Economic Integration, Wage Policies and Social Policies

Transcription

1 DISCUSSION PAPER SERIES IZA DP No. 748 Economc Integraton, Wage Polces and Socal Polces Mchael Pflüger Aprl 2003 Forschungsnsttut zur Zukunft der Arbet Insttute for the Study of Labor

2 Economc Integraton, Wage Polces and Socal Polces Mchael Pflüger DIW Berln and IZA Bonn Dscusson Paper No. 748 Aprl 2003 IZA P.O. Box 7240 D Bonn Germany Tel.: Fax: Emal: Ths Dscusson Paper s ssued wthn the framework of IZA s research area Internatonalzaton of Labor Markets. Any opnons expressed here are those of the author(s) and not those of the nsttute. Research dssemnated by IZA may nclude vews on polcy, but the nsttute tself takes no nsttutonal polcy postons. The Insttute for the Study of Labor (IZA) n Bonn s a local and vrtual nternatonal research center and a place of communcaton between scence, poltcs and busness. IZA s an ndependent, nonproft lmted lablty company (Gesellschaft mt beschränkter aftung) supported by the Deutsche Post AG. The center s assocated wth the Unversty of Bonn and offers a stmulatng research envronment through ts research networks, research support, and vstors and doctoral programs. IZA engages n () orgnal and nternatonally compettve research n all felds of labor economcs, () development of polcy concepts, and () dssemnaton of research results and concepts to the nterested publc. The current research program deals wth (1) moblty and flexblty of labor, (2) nternatonalzaton of labor markets, (3) welfare state and labor market, (4) labor markets n transton countres, (5) the future of labor, (6) evaluaton of labor market polces and projects and (7) general labor economcs. IZA Dscusson Papers often represent prelmnary work and are crculated to encourage dscusson. Ctaton of such a paper should account for ts provsonal character. A revsed verson may be avalable on the IZA webste ( or drectly from the author.

3 IZA Dscusson Paper No. 748 Aprl 2003 ABSTRACT Economc Integraton, Wage Polces and Socal Polces Ths paper uses a two country trade and geography model of monopolstc competton to study the effects of wage polces and socal polces on the locaton of ndustry. It s frst shown that a unon wage push n one of two otherwse dentcal countres nduces a relocaton of frms whch ncreases wth the level of economc ntegraton as measured by trade costs. Ths 'tradtonal vew' s then contrasted wth a 'new economc geography vew' n whch one of the countres has hstorcally emerged as the core. The agglomeraton rent whch accrues to the moble factor gves unons and governments n the core scope to set hgher wages and to choose more generous welfare polces than ther counterparts n the perphery wthout havng to encounter an ext of frms. The relatonshp between the maxmum nternatonal unon wage dfferental and the level of ntegraton s shown to be bellshaped. JEL Classfcaton: F12, F15, F16, F21, F22, R12 Keywords: ntegraton, wages, ndustry locaton, agglomeraton, monopolstc competton Mchael Pflüger DIW Berln Köngn-Luse-Straße Berln Tel.: Fax: Emal: mpflueger@dw.de Presented at the Congress of the European Economc Assocaton n Vence, at the congress of the European Assocaton of Labour Economsts n Pars, at the annual congress of the Veren für Socalpoltk n Innsbruck, at a CORE/FUNDP research workshop n Namur, an IZA workshop n Bonn, at the Unversty of Grefswald and at the Free Unversty Berln. I would lke to thank Torben Andersen, elmut Bester, Lucas Bretschger, Björn Frank, Andreas aufler, Wnfred Koennger, Ka Konrad, Jörg Lngens, Jochen Mchaels, Markus Pannenberg, Phllp Schröder, Jan R. Skaksen and Erc Toulemonde for comments and suggestons. Specal thanks to two anonymous referees for very careful reports.

4 1 1. Introducton The labour market mplcatons of economc ntegraton, both on a regonal level (e.g. n the European Unon) and on a global level, have conquered a central place n the publc debate n ndustralsed countres n recent years (see e.g. Rodrk 1997). Fears are wdespread that the ntensfed competton assocated wth the process of ntegraton severely constrans (unon) wages and socal polces. Some argue that a levellng of wages and welfare systems, f not a race to the bottom, s nevtable. In analysng these fears, economsts have studed a varety of channels through whch ncreased product market ntegraton affects labour markets (see Andersen et al for a recent survey). The present paper extends these consderatons by studyng the relocaton ncentves of frms n a model of monopolstc competton wth frm and captal moblty. The central contrbuton of ths paper s to show that unons and governments n an agglomeratve core have scope to choose hgher wages and more generous welfare polces than ther counterparts n the perphery wthout nducng a relocaton of ndustry. Ths result s based on the explotaton of the agglomeraton rent accrung to moble entrepreneural captal owners. The result provdes a ratonale for the emprcal fndng of pronounced dfferences n wages and socal polces even n well-ntegrated regons such as the European Unon. Another contrbuton of ths paper s to provde a unfyng framework both for the publc vew of a large-scale ext of frms as well as for the dea of a relocaton shelter dervng from an ntal agglomeraton. A key to most of the results n the prevous lterature s that the ntegraton of product markets affects the elastcty of labour demand and so the trade unons wage mark-up. Several channels mply that the elastcty of labour demand ncreases wth product market ntegraton so that the wage mark-up s reduced. 1 Other channels reverse ths concluson. 2 Moreover, the ntegraton of product market has effects on nsttutons (e.g. the tax system), productvty and techncal progress whch, n turn, nfluence wages n dverse ways. Emprcally, a tendency for wage convergence and an ncreasng senstvty of domestc to foregn wages has been found, 1 Ths may be due to a removal of trade and FDI barrers whch allows frms to move to areas where producton costs are low (e.g. Drffll and Van der Ploeg 1995) or due to the fact that lower export costs shft the sectoral composton of an economy towards the traded goods sector whch s affected by foregn (wage) competton (e.g. Andersen et al. 2000, uznga 1993). 2 One lne of argument concerns the degree of centralzaton n wage formaton whch may decrease (procompettve response) or ncrease (nternatonal cooperaton of unons) n the face of product market ntegraton and whch n tself s related to wages n a non-monotonous way (e.g. Drffll and Van der Ploeg 1993, 1995; Danthne and unt 1994). Buldng on a model of recprocal nternatonal trade n homogeneous goods, Naylor (1998) provdes another argument whch exemplfes that reductons n trade costs lead to hgher wages.

5 2 notably n the EU (Andersen et al. 2000). owever, dfferences n wage levels and socal polces among ntegratng countres are stll szeable and there s no emprcal support for the race to the bottom hypothess. Ths seems to justfy the concluson that there s stll a role for domestc wage and socal polces. The focus of ths paper s on the moblty of captal and frms and notably on the ncentves of ndustry to relocate between regons n the process of economc ntegraton. The queston posed n ths paper s thus related but slghtly dfferent from the one tackled n the prevous lterature. In fact, n order to focus on the relocaton ncentves of frms, the desre of unons to adjust ther wages, s delberately dsallowed for n the present analyss. In the theoretcal model used below ths s acheved by postulatng a rgd unon wage whch s ted up wth the tax-beneft system. Moreover, n contrast to the prevous lterature, the present paper draws on the new economc geography. 3 The theoretcal analyss s bult around a smple model of monopolstc competton wth frm and captal moblty whch has the vrtue that t s analytcally tractable and yelds closed-form solutons. There are two ntally dentcal countres whch are connected through ntra-ndustry trade and frm and captal moblty. The analyss starts wth the assumpton that frms and captal are moble nternatonally but ther entrepreneural owners are not. ence, all rents from captal nvested abroad are repatrated. The proportonate ndustry relocaton effect assocated wth a unon wage push n one of the two countres s then derved analytcally. It s shown that ths effect ncreases wth the level of economc ntegraton as measured by trade costs. Ths confrms the tradtonal vew whch underles the publcs fears and some of the results n the prevous lterature. By assumng that entrepreneural owners of captal are moble nternatonally, the agglomeratve forces stressed n the new economc geography are then brought nto play n the model. When, ntally, the two countres are dentcal and the ndustry s splt evenly, or when there s only a partal agglomeraton of ndustry, a wage push n one country s shown to have a smlar (strong) relocaton effect as expressed n the tradtonal vew. owever, a new economc geography perspectve suggests another possblty, too. Suppose that, for some reason, one country has hstorcally emerged as the core,.e. s the host of all ndustry. Then, an agglomeraton rent accrues to the moble factor. Ths rent s the focus of recent publc fnance analyses whch show that t can be taxed drectly by governments by means of captal taxes. 4 A central element of the present analyss s to lnk ths nsght to the debate about the 3 See Fujta et al. (1999) and Neary (2001) for recent expostons and assessments. 4 See Andersson and Forsld 1999, Baldwn and Krugman 2000, Knd et al and Ludema and Wooton 2000.

6 3 labour market effects of product market ntegraton. In partcular, t s shown that the mmoble unonsed labour force s able to extract ths rent ndrectly. Indeed, unons n the core can have hgher wages and governments n the core can choose more generous welfare polces - than ther counterparts n the perphery. More specfcally, I derve the maxmum nternatonal unon wage dfferental that can be mantaned at each stage of product market ntegraton (.e. at each level of trade costs) wthout nducng an ext of ndustry. Ths relatonshp s bell-shaped and acheves ts peak at an ntermedate stage of product market ntegraton. ence, ths result provdes a ratonale for the persstence of szeable dfferences n wage levels and socal polces even n regons whch are well-ntegrated. Moreover, ths result mples that, nternatonal ntegraton notwthstandng, there s some natonal autonomy n wage and socal polces when hstorcal agglomeraton advantages are taken nto account. The paper s organsed as follows. The monopolstc competton model whch underles the analyss s ntroduced n Secton 2. Secton 3 studes the ncentve of entrepreneurs to relocate ther frms and captal when tradtonal moblty assumptons are met. Secton 4 turns to the new economc geography perspectve. The dscusson n secton 5 sets the results nto a wder perspectve and puts the model s assumptons under closer scrutny. Secton 6 concludes. 2. The Model 2.1. Basc structure I consder a trade and geography verson of the Spence-Dxt-Stgltz model of monopolstc competton (Spence 1976, Dxt and Stgltz 1977, Krugman 1980, 1991) whch obtans ts analytcal ease by the mposton of two assumptons. Frst, by assumng that one unt of captal s needed to run a frm as n Flam and elpman (1987), the model becomes analytcal tractable even n a geography context (Forsld 1999). Second, a sgnfcant further smplfcaton s obtaned by assumng that consumers upper-ter utlty s quas-lnear so that all ncome effects are channelled to the compettve outsde sector (as e.g. n Drffll and Van der Ploeg 1995, Martn and Rogers 1995 and Pflüger 2001, 2003). 5 The world s composed of two countres, home and foregn (denoted by an astersk *) whch are assumed to have dentcal preferences, technology, factor endowments and trade costs but possbly dfferent beneft-tax systems. There are three factors of producton, non-unonsed 5 Ottavano and Thsse (1998) provde another analytcally solvable model whch assumes a quadratc utlty functon.

7 4 labour ( L ), unonsed labour ( ) and entrepreneural captal ( K ), and two sectors, manufacturng ( X ) and an outsde sector ( A ). Non-unonsed labour s ntersectorally moble at the natonal level but mmoble at the nternatonal level. Unonsed labour s nternatonally mmoble and ether employed n the manufacturng sector or unemployed. Captal s assumed to be nternatonally moble throughout. owever, ts entrepreneural owners are so only n the geography verson of the model (secton 4). Ths ntroduces the agglomeratve forces whch are the hallmark of the new economc geography. The outsde good s homogeneous, traded wthout costs and produced perfectly compettvely under constant returns wth non-unonsed labour only. Ths good s the numérare and assumed to be produced n both countres after trade. The monopolstcally compettve X sector employs all three factors to produce dfferentated goods wth a lnear cost functon. The two types of labour are varable nputs. Entrepreneural captal enters only the fxed cost. One unt of t s needed to produce at all. Trade n X s nhbted by ceberg costs. Whereas the compensaton of non-unonsed labour and of entrepreneural captal are determned on compettve markets, unonsed labour s wage s rgd and related to the beneft-tax system Utlty and demand There are L K households, each of whom supples one unt of non-unonsed or unonsed labour and entrepreneural captal, respectvely. Ther compensaton and ncome before taxes are denoted by W, L W and R, respectvely. Those parts of unonsed labour whch do not fnd a job are elgble for a per-capta unemployment beneft, B. Each household's preferences are charactersed by a quas-lnear upper-ter utlty as n Dxt (1990): U lnc X C A, N 1 N* 1 1 x x j 0 N C X, 0, 1 (1) where C X s the manufacturng aggregate as n Dxt and Stgltz (1977), C s the A consumpton of the outsde good, x (x ) s the quantty consumed of a domestc varety j (foregn varety j ), N and N * are the number of varetes produced n home and foregn and s the elastcty of substtuton between manufacturng varetes. The budget constrant s gven by PC X C Y, P NP N P 1, 1 (2) A * j

8 5 where Y denotes a household s ncome, P s the perfect CES-prce ndex, P (P ) s the j prce set by a domestc (foregn) frm. Iceberg transport costs are formalsed by the constant and mply that only 1 / of a unt of a foregn varety arrves for consumpton. The consumer prce of an mported varety s gven by P. Utlty maxmsaton yelds the demand functons and ndrect utlty, V, as expressed n the followng: j C X P, C A Y, x P P, x j Pj P (3) ln 1 V ln P Y (4) 2.3. Producton The producton functon of the outsde good s X A L A where X A and LA denote ouput and the nput of non-unonsed labour, respectvely. Perfect competton ensures that ths good s prced at margnal (and also average) cost. Snce ths good s the numérare, the wage rate of non-unonsed labour s unty, W 1. L Each product type s suppled by a sngle frm whch ncurs fxed costs R due to the requrement of one unt of entrepreneural captal n order to be able to produce at all. Producton then takes place accordng to the Cobb-Douglas functon X l h 1 where X denotes the frm s producton level and l and labour, respectvely. Margnal costs are gven by ts nputs of non-unonsed and unonsed 1 1 W varable costs attrbutable to non-unonsed labour (unonsed labour) s gven by h c 1. The share of ( 1 ). The frm s condtonal nput demand for non-unonsed labour and unonsed labour s gven by l and h 1 cx, respectvely. Let the producer prces charged to cx W 1 domestc (foregn) households be denoted P * (P ). Notng that part of demand s ndrect, caused by transport losses, and denotng the demand of a foregn household by * x, profts of the representatve frm n the home regon,, are gven by: * P cl K x P cl * * K * x R * (5) Wth the Chamberlnan large group assumpton, proft maxmsng prces on both markets are constant mark-ups on margnal costs (whch are ncreasng n unon wages)

9 P 6 / 1 P * c (6) The compensaton of human captal adjusts so as to ensure zero proft equlbrum. Usng the * * * * K x market clearng condton, X L K x L for domestc varety, the followng relatonshp between frm scale X and fxed costs R s obtaned: X R 1/ c (7) 2.4 Wages The compensaton of both non-unonsed labour and of entrepreneural captal s determned on compettve markets as just descrbed. In contrast, the compensaton of unonsed labour, W, s rgd and determned by purely exogenous parameters of the tax-beneft system 6 W 1 t 1 t B W B (8) where t W denotes the proportonal tax rate on wage ncomes of unonsed and non-unonsed labour and t denotes the proportonal tax rate on unemployment benefts. The parameter s constant and provdes a measure for the net replacement rato (.e. 1 t B B / 1 t W W ). Eq. (8) captures n a smple way that wages are ted up wth socal polces, the constancy of the net replacement rato beng reasonable n the longer-run (e.g. Nckell and Layard 1999, Layard et al. 1991). It follows mmedately from (8) that, ceters parbus, the unon wage s hgher, the hgher the beneft level and the wage tax and the lower the tax on unemployment benefts. The level of unemployment s gven by Nh where the number of frms equals the captal stock nstalled domestcally. Government budget balance, B Nh 1 t B B t W L W h N W unemployed are fnanced by the taxes on wage ncomes and benefts. L, commands that the outlays for the 6 Eq. (8) can be gven smple mcrofoundatons by assumng that unons choose wages so as to maxmse a Stone- Geary utlty functon, h W 1 t B1 t W B ln h / lnw (e.g. Booth 1995) and by notng that the wage elastcty of labour demand s gven by 1 wth Cobb-Douglas technologes and a constant prce elastcty of demand for each varety gven by. The constant parameter n (8) s then gven by 1 / 1. Thanks to Erc Toulemonde for pontng out to me ths way to arrve at (8).

10 7 3. Revstng the tradtonal vew Suppose that the two countres whch share dentcal preferences, technology, factor endowments and trade costs but have possbly dfferent beneft-tax systems engage n free trade and allow entrepreneural captal to move freely across borders. owever, assume that the entrepreneural owners of captal are mmoble nternatonally, as tradtonally assumed. All rents from captal nvested abroad are therefore repatrated. The nternatonal equlbrum under these moblty assumptons s determned by the condtons that captal rents are equalsed nternatonally, R R*, that there s captal market clearng nternatonally, N N* 2K, and that product markets clear and profts of domestc and foregn frms vansh n long-run equlbrum as stated n the followng two condtons:7 L K L * * K * R N * N * N N * L K L * * K * * R* (9) N * N * N N * 1 1 where W W * W * / W Solvng these condtons yelds: / 1 1 * N K 1 2 * * 1 1 * 1 2 * 1 1 * N * K R L / K (10) The output level of frms can then be derved from (7) and all other endogenous varables follow mmedately. If the regons are equal wth respect to ther tax-beneft systems and, hence, f regonal unon wages are equal, then t follows mmedately that * and from (10), N N* K. Each regon has half of the entrepreneural captal and frms, accordngly. Suppose now that unons n one regon demand a hgher wage, say because hgher unemployment benefts are granted whch n turn are fnanced by an ncrease n wage taxes. Ths wage polcy wll clearly have an effect on the locaton of frms. Around the symmetrcdentcal equlbrum ths relocaton effect can be derved from (10) as: N / N W / W (11) 7 To arrve at these condtons one has to mpose the zero proft condton on (5), and to use (2), (3), (6) and the expresson for margnal costs together wth ther foregn counterparts.

11 8 The sze of ths relocaton effect ncreases wth the cost share of unonsed labour n varable costs, ( 1 ), and wth the nverse of the elastcty of substtuton, 1 /, whch s an ndcator for the returns to scale n zero proft equlbrum n ths type of model 8 More mportantly, the relocaton effect s negatvely related to the level of trade costs,. The lower these are, the stronger the movement of frms out of ths regon and nto the other for any gven proportonate ncrease n unon wages. The relocaton effect goes to nfnty as trade costs vansh ( 1). ence, n a world where frms produce wth ncreasng returns to scale and where captal and frms are moble, a unon wage push n one regon has a strong effect on ndustry locaton. Note n partcular, that ths effect ncreases wth the level of economc ntegraton as proxed by trade costs. Ths result underles the tradtonal vew whch stpulates that countres wll necessarly have to stck to dentcal (or, at least, smlar) wage and socal polces or they wll experence a massve relocaton of frms. 4 The new economc geography vew 4.1 Moble entrepreneurs Suppose now, n contrast to the prevous secton, that the entrepreneural owners of captal are moble nternatonally, too. When entrepreneurs move across borders along wth ther captal, ther captal rents (ncomes) wll no longer be repatrated but accrue to the regon where the entrepreneurs locate. Ths opens up agglomeratve forces of the new economc geography. To study the ncentve of entrepreneurs to move across borders, t s necessary to derve and compare ther ndrect utlty levels assocated wth lvng n one or the other country for a gven short-run allocaton of the world-wde captal stock. By symmetry, N* K *, n the short-run. Equatons (9) can therefore be reformulated to read: N K and L K L * * K * R (12) K * K * K K * * R* K * K * L K L * * K * K K * 8 As Neary (2001: 548) notes, t s a qurk of the Dxt-Stgltz model that a standard measure of the degree of returns to scale, the rato of average costs to margnal costs, depends on the elastcty of substtuton. To see ths, note that prces equal average costs n zero proft equlbrum and prces are constant markups on margnal costs accordng to (7). ence, the rato of average to margnal costs s gven by / 1. See also elpman and Krugman (1985, Chapt. 7) on ths.

12 9 Because of the quas-lnear upper ter utlty functon, the compensaton for entrepreneural captal n home and foregn, R and R*, n such a short-run equlbrum can drectly be read of (12). In the long run, entrepreneurs move to the country where they obtan the hgher level of ndrect utlty. The utlty dfferental, V V* ln P * / P R R *, can be derved analytcally for general trade costs n ths model. If t s assumed that countres are dentcal except for unon wages and beneft-tax systems and f all endowments are normalsed to unty, L L* * K K* 1, ths utlty dfferental reads 1 1 * 1 * 1 1 / 2 1 V V* ln ln (13) 1 1 * 1 where K / K K * denotes the share of the world-wde captal stock located n home. A long-run equlbrum n whch both regons produce dfferentated goods s gven when V V* 0. When regons are dentcal n all respects, 1/ 2 s such an equlbrum. owever, ths symmetrc equlbrum s not necessarly stable because the model contans two agglomeratve forces, a supply lnkage and a demand lnkage. The supply lnkage arses because the country wth the hgher share of entrepreneural captal has a larger manufacturng sector and therefore a lower prce ndex. Ths s captured n the frst term n (13). There s a demand lnkage because ncreasng the share of entrepreneural captal n one country nduces an ncrease n ncome and, hence, n market sze. Ths rases the proftablty of frms n ths country as expressed n R R* and thus attracts even more entrepreneural captal. The demand lnkage s captured n the second term n (13). owever, shftng frms from the foregn to the domestc economy ncreases competton among frms for gven expendtures on domestc products whle lowerng competton n the foregn market, thereby reducng the proftablty of the domestc market n relaton to the foregn market. Ths local competton effect s a deglomeratve force. Trade costs affect the balance between the two types of forces. When they are large enough, they render the symmetrc equlbrum stable. owever, when trade costs are contnuously reduced, the symmetrc equlbrum becomes unstable. Then two stable and ncreasngly asymmetrc equlbra emerge n whch a larger part (and, fnally, all) of the dfferentated goods ndustry s located n one or the other country. Asymmetrc equlbra emerge at trade costs 2 / The level of trade costs at b whch full specalsaton arses s mplctly gven by the condton ln f 2 1 f 1 f 1 / 3. ence, the model exhbts a smooth 'supercrtcal ptchfork f

13 10 bfurcaton'. 9 Ths s llustrated n Fgure 1 whch graphs the equlbrum shares of captal n each country aganst the level of trade costs. Fgure 1: Bfurcaton dagram 1,2 1 0,8 0,6 0,4 0,2 0-0,2 f b 9 The standard core-perphery model of Krugman (1991) exposed n great detal n Fujta et al (1999), n contrast, exhbts a chaotc subcrtcal ptchfork bfurcaton,.e. at a certan level of trade costs (the so-called sustan pont) two stable fully agglomeratve equlbra emerge n addton to the ntally stable symmetrc equlbrum. At a stll lower level of trade costs (the so-called break pont) the symmetrc equlbrum becomes unstable. The dfferent bfurcaton pattern n the present model derves from the fact that the logarthmc quaslnear utlty functon elmnates ncome effects from the manufacturng sector and therebye weakens the demand lnkage as compared to the Cobb-Dougals upper ter utlty used n the standard core perphery model. A detaled dscusson of ths dfference s provded n Pflüger (2003). See Fujta et al. (1999) for an accessble ntroducton to bfurcatons and Grandmont (1988) for a techncal treatment.

14 11 Fg. 1 llustrates that an analyss of a domestc wage hke, startng from an nternatonal unon wage dfferental of unty, can be performed for three ntal stuatons, a symmetrc equlbrum, an equlbrum wth partal agglomeraton and an equlbrum wth full agglomeraton. Start wth a symmetrc equlbrum. A domestc wage hke whch rases the nternatonal unon wage dfferental above unty leads to a relocaton of frms to the foregn country. The same result holds true for an ntal equlbrum wth partal agglomeraton of frms. These results can be establshed analytcally by mposng the condton V V* 0 on (13) and then by dervng ln / lnw 0 through mplct dfferentaton. 10 In ths sense, the tradtonal vew s confrmed n ths new economc geography settng. 4.2 A core advantage Suppose now, however, that for some reason an agglomeraton process has already occurred hstorcally and that the whole ncreasng returns to scale ndustry s located n, say, the home country. Then, an agglomeraton rent accrues to the nternatonally moble entrepreneurs n the core. For the present model, ths agglomeraton rent can be calculated by takng the * effect of the nternatonal unon wage dfferental W / W on and * nto account, and by usng 1 n (13): V V* ln (14) 1 Intutvely, the sze of ths agglomeraton rent s postvely related to, the ncome fracton devoted to dfferentated goods (see eq. (3)), to, the share of varable costs attrbutable to non-unonsed labour and to 1 /, the degree of returns to scale n the dfferentated goods ndustry. The relatonshp between the agglomeraton rent and trade costs s bell-shaped. The agglomeraton rent s zero at the border value of trade costs at whch the agglomeraton n one country swtches from partal to full (e.g. to 1). Ths rent s zero, too, when trade costs vansh ( 1) and locaton does not matter anymore. At some ntermedate trade costs, the agglomeraton rent s hghest. Crucally, (14) shows that the agglomeraton rent s related to the nternatonal unon wage dfferental,. Ths suggests that unons and governments whch are located n the core can explot the agglomeraton rent and therefore choose hgher benefts, taxes and wages than ther counterparts n the perphery. Ths scope s lmted by the condton that the utlty dfferental V V * n (14) be non-negatve. Imposng ths condton 10 The analytcal expresson s rather unweldy and for that reason not reported here.

15 yelds a closed-form soluton for the maxmum nternatonal unon wage dfferental whch can be mantaned wthout havng to fear an ext of ndustry: M M 1 2 (15) 1 3 ln The maxmum wage dfferental s ncreasng n the degree of returns to scale, 1 /, and n non-unonsed labour s share of varable costs,. Fgure 2 plots the maxmum unon wage dfferental aganst the level of trade costs. As ths graph reveals, the dfferental s bellshaped. Ths suggests that, n the process of economc ntegraton, the ablty of unons n the core to charge a hgher wage than unons n the perphery frst rses, then reaches a peak and fnally dssolves when trade costs are nl. The level of trade costs at whch the maxmum wage dfferental tself reaches a maxmum s mplctly determned by the condton ln / 21 3 ln. Intutvely, the bell-shaped relatonshp between trade costs and the wage dfferental reflects the behavour of the agglomeraton rent whch constrans the core unon s ablty to acheve hgher wages. Fgure 2: Maxmum nternatonal unon wage dfferental and trade costs M 0 f 11 The mgraton rule used n ths paper s based on the assumpton that entrepreneurs are myopc,.e. have statc assumptons and thus base ther mgraton decson on the current ndrect utlty dfferental. ence, no ndvdual entrepreneur would have an ncentve to move to the perphery wth the wage dfferental specfed n (15). owever, ths stuaton offers a proft opportunty for large forward lookng agents (smlar as n enderson s analyss of urban systems (e.g. Fujta et al. Chapt. 2)). By recurrng to the assumpton of myopc agents and (unmodelled) transacton costs the present analyss precludes that the entrepreneurs are able to coordnate on such a jont move to the other country. I am grateful to a referee to makng me aware of ths pont.

16 13 Wth parameter values of 0. 5 and 6, unons n the core can mantan a maxmum wage premum of 10% (at ntegraton level ) vs-à-vs unons n the perphery. When economes of scale are as strong as assumed n Baldwn and Krugman s (2000) baselne scenaro ( 2.5 ), ths maxmum premmum rses to 55% (at ). Ths analyss suggests that all nternatonal ntegraton notwthstandng, there s some natonal autonomy n wage and socal polces when hstorcal agglomeraton advantages are taken nto account. 12 ence, ths paper provdes an explanaton for the persstence of szeable dfferences n wage levels and socal polces even n regons whch are well-ntegrated. Trade costs n ths model can be thought of as comprsng both artfcal trade barrers, such as tarffs or product standards, and natural barrers, such as transport costs or cultural dfferences. Even though artfcal barrers could possbly be elmnated altogether between ntegratng countres, some natural barrers wll, all past and future transport cost revolutons notwthstandng, surely reman. A complete levellng of wage and socal polces n an ntegratng world n whch countres start wth dfferent ntal condtons s therefore unlkely. Even f the focus of the paper s on wage and socal polces, what can be sad about unemployment? In the perphery, all unonsed labour s unemployed. In the core, unemployment s gven by Nh where N K K * and where the demand for labour of the representatve domestc frm s accordngly gven by h / / W. 13 In a process of economc ntegraton, ths demand s governed by two forces: trade costs are fallng whch per se ncreases the demand for unonsed labour by ncreasng foregn demand (through (3)) whereas W frst ncreases and then decreases accordng to (15) and so frst reduces and then ncreases the demand for unonsed labour. ence, ntally the two forces are counteractng. When the peak of the wage dfferental n fgure 2 s reached they jontly contrbute to an ncrease n demand for unonsed labour and hence a reducton of unemployment n the core. 12 A smlar result s derved by Munch (1999) n an nternatonal Cournot model of the Brander-Krugman-type. 13 Ths dervaton uses the market clearng condton X and (2), (3), (6) L K K x L * * x * * and the normalsaton of all endowments to unty to determne the producton level, X, gven margnal cost. Applyng the condtonal demand for unonsed labour as stated n secton 2.3 gves the result above.

17 14 5 Dscusson Accordng to ths analyss the most sgnfcant dfference between a world whch takes the exstence of agglomeraton forces nto account and a world whch does not s that the former allows for some natonal autonomy n wage and socal polces. Ths result requres that the manufacturng sector has agglomerated n one country, or, to put t dfferently, that the perphery s completely dendustralsed. Accordngly, all unonsed labour n the perphery s unemployed, clearly an extreme case. Ths s a consequence of the parsmonous specfcaton of the model, however. The followng dscusson sets ths result n a wder perspectve and puts some assumptons under closer scrutny. The complete dendustralsaton of the perphery results from the assumpton of only two producton sectors and, n a sense, t parallels the famlar result of extreme specalsaton n the smple two-good Rcardo model. As n extensons of that model (e.g. Bhagwat et al. 1998), t would be natural to allow for many manufacturng sectors. In the geography context ths s done by Venables (1999) and Fujta et al (1999, Chapt. 16). In these works t s shown that wth such extensons the emergence of ndustral clusters can be explaned,.e. countres then specalse on a selecton of manufacturng sectors and the dstrbuton of sectors between countres need not be equal (.e. need not necessarly be half and half). If we consder such a mult-sector world, and thnk of entrepreneural captal as beng sector-specfc and of unons beng sectoral ones too, our result would suggest that there can be cross-country wage dfferentals n partcular sectors. The country whch, n a process of economc ntegraton, obtans the larger share of manufacturng sectors then ends up wth the hgher average unon wage. Wth respect to unemployment, t should be noted that underlyng the analyss s a presumpton that the supply of unonsed labour s always larger then the (aggregate) demand for labour at the varous levels of trade costs and assocated wage rates. If ths assumpton, a smplfcaton to avod the ssue of ratonng, was not met, the rgdty of wages, could not be mantaned and one would have to face the fact that the wage would be bd up from the labour demand sde. Ths would create dffcult nteractons wth the man analyss. owever, there s a smple alternatve to exclude ratonng. Ths alternatve s to follow Drffl and van der Ploeg (1995) and to assume that unemployed unon members can work n the constant returns L outsde sector at the wage W 1 prevalng there. Such a model would dspose of the taxbeneft system n (8) and so place the focus on wage polces only but assume nstead that

18 15 unon wages are smple mark-ups on ths outsde opton. Ths would leave the remanng analyss ntact. In such a model there would be no unemployment snce the Nh unonsed workers not employed n the manufacturng sector could resort to a job n the outsde sector. owever, they would clearly experence a welfare loss by beng forced to work for a lower wage. Fnally, note that the ssue of unemployment would be a much less severe one (notably for the perphery) n a mult-sector extenson of the present model as suggested above. A fnal concern s trade costs. The model assumes that the outsde-good can be traded wthout cost whereas ceberg costs have to be ncurred n cross border trade of dfferentated goods. Ths may not seem to be an nnocuous assumpton because, for a smlar model, Davs (1998) has shown that the ntroducton of such costs for the good produced under constant return to scale elmnates the market sze effect on whch the demand lnkage n geography models s based (e.g. Neary 2001). owever, as shown by Fujta et al. (1999 Chapt. 7.4), when the output n the constant returns sector s dfferentated across countres, the market sze effect and hence the demand lnkage s re-establshed even wth trade costs n the constant returns sector. In lght of ths result, the assumpton of no trade costs n the outsde-good can be defended as smply beng a convenent modellng short-cut. 6 Conclusons In ths paper, a smple model of monopolstc competton wth moble captal and frms s set up as an ntegratve theoretcal framework for two vews n the debate about the vablty of natonally ndependent wage and socal polces. The domnant vew n the publc debate holds that those countres whch mantan hgher unon wages and more generous welfare systems than others (all other factors equal) experence an ncreasng ext of frms as product market ntegraton proceeds. Ths vew s at odds wth emprcal fndngs whch show that szeable dfferences reman even n regons whch have acheved hgh levels of ntegraton. Ths paper ponts to one theoretcal ratonale for ths fndng. Agglomeraton advantages whch have bult up n the past allow unons n the core to have hgher wages than unons n the perphery. The scope to have hgher wages s constraned, however, and related to the level of trade costs n a bell-shaped way. In order to work out ths argument n the possbly smplest way, ths paper has used the Flam- elpman (1987) specfcaton of the model of monopolstc competton and has followed the Krugman (1991) factor moblty tradton to model agglomeratve forces. Due to these

19 16 assumptons, smple closed-form solutons could be provded for all results. The arguments of ths paper could alternatvely have been presented n a model followng the tradton establshed by Venables (1996) and Krugman and Venables (1995). They provde a new economc geography model where the agglomeratve forces are due to the proxmty of an ntermedate goods ndustry to a fnal goods ndustry. owever, ths alternatve framework would have made t necessary to rely on numercal solutons. There are several avenues for future work some of whch were already spelled out n the prevous secton. Relaxng the assumpton of an exogenously determned wage, s arguably the most mportant one. Ths would allow to address the compettve choce of wage and socal polces n the core and the perphery.

20 17 References Andersen, T.M., aldrup, N. and Sorensen J.R. (2000). Labour market mplcatons of EU product market ntegraton, Economc Polcy, 15, Andersson, F. and Forsld, F. (1999). Tax competton and economc geography, CEPR Dscusson Paper 2220 Baldwn, R.E. and Krugman, P. (2000). Agglomeraton, ntegraton, and tax harmonzaton, CEPR Dscusson Paper 2630 Bhagwat, J. N., Panagarya, A. and Srnvasan, T.N. (1998). Lectures on nternatonal trade. Second Edton. The MIT Press, Cambrdge, Massachusetts Booth, A. (1995). The economcs of the trade unon. Cambrdge Unversty Press Danthne, J.-P. and unt, J. (1994). Wage barganng structure, employment and economc ntegraton, Economc Journal, 104, Davs, D. (1998). The home market, trade, and ndustral structure, Amercan Economc Revew, 88, Dxt, A. (1990). Optmzaton n economc theory. Second Edton, Oxford Unversty Press. Dxt, A. and Stgltz, J.E. (1977). Monopolstc competton and optmum product dversty, Amercan Economc Revew, 67, Drffl, J. and Van der Ploeg, F. (1993). Monopoly unons and the lberalsaton of nternatonal trade, Economc Journal, 103, Drffl, J. and Van der Ploeg, F. (1995). Trade lberalzaton wth mperfect competton n goods and labour markets, Scandnavan Journal of Economcs, 97, Flam,. and elpman, E. (1987). Industral polcy under monopolstc competton, Journal of Internatonal Economcs, 22, Forsld, R. (1999). Agglomeraton wth human and physcal captal: An analytcally solvable case, CEPR Dscusson Paper 2102 Fujta, M., Krugman, P.R. and Venables, A.J. (1999). The spatal economy. Ctes, regons, and nternatonal trade, MIT-Press, Cambrdge, Mass. Grandmont, J.M. (1988). Nonlnear dfference equatons, bfurcatons and chaos: An ntroducton. Lecture Notes No.5, May, IMSSS-Economcs Lecture Notes Seres, Insttute for Mathematcal Studes n the Socal Scences, Stanford, CA elpman, E. and Krugman, P.R. (1985). Market structure and foregn trade, The MIT- Press, Cambrdge, Mass. uznga,. (1993). Internatonal market ntegraton and unon wage barganng, Scandnavan Journal of Economcs, 95, Knd,.J., Mdelfart Knarvk, K. and Schjelderup, G. (2000). Competng for captal n a lumpy world, Journal of Publc Economcs, 78, Krugman. P.R. (1991). Increasng returns and economc geography, Journal of Poltcal Economy, 99,

21 18 Krugman, P.R. and Venables, A.J. (1995). Globalzaton and the nequalty of natons, Quarterly Journal of Economcs, 60, Layard, R., Nckell, S. and Jackman, R. (1991). Unemployment. Macroeconomc performance and the labour market, Oxford Unversty Press Ludema, R. and Wooton, I. (2000). Economc geography and the fscal effects of regonal ntegraton, Journal of Internatonal Economcs, 52, Martn, P.J. and Rogers, C.A. (1995). Trade effects of regonal ad, n: R. Baldwn, P. aaparanta and J. Kander (eds), Expandng Membershp of the European Unon, Cambrdge Unversty Press, Munch, J. R. (1999). The Locaton of Frms n Unonzed Countres, Aarhus Department of Economcs Workng Paper 1999:12 Naylor, R. (1998). Internatonal trade and economc ntegraton when labour markets are generally unonsed, European Economc Revew, 42, Neary, P.J. (2001). Of hype and hyperbolas: Introducng the new economc geography, Journal of Economc Lterature, 39, June, Nckell, S. and Layard, R. (1999). Labor market nsttutons and economc performance, n: O. Ashenfelter and D. Card (eds), andbook of Labor Economcs, Vol. 3, Ottavano, G. and Thsse, J.-F. (1998). Agglomeraton and trade revsted, CEPR Dscusson Paper 1903 Pflüger, M. (2001). Ecologcal dumpng under monopolstc competton, Scandnavan Journal of Economcs, 103, Pflüger, M. (2003). A Smple, Analytcally Solvable, Chamberlnan Agglomeraton Model, revsed verson of IZA DP 359, forthcomng n: Regonal Scence and Urban Economcs Rodrk, D. (1997). as globalzaton gone too far? Insttue for Internatonal Economcs, Washngton D.C. Spence, M.E. (1976). Product selecton, fxed costs, and monopolstc competton, Revew of Economc Studes, 43, Venables, A.J. (1996). Equlbrum locaton of vertcally lnked ndustres, Internatonal Economc Revew, 37, Venables, A.J. (1999). The nternatonal dvson of ndustres: Clusterng and comparatve advantage n a mult-ndustry model, Scandnavan Journal of Economcs, 101,

22 IZA Dscusson Papers No. Author(s) Ttle Area Date 733 S. Wolter S. Denzler Wage Elastcty of the Teacher Supply n Swtzerland 2 03/ S. Wolter Sblng Rvalry: A Sx Country Comparson 2 03/ R. Desmet A. Jousten S. Perelman P. Pesteau Mcro-Smulaton of Socal Securty Reforms n Belgum 3 03/ I. Bolvg P. Jensen M. Rosholm The Employment Effects of Actve Socal Polcy 6 03/ A. L. Booth M. Francescon G. Zoega Unons, Work-Related Tranng, and Wages: Evdence for Brtsh Men 3 03/ V. Grossmann Manageral Job Assgnment and Imperfect Competton n Asymmetrc Equlbrum 739 M. Fertg Who's to Blame? The Determnants of German Students' Achevement n the PISA 2000 Study 1 03/ / B. T. rsch S. L. Mehay 741 B. T. rsch D. A. Macpherson Evaluatng the Labor Market Performance of Veterans Usng a Matched Comparson Group Desgn Wages, Sortng on Skll, and the Racal Composton of Jobs 6 03/ / R. A. Easterln Buldng a Better Theory of Well-Beng 3 03/ G. eneck J. Schwarze 744 M. R. West L. Woessmann Substance Use and Earnngs: The Case of Smokers n Germany Whch School Systems Sort Weaker Students nto Smaller Classes? Internatonal Evdence 5 03/ / L. Woessmann Educatonal Producton n East Asa: The Impact of Famly Background and Schoolng Polces on Student Performance 2 03/ A. Ammermueller. ejke L. Woessmann Schoolng Qualty n Eastern Europe: Educatonal Producton Durng Transton 4 03/ J. Messna Sectoral Structure and Entry Regulatons 3 03/ M. Pflüger Economc Integraton, Wage Polces and Socal Polces 2 04/03 An updated lst of IZA Dscusson Papers s avalable on the center s homepage