmen for Ausralian manufacuring as a whole, bu effecs vary from a 1.2% ne increase in invesmen for prining, publishing and recording media o a 15% decrease for meal producs. These resuls conras wih hose repored by Campa and Goldberg (1999) in heir cross-

15 counry survey. They esimaed a ne increase in invesmen in oal manufacuring in boh he USA and Japan following a 10% appreciaion of he domesic currency, wih a wider variaion in indusry response han ha found here. Table 2: Esimaed effecs of a 10% real appreciaion of he AUD on invesmen in Ausralian Manufacuring Indusries Invesmen in Toal Invesmen Equipmen, Plan and Machinery Impored Impored Indusry (ANZSIC code) Expor channel inpu channel Expor channel inpu channel Div. C: Toal Manufacuring - 8.0% 3.8% -7.3% 3.1% Sub-Divisions: 21 Food, Beverages and Tobacco - 8.9% % 1.2% Manufacuring 22 Texile, Clohing, Foowear and - 8.3% 4.4% -7.5% 3.6% Leaher Manufacuring 23 Wood and Paper Produc - 2.7% 2.7% -2.5% 2.2% Manufacuring 24 Prining, Publicaion and Recording - 1.0% 2.2% -0.9% 1.8% Media 25 Peroleum, Coal and Chemical - 5.1% 5.0% -4.6% 4.0% Manufacuring 26 Non-Meallic Mineral Produc - 1.0% 1.5% -0.9% 1.2% Manufacuring 27 Meal Produc Manufacuring % 3.3% -15.1% 2.6% 28 Machinery and Equipmen - 8.9% 7.0% -8.0% 5.6% Manufacuring 29 Oher Manufacuring - 4.8% 3.6% -4.3% 2.9% Noe: The esimaes are calculaed using period averages of expor share, impored inpu shares and markups for each indusry, combined wih he relevan coefficiens from Table 1. However, he four Ausralian indusries showing he larges effecs in Table 2 (21, 22, 27 and 28) are hose wih he larges expor shares, averaging over 20% expor share over he period, and in he case of meal producs, over 40%. They are also among he larges subdivisions overall, ogeher averaging more han 60% of he oal sales of he manufacuring indusries. These resuls highligh he coninued imporance of he expor channel for

16 exchange rae effecs for Ausralia. They also emphasise he imporance of counry and indusry-specific influences in deermining he direcion and size of he response o exchange rae changes. The remaining indusries, more han half he indusry sub-divisions (23, 24, 25, 26 and 29), experience ne changes of around 1% or less, indicaing ha even quie subsanial reducions in invesmen hrough he expor price channel can be counerbalanced by gains hrough he prices of impored inpus, as for example in peroleum, coal and chemical producs. 5. Conclusion This paper provides quaniaive measures of he size and direcion of he effecs of exchange rae changes on invesmen in Ausralian manufacuring indusry beween 1988 and The empirical resuls confirm he conclusions of he heoreical model ha he magniude of he effecs varies wih he exernal exposure of indusry, posiively wih expor share of sales and negaively wih he share of impored inpus ino producion. For individual manufacuring subdivisions, he expeced response o a 10% appreciaion of he AUD varies from a 1.2% ne increase in oal invesmen o a 15% ne decrease. For he manufacuring division as a whole, he decrease of around 4.2% is a resul of an 8% decrease in invesmen hrough he expor price channel and a 3.8% increase hrough he impored inpu price channel. Almos all of he invesmen response occurs hrough invesmen in equipmen, plan and machinery. For Ausralia, he esimaes sugges ha expor share remains he mos imporan channel for exchange rae effecs overall, alhough here are significan differences beween indusries. The resuls found here demonsrae he need for furher research o address he

17 deerminans of differences in exchange rae effecs boh beween indusries and beween counries. Appendix 1: Daa Invesmen (I ): Daa on oal invesmen, and invesmen in equipmen, plan and machinery, and buildings and srucures, a curren prices are aken from ABS, Privae New Capial Expendiure and Expeced Expendiure, Ausralia, Ca. no They are deflaed by he chain price indices for oal privae gross fixed capial formaion from ABS, Ausralian Naional Accouns: Naional Income, Expendiure and Produc, Ca no Exchange rae (e ): Daa on he real rade-weighed index are aken from Real Exchange Rae Indices Quarerly on he RBA websie ( Ineres rae (r ): The real ineres rae is calculaed as he quarerly average of he ineres rae on wo-year Treasury bonds from RBA, Ineres Raes and Yields: Money Marke and Commonwealh Governmen Securiies, Occasional Paper no. 10 (updaed), minus he expeced change in manufacuring selling prices from ABS, Ausralian Business Expecaions, Ca. no Expor share (X ) and Sales (sales ): The value of expors is aken from ABS, Inernaional Trade in Goods and Services, Ausralia, Ca. no Daa on sales of goods and services are aken from ABS, Business Indicaors, Ausralia, Ca. no Impored Inpu Shares (α i ): Daa on he inpu-oupu srucure of producion are aken from he ABS, Ausralian Naional Accouns: Inpu-Oupu Tables, Ca. no for 1996/97. Impor shares for each inpu indusry are calculaed as he raio of he value of impors from ABS, Inernaional Trade in Goods and Services, Ausralia, Ca. no , o he value of sales of goods and services for each indusry from ABS, Summary of Indusry Performance, Ausralia Ca. no Labour coss are aken from ABS, Wage and Salary

18 Earners, Ausralia, Ca. no , and deflaed by he price index for maerials used in manufacuring from ABS, Producer Price Index, Ca. no Average Markup (AMKUP ): Daa on company gross operaing profis, sales and invenories are aken from ABS, Business Indicaors, Ausralia, Ca. no References Campa, J. and Goldberg, L. S. 1995, 'Invesmen in Manufacuring, Exchange Raes and Exernal Exposure', Journal of Inernaional Economics, vol. 38, no. 3-4, pp Campa, J. M. and Goldberg, L. S. 1999, 'Invesmen, Pass-Through and Exchange Raes: A Cross-Counry Comparison', Inernaional Economic Review, vol. 40, no. 2, pp Chirinko, R. S. 1993, 'Business Fixed Invesmen Spending: Modeling Sraegies, Empirical Resuls, and Policy Implicaions', Journal of Economic Lieraure, vol. XXXI (December), pp Domowiz, I., Hubbard, R. G. and Peersen, B. C. 1986, 'Business Cycles and he Relaionship beween Concenraion and Price-Cos Margins', Rand Journal of Economics, vol. 17, no. 1, pp Dornbusch, R. 1987, 'Exchange Raes and Prices', American Economic Review, vol. 77, no. 1, pp Driver, C. and Dowrick, S. 1997, 'Invesmen Inenions as an Indicaor of Acual Invesmen in Ausralian Manufacuring', Ausralian Economic Papers, vol. 36, no. 68, pp Goldberg, P. K. and Kneer, M. M. 1997, 'Goods Prices and Exchange Raes: Wha Have We Learned?', Journal of Economic Lieraure, vol. 35, no. 3, pp

19 Newey, W. and Wes, K. 1987, 'A Simple Posiive-Definie Heeroskedasiciy and Auocorrelaion Consisen Covariance Marix', Economerica, vol. 55, no. 3, pp Nucci, F. and Pozzolo, A. F. 2001, 'Invesmen and he exchange rae: An analysis wih firmlevel panel daa', European Economic Review, vol. 45, no. 2, pp Swif, R. F. 2005, 'The Changing Exernal Exposure of Ausralian Indusry', Ausralian Economic Review, forhcoming.