Interregonal Trade, Industral Locaton and Import Infrastructure* Toru Kkuch (Kobe Unversty) and Kazumch Iwasa (Kyoto Unversty)** Abstract The purpose of ths study s to llustrate, wth a smple two-regon, two-good, two-factor model, how an mprovement n one regon s mport nfrastructure can affect frms locaton decsons and the nature of the tradng equlbrum. It s shown that, through mprovements n mport nfrastructure, one regon mght dvert hgh-tech ndustres to another regon. Ths effect reduces the ncentve to mprove mport nfrastructure. *We would lke to the anonymous referees for constructve comments. ** kkuch@econ.kobe-u.ac.jp (Kkuch), wasa@ker.kyoto-u.ac.jp (Iwasa)
Introducton In the last decade the role of publc nfrastructure n regonal economes, such as ports, ralway networks, and new telecommuncatons networks, has been wdely dscussed. It s ncreasngly recognzed that the growng connectvty of ndvduals and organzatons s acheved through mprovements n the qualty of publc nfrastructure whch reduces transacton costs between regons, and a consequent ncrease n the flow of goods and servces across regons. Related to ths, competng new economc geography theores mply that regonal ntegraton mght serve to greatly reduce regonal nequaltes. 2 In a recent nfluental contrbuton, Martn and Rogers (995) proposed a new way to model varous types of publc nfrastructure, whch allows the analyss of ts mpact on trade patterns, ndustral locaton, and welfare. They show that frms wth ncreasng returns wll tend to locate n the countres (or regons) wth the best nfrastructure n order to take advantage of economes of scale. In partcular, they show that an mprovement n domestc nfrastructure n one country (or regon) wll mply a relocaton of frms to ths country (or regon) (Martn and Rogers, 995, p. 344). The ntuton s that, as the qualty of nfrastructure mproves, the transacton costs of goods produced and consumed n that country (or regon) decrease, ncreasng the effectve demand. However, Martn and Rogers (995) do not dstngush between export nfrastructure and mport nfrastructure: an mprovement n the qualty of nfrastructure mples a symmetrc reducton n transacton costs for both exports and mports. Ths assumpton s justfed for smplfcaton. However, nfrastructure mprovements often cause asymmetrc reductons n transacton costs. For example, an mprovement n the qualty of a regon s local transport networks affects the regon s mports more than See, for example, Costa-Font and Rodrguez-Oregga (2005), Lmao and Venables (200), and Mor and Nshkm (2002). 2 See, for example, Fujta et al. (999), Ottavano and Thsse (2004), and Behrens et al. (2007).
ts exports. 3 Gven these observatons, the present study focuses on the role of mport nfrastructure n facltatng mport transactons. The purpose of ths study s to llustrate, wth a smple two-regon, two-good (homogeneous good/dfferentated hgh-tech products), two-factor (labor/captal) model, how an mprovement n one regon s mport nfrastructure can affect frms locaton decsons and the nature of the tradng equlbrum. In contrast to Martn and Rogers (995), t wll be shown that, through mprovements n mport nfrastructure, one regon mght dvert hgh-tech ndustres to another regon. The man result of the current study, whch captures the negatve effect of the domestc nfrastructure, has not appeared n the exstng lterature. Secton 2 presents the model. Secton 3 analyzes the mpact of an mprovement n the qualty of mport nfrastructure on the locaton of ndustres. 2 The Model Suppose that there are two regons (Regon and Regon 2), each wth two factors (captal, K and labor, L ) and two types of goods (a homogeneous good and a large varety of dfferentated hgh-tech products). Assume that the regons are dentcal n regard to tastes, sze, and technology, but dffer wth respect to the costs of mportng dfferentated hgh-tech products. We assume these costs are drectly related to the qualty of mport nfrastructure. It s mportant to note that the present model s a varant of Kkuch (2005). There are two man dfferences between the present setup and Kkuch (2005) s one: () the former labels two factors as captal and labor, whle the latter labels those as sklled and unsklled labor; (2) the former assumes that the dstrbuton of factor endowments s dentcal among regons, whle the latter assumes that the dstrbuton of factor endowments s dfferent between countres. Although the assumpton of the 3 See World Bank (2004) for dscusson.
dentcal factor endowments s qute strong, t s n order for smplfyng the analyss. Consumers have Cobb-Douglas preferences over both categores and spend fracton m of ther ncome on hgh-tech products. Regon s prce ndex for hgh-tech products s represented by the Dxt-Stgltz form: /( -s ) -s -s é ù j j P = ë n ( p ) + n ( t p ) û, s >, () where s s the degree of substtuton among all products, p s the producer prces for hgh-tech products produced n regon, and n s the number of varetes produced n regon, respectvely. Import costs t ( t > ) for the hgh-tech products are n the form of ceberg costs. We assume these costs are drectly related to the qualty of regon s mport nfrastructure: changes n these costs represent changes n mport nfrastructure. Thus, the consumer demand functons n regon for a regon (.e., local) varety and a regon j (.e., mported) varety are respectvely c = p -s P s - me, (2) c j = ( t p j ) P -s s - me, (3) where E s the total ncome n Regon (, j=, 2). The homogeneous good s produced wth constant returns, usng only labor as an nput. Unts are chosen so that one unt of labor produces one unt of output. As usual n new geography models, no transport costs exst for the homogeneous good, whch serves to te down the wage rate. Also assume that the parameters of the model are such that both regons produce the homogeneous good; thus, constant, dentcal wages for labor hold (hereafter set to unty). The producton of each varety of hgh-tech product requres one unt of captal to develop the product and b unts of labor per unt of output. As n Martn and Rogers (995) and Martn and Ottavano (999), the central assumpton of the present analyss s that the captal s frm specfc, but t moves
freely between regons: f a varety developed by Regon s captal s produced n Regon 2, the operatng profts are repatrated to Regon. Gven a Dxt-Stgltz specfcaton wth constant elastcty s, each frm sets ts prce as p = p2 = ( bs ) / ( s - ). In order to smplfy the analyss, we choose unts such that b = ( s - ) / s to have p = p2 =. (4) Gven that one unt of captal s requred to develop a product, the payment for each unt of captal employed n Regon (=, 2), r, must satsfy r = p x - b x = x / s, (5) where x s the output of a representatve frm n Regon. When captal moblty s unrestrcted, the payment for captal wll be equalzed between regons, whch mples that r = r2 and thus x = x. (6) 2 3 Import Infrastructure and Industral Locaton Now consder the frms locaton decsons. The product market equlbrum n Regon requres that supply equals demand for each varety: x = c + t jc j. Substtutng (2), (3), and (4) nto ths condton yelds the followng equlbrum condton: æ t ö 2 x = ç + me, (7) è n + tn2 t 2n + n2 ø x æ t ö ç me, (8) è ø 2 = + n + tn2 t 2n + n2 where E º r K + L and t º t ( t < ) measures the freedom of trade, whch s drectly related to the -s qualty of Regon s mport nfrastructure.
Usng (6), (7) and (8), the equlbrum share of Regon frms, s can be obtaned: n é t -t ù 2 s º = ê - ú. (9) n + n2 2 ë ( -t)( -t 2) û Equaton (9) mples the surprsng feature of nfrastructure polces, whch s hghly contrasted wth the result of Martn and Rogers (995) (.e., frms relocaton to the regon wth a better nfrastructure). Proposton: An mprovement n the qualty of mport nfrastructure n a regon wll nduce a dverson of hgh-tech products away from that regon. Fgure llustrates the mplcatons of ths proposton. The horzontal axs shows the share of hgh-tech frms n Regon, s, and the vertcal axs shows the equlbrum output level n each regon, x. A decreasng (resp. ncreasng) curve corresponds to (7) [resp. (8) ]: the equlbrum share of frms s obtaned from the ntersecton of these curves. To smplfy the argument, let us assume that ntally t = t 2 holds and s = / 2. Now, suppose that the qualty of Regon s mport nfrastructure mproves (.e., t ncreases). Ths change nduces two effects. Frst, t shfts the curve representng Regon s equlbrum condton downward: lower trade costs mply an ncrease n the effectve number of mported varetes, t n2, whch leads to a fall n local demand for locally produced varetes n Regon. Second, t shfts the curve representng Regon 2 s equlbrum condton upward: easer access to the Regon market ncreases the advantage of locatng n Regon 2. These two effects renforce each other and nduce hgh-tech frms (.e., captal) to flow out of Regon. Ths result has mportant polcy mplcatons for regonal economes. Improvements n mport nfrastructure n one regon can dvert frms n hgh-tech ndustres over to another regon. Ths lowers the ncentve to mprove mport nfrastructure. Although better mport nfrastructure reduces mport
transacton costs, t also nduces ndustral dverson and rases the transacton costs of recevng products from those ndustres that relocate elsewhere. The possblty that ndustres wll be dverted provdes some theoretcal grounds for the coordnaton of nfrastructure nvestments among regonal economes. Further research should focus on these polcy mplcatons.
References [] Behrens, K., Gagne, C., Ottavano, G.I.P., and J.-F. Thsse (2007) Countres, Regons and Trade: On the Welfare Impacts of Economc Integraton, European Economc Revew, Vol. 5, pp. 277 30. [2] Costa-Font, J., and E. Rodrguez-Oregga (2005) Trade and the Effect of Publc Investment on Regonal Inequaltes n Heterogeneously Integrated Areas, World Economy, Vol. 28, pp. 873 89. [3] Dxt, A. K., and J. Stgltz (977) Monopolstc Competton and Optmum Product Dversty, Amercan Economc Revew, Vol. 67, pp. 297 308. [4] Fujta, M., P. Krugman and A. J. Venables (999) The Spatal Economy: Ctes, Regons and Internatonal Trade, Cambrdge, Mass.: MIT Press. [5] Lmao, N., and A. J. Venables (200) Infrastructure, Geographcal Dsadvantage, Transport Costs, and Trade, World Bank Economc Revew, Vol. 5, pp. 45 479. [6] Kkuch, T. (2005) Interconnected Communcatons Networks and Home Market Effects, Canadan Journal of Economcs, Vol. 38, pp. 870 882. [7] Martn, P., and G. Ottavano (999) Growng Locatons: Industry Locaton n a Model of Endogenous Growth, European Economc Revew, Vol. 43, pp. 28 302. [8] Martn, P., and C. A. Rogers (995) Industral Locaton and Publc Infrastructure, Journal of Internatonal Economcs, Vol. 39, pp. 335 35. [9] Mor, T., and K. Nshkm (2002) Economes of Transport Densty and Industral Agglomeraton, Regonal Scence and Urbanl Economcs, Vol. 32, pp. 67 200. [0] Ottavano, G., and J.-F. Thsse (2004) Agglomeraton and Economc Geography, n Henderson, J. V., and J.-F. Thsse (eds.) Handbook of Regonal and Urban Economcs, Vol. 4, Amsterdam:
Elsever, pp. 2563 2608. [] The World Bank (2004) Reformng Infrastructure: Prvatzaton, Regulaton, and Competton, Berln: Sprnger.
Fgure x 2 2 O / 2 s