Government spending, location and the exchange rate

Transcription

1 Tezukayama RIEB Discussion Paper Series No. overnmen spending, locaion and he exchange rae Waaru Johdo Faculy of Economics, Tezukayama Universiy Ocober 5 Tezukayama Universiy Research Insiue for Economics and Business 7-- Tezukayama, Nara 63-85, Japan

2 overnmen spending, locaion and he exchange rae Waaru Johdo * Tezukayama Universiy Ocober 5 Absrac In his paper, we exend a new open economy macroeconomics model o a hree-counry framework including relocaion of firms and explore he effecs of governmen spending in each counry on relaive consumpions and he exchange raes. From his analysis, i is found ha a governmen spending rise in a counry always depreciaes is currency, causes firms locaed abroad o relocae o he counry and consequenly decreases he counry s relaive consumpion. In conras, he governmen spending can be beneficial for neighboring counries in spie of he ouflows of firms. JEL classificaion codes: E; E6; F; F3; F4 * Waaru Johdo, Deparmen of Economics,Tezukayama Universiy, 7--, Tezukayama, Nara 63-85, Japan. Tel.: ; fax: address: johdo@ezukayama-u.ac.jp

3 . Inroducion In he new open economy macroeconomics (NOEM) lieraure, he heoreical relaionship beween governmen spending shocks and aggregae economic aciviy has been sudied exensively (see, for example, Obsfeld and Rogoff 995, 996; Bes and Devereux ; aselli ; orsei and Peseni ; avallo and hironi ; anelli 5; hu 5). This lieraure has focused on how he macroeconomic aciviy of each counry and he exchange rae are influenced by unanicipaed fiscal shocks under monopolisic disorions and price rigidiies. However, since he publicaion of he Obsfeld and Rogoff (995, 996) papers, mos NOEM models have assumed ha firms are immobile across counries. Alhough i is feasible o explore he effecs of a governmen spending shock in his framework under he assumpion of a fixed inernaional disribuion of firms, recen empirical evidence suggess ha exchange raes affec he producion locaions of firms (see, for example, ushman 985, 988; Froo and Sein 99; ampa 993; Klein and Rosengren 994; oldberg and Kolsad 995; Blonigen 997; oldberg and Klein 998; Bénassy-quéré e al ; hakrabari and Scholnick ; Farrell e al 4). In addiion, recenly, he movemen of inernaional firms aiming for higher profis has been expanding rapidly beween emerging counries (for example, hina, India, Brazil, ec.) and developed counries (for example, he Unied Saes, Japan, Souh Korea, ec.). This is because he nominal exchange rae affecs he relaive price of goods produced in boh counries, hereby changing relaive real profis across counries. I is, herefore, imporan o For a survey, see Lane and anelli (3). For a survey of he lieraure examining deerminans of foreign direc invesmen, see Blonigen (5). 3

4 invesigae he effecs of a governmen spending shock by one counry in an open economy in which he inernaional relocaion of firms maers. In he heoreical lieraure on he NOEM, here has been lile sudy of how allowing for inernaional relocaion of firms affecs he macroeconomic impacs of governmen spending shocks. One excepion is Johdo (5), who aemps o presen a new NOEM model wih inernaional relocaion of firms and succeeds in showing explicily he effecs of one counry s governmen spending on he consumpion of he wo counries and he exchange rae. However, because Johdo (5) begins wih he assumpion of a wo-counry economy, he canno consider how allowing for a hird counry affecs he impacs of a governmen spending shock on inernaional relocaion and oher macroeconomic variables, including consumpion and he exchange rae. Recenly, mulinaional firms have very acively invesed across naional borders: American, Japanese, Sweden, and France s mulinaional firms are increasingly making heir way no only ino each oher s markes bu also ino Singapore, hina, Brazil, India, and Vienam. I is, herefore, appropriae ha a mulicounry model be adoped o examine how allowing for inernaional relocaion of firms affecs he impacs of a governmen spending shock on consumpion and exchange raes. iven his moivaion, his paper invesigaes he impacs of governmen spending shocks on he inernaional disribuion of firms, he exchange rae, and consumpion by exending he wo-counry model of Johdo (5) o a hree-counry model. From his analysis, we show explicily he macroeconomic effecs of governmen spending shocks, which lead o firm relocaion among hree counries, and i is found ha a governmen spending shock in one of he hree counries always depreciaes is currency, causes a firm locaed abroad o relocae o he counry and consequenly decreases he counry s 4

5 relaive consumpion in spie of he inflows of foreign firms. In conras, i is also found ha he governmen spending shock can be beneficial for neighboring counries in spie of he ouflows of firms. The remainder of his paper is srucured as follows. Secion oulines he feaures of he model. Secion 3 describes he equilibrium. In Secions 4 and 5, we examine he impacs of governmen spending shocks on he disribuion of firms across he hree counries, he nominal exchange rae, and consumpion. The final secion summarizes he findings and concludes.. The model In his secion, we consruc a perfec-foresigh, hree-counry model wih inernaional relocaion of firms. The hree counries are denoed by A, B, and, respecively. For simpliciy, he size of he world populaion is normalized o uniy, and households in counries A and B inhabi he inervals 3 and 3 3, respecively, and hose in counry inhabi he inerval 3. Therefore, he shares of households in A, B, and are 3, 3, and 3, respecively. There is monopolisic compeiion in he markes for goods and labor, whereas he markes for money and inernaional bonds are perfecly compeiive. On he producion side, monopolisically compeiive producers exis coninuously in he range, each of which produces a single differeniaed produc ha is freely radable. This implies ha producive aciviy canno be carried ou in more han one locaion. In his model, counry A consiss of hose producers in he inerval m, counry B consiss of hose producers in he inerval [m n, and he remaining [n producers are in counry, where m and n are 5

6 endogenous variables. Finally, we assume ha firms are mobile inernaionally bu heir owners are no. Therefore, all profi flows from firms are disribued o heir immobile owners according o heir share of holdings... Households The ineremporal objecive funcion of represenaive household x in counry h a ime, wih h A, B,, is: U h x (log h x log(m h xp h ) sh x ) () where is a consan subjecive discoun facor; h x is he consumpion index ha is defined laer; M h xp h is real money holdings, where M h x denoes nominal money balances held a he beginning of period, and P h is he consumpion price index of counry h; and sh x is he amoun of labor supplied by household x. A each poin in ime, households receive reurns on risk-free nominal bonds, earn wage income by supplying labor, and receive profis from all firms equally. Therefore, a ypical domesic household faces he following budge consrain: E h B h +x M h x i E h B h x M h x W h x sh x P h h x P h h E h E A m A zdz E h E B n m B zdz E h zdz () n where E h denoes he nominal exchange rae, defined as counry h s currency per uni of counry s currency (so ha E ); B h +x denoes he nominal bond denominaed in he counry s currency held by counry h s agen x in period ; i denoes he 6

7 nominal yield on he bond in erms of he counry s currency; W h x sh x is nominal labor income, where W h x denoes he nominal wage rae of labor supplied by m household x in period ; A n zdz, m B zdz, and zdz represen he oal n nominal profi flows of firms locaed in counries A, B, and, respecively; P h h x represens nominal consumpion expendiure; and h denoes real lump-sum axes. Noe ha all variables in () are measured in per capia erms. In he governmen secor, we assume ha governmen spending is purely dissipaive and ha i is financed by lump-sum axes and by seigniorage revenues derived from prining he naional currency. Hence, he governmen budge consrain in counry h is h s h h [M h + M h P h ], where h denoes he governmen spending in counry h, M h is aggregae money supply, and s h denoes he populaion share of counry h in he world populaion. ounries B and have an analogous governmen budge consrain. Here, we assume ha any monopolisically compeiive firm ha operaes in every counry employs he same producion echnology. In wha follows, we mainly focus on he descripion of counry A, because oher counries are described analogously. In counry A, firm z m hires a coninuum of differeniaed labor inpus domesically and produces a unique produc in a single locaion according o he ES producion funcion: y A z ((3) / 3 z x A, / dx) / (3) where y A z denoes he producion of firm z in period ; A z, x is he firm z s inpu of labor from household x in period ; and is he elasiciy of inpu subsiuion. 7

8 iven he firm s cos minimizaion problem, firm z s opimal demand funcion for labor x is as follows: A z, x (3) (W A zw A ) y A z (4) where W A ((3) 3 W A x dx) / is a price index for labor inpu. Similarly, he oher counries firms have an opimal demand funcion for labor x ha is analogous o equaion (4).... Definiion of consumpion baske The consumpion baske of household x living in counry h a period is: h m ( ) / n ( ) / ( h h x c z, x dz c z, x dz c z, x h A m B n ) / dz (5) where is he elasiciy of subsiuion among varieies produced wihin each counry; and c h jz, x denoes consumpion by household x locaed in counry h of he good produced by firm z locaed in counry j. From (5), he consumpion-based price indexes is defined as: P h m h P A z dz n m h P B z dz z h P dz where P h jz is he price in counry h of he good produced by firm z in counry j, j A, B,. n... Household decisions 8

9 Households maximize he consumpion index h x subjec o a given level of expendiure by opimally allocaing differeniaed goods produced in he hree counries c h jz, x, j A, B,. From his problem, we obain he following privae demand funcions: c h jz, x P h jzp h h x (6) As in he NOEM lieraure, we assume ha he governmen s consumpion index is he same as he household secor s, given by (5). Therefore, he governmen s demand funcions for good j in he home and foreign counries are he same as hose of he household secor. Summing he privae and public demand funcions and equaing he resuling equaion o he produc of firm z locaed in counry j yields he following marke-clearing condiion for any produc z produced in counry j: y j z P A jzp A A A P B jzp B B B P jzp (7) where A 3 A xdx 3 3, B B xdx, xdx, A A xdx, B B xdx 3, and xdx. From he law of one price and he purchasing power pariy arising from symmeric preferences, (7) is rewrien as: y j z P j jzp j w w (8) where w A B, w A B. In he second sage, households maximize () subjec o (). The firs-order condiions for his problem wih respec o B h +x and M h x can be wrien as: 9

10 h +x h xi + P h E h P h +E h + (9) M h xp h h x i + E h + i + E h + E h () Equaion (9) is he Euler equaion for consumpion, and () is he one for money demand. Following orsei and Peseni (), we inroduce nominal rigidiies ino he model in he form of one-period wage conracs under which nominal wages in period are predeermined a he end of period. In he monopolisic labor marke, each household provides a single variey of labor inpu o a coninuum of domesic firms. Hence, in counry A, he equilibrium labor-marke condiions imply ha sa (x) m A z, xdz, x,3, where he lef-hand side represens he amoun of labor supplied by household x, and he righ-hand side represens firms oal demand for labor x. By aking W A, P A, and m as given, subsiuing sa (x) A z, xdz and equaion (4) ino he budge consrain given by (), and maximizing he lifeime uiliy given by () wih respec o he nominal wage W A (x), we obain he following firs-order condiion for he opimal nominal wage, W A (x): m sa (x) W A (x)p A sa (x) A () The righ-hand side of () represens he marginal consumpion uiliy of addiional labor income resuling from a decrease in he nominal wage rae. This erm is posiive because. The lef-hand side represens he marginal disuiliy of an associaed increase in labor effor. Hence, each monopolisically compeiive household uses ()

11 o se is wage rae. The labor suppliers of counries B and have analogous opimal wage condiions... Firm s decision Since he counry A-locaed firm z hires labor domesically, given W A, P A A, and P A, w, w, m, (3), and subjec o (8), he counry A-locaed firm z faces he following profi-maximizaion problem: max A z P A 3 A Azy A z W z z x P A A z, A dx P A Az W A y A z subjec o y A z P A AzP A w w iven he above, he price mark-up is chosen according o: P A Az W A () Since W A is given, () yields P A Az P A A, z, m. These relaionships imply ha each firm locaed in counry A supplies he same quaniy of goods. Similarly, oher firms locaed in differen counry have he price mark-up ha is analogous o equaion (). Denoing he maximized real profi flows of counry j-locaed firms by j zp j, and subsiuing (8) and () ino j z yields: j zp j P j jzp j w w (3).3. Relocaion behavior

12 The driving force for relocaion o oher counries is differences in curren real profis beween wo bounded counries. 3 In addiion, following he formulaion in Johdo (5), we assume ha all firms are no allowed o relocae insananeously even if here is he profi gap. A each poin in ime, his adjusmen mechanism for relocaion beween counries A and B is formulaed as follows: m m A zp A B zp B A zp A E A E B B zp A (4) Analogously, he adjusmen mechanism for relocaion beween counries B and is formulaed as follows: n n B zp B zp B zp B E B zp B (5) where is a consan posiive parameer ha deermine he degree of firm mobiliy beween wo bounded counries: a larger value of implies higher firm mobiliy beween counries. Inuiively, he parameer reflecs he coss falling on mobile firms in heir new locaions. Examples include he coss of finding appropriae plans, raining he local workforce, and adaping o he local legal sysem. Because of hese coss, firms canno move insananeously o a beer locaion even if a profi gap beween wo counries provides he moivaion..4. Marke condiions The equilibrium condiion for he inegraed inernaional bond marke is given by: 3 In he lieraure on mulinaional firms, Helpman e al. (4) and Eckel and Egger (9) derive he share of mulinaional firms endogenously by using his ype of profi differenial beween exporing and mulinaional aciviy.

13 3 B A xdx 3 3 B B xdx 3 B xdx (6) This means ha he ne supply of bonds worldwide is zero. In addiion, he money markes are assumed always o clear in all counries, so ha he equilibrium condiions are given by M A respecively. 3 M A xdx 3 3, M B M B xdx, and M M xdx, 3 3. Seady sae values In his secion, we derive he soluion for a symmeric seady sae in which all variables are consan, iniial ne foreign asses are zero (B h ) and h, h A, B,. 4 Henceforh, we denoe he seady-sae values by using he subscrip ss. In he symmeric seady sae, given he Euler equaion for consumpion (equaion (9)), he consan real ineres rae is given by: r ss (7) where is he rae of ime preference. Because symmery, which implies h ss w ss, holds, he seady-sae inernaional allocaions of firms are: m ss 3 (8) n ss 3 (9) The seady sae oupu levels are: 4 In he symmeric seady sae, we drop he index value x from all variables in order o simplify noaion. 3

14 y jss sh ss h ss w ss, j, h A, B, () Equaion () shows ha no only do all firms worldwide produce he same amoun of oupu, i also shows ha all households worldwide consume his oupu and supply he labor required o produce his oupu. Subsiuing w ss from equaion () ino equaion (3) yields he following seady-sae levels of real profi flows of counry j-locaed firms, which are equal: jss P j ss, j A, B, () 4. A log-linearized analysis To examine he macroeconomic effecs of an unanicipaed permanen governmen spending shock, we solve a log-linear approximaion of he sysem around he iniial, zero-shock seady sae wih B h ss,, h A, B,, as derived in he previous secion. For any variable X, we use X o denoe shor-run percenage deviaions from he iniial seady-sae value, i.e., X dx X ss,, where X ss, is he iniial, zero-shock seady-sae value, and subscrip denoes he period in which he shock akes place. These shor-run percenage deviaions are consisen wih he lengh of nominal wage j j conracs. Thus, nominal wages and goods prices can be deermined as W P z j A, B,, in he shor-run log-linearized equaions. In addiion, we use X o denoe long-run percenage deviaions from he iniial seady-sae value, i.e., j, X dx X dx X ss, ss ss,, which is consisen wih flexible nominal wages. Noe 4

15 ha X X ss because he new seady sae is reached a period (see Appendix for he derivaion of shor-run and long-run fiscal policy effecs). By log-linearizing equaions (4) and (5) around he symmeric seady sae and j j seing W P z j for he inernaional disribuion of firms:, j A, B,, we obain he following log-linearized expression A B E E 3 m 3 () B 3 3 E n (3) Equaion () shows ha under a given E B, an exchange rae depreciaion of counry A s currency ( E A E B ) induces relocaion of firms locaed in counry B owards he counry A. 5 Inuiively, wih fixed nominal wages, which cause nominal produc prices o be sicky because of mark-up pricing by monopolisic produc suppliers, he depreciaion in counry A s currency increases relaive producion of counry A s goods A B A B hrough he expendiure-swiching effec ; i.e., y y E E. 6 This increases he relaive profis of counry A-locaed firms, and consequenly, firms locaed in counry B relocae o he counry A. Equaion () also shows ha nominal exchange rae changes have greaer effecs he greaer is he flexibiliy of relocaion (he larger is 5 This resul is consisen wih he evidence found in he empirical lieraure on he relaionship beween exchange raes and foreign direc invesmens (see, for example, ushman 988; aves 989; Froo and Sein 99; ampa 993; Klein and Rosengren 994; Blonigen 997; oldberg and Klein 998; Bénassy-quéré e al ; hakrabari and Scholnick ; Kiyoa and Uraa 4; Bolling e al 7). 6 The expendiure-swiching effec arises inuiively because exchange rae depreciaion causes a decrease in he relaive real price of counry A s goods for households in all counries so ha world consumpion demand swiches oward counry A s goods. orsei e al (5) also define his as compeiive effec. 5

16 6 ). By conras, when relocaion coss are high ( ), nominal exchange rae changes have a negligible effec on he relocaion of firms. The inuiion behind he impac of E B in equaion (3) on he inernaional relocaion of firms beween counries B and can be explained analogously. 5. overnmen spending shocks Now, we consider he effecs of an unanicipaed permanen governmen spending shock in each counry. 5.. The case of, B B A A In his subsecion, we focus on he impacs of a permanen governmen spending shock in counry A. In his case, he closed-form soluions for he six key variables are as follows: ~ A B A E E (4) ~ A B E (5) ~ 3 A m (6) ~ 3 A n (7)

17 ~ A B A (8) ~ A A (9) where ~ ~ 9 ~ ~ ~ 6 (3) ~ 6 ~ 3 (3) (3) ~ ~,, ~ Equaions (4) and (5) indicae ha an unanicipaed governmen spending shock in counry A leads o exchange rae depreciaion in E A E B and E B, respecively. Equaion (6) shows ha an unanicipaed governmen spending shock in counry A causes counry B firms o relocae o counry A. Equaion (7) shows ha an unanicipaed governmen spending shock in counry A causes counry firms o relocae o counry B. Equaions (8) and (9) show ha he relaive consumpion levels of counry A decrease when here is an unanicipaed governmen spending shock in counry A. The above resuls can be explained inuiively as follows. Firs, a rise in governmen spending in counry A resuls in crowding-ou of counry A consumpion, because 7

18 counry A s governmen-spending rise does no increase counry A s oupu sufficienly o offse he rise in axes. Hereafer, we shall call his he crowding-ou effec. Under a given E B, he reducion in counry A s consumpion hen leads o a depreciaion of is currency ( E A E B, see equaion (4)). This happens because, given ha he demand for real money balances is increasing wih consumpion, counry A s currency mus depreciae and decrease he supply of real money balances in counry A o resore money marke equilibrium. Furhermore, he exchange rae depreciaion causes consumpion swiching as world consumpion demand shifs oward counry A s goods because of he fall in he relaive price of counry A s goods. This in urn causes counry B s firms o relocae o counry A because of he increase in relaive profis of firms locaed in counry A ( m, see equaion (6)). This relocaion hen increases labor demand in counry A and decreases labor demand in counry B, which in urn raises he labor income of counry A and decreases he labor income of counry B. Hereafer, we shall call his he AB relocaion effec. As a resul, he consumpion increases in counry A while he consumpion decreases in counry B. Thus, A B is deermined by he wo conflicing mechanisms of he crowding-ou effec and he AB relocaion effec. However, from equaion (8), such a governmen spending rise unambiguously leads o a decrease (rise) in he relaive consumpion of counry A (B), A B. In addiion, from he decrease in he consumpion of counry B hrough he AB relocaion effec, counry B s currency mus depreciae and decrease he supply of real money balances in counry B o resore money marke equilibrium ( B E, see equaion (5)). This in urn causes counry s firms o relocae o counry B because of he increase in he relaive profis of firms locaed in counry B ( n, see equaion 8

19 (7)). This relocaion hen increases labor demand in counry B and decreases labor demand in counry, which in urn raises labor income in counry B and decreases labor income in counry. Hereafer, we shall call his he B relocaion effec. As a resul, he consumpion increases in counry B while he consumpion decreases in counry. Therefore, A is deermined by he hree conflicing mechanisms of he crowding-ou effec, he AB relocaion effec, and he B relocaion effec. However, from equaion (9), such a governmen spending rise unambiguously leads o a decrease (rise) in he relaive consumpion of counry A (), A. In sum, a permanen governmen spending shock in counry A is derimenal o counry A in erms of he relaive consumpion level. In oher words, a permanen governmen spending shock in counry A always benefis no only counry B bu also counry in erms of relaive consumpion. 5.. The case of B B, A A In his subsecion, we focus on he impacs of a permanen governmen spending shock in counry B. In his case, he closed-form soluions for he six key variables are as follows: ~ A B B E E (33) ~ B B E (34) 9

20 ~ 3 B m (35) ~ 3 B n (36) ~ A B B (37) ~ B B (38) The above resuls can be explained inuiively as follows. Firs, a rise in he governmen spending in counry B resuls in he crowding ou of counry B s consumpion, because counry B s governmen spending rise does no increase counry B s oupu sufficienly o offse he rise in axes (he crowding-ou effec). Under a given E A, he decrease in he consumpion of counry B hen leads o exchange rae depreciaion of counry B s currency ( E A E B, B E, see equaions (33) and (34)). Furhermore, he exchange rae depreciaion causes consumpion swiching as world consumpion demand shifs oward counry B s goods because of he fall in he relaive price of counry B s goods. This in urn causes firms locaed in counries A and o relocae o counry B because of he increase in he relaive profis of firms locaed in counry B ( m, n, see equaions (35) and (36)). This relocaion increases labor demand in counry B and decreases labor demand in counries A and, which in urn raises labor income in counry B and decreases labor income in counries A and. As a resul, he relocaion increases he consumpion in counry B, while i decreases he

21 consumpion in counries A and. Thus, he governmen spending effec on A B is deermined by he hree conflicing mechanisms of he crowding-ou effec, he AB relocaion effec, and he B relocaion effec. However, from equaion (37), such a governmen spending rise unambiguously leads o a rise (decrease) in he relaive consumpion of counry A (B), A B. Similarly, he impac of an increase in governmen spending in counry B on B is ambiguous. This is because he impac of an increase in governmen spending is also deermined by hree conflicing mechanisms: he crowding-ou effec, he AB relocaion effec, and he B relocaion effec. However, from (38), such a governmen spending rise unambiguously leads o a decrease (rise) in he relaive consumpion of counry B (), B. In sum, a permanen governmen spending rise in counry B is derimenal o counry B in erms of he relaive consumpion level. In oher words, a permanen governmen spending rise in counry B always benefis counry A bu also counry in erms of relaive consumpion level The case of, A A B B In his subsecion, we focus on he impacs of a permanen governmen spending shock in counry. In his case, he closed-form soluions for he six key variables are as follows: ~ A B E E (39)

22 ~ B E (4) ~ 3 m (4) ~ 3 n (4) ~ B (43) ~ A (44) The above resuls can be explained inuiively as follows. Firs, a rise in governmen spending in counry resuls in he crowding ou of counry s consumpion, because counry s governmen spending rise does no increase counry s oupu sufficienly o offse he rise in axes (he crowding-ou effec). The decrease in he consumpion of counry hrough he crowding-ou effec hen leads o exchange rae depreciaion of is currency ( B A E E, see equaion (4)). However, a his sage, counry A s currency relaive o B s remains unchanged, because B A E E. Furhermore, he exchange rae depreciaion causes consumpion swiching as world consumpion demand shifs oward counry s goods because of he fall in he relaive price of counry s goods. This in urn causes counry B s firms o relocae o counry because of he increase in he relaive profis of firms locaed in counry ( n, see equaion (4)). This relocaion increases labor demand in counry and decreases labor

23 demand in counry B, which in urn increases labor income in counry and decreases labor income in counry B (he B relocaion effec). As a resul, he relocaion increases he consumpion in counry and decreases ha of counry B. Thus, B is deermined by he wo conflicing mechanisms of he crowding-ou effec and he B relocaion effec. However, from equaion (43), such a governmen spending rise unambiguously leads o a rise (decrease) in he relaive consumpion of counry B (), B. Furhermore, as discussed in he definiion of he B relocaion effec, he rise in counry s governmen spending also decreases counry B s consumpion hrough firm relocaion from counry B o counry. From his resul, counry B s currency mus depreciae o resore equilibrium in he marke for real balances. This depreciaion of counry B s currency weakens he iniial appreciaion of is currency, and consequenly he change in counry A s currency relaive o B s is negaive ( E A E B, see equaion (39)). Furhermore, his leads o reducion of he real prices of counry B s goods relaive o counry A s goods, which causes world demand o swich from counry A s goods o counry B s goods. These demand shifs increase he relaive profis of firms locaed in counry B, which cause firms locaed in counry A o relocae o counry B ( m, see equaion (4)). This relocaion increases labor demand in counry B and decreases labor demand in counry A, which in urn increases labor income in counry B and decreases labor income in counry A (he AB relocaion effec). As a resul, he relocaion decreases he consumpion in counry A. Thus, A is deermined by he hree conflicing mechanisms of he crowding-ou effec, he AB relocaion effec, and he B relocaion effec. However, from equaion (44), such a governmen spending 3

24 rise unambiguously leads o a rise (decrease) in he relaive consumpion of counry A (), A. This is because he decrease in A hrough he AB and B relocaion effecs is dominaed by he counry s consumpion reducion hrough he crowding-ou effec. In sum, a permanen governmen spending shock in counry is derimenal o counry in erms of he relaive consumpion revel. In oher words, a permanen governmen spending shock in counry always benefis no only counry A bu also counry B in erms of relaive consumpion. 6. onclusion In his paper we considered he quesion of how allowing for inernaional relocaion of firms among hree counries affecs he impacs of governmen spending shocks on relaive consumpion and exchange raes. From his analysis, we showed explicily he macroeconomic effecs of governmen spending shocks ha lead o firm relocaion among hree counries, and i was found ha a governmen spending shock in one of he hree counries always depreciaes is currency and decreases is relaive consumpion levels, while i can be beneficial for he neighboring counries in spie of he ouflows of firms. However, he model developed here is raher simple in a number of respecs. This suggess many direcions for fuure research. Firs, his paper may yield resuls ha are more ineresing if he curren model is modified o include sunk coss, as in Russ 4

25 (7). 7 Also of ineres is exending he model o accoun for rade impedimens such as ariffs and ranspor coss as in Fender and Yip (). 8 Furher, he consideraion of he effecs of oher fiscal policy (for example, consumpion ax) in our model is noeworhy 9. These issues remain for fuure research. 7 ampa (993) finds he negaive effec of sunk coss (e.g., invesmen in adverising and media promoion) on indusry enry ino he US during he 98s. Brainard (997) also finds ha overseas producion by mulinaionals decreases wih he fixed coss of producion. 8 Empirical evidence shows ha higher ariff has an imporan effec on foreign direc invesmen of firms based in developed counries (see Brainard, 997, and Blonigen, ). 9 Johdo (3) sudies he welfare effecs of a consumpion ax rise based on he wo-secor small open economy model of Obsfeld and Rogoff (995), bu do no allow for he endogenous deerminaion of he disribuion of firms. 5

26 References Bénassy-quéré, A., Fonagné, L., Lahréche-révil, A.,. Exchange-rae sraegies in he compeiion for aracing foreign direc invesmen. Journal of he Japanese and Inernaional Economies 5, Bes,., Devereux, M.,. Exchange rae dynamics in a model of pricing-o-marke. Journal of Inernaional Economics 5, 544. Blonigen, B.A., 997. Firm-specific asses and he link beween exchange raes and foreign direc invesmen. American Economic Review 87, Blonigen, B.A.,. Tariff-jumping anidumping duies. Journal of Inernaional Economics 57, 35. Blonigen, B.A., 5. A review of he empirical lieraure on FDI deerminans. Universiy of Oregon and NBER. Bolling,., Shane, M., Roe, T., 7. Exchange raes and U.S. foreign direc invesmen in he global processed food indusry. Agriculural and Resource Economics Review 36, 338. Brainard, L.S., 997. An empirical assessmen of he proximiy-concenraion rade-off beween mulinaional sales and rade. American Economic Review 87, ampa, J.M., 993. Enry by foreign firms in he unied saes under exchange rae uncerainy. Review of Economics and Saics 75, 646. aselli, P.,. Fiscal consolidaion under fixed exchange raes. European Economic Review 45, avallo, M., hironi, F.,. Ne foreign asses and he exchange rae: Redux revised. Journal of Moneary Economics 49,

27 aves, R., 989. Exchange-rae movemens and foreign direc invesmen in he unied saes, in: Audresch, D.B., and laudon, M.P. (Eds.), The Inernaionalizaion of U.S. Markes, New York Universiy Press, 998. hakrabari, R., Scholnick, B.,. Exchange rae expecaions and foreign direc invesmen flows. Welwirschafliches Archiv 38,. hu, H., 5. Welfare and public: The role of inernaionalized producion. Economic Inquiry 43, orsei,., Peseni, P.,. Welfare and macroeconomic inerdependence. Quarerly Journal of Economics 6, orsei,., Marin, P., Peseni, P., 5. Produciviy, erms of rade, and he home marke effec. Journal of Inernaional Economics 73, 997. ushman, D.O., 985. Real exchange rae risk, expecaions, and he level of foreign direc invesmen. Review of Economics and Saisics 67, ushman, D.O., 988. Exchange rae uncerainy and foreign direc invesmen in he Unied Saes. Welwirschafliches Archiv 4, Eckel,., Egger, H., 9. Wage bargaining and mulinaional firms. Journal of Inernaional Economics 77, 64. Farrell, R., ason, N., Surm, J.M., 4. Deerminans of Japan s foreign direc invesmen: An indusry and counry panel sudy, The Journal of he Japanese and Inernaional Economies 8, 68. Fender, John., Yip,.K.,. Tariffs and exchange rae dynamics redux. Journal of Inernaional Money and Finance 9,

28 Froo, K.A., Sein, J.., 99. Exchange raes and foreign direc invesmen: An imperfec capial markes approach. Quarerly Journal of Economics 6, 97. anelli,., 5. The new open economy macroeconomics of governmen deb. Journal of Inernaional Economics 65, oldberg, L., Klein, M., 998. Foreign direc invesmen, rade, and real exchange rae linkages in developing counries, in: lick, R. (Eds.), Managing apial Flows and Exchange raes: Perspecives from he Pacific Basin, ambridge, UK: ambridge Universiy Press, 73. oldberg, L., Kolsad,., 995. Foreign Direc Invesmen, Exchange rae variabiliy and Demand Uncerainy. Inernaional Economic Review 36, Helpman, E., Meliz, M.J., Yeaple, S.R., 4. Expors versus FDI. American Economic Review 94, 336. Johdo, W., 3. onsumpion Tax, Nonraded oods and Welfare. Theoreical Economics Leers 3, 666. Johdo, W., 5. Fiscal Policy and Relocaion of Firms in he New Open Economy Macroeconomics. Tezukayama RIEB Discussion Paper Series No. 5. Kiyoa, K., Uraa, S., 4. Excahnge rae, exchange rae volailiy and foreign direc invesmen. The World Economy 7, Klein, M.W., Rosengren, E.S., 994. The real exchange rae and foreign direc invesmen in he Unied Saes: Relaive wealh vs. relaive wage effecs. Journal of Inernaional Economics 36,

29 Lane, P.R., anelli,., 3. Dynamic general equilibrium analysis: he open economy dimension, Dynamic Macroeconomic Analysis: Theory and policy in general equilibrium. ambridge Universiy Press. Obsfeld, M., Rogoff, K., 995. Exchange rae dynamics redux. Journal of Poliical Economy 3, Obsfeld, M., Rogoff, K., 996. Foundaions of Inernaional Macroeconomics. MIT press, ambridge, MA. Russ, K.N., 7. The endogeneiy of he exchange rae as a deerminan of FDI: A model of enry and mulinaional firms. Journal of Inernaional Economics 7,