European integration and inequality among countries: a lifecycle income analysis

Transcription

1 MPRA Munich Personal RePEc Archive European inegraion and inequaliy among counries: a lifecycle income analysis José M. Pasor Monsálvez and Lorenzo Serrano Insiuo Valenciano de Invesigaciones Económicas, Universia de València 2009 Online a hps://mpra.ub.uni-muenchen.de/18510/ MPRA Paper No , posed 11 November :00 UTC

2 European inegraion and inequaliy among counries: a lifecycle income analysis * José M. Pasor Lorenzo Serrano (Universia de València & Ivie) Absrac We analyze he effecs of he expansions of he European Union on inequaliy using an approach based on individuals' lifecycle incomes. This allows us o consider he effec of differen raes of growh and survival raes. This differs form he usual analyses of inequaliy ha focus on he evoluion of curren per capia income for he period. Our resuls show ha inequaliy in erms of permanen income was subsanially less han in curren per capia income a he ime of all he expansions excep hose of he las en years. The resuls poin o he key role of policies ha simulae growh in he less developed counries. Wih an annual β-convergence of 2% in curren income, inequaliy in permanen income would be one hird lower. *The auhors wish o hank he Miniserio de Educación y Ciencia-FEDER (SEC /ECON) for heir financial suppor.

3 1. Inroducion Wih he enry of Rumania and Bulgaria in 2007 he European Union (EU) reached he figure of 27 members and culminaed, for he momen, a process of expansion begun half a cenury earlier when a group of six counries ook he firs seps in he projec of European uniy, from which subsanial economic and social benefis were expeced. The road ravelled has no been wihou difficulies, bu i is fair o acknowledge ha he resul of he successive expansions, always a he volunary reques of he candidae counries, shows ha he advanages have clearly been greaer han he possible disadvanages. The economic dimension of he projec of European uniy has been fundamenal from he sar. Indeed, he inegraion projec has frequenly been criicised for he excessive weigh of economic aspecs o he derimen of oher maers. This is undoubedly due in par o he subsanial economic benefis ha he counries expeced (and sill expec) o obain in he long erm as a consequence of he process of inegraion. A he presen ime hese economic advanages associaed wih inegraion coninue o be a consan incenive o progress along he pah underaken, in spie of he difficulies. 1 One of he reasons given o jusify he benefis of economic inegraion is ha he progressive expansion of he markes for goods, services and facors of producion (capial and labour) will generae a greaer and more efficien use of European counries' resources, simulaing he economic developmen of he area. Few doub he long erm advanages of economic inegraion for he member counries; however, he process of inegraion raises a series of quesions ha deserve an adequae response. In paricular: 1) Do all member counries benefi equally? 2) Can he more developed counries beer exploi he advanages of a wider marke? Or, on he conrary, 3) Can he less developed counries, due o heir lower labour coss, benefi o a greaer exen from he enry of capial, foreign invesmen and he ransfer of echnology? 1 In fac, afer he 2004 enlargemen, in mid-2005, French and Duch voers rejeced he Treay esablishing a Consiuion for Europe in naional referendums, hus hrowing ino doub he EU s abiliy o work effecively and drive inegraion forward. A he same ime, public scepicism abou fuure enlargemens began o moun and his opposiion was indeed one of he reasons cied by Duch and French opponens of he Consiuional Treay. Many people in he old EU Member Saes hink ha he EU has no ye successfully digesed he 2004 enlargemen. They feel ha he addiion of he Cenral and Easern European counries has changed he naure of he Union

4 Answering hese quesions implies analyzing he economic convergence of he member counries of he EU. The fundamenal quesion is wheher European inegraion gives rise o a paern of growh ha generaes among he member counries convergence, greaer differences, or has no appreciable effec in his regard. The analysis of his phenomenon is of special relevance since one of he explici objecives of he EU is convergence among is counries and regions, and o his end i has devoed and inends o devoe a large par of is budge hrough insrumens such as he FEDER or he Cohesion Fund. The successive expansions, analyzed in greaer deail laer, have affeced boh counries already highly developed (e.g., Unied Kingdom, Denmark or Sweden) and ohers iniially less developed (e.g. Ireland, Greece, Spain or he counries of Easern Europe). Obviously, hough perhaps for differen reasons, boh ypes of counries expec o benefi from joining he EU. A he empirical level here is lile lieraure abou he effecs of he process of European inegraion on he convergence of he economies of he EU, and wha here is no dedicaed explicily o he analysis of he repercussions of he process on inequaliy in per capia income of he member counries. Thus, a firs group of sudies are dedicaed o analyzing he evoluion of he economies in he fulfilmen of he crieria for forming par of he Euro area (e.g. Guldager 1997 and Cappelen e al and Erur and Koch 2006). Anoher group of sudies analyze convergence a regional level, raher han counry level (e.g. Quah 1996), and in mos cases do no refer o he explici analysis of he effecs of he process of European inegraion, bu simply es for he exisence of convergence among he economies analyzed. Finally, a hird group of sudies, hough dedicaed o he analysis of he convergence of counries, do no analyze he collecive of EU counries, bu sudy he individual experiences of counries, he counries of he OCDE as a whole, or all he economies of he world (e.g. Barro, Sala-i-Marin 1992, Shulz 1998, ec.). Mos of hese sudies use measures of dispersion, such as he sandard deviaion of he logarihm of he variable (σ-convergence), or analysis of β-convergence so as o analyze he evoluion of he income dispariies among economies. However, regardless of he indicaor used and he ype of analysis of convergence, i should be poined ou ha hese sudies are usually based on curren per capia income. Alhough his mehodology provides useful informaion, his approach could be enriched wih a mehodology ha also akes ino accoun he whole life cycle dimension. Some recen sudies have ried o consider his issue by he use of alernaive measures. Dowrick e al. (2003) propose heir own index based on consumpion and life expecancy, avoiding arbirary

5 weighings by means of he revealed preferences; Becker e al. (2001 and 2005) analyse inequaliy of welfare by giving an economic value o he gains achieved in erms of life expecancy; likewise Philipson and Soares (2001) propose and analyse he properies of a measure of oal income (Full Income Measure of Human Developmen). In his sudy we aim o use an approach ha is also differen from he convenional one, hough complemenary o i. I is a ype of approach similar o ha proposed by Serrano (2006) 2. The main characerisic of he mehod is ha i considers he lifecycle income of individuals (presen value of fuure income) and no only he income of a specific period. By using he presen value (permanen income or lifecycle income) we coninue o ake ino accoun he level of curren income of he period, bu we also value aspecs such as he differen life expecancies in each counry and he differen pace of fuure growh of per capia income (which may converge much, lile or no a all). The idea is ha o judge inequaliy i may be appropriae o consider heir lives as a whole, using well known ools of economic analysis like permanen income or lifecycle income which are basic o he modern heory of consumpion (Modigliani 1986, Friedman 1957, Modigliani, Brumberg 1954). A counry's enry ino he EU is a srucural change in is economy whose effecs can only be valued from a long erm perspecive. For his reason, i is appropriae o use a lifecycle approach o analyze he effecs of inegraion on he ciizens' lifecycle income, and no only on he incomes of specific periods. The proposed approach enriches he radiional approach in ha i considers ha he differen counries may grow a differen raes, which will deermine he fuure incomes of heir inhabians. I also allows us o consider differen raes of survival of he individuals of each counry, which influences he number of years during which incomes are generaed, and herefore he oal incomes ha hey will obain in he course of heir lifeime. Boh aspecs may have imporan implicaions when judging he degree of inequaliy of he EU economies and heir evoluion following he successive expansions. The paper is organised as follows. The nex secion reviews he posiive effecs on economic growh associaed wih inegraion and, briefly, he hisory of he process of European inegraion. Secion hree reviews he exising lieraure on European convergence. Secion four presens he general 2 This approach has been used recenly by Pasor and Serrano (2008).

6 formulaion of he approach. Secion five presens he daa and he resuls regarding inequaliy among EU counries a differen momens in ime. Finally, he las secion presens he main conclusions. 2. The process of European inegraion: expeced effecs on economic growh The main economic objecive of European inegraion is o increase he rae of growh of he incomes of he paricipaing economies, of per capia income, and in he final insance, of he wellbeing of heir ciizens. The underlying idea is ha he consrucion of a wider European economic and social area will benefi he paricipaing economies. The sources of he poenial benefis associaed wih inegraion are very diverse (Viner 1950). Thus, inegraion favours greaer specializaion and beer use of he comparaive advanages of he economies as well as permiing more exensive exploiaion of economies of scale (Harris 1984, Gasiorek, Smih & Venables 1992, Francois, McDonald & Nordsröm 1994). Also, he opening-up of markes among he member counries (eliminaion of legal and cusoms barriers) increases compeiion, exering furher pressure o increase he efficiency of producion (MacDonald 1994, Caves, Baron 1990) and counries can purchase raw maerials and inermediae goods on beer erms, wih he consequen increase in producive efficiency (Lee 1992). Finally, changes in he quaniy and qualiy of he facors of producion used can also be expeced, due o he greaer mobiliy of he facors of producion wihin he area and o increased echnical progress (Maudos, Pasor & Serrano 1999). These poenial advanages have been a coninued simulus owards an ever greaer degree of inegraion among European counries. I has been a long and many-saged process whose ulimae oucome is he European Union of 27 members, and whose successive phases i would be appropriae o review before underaking he empirical analysis. The European Union has is origin in he European Coal and Seel Communiy (ECSC) founded in 1951 and formed of six counries (France, he Federal Republic of Germany, Ialy, The Neherlands, Belgium and Luxemburg). These six counries, signaories of he Treay of Rome in 1957, were he founders of he European Economic Communiy (EEC), a much more ambiious projec no longer limied o he coal and seel indusries. This led o he oal aboliion of cusoms ariffs on indusrial

7 producs in 1968 and o he developmen of common policies (as in he case of agriculural and rade policies). The success of he projec araced new candidaes and in 1973 he EEC expanded o nine members wih he enry of Denmark, Ireland and he Unied Kingdom. In he 1970s he EEC's field of acion expanded wih he developmen of social, environmenal and regional policies, wih he creaion of he European Regional Developmen Fund (ERDF) in In he 1980s he process of expansion coninued saw he enry of Greece and 1986 ha of Spain and Porugal. This brough wih i a greaer role for regional policies, wih greaer budge allocaions for he srucural funds, wih he aim of reducing he dispariies of economic developmen among he welve member counries. Wih he signing in 1986 of he Single European Ac he creaion of a grea single marke was agreed, becoming realiy in early A he end of 1990 he reunificaion of Germany ook place, so ha he länder belonging o he former GDR came o form par of he EEC. In 1993 came ino force he European Union reay planning he creaion of he Moneary Union for 1999, as well as seing in moion various insiuional reforms and once again expanding he EU's field of acion wih common policies on ciizenship, he common securiy and foreign policy (CSFP) and disposiions regarding homeland securiy. A he sar of 1995 hree oher counries joined (Ausria, Finland and Sweden), raising he oal membership o 15. The single currency (he euro) was creaed on 1 s January 1999, and welve of he EU's fifeen member counries joined i (all excep he Unied Kingdom, Denmark and Greece, hough he laer counry would join i in 2001). In 2002 his currency physically enered ino circulaion. In he mid-1990s he former people's democracies of he Sovie bloc knocked a he EU's door. As a consequence of he negoiaions begun in he laer years of ha decade, in May 2004 here was a new expansion wih he enry of en new members (Mala, Cyprus, Czech Republic, Slovakia, Hungary, Poland, Slovenia, Esonia, Lavia and Lihuania). Finally in early 2007 here occurred wha for he momen is he las expansion, wih he enry of Rumania and Bulgaria ino he European Union of 27 members.

8 The reducion of he differences in sandard of living among he member saes and among heir differen regions is one of he explici objecives of he EU. Regional policy achieves his by ransferring resources from he riches regions o he poores ones, being an insrumen of economic solidariy and a ool of economic inegraion. The join resul of he dynamic effecs of EU enry combined wih he acion of he EU's regional policy, which i is inended o make more powerful and wih more clearly defined arges afer successive reforms of is operaion, will ranslae ino greaer equaliy. The Commission iself poined wih saisfacion o he case of Ireland (currenly wih one of he highes per capia GDPs when a he ime of is enry i barely reached 64% of he EU average) as an example of wha can be achieved. The financial effor has been subsanial. In he period he srucural funds dedicaed o his purpose amouned o 213,000 million euros (one hird of he EU's oal spending in ha period). For he period he resources budgeed are approximaely 308,000 million euros, or 36% of he oal expendiure planned. 3. Review of he lieraure on European convergence There is surprisingly lile scienific lieraure abou he effecs of he European inegraion process on he convergence of he economies of he EU, paricularly wih regard o explici analysis of he repercussions of he inegraion process iself on convergence and inequaliy in he per capia income of he economies of he member counries following he successive expansions of he EU. The small amoun of empirical evidence available on he process of convergence in he European Union can be classified ino hree groups of sudies. A firs group, perhaps he mos numerous, conains sudies ha in realiy focus on he analysis of he crieria for enering and forming par of he euro zone. This ype of sudies analyze macroeconomic convergence in erms of public defici, ineres raes or inflaion rae (e.g. Guldager 1997). Anoher considerable par of he lieraure analyzes convergence in per capia income bu a regional level raher han by counries (e.g. Quah 1996) and only someimes direcly linked wih he process of European inegraion (e.g. Erur, Koch 2006). Finally, in a hird group of sudies, long erm economic growh and he convergence of he counries of Europe considered as counries is habiually subsumed in wider collecives such as he OECD or he world as a whole (e.g. Barro, Sala-i-Marin 1992), or analyzed only for he case of individual experiences (recenly, for example, he case of Ausria in Sockhammer (2009)).

9 The empirical lieraure on he counries of he European Union indicaes ha convergence is no seady over ime. Various facors seem o ac in opposing direcions, some generaing convergence and ohers divergence. Furhermore hese facors seem o have differen imporance for differen ypes of counries, so he effec of inegraion can vary significanly among counries. Thus, he resuls obained in Henrekson e al. (1997) sugges ha European inegraion may affec no only saic efficiency hrough changes in resource allocaion, bu also long erm growh raes. Their basic resul is a fairly robus associaion beween European inegraion and growh. The growh effec would be of he order of percenage poins p.a. The resuls also sugges ha echnology ransfer is he main mechanism hrough which EC and EFTA membership affec growh. Delgado-Rodriguez and Alvarez-Ayuso (2008) analyze he evoluion of labour produciviy among EU-15 counries over he period Using β-convergence echniques hey idenify periods of non-significan convergence ( and ), as well as ohers of rapid and significan convergence ( and ) in which less producive economies end o grow faser han more producive economies. For he whole period resuls are no saisically significan. Physical and human capial accumulaion appeared o be he main driving force behind he process. On he oher hand, echnological progress ended o conribue o divergence, alhough a change in he rend was observed a he end of he period. Maudos e al. (1999) analyze he evoluion of he counries of he European Union and he impac on efficiency and produciviy of he successive expansions during he period The resuls show ha unil 1990 inegraion was beneficial for all he paricipans. The counries ha joined experienced subsanial relaive gains in efficiency, greaer in all cases han hose regisered in he period prior o heir enry. Also he growh rae of TFP in he founder counries received a posiive impulse wih each new expansion. Kaiila (2004) analyzes boh σ and β convergence and discusses he impac of EU membership. According o he resuls, he EU15 counries real per capia GDP levels, adjused for purchasing power, converged in Convergence occurred in wo spells, in and , wih an inerim period of sagnaion. Abiad e al (2007) sress ha, due o increasing financial inegraion, capial in Europe has ravelled downhill from rich o poor counries, and has done so wih gahering srengh. These inflows would have been associaed wih a significan acceleraion of income convergence. In Reza e al (2008) real convergence of he

10 en new members' economies o he EU average income is esed by using quarerly real GDP per capia daa from 1995 o Applicaion of he uni roo ess for esing absolue convergence and caching-up make i possible o conclude ha he 10 new members of he EU in 2005 end o converge owards he EU average income. In any case, hese sudies are dedicaed o he analysis of convergence and inequaliy in curren per capia income and/or labour produciviy. However, he effecs of European inegraion are long erm, so i is naural o use also a lifecycle approach like ha described in he nex secion. 3. General formulaion As we have already indicaed, his paper analyzes inequaliy wihin he EU hroughou is exisence, using as he key variable he per capia permanen income or lifecycle income. The per capia permanen income ( PI i 0) of economy i a ime 0 is he discouned value of he presen and fuure per capia curren income ( y i ), aking ino accoun he survival rae in each period. We define Si(, -1) as he probabiliy ha a person who is alive in period -1 will sill be alive in period. Thus, he permanen income a ime 0 is defined as indicaed in he following expression PI i0 120 yi = Si(, 1) (1 + r) [1] = 0 in which we assume a common and consan ineres rae, r, and ha he maximum life of an individual is 120 years (an unresricive assumpion). Analysis of inequaliy and convergence in permanen income provides an analyical framework permiing us o idenify heir deerminans and heir sources of possible convergence. A more formal analysis of his ype of approximaion and is relaionship o heoreical models of growh is offered in Serrano (2006). The approach proposed is more complee han he radiional one because i permis he

11 consideraion ha counries grow a differen raes and heir individuals have differen survival raes. All his influences (i) heir inhabians' capaciy o obain fuure incomes and (ii) he number of years during which such incomes can be generaed. To illusrae he proposed approach le us imagine wo EU counries. Currenly counry 1 has a lower per capia income han counry 2. If counry 1 grows a a faser rae han counry 2, he income of counry 1 will herefore be closer o ha of counry 2 in he fuure. If we were o measure inequaliy using no only curren income, bu all incomes obained hroughou he lifeime of individuals, he inequaliy would be less han ha observed when we use only curren income. A similar argumen can be applied o he siuaion where he counries differ in heir survival raes. If individuals have a higher rae of survival in counry 2, which is richer, he number of years during which incomes are generaed will be greaer, and hus, ceeris paribus, heir fuure flow of incomes will also be greaer. If we measure inequaliy again using no only curren income, bu all he incomes ha will be obained hroughou individuals lives, he inequaliy in his case will be greaer han ha observed when using only curren income. 3 Noe ha, ceeris paribus, according o expression [1], counries will have higher levels of permanen income: - he higher heir iniial per capia incomes ( y i0), since he higher he iniial income, he greaer he fuure income flows ( y i ) given he raes of growh ( g i ), [ y = y (1 + g ) ]; i io i - he higher heir raes of growh ( g i ), since he higher he rae of growh, he higher heir fuure per capia incomes ( y i ) given he iniial levels of per capia income ( y i0), [ y = y (1 + g ) ]; i io i - he greaer he survival raes, [ Si(, -1)], since his will deermine ha incomes will be obained for more years and ha he presen value of hose income flows will increase; - he lower he rae of discoun (r), since his increases he presen value of fuure incomes. 3 Of course, alhough he approach proposed is more complee han he radiional one based on curren income, he permanen income approach poses some problems: 1) I is no such a sraighforward concep as curren income and 2) a number of addiional assumpions (on fuure curren incomes, life expecancy and discoun raes) are needed o esimae i. For hese reasons, we do no hink ha permanen income supersedes curren per capia income, which is a very useful and informaive way o look a inequaliy issues. However, we do believe ha permanen income is a complemenary, useful and suiable way o look a inequaliy beween economies because i ries o ake ino accoun whole life cycles of represenaive individuals. This approach, we hope, may provide us wih addiional insighs on he problem.

12 A empirical level, hree facors will influence he inequaliy in permanen income: - he iniial levels of per capia curren income, - he per capia fuure income flows, - he survival raes of individuals. Bearing his in mind, in he nex secion we will consider differen counerfacual scenarios o evaluae separaely he role of each of hese deermining facors. This will enable us o value he effec of each of hese facors on inequaliy and convergence in he EU. 4. Daa and resuls In his secion we presen he resuls regarding he inequaliy among EU counries a wo momens in ime, 1960 and The comparisons were always made using he Unied Saes as benchmark 4. All he daa are aken from World Bank Developmen Indicaors A deailed discussion of how he survival raes were obained can be found in he appendix. Table 1 offers he deailed daa regarding life expecancies and per capia curren incomes. Since we are ineresed in analyzing precisely he relaive differences beween counries, he daa are shown relaive o he USA. Addiionally we presen he relaive posiions of each counry in erms of curren and permanen per capia income in boh periods. 4 Using he USA as he benchmark economy is habiual in inernaional work. This allows us o avoid using any EU average which could be a problem given he changing composiion of he EU over ime. Furhermore, we avoid having o choose any specific European counry. A he same ime, we also obain resuls on inequaliy wihin he EU as well as on he relaive performance compared wih he world leader economy. This has he addiional advanage of using as a benchmark a counry which is no direcly affeced by he European inegraion projec. 5 Cyprus and Luxembourg have been excluded because of informaion problems.

13 Table 1: Curren income, permanen income and life expecancy in he EU counries. Benchmark counry=100. Year of Life Curren Permanen income * EU Expecancy (years) Per capia income (Hisorical scenario) Enry Counry Belgium France Germany Ialy The Neherlands Denmark Ireland Unied Kingdom Greece Porugal Spain Ausria Finland Sweden Czech Republic Esonia Hungary Lavia Lihuania Mala Poland Slovak Republic Slovenia Bulgaria Romania Unied Saes Deviaion coefficien EU6 (1951) EU9 (1973) EU10 (1981) EU12 (1986) EU15 (1995) EU25 (2004) EU27 (2007) * Permanen income calculaed using individuals' raes of growh (gi), individual income per capia (yi) and individuals' survival raes (Si(,0)). Discoun rae = 2%. Source: World Bank and own preparaion.

14 In respec of life expecancy a birh, he daa show modes bu significan differences. Thus, in 1960 a counry such as The Neherlands had a life expecancy 5.2% longer han he USA, while Porugal ranked 9.1% below ha level. This is a subsanial difference, of he order of 15%. Among he res of he counries he differences were smaller bu appreciable. In 2005 he greaes difference was beween Sweden (4% above he USA) and Rumania (7.9% below he USA), a difference of 12%. There were significan changes during his period, such as he relaive improvemens of counries such as France, Ialy, Porugal, Spain, Ausria and Finland; on he oher hand in oher cases he evoluion was less saisfacory (The Neherlands, Denmark and nearly all he easern European counries). These differences in erms of life expecancy and he changes occurring during he period make his an aspec o be aken ino accoun when valuing he levels of inequaliy in he EU and convergence among counries from a long erm perspecive. In erms of per capia income he differences are of greaer magniude. In 1960 Denmark and Bulgaria were he exreme cases wih per capia incomes equivalen respecively o 88.7% and 4.5% of he per capia income in he USA. In 2005 hese wo counries sill showed he maximum and minimum values wihin he group of counries currenly forming he EU27, Denmark wih a per capia income of 84.1% ha of he USA and Bulgaria barely reaching 5.5%. As well as he size of he differences, he changes occurring during he period should also be aken ino accoun. The exreme values show a sabiliy which i would be decepive o consider somehing general. Thus, counries like Ireland, Belgium, Ialy, Greece, Porugal, Ausria, Finland, Hungary, Mala or Slovenia achieved subsanial improvemens in relaive erms. Oher counries like Denmark, The Neherlands, Unied Kingdom, Czech Republic or Sweden, however, presened a less saisfacory evoluion. This dispariy of behaviours over ime makes i imporan when valuing he inequaliy among hose counries a each momen in ime o ake ino accoun no only he per capia incomes a ha momen, bu also he presen value of he per capia incomes expeced in he fuure. Wih such disparae growh raes of per capia income, inequaliy in erms of lifecycle income can vary significanly from inequaliy in income of he period. The esimaions of per capia permanen income based on he hisorical scenario are offered in he hird column of able 1. In his scenario, according o expression [1], he iniial per capia income and he survival raes have been used for each counry. Also, we have used a long erm growh rae based on he assumpion ha per capia income grows a he mean rae achieved during he period

15 Finally, o conver fuure incomes o presen values we have used a discoun rae of 2%. In 1960 permanen income varied beween he value for Denmark (87.6% of he USA's) and ha for Bulgaria (5.4% of he USA's). I would seem, herefore, ha considering he lifecycle is no oo imporan given ha he resuls for boh exreme cases are similar o hose obained using simply he curren income for However, when we observe wha happens in he oher counries and no only in he exreme cases we see imporan changes. Though all he counries are below he USA, heir relaive posiions change subsanially if curren income or permanen income is considered. Among ohers, Belgium, France, Ialy, Greece, Porugal, Spain, Ausria, Finland, or especially Mala and Ireland, improve considerably when permanen income is considered 6. Thus, for example, Ireland goes from 34.6% of he USA o 79.5%; Spain from 26.6% o 40.5%, or Ialy from 41.6% o 50.3%. On he oher hand, here are counries such as he Unied Kingdom or Denmark where he opposie occurs and ohers where he improvemen is unimporan (e.g. Germany, The Neherlands or Poland). In 2005 somehing similar occurs and i is easy o see ha he ranking of counries would change considerably if permanen income were considered insead of curren income. The above resuls show he exisence of differences in erms of life expecancy and economic growh raes among European counries which jusify he ineres in adoping a permanen income approach o analyze inequaliy in he EU. Also, he esimaions of permanen income indicae ha he resuls can differ in many counries from hose habiual in exercises based on curren income. For his reason, we will analyze he evoluion of inequaliy among counries wihin he EU hroughou he period, and he possible changes associaed wih he various expansions hereof, wih his lifecycle perspecive always in mind. Our analysis of inequaliy in he EU will be based on he use of he coefficien of variaion of per capia income, a dispersion saisic habiual in his ype of inequaliy analysis. Figure 1 shows he level of inequaliy in differen periods (including hose years when expansions of he EU occurred) for he counries forming he EU a ha ime, boh in erms of curren income and of permanen income. Le us firs examine he resuls in erms of curren income. The coefficien of variaion of curren income of he EU6 was in Following he enry in 1973 of Denmark, he Unied 6 The resuls for Mala and Ireland (boh counries showing an esimaed permanen income greaer han he US in 2007) is due o applying he previous long run rae of growh in his scenario, which is very high in boh counries. Recen developmens show a less exraordinary performance. Oher scenarios esimaed in he paper allow for differen hypoheses abou fuure raes of growh.

16 Kingdom and Ireland he inequaliy increased o In 1981, he year of Greece's enry, i was moderaed o 0.249, growing subsanially o in 1986 (enry of Porugal and Spain). From hen onwards here was a gradual descen, o in 1990 (reunificaion of Germany) and in 2004 (expansion of he EU o 15). I has o be said ha he effec of he mos recen expansions has been a very significan increase of inequaliy among counries in he EU. In 2004 (expansion o 25 member counries) he coefficien reached and if we include Bulgaria and Rumania (which joined in 2007) he coefficien would be a levels around Figure 1. Inequaliy in per capia curren income and permanen income in he EU counries (Deviaion coefficien) 0,90 0,80 0,70 0,60 0,50 0,40 0,30 0,20 0,10 0,00 0,161 0, EU6 0,254 0, EU9 0,249 0, EU10 0,359 0, EU12 0,315 0, EU12 0,282 0, EU15 0,702 0, EU25 0,653 0, EU27 Curren Income Permanen Income In sum, we observe a progressive increase in inequaliy unil he mid-1980s, a reducion over he nex 20 years and a sharp rise as a consequence of he laes expansions owards Easern Europe. The curren levels of relaive inequaliy, hen, wihin his much more heerogeneous 27-member EU, muliply several imes hose exising among he original members of he projec of European inegraion. Le us now consider inequaliy from a lifecycle perspecive, using once again he coefficien of variaion, bu in his case ha of permanen per capia income. The resuls appear in he same figure 1. The emporal profile now shows a coninued growh which, indeed, acceleraes quie visibly wih he laes expansions. Thus he coefficien sars a a level of in 1960 (EU6), increases o in 1973, in 1981, in 1986, in 1990, in 1995, in 2004 and 0.770

17 in Also when permanen income is used we observe ha levels of inequaliy are a maximum values for he EU, much higher han he iniial ones. When he resuls obained from hese wo perspecives (curren and permanen income) are compared, we observe some ineresing differences. During he period prior o he laes expansion, inequaliy was much lower if permanen income is considered, especially for he years before Tha is o say, aking ino accoun he fuure and no only he income of he period, he levels of inequaliy wihin he EU were subsanially less han hose indicaed by he curren per capia income of he period. However, afer he laes expansions aking place his cenury he image is quie he conrary. Inequaliy is greaer in erms of permanen per capia income. The differences of life expecancy and of incomes foreseeable in he fuure end o magnify he inequaliy among he members of he presen EU, whereas in he pas he opposie occurred. This means ha, unless European cohesion policies conribue more acively o changing his panorama, he levels of inequaliy in he EU will coninue o be high. In his sense, i has o be poined ou ha he laes reform of he EU's regional policy (European Commission 2004) moivaed o a large exen by he laes expansions wih he enry of more heerogeneous counries, is moving in he direcion of concenraing is acions mos on he unequivocally less developed areas. The aim of hese reforms is o make he regional policy more effecive in boosing he developmen of he less developed regions, essenially he majoriy of new member counries. Once we have analyzed inequaliy and convergence in permanen income (scenario 1 or hisorical scenario), our nex sep is o evaluae separaely he role of each of heir deermining facors, such as survival raes, per capia curren income and rae of convergence. In order o do his, he effec of each of hem will be isolaed sep by sep, i.e. allowing for changes only in one of hese variables each ime. These correspond o differen assumpions as o survival raes, he iniial levels of per capia curren income and he exisence or oherwise of convergence among counries in erms of per capia curren income. On he basis of hese scenarios, we simulae he inequaliy of permanen income of he counries by building some counerfacual scenarios. Paricularly, he following scenarios will be considered:

18 - Scenario 1 (Hisorical base scenario): In his scenario i is considered ha he per capia income of each economy in he iniial period ( y i 0) grows a he individual average rae of growh ( g i ) during he period The series of per capia incomes ( y i ) obained in his way [ y = y (1 + g ) ] is used o calculae permanen income ( PI i0) according o expression [2]. i io i y y (1 + g ) PI S (, 0) S (, 0) i io i i0 = i = i = 0 (1 + r) = 0 (1 + r) [2] - Scenario 2 (wihou convergence in curren per capia income): In his scenario i is considered ha he per capia income of each economy in he iniial period ( y i 0) grows a he average rae of growh of he benchmark economy ( g US ) for he period The series of per capia incomes ( y i ) obained in his way [ yi = yio(1 + gus ) ] is used o calculae permanen income ( PI ) according o expression [3]. i0 y y (1 + g ) PI S (,0) S (,0) i io US i0 = i = i = 0 (1 + r) = 0 (1 + r) [3] - Scenario 3 (Scenario wih idenical survival raes): This scenario is he same as scenario 1 wih he sole excepion ha he per capia incomes of each economy ( y i ) obained as described [ y = y (1 + g ) ] are combined wih he survival raes of he benchmark economy ( S ), i i0 i obaining he permanen income ( PI i 0) according o expression [4]. In his way we can evaluae he differences in permanen income ha would persis even if he survival raes did no differ beween economies. US y y (1 + g ) PI S (,0) S (,0) i io i i0 = i = US = 0 (1 + r) = 0 (1 + r) [4] - Scenario 4 (Scenario of idenical iniial per capia incomes): In his scenario i is considered ha he per capia income of each economy in he iniial period is equal o ha of he benchmark economy ( yi0 = yus0) and grows a he individual average rae of growh ( g i ) during he period

19 The series of per capia incomes ( y i ) obained in his way [ yi = yuso (1 + gi ) ] is used o calculae permanen income ( PI i 0) according o expression [5]. y y (1 + g ) PI S (,0) S (,0) i USo i i0 = i = i = 0 (1 + r) = 0 (1 + r) [5] - Scenario 5 (considering he raes of growh since inegraion): In his scenario i is considered ha he per capia income of each economy in he iniial period ( y i 0) grows a he individual POST average rae of growh ( g ) during he pos-inegraion period o The series of per i POST capia incomes ( y i ) obained in his way [ y = y (1 + g ) ] is used o calculae permanen income ( PI i 0) according o expression [6]. i io i y y (1 + g ) PI S (,0) S (,0) POST i io i i0 = i = i = 0 (1 + r) = 0 (1 + r) [6] - Scenario 6 (Scenario wih convergence in curren per capia income): In his scenario i is considered ha he per capia income of each economy in each period ( y i ) converges owards ha of he benchmark economy ( y US ) a a speed of convergence β. If we define he per capia income of an economy a momen, relaive o he benchmark economy, as η = y i y US and we furher assume ha here are no differences in seady sae, hen convergence a an annual rae of β makes η = 1 e β ( 1 η ) o and yi yus ( 1 (1 η0) e β ) =. The series of per capia incomes ( y i ) obained in his way is used o calculae permanen income ( PI i 0) according o expression [7]. In his scenario hree raes of convergence are considered, β=2%, β=3% and β=5% y yus ( 1 (1 η0) e ) i β (,0) (,0) [7] PI = S = S i0 i i = 0 (1 + r) = 0 (1 + r) Table 2 shows he coefficiens of variaion of permanen per capia income corresponding o hese new counerfacual scenarios.

20 Table 2: Inequaliy in curren income and permanen income in he EU counries. Differen scenarios. (Deviaion coefficien) Curren income Scen. 1 Scen. 2 (g USA ) Permanen income (Scenarios) Scen. 4 Scen. 5 (Ypc USA ) (g pos ) Scen. 3 (S USA ) Scen. 6 ( β=2%) Scen. 6 (β =3%) Scen. 6 (β =5%) 1960 EU EU EU EU EU EU EU EU As we have shown above, one of he poenially imporan facors in deermining permanen income is he fuure rae of growh of curren per capia income. To esimae is impac on inequaliy we define scenario 2. In i we have obained he permanen incomes of each counry on he assumpion of a common fuure growh rae. Specifically, he growh rae of per capia income of he USA from 1960 o he presen has been used for all counries. The iniial levels of per capia income and life expecancies, on he oher hand, coninue o be hose of each counry. The resuls of scenario 2 show increasing inequaliies unil he creaion of he EU12, followed by sligh reducions in he 1990s and a sharp rise wih he laes expansions. Of more ineres is he comparison beween hese resuls and hose of scenario 1, as he differences beween he wo indicae he par of he inequaliy in permanen income associaed wih he differen long erm growh raes of he curren per capia income of each counry. As can be observed, he differen raes of growh of per capia income during he period helped o reduce inequaliy, wih he excepion of 1960 and In he successive expansions from 1973 o 1995 we can observe ha he inequaliy in permanen income of he hisorical scenario (scenario 1) is always less han ha which would have been obained wih a common rae of growh. Thus, in 1986 he coefficien of variaion of he EU12 counries was 0.218, bu applying he common growh rae (scenario 2) his coefficien would be 0.313, indicaing nearly 50% more inequaliy in permanen income. The reason is ha, when he member counries of he EU9, EU10, EU12 and EU15 are considered as a whole, he counries wih lowes per capia income showed faser long erm growh during he period However, he laes expansions again show differenial characerisics. For he EU25 inequaliy sands a and for he EU27 a 0.77, according o scenario 1. Using a common growh rae he inequaliy falls o and respecively. This is because he laes expansions have brough in counries ha are less developed and which have in he pas shown less capaciy for long erm growh.

21 The second imporan facor for explaining inequaliy among counries is he difference in life expecancy. Scenario 3 has been defined for he purpose of evaluaing he imporance ha differences in life expecancy have had for he levels of inequaliy among he counries of he EU hroughou is hisory. The resuls of his scenario were obained under he assumpion ha all he counries had a common life expecancy, specifically ha of he USA. Therefore, he differences beween he inequaliy levels of his scenario 3 and hose of scenario 1 (hisorical scenario) have o be aribued o he differences in life expecancy of each counry. Thus, in 1960 he differences in life expecancy among he member counries of he EU6 explain a significan par of he inequaliy in permanen income (wih a common life expecancy like ha of he USA he coefficien of variaion of permanen income would have been insead of 0.101). On he oher hand during he 80s and 90s he effec was he opposie, helping o reduce slighly he inequaliy in permanen income (hus, in 1995 he coefficien of variaion among he counries of he EU-15 wih a common life expecancy would have been insead of 0.258). The impac of life expecancy is currenly very low, and he levels of inequaliy would barely change even hough he differences in life expecancy disappeared. Scenario 4 corresponds o he esimaions of permanen income obained by assuming ha all he counries of he EU sar wih he same iniial per capia income, bu mainaining he life expecancies and long erm growh raes of each counry. In 1960 his gives an inequaliy of permanen income of 0.075, clearly below he inequaliy in curren per capia income and also below he inequaliy observed in he hisorical scenario (scenario 1). This confirms ha differences in life expecancy and, paricularly, in growh raes, were conribuing significanly o he inequaliy among he member counries of he original EU. From ha momen onwards, he impac of hese wo facors becomes more and more subsanial. Inequaliy in permanen income is greaer han ha obained in scenario 1 in 1973, 1981, 1986 and Wih he laes expansion his rend has been broken since mos of he new member counries presen a subsanial gap beween heir iniial curren per capia incomes and hose of he exising member counries. The impac of growh raes and of differences in life expecancy is subsanial (hus he inequaliy in scenario 4 rises from for he EU15 in 1995 o for he EU25 in 2004), bu he differences are smaller han in scenario 1 (hus, for he EU25 in 2004 inequaliy in permanen income is in scenario 4 and in scenario 1). The evidence for he growh rajecories of he European counries afer each expansion are no

22 especially encouraging. The resuls of scenario 5 were obained by esimaing in each year he fuure curren incomes on he basis of he growh raes experienced by each counry since ha year. Le us recall ha he esimaions of scenario 1 are based always on he growh raes measured from 1960 o he presen. The comparison beween scenario 5 and scenario 1 is clear. We can leave aside he resul for 1960 which, naurally, has o coincide. In he res of he years he inequaliy esimaed in permanen income is significanly greaer because pos-expansion growhs are used. This already occurs in 1973 (0.241 and 0.141) and he phenomenon persiss wih increasing inensiy. In 1995 he level of inequaliy would be muliplied by four (1.078 and 0.258) and wih he laes expansion he resul is similar (2.85 and 0.77). Naurally, we have o bear in mind ha he more recen he year analyzed, he shorer he period ha serves o calculae he growh raes, which may be subjec more o immediae facors han o long erm ones. In sum, he resuls indicae ha if he curren posinegraion growh raes are mainained, subsanial levels of inequaliy will remain. Le us recall ha he base esimaion (scenario 1) was obained from he raes of growh of per capia long erm income ( ) of each counry in he pas. A differen paern of growh in he fuure could have a subsanial impac. In scenario 6 hese growh raes have been replaced by ohers ha correspond o a siuaion of convergence in per capia curren incomes among he counries of Europe. Under his hypohesis he counries wih lowes per capia income would grow mos and would do so faser due o heir relaive backwardness. Three differen annual convergence raes have been posied: 2%, 3% and 5%. These hree convergence scenarios correspond o he hypohesis ha each counry manages o reduce he gap in curren per capia income by 2%, 3% or 5% (respecively) each year. The resuls indicae ha his would lead o a seep reducion of he inequaliy in permanen income among counries of he EU. The resuls of scenario 1 indicae ha he inequaliy in permanen income in he EU is currenly However, if insead of mainaining he pas growh raes of each counry we assume ha in he fuure here will be an annual convergence of 2%, he esimaed inequaliy in permanen income would currenly be only 0.223, less han one hird. If a somewha greaer annual convergence in curren per capia income (3%) were achieved, i would be only Finally, wih a convergence rae of 5%, he inequaliy in permanen income would be barely In oher words, if hanks o he process of economic inegraion iself or o he EU's cohesion policies, a convergence rae of 2% were achieved, he inequaliy in curren income in 2005 of 0.653

23 would be compaible wih a lifecycle inequaliy wo hirds lower (0.223). I should be emphasized ha raes of convergence beween economies of 2% are perfecly feasible. Numerous sudies have esimaed similar convergence raes among he counries of he OCDE, he saes of he USA, he prefecures of Japan, he regions of Germany, Spain, ec. (Barro, Sala-i- Marin 1995). Furhermore, any increase in ha rae of convergence would have is reward in an appreciable reducion of he inequaliy in permanen income. Afer analyzing he influence of he deerminans of permanen income on inequaliy, we may wonder abou he evoluion of inequaliy following he successive expansions, boh for he old member counries, and for he counries ha are joining. Table 3 permis us o observe he phenomenon of inequaliy in his muliple dimension, boh in erms of curren per capia income (panel a) and in erms of permanen per capia income under he scenario 1 (panel b).

24 Table 3: Inequaliy in curren income and permanen income in he EU counries. (Deviaion coefficien) a) Per capia income EU6 (1951) EU9 (1973) EU10 (1981) EU12 (1986) EU12 (1990) EU15 (1995) EU25 (2004) EU27 (2007) b) Permanen income EU6 (1951) EU9 (1973) EU10 (1981) EU12 (1986) EU12 (1990) EU15 (1995) EU25 (2004) EU27 (2007) The rows of he upper panel permi us o see he evoluion of he inequaliy in curren per capia income of he successive groups of counries ha have come o form he EU hroughou he period In he firs row we can observe ha he inequaliy among he founding counries a he sar of he EU (EU6) decreased progressively from o less han half ha in 1995 (0.078), rising slighly hereafer (0.085 in 2005), hough remaining well below he iniial levels. For he expanded group of counries ha formed he EU9 in 1973 somehing similar occurred, wih inequaliy decreasing from in 1973 o in Wih he enry of Greece in 1981 he EU10 was formed, he reducion of inequaliy for his group being weaker (from in 1981 o in 2005). More posiive was he experience of he EU12 wih he enry of Spain and Porugal in 1986: inequaliy was reduced from a level of in ha year o in Sligh reducions are also observed for he expanded collecives EU15 and EU25 as from heir respecive creaions. Tha is o say ha he general one is one of he successive expansions being accompanied by reducions in he inequaliy among he old members.