IMPACT OF IMMIGRATION ON THE JAPANESE ECONOMY: A MULTI-COUNTRY SIMULATION MODEL

Transcription

1 IMPACT OF IMMIGRATION ON THE JAPANESE ECONOMY: A MULTI-COUNTRY SIMULATION MODEL Absrac To quanify he impacs of immigraion and fiscal reconsrucion on he Japanese economy, we presen a dynamic compuable general equilibrium OLG model wih an over lapping generaions srucure. We use a oal of 16 counries and regions, boh including hose ha are indusrialized, such as Japan, he US, and he EU, and developing counries, such as China, Brazil, he Philippines, and Peru. Our simulaion resuls show ha a permanen immigraion flows of 150,000 will improve he Japanese economy and he welfare of curren and fuure generaions. On he oher hand, a sandalone increase in he consumpion ax will no improve long-run welfare. The resuls indicae ha subsanially increased inflows of working-age immigrans would alleviae he need for fuure fiscal reform and also help o dramaically reduce he public pension burden on he working generaions. JEL classificaion: C68; D9; E62; H5; H6; H55 Keywords: Compuable general equilibrium (CGE) model; overlapping generaions (OLG); aging; immigraion Manabu Shimasawa Associae Professor, Akia Universiy mshimasawa#ed.akia-u.ac.jp ( # sands ) Kazumasa Oguro Consuling Fellow, Research Insiue of Economy, Trade and Indusry, Minisry of Economy, Trade and Indusry and Associae Professor, Hiosubashi Universiy ZVU07057#nify.com ( # sands ) 1

2 1. Inroducion Indusrialized counries are now facing unprecedened demographic changes which require exensive reform in fiscal sysems, social securiy sysems, and oher relaed programs. However, due o conflicing ineress beween younger and older generaions, reform will be poliically difficul. As an example from pay-as-you-go pension sysems, in order o improve he susainabiliy of he sysem, he governmen has he opion of eiher reducing he benefis o he elderly or increasing he burden on he working generaion. Obaining agreemen on reform by boh generaions is ofen oo difficul for he governmen o achieve. Some have suggesed ha immigraion may ease hese radeoffs. The ruling Liberal Democraic Pary (LDP) in Japan recenly proposed an explici immigraion policy o receive 10 million immigrans in he nex 50 years (200,000 annual immigran flows), a number significanly greaer han he annual 68,054 ne flow in The proposed policy increases immigraion in order o improve he welfare of boh curren and fuure generaions. Moreover, some economic organizaions, including he larges of hem, Nippon Keidanren (Japan Business Federaion), also insis ha receiving immigrans will mainain he susainabiliy of he social securiy sysem and aver he expeced decline in economic growh. To dae, however, he economic effecs of hese proposals have no been rigorously sudied. According o saisics published by he Minisry of Jusice in 2008, he number of regisered foreign naionals residing in Japan has increased every year since 1987 and oaled 2,152,973 as of he end of 2007 (See Figure 1) 1. Regisered foreign naionals share of he oal Japanese populaion has also increased every year o 1.69% in By place of origin, China exceeded he Korean peninsula for he firs ime in 2007, o become he op source of immigrans, followed by he Korean peninsula, Brazil, he Philippines, and Peru. The aim of his paper is o quanify he impacs of immigraion on he Japanese economy. To his end, we use a large-scale dynamic compuable general equilibrium model incorporaing overlapping generaions (OLG) and muliple counries or regions. The 16 counries observed include boh indusrialized counries and regions such as he US and EU as well as developing counries such as China, Brazil, he Philippines, and Peru. The applicaion of OLG models o fiscal impacs was pioneered by Auerbach and Kolikoff (1987). Since hen, OLG models have been used exensively o sudy he impac of populaion aging and o evaluae various policy changes including ax, pension, and public deb policy. In he lieraure, we find only hree A-K ype OLG simulaion models using muliple counries 1 According o he esimaion by he Minisry of Economy, Trade and Indusry, he number of foreign sudens sudying is abou 100,000 and he number of rainees is abou 90,000 as of he end of And he Minisry of Jusice repored ha he number of illegal immigrans is abou 113,000 as of he beginning of Illegal immigrans comprise only 0.09% of he oal populaion. Moreover, he number has decreased consisenly from a high of 299,000 in 1993 by 186,000 o he 113,000 number in early The large majoriy are Asians, mosly from he Korean peninsula (21.4%), China (16.3%), he Philippines (15.3%), Thailand (5.3%), and Taiwan (4.4%). 2

3 o evaluae he effecs of global aging on inernaional capial flows and he worldwide economy. These are models by Fehr e al. (2004), Börsch-Supan e al. (2006), and Agliea e al. (2007). Fehr e al. (2004) and Börsch-Supan e al. (2006) incorporae only indusrialized counries in heir model. Alhough Agliea e al. (2007) include boh indusrialized and developing counries, hey cover only 19 generaions in heir model. Moreover, each of hese models use he same values of srucural parameers such as uiliy and producion for every counry. This sudy, however, models a 16 indusrialized and developing counries covering 65 generaions wih differen values of srucural parameers for counries, depending on heir developmen saus. Soresleen (2000) calibraed a general equilibrium OLG model of he US economy, explicily aking ino accoun he differences beween immigrans and naives, o esimae he long-run fiscal impac of immigrans. He found ha he fiscal impac of immigraion on he hos counry is posiive, even when considering immigran age and skill level. Meanwhile, Fehr e al. (2004) developed a hree-region (US, Japan, and EU) dynamic general equilibrium OLG model o analyze wheher immigraion can miigae he negaive impacs of demographic ransiion on an economy. They concluded ha immigraion will no aler he major negaive impacs regardless of he immigrans skill level. However, our mehod is differen from ha of Fehr e al. (2004). Their model is based on immigrans coming from counries ouside of he model, and does no include he effec ha he demographic changes in he ou-migraion counries have on he in-migraion regions (US, Japan, and EU) hrough he inernaional capial flows. Capial flows are shifed in response o he change in he capial labor raio wihin he ou-migraion counries. To illusrae, consider wo cases wih only counries A and B. In he firs case, par of he working generaion emigraes from counry A o counry B, and in he second case par of he working generaion emigraes from ouside of boh counries o counry B. In he firs case, he capial labor raio of A rises and ha of B falls, all else equal. In he second case, he capial labor raio of A does no change and he capial labor raio of B falls again. I is possible ha he pos immigraion capial labor raio of B in he firs case would be higher han in he second case. Even if he size of immigran flows are he same in boh cases, changes in he capial labor raio in ou-migraion counries will be differen, which will affec he local use of capial, and hence he worldwide cos of capial differenly. If he pos immigraion worldwide cos of capial is lower in he firs case, capial will be used more inensely in counry B and he capial labor raio be higher han he second case. I is also possible ha he capial labor raio of B afer immigraion could be close o ha before immigraion in he firs case. In addiion, because he Governmen will consider differen policy choices, i is imporan o analyze he difference beween he effecs of a emporary and a permanen increase in immigraion as well as he difference beween an immediae and a delayed increase in immigraion. Fehr e al. 3

4 (2004) did no consider he effecs of hese policy variaions. A vas number of sudies have applied he OLG model o he problem of aging in Japan, such as Homma e al. (1987), Kao (1998, 2002), Sadahiro and Shimasawa (2001, 2003), Okamoo (2005), and Ihori e al. (2006). In each of hese models, he life-cycle hypohesis of consumpion serves a crucial role, jus as in he A-K ype OLG model. These sudies found ha populaion aging leads o a sharp reducion in he savings rae, which affecs capial formaion and facor prices. Therefore, he naional economy mirrors he shrinking of he size of he working-age populaion. Alhough policy reforms can alleviae he economic burden of an aging populaion, heir simulaions showed ha here is no easy way o reduce he economic burden from an increase in he raio of he reiremen age o working age populaion (reiree age raio). Those sudies consider various policy reforms o cope wih aging reduced pension benefis, increase in axes, and public deb reducion bu, no research has ye quanified he effecs of immigraion on he Japanese economy using a general equilibrium model wih an OLG srucure. To dae, immigraion has no been sudied as a policy insrumen o cope wih aging in Japan. Therefore, in his paper, we use a dynamic general equilibrium OLG model calibraed o he Japanese economy in order o quaniaively evaluae he effecs of immigraion policies inended o cope wih populaion aging in Japan. Muliple counries, including he immigran suppliers China, Souh Korea, Philippines, and Peru, are included in he model. By doing his, we aemp o idenify which immigraion policy changes migh resul in significan posiive effecs on he Japanese economy, in erms of he governmen and public pension budge and ciizens welfare. The resuls show ha subsanial increased inflows of working-age immigrans would reduce he need for fuure fiscal reform. Furher, findings indicae ha immigraion can also help o dramaically reduce he public pension burden on he working generaions. Our resuls have several policy implicaions. In Japan, i is ofen argued ha an increase in immigraion would negaively affec he economy by causing a downward spiral of wages for he naive workers. According o our findings, hese concerns urn ou o be largely unnecessary because an increase in immigraion reduces he pension ax on wages, increasing he lifeime resources available o workers. Finally, he posiive effecs are greaer if he immigraion is permanen raher han emporary, larger in volume, and implemened sooner raher han laer, as he LDP and Nippon Keidanren insis. Therefore, he Japanese governmen should discuss he inroducion of an explici permanen immigraion policy as early as possible argeing a large volume of immigans. In addiion, he governmen is presenly promoing Economic Parnership Agreemens (EPAs) wih Eas Asian counries in order o share in heir dynamic growh. The EPAs include an increase in he enrance of foreign skilled workers (e.g. nurses for elderly assisance ec.). A he same ime, he governmen is also promoing a program aimed a increasing he number of foreign sudens o 300,000 by around This foreign suden program is being implemened as par of Japan's 4

5 global sraegy o enhance he naion's inernaional compeiiveness in educaion and research, by aracing excepional foreign sudens o sudy in Japan. The governmen has also been seeking o expand heir employmen opporuniies in Japan, aking seps o mach foreign graduaes from Japanese universiies wih he recruimen needs of Japanese firms. Thus, if foreign graduaes have a smooh ransiion from graduaion o employmen in Japan, Japanese companies can more easily recrui high qualiy foreign workers. Therefore, when an explici immigraion policy is discussed, i is imporan for he governmen o consider he relaionship beween immigraion policy, EPAs, and he foreign sudens programs. The srucure of his paper is as follows: In Secion 2, we describe he model srucure. Secion 3 presens he calibraion sraegy and he findings. Secion 4 describes simulaion resuls, and Secion 5 conains concluding remarks and policy implicaions. 2. The Model Srucure In his secion, we describe he demographic and economic srucure of our model. The model used here is a dynamic compuable general equilibrium OLG model wih perfec foresigh agens, muliple periods, and muliple counries. In our model, here is a represenaive individual for each generaion in he household secor. Each individual a age 20 maximizes his/her ineremporal uiliy funcion wih consumpion and beques. The represenaive compeiive firm has a sandard Cobb-Douglas producion echnology and is profi maximizing. In our model, no only he goods marke bu also facor markes are perfecly compeiive. The model has five main pars: 1) demographic projecion, 2) household behavior, 3) firm behavior, 4) he Governmen, and 5) he public pension. Deails of each par are below. To limi noaion, we suppress regional indices o he exen possible. Iniial values and parameers for hese building blocks will be described in Secion 3. (1) Demographic projecion In our model, he demographic projecion is exogenous. In each region, he size of oal populaion of age j in he period, where N, j N, j N 1, j1 M, j j 0 is given recursively by: 50 for, N, 0 f j jn 16, 1, j and MM M j, j (1) M, j denoes he migraion in j age-cohor a he ime, f, j MM he aggregae ne migraion in he period. he age-specific feriliy rae, and In addiion, o simplify calculaions in our model, we don disinguish beween naives and immigrans in he model once he immigrans have enered he age specific group of he hos counry. 5

6 (2) Household behavior There is a represenaive individual for each generaion in he household secor. We assume ha preferences are he same for all agens in all generaions. Moreover, each individual lives for a fixed number of periods. In each period of he model, he oldes generaion dies and a new one eners. The represenaive individuals maximize heir ineremporal uiliy funcion wih consumpion and beques dependen on o heir lifeime income. They are raional, and possess perfec foresigh. In he m h region, each generaion eners he labor marke a age 21, reires a age graned a pension a Q +1, and dies a age m Q, is m Z. In addiion, each supplies labor inelasically. The m wihin-period uiliy funcion exhibis consan relaive risk aversion, and preferences are addiive and separable over ime. In each region, he uiliy funcions of he i h generaion are specified as: U i j-1 Z 20 1 j ci, j Z qi 1- where j refers o he j h period of life, he rae of ime preference, he inverse of he elasiciy of ineremporal subsiuion, and he beques moive. The argumens of he uiliy funcion are he consumpion per period ( c i, j ) and he beques a he deah period ( q i ). Leisure does no ener he uiliy funcion since he individual s labor supply is assumed o be exogenous. The echnological progress is exogenous and embodied in labor. We model age-specific labor produciviy by assuming a hump-shaped age-earnings profile, i.e., a quadraic form of is age ( j+20), so is age-wage profile e j akes he following form: e j j j,, 0 and 0 (3) The ineremporal budge equaion of each generaion may be described as follows: Q-20 j1 PDV Z-20 i, j(1- w -p )w (1 ) e j PDVi, jpi, j q i ( Z21) / jq-19 i (2) where Z-20 PDV j1 i, jci, j(1 c ) PDVi, Z-20(1 b)q i j PDV i, j 1 / (1 r k 1 i k 1) refers o he facor of he presen discouned value, w is he wage rae a ime i+j-1, e j is he wage profile a age ( j+20), w is he labor income ax rae a ime, p is he public pension conribuion rae a ime, measures he rae of echnical progress, i is he populaion raio of generaion i o generaion i-(z-21), (4) c is he consumpion ax rae a ime, b is he inheriance ax rae, and p i,j sands for pension benefi of generaion i a age ( j+20). Each generaion maximizes is uiliy funcion (2) under he budge consrain (4). Maximizing uiliy, i.e. differeniaing he household uiliy funcion (2) wih respec o ci, j and q i, subjec o he individual s life-ime budge consrain (4), yields he following Euler equaions concerning consumpion per period. 6

7 c i, j 1 r (1- r ) 1 c -1 ci 1 1c 1 1, j-1, C Z-20 j1 N, j20 c i, j (5) where r is he ineres rae a ime, r is he ax rae on ineres income a ime, N,j measures he number of he people of age j a period, and C is he aggregaed consumpion a ime. This Euler equaion dicaes, as in any life-cycle model, ha he rade-off beween curren and fuure is deermined by he raio of he ineres rae and he ime preference rae, and by he degree of risk aversion. We can also derive he following physical wealh accumulaion equaion: a i, j a i, j-1 N, j20a i, j j1 (6) where a i,j is physical wealh asse of generaion i a age ( j+20) and PA is he aggregaed privae asse a period. (3) Firm behavior The privae secor producion is represened by he Cobb-Douglas producion funcion wih consan reurn o scale. The firm decides he demand for physical capial and effecive labor in order o maximize is profi wih he given facor prices of wage and ren, which are deermined in he perfec compeiive markes. 1- Y AK L (7) e, -1 where Y is oupu, sands for capial income share, A is a scale parameer, is he depreciaion of physical capial, K is he physical capial sock, and L e is he effecive labor. We can derive wo facor prices, he rae of reurn r and he wage rae per uni of effecive labor w, by he firs-order condiions for firm s maximum profi: Z-20 1 r (1- r ) (1- w -p )w (1 ) e j - (1 c )ci, j, PA K I (1- ) K (8) r AK L -, w (1- )AK L (9) e, - e, (4) The Governmen The governmen secor has revenue hree ypes of axes, wage ax, consumpion ax and capial ax, as well as he public deb issuance income, and consumpion, invesmen, and ineres paymens as expendiures. Q-20 Z-20 Z-20 w (1 ) e j N, j20 c ci, jn, j20 r T w a N (10) j1 j1 We keep all ax raes consan. The role of he governmen is o endogenously deermine he rae of he public deb issue as a residual of governmen expendiure and revenue. j1 i, j, j20 7

8 Bond G - T D (1 r )D -1 Bond where Bond is he public deb issue a ime, G sands governmen expendiure a ime, T denoes ax revenue a ime, D denoes public deb a ime. In addiion, Dekle (2004) has clarified he fac ha educaion and healhcare impacs o GDP wih immigraion are close o hose wih no immigraion, bu he impac of social securiy o GDP wih immigraion is dramaically changed. Therefore, we assume ha he governmen expendiure, excep governmen subsidies o public pension, in our model is proporional o GDP. As menioned above, he public deb issue is se endogenously due o he difference beween expendiure and ax revenue. I should be noed ha he public deb issue o GDP raio will change over ime as a resul of possible imbalances beween revenues and expendiures. Thus we do no know wheher he fiscal policy of a counry is susainable and if he governmen s ineremporal budge consrain mus be saisfied. (11) (5) The public pension The pension secor grans a pension o he reiremen generaions while pension conribuion is colleced from he working generaions. P Q-20 j1 p w (1 ) e N (12) j, j20 where P sands for he aggregaed pension conribuion. The aggregaed pension benefis a ime is given by he produc of he populaion of reiremen age, replacemen rae, and average earnings during heir working ime Z-20 jq-19 j, j20 W j. B W N (13) where denoes replacemen rae and B is he aggregaed pension benefi. We explicily model he public pension sysem as pay-as-you-go. The budge consrain of he pension secor can be shown as follows: P B (14) (1 sp) where sp denoes public subsidy o pension, which is financed by governmen expendiure G. Moreover, we assume ha he public pension secor mainains a fixed replacemen rae exogenously. As a resul, in our model, he pension conribuion rae is endogenously deermined in order o keep he budge consrain (14). (6) World equilibrium Our model of an open economy in global equilibrium requires several condiions. We assume perfec capial mobiliy across regions. Therefore, he rae of reurn on financial invesmen is 8

9 equalized across all counries: m m ~ ( 1- r )r r (15) where m denoes he m h region. Second, he aggregae value of world asses equals he marke value of he worldwide capial socks plus he value of all ousanding regional governmen bonds: n m1 m1 m n m n m PA K D (16) m1 where m denoes he m h region, and n he oal number of regions which is 16 in our model. Third, he sum of he ne migraion across all 16 world regions equals zero: n m 1 MM m 0 (17) Finally, in order o close he model srucure logically, he good marke mus reach equilibrium. Aggregae worldwide supply be equal o aggregae worldwide demand. n m n m m m C G K m1 m1 m m Y 1 -(1- ) K (18) 3. The Daa, Calibraion, and Scenarios 3.1. Daa and calibraion We disinguish 16 world counries/regions: Japan, Unied Saes of America, China, Souh Korea, India, Indonesia, Bangladesh, Philippines, Thailand, Vienam, Brazil, Peru, Ausralia, Canada, Europe, and Res of World. All regions excep he Unied Saes, Europe, and Res of World are ne immigraion suppliers o Japan according o he saisics of regisered foreign naionals published by he Japanese Minisry of Jusice. Our demographic projecion model for hese regions is calibraed o fi he Unied Naions (UN) (2007) projecions. The UN publishes daa for each year over for five-year age cohors. However, he daa such as age-specific feriliy raes are given only a quinquennial frequency. Therefore, we inerpolae beween age groups and ime inervals in order o fi our populaion model o he UN populaion daa for To simplify he calculaion, we assume ha age-specific feriliy raes are consan a he curren level. In addiion, we presen he values of he main parameers and exogenous variables of he model in Table 1. The srucural parameers for he households and firms behaviors are derived from Auerbach and Kolikoff (1987) and various early OLG simulaion sudies in Japan. 2 These parameers, such as echnological progress and preference parameers, are assumed o be consan and equal across all regions, excep capial share parameers which are differen in developed 2 See Homma e al. (1987), Kao (1998, 2002), Sadahiro and Shimasawa (2001, 2003), Uemura (2002), Okamoo (2005), Ihori e al. (2006). 9

10 regions and developing regions. The exogenous variables such as he macroeconomic, fiscal and public pension variables, are mainly derived from OECD (2007) Tax Daabase, ISSA ( ) Social Securiy Programs Throughou he World, and Whiehouse (2007) Pensions Panorama. 3 We sar our calculaions wih a phase-in period of abou 100 years in order o relax he unrealisic assumpion of a seady sae in 2006, he base year of our simulaion. Moreover, since he model is simulaed over 500 periods, we ensure a sufficienly long period for a achieve seady sae equilibrium. Table 2 repors he acual values of some key variables in 2005 and he prediced values from he model. Noe ha acual and prediced values closely correspond Scenarios Nex we presen simulaed policy scenarios. The scenarios are classified ino four caegories. Scenario 1 assumes a no immigraion baseline case, and Scenarios 2 o 4 assume he arrival of 150,000 immigrans annually. Scenario 5 assumes arrival of 75,000 immigrans annually, afer Scenarios 6 and 7 assume no immigraion, bu a permanen increase in he consumpion ax o 20% and 30%, respecively, beginning in Finally, Scenario 8 is he policy-mix of Scenario 2 (permanen immigraion) and Scenario 7 (30% consumpion ax reform). Scenarios wih immigraion, 1 o 5 and 8, assume ha he age srucure of new immigrans is he same as he age srucure of he curren immigraion populaion regisered wih he Japanese Minisry of Jusice. To evaluae variaions in immigraion policy, each scenario is slighly differen. In Scenarios 2, 5, and 8, he increase of immigraion o Japan is permanen beginning in In Scenario 3, he increase of immigraion is emporary, beginning in 2015 and ending in In Scenario 4, he permanen increase of 150,000 million is delayed, beginning in We do no disinguish beween naives and immigrans once hey have joined he naive Japanese populaion. 4. Simulaion resuls We now urn o describe he simulaion resuls repored in Figures 2 and 3 and Table 3. For his sudy, we focus on Japan raher han oher developed regions facing he effecs of severe populaion aging such as Germany and Ialy. In conras o oher counries, Japan has a very low immigraion rae, and has radiionally admied few immigrans. We presen he resuls of he immigraion scenarios in comparison o he baseline,, consumpion ax reform, and policy-mix (immigraion and consumpion ax reform) scenarios in Japan. 3 In our model, for he reason ha deailed daa of capial ax in some developing regions were unavailable, capial ax variables are assumed o be equal o 20% across all regions. 10

11 (1) Macroeconomic effecs Immigraion conribues o he rise in he working age populaion rae and o he reducion in he raio of older people (65 years old and above). In Scenarios 2 o 4 and 8, we consider an annual flow of 150,000 immigrans. In 2100, he proporion of immigrans reaches 37% of he oal populaion of Japan in Scenarios 2 and 8, 16% in Scenario 3, and 29% in Scenario 4 (Figure 2). In Scenario 5, wih an annual flow of 75,000 immigrans, he proporion of immigrans reaches 21% in We conclude ha increasing he size of annual immigraion makes a subsanial difference in he long-run demographic rend of Japan. Figure 3 shows he reired populaion raio. Because he age srucure of immigrans is generally younger han ha of he hos populaion, immigraion can increase he working-age populaion of he receiving counry. This is rue for our scenarios. While he raio coninues o increase o 37% in he long run wihou immigraion (Scenarios 1, 6, 7), in each immigraion scenario he reired populaion increases a a lower rae. Immigraion lowers he reired populaion raio by a several percenage poins in 2100 relaive o he no immigraion scenario. High immigraion Scenarios 2 and 8 lower he raio by 10 percenage poins. Temporary immigraion (scenario 3) and delayed immigraion (Scenario 4) lower he raio by 3 and 9 poins while low immigraion (Scenario 5) lowers he raio by 5 percenage poins. Thus, he inflow of immigrans reverses he reired populaion raio, implying ha here are more curren workers o suppor he reired populaion. Addiionally, he effec of he delayed, high, and permanen policy is superior o emporary or he immediae, low policy. Because we adop he lifecycle consumpion hypohesis, he savings rae (ne naional savings o GNP) is severely affeced by he rise of he elderly populaion, which is direcly relaed o he reired populaion raio. Immigraion and baseline scenarios each have similar rend in he savings rae; naional savings raes increases in 2010, hen declines o abou -2.4% in However, in every year, he savings raes are higher under he immigraion scenarios. In 2010, he savings raio under he full, permanen, immediae scenario 2 is 14% higher han he no immigraion scenario. Hence, able 3 shows ha he rejuvenaion of he populaion srucure caused by he inflow of immigraion raises he savings raio. In addiion, Table 3 shows ha, in scenarios wih consumpion ax reform (Scenarios 6, 7 and 8), he savings raio is subsanially higher. The raio under hese scenarios, however, also goes down in In order o achieve a higher level of savings, we require more fundamenal fiscal reform such as a drasic cu of governmen deb or immigraion. In fac, he savings raio of Scenario 7 (30% consumpion ax reform) or Scenario 8 (policy-mix) does no go more down in 2050 han ha oher scenarios. Because of he echnology and lifecycle hypohesis assumpions, he GNP is deermined mainly by working-age populaion dynamics. In he baseline scenario, he GNP growh is sagnan, so he level varies lile over 50 years. I rises slowly over he firs wo decades, 2010 o 2030, bu hen 11

12 declines markedly from 2030 o 2050, reflecing he declining labor force. By 2050, GNP declines o 94.56% of 2005 GNP. In all oher scenarios, GNP in 2050 is higher because of he effecs of immigraion or fiscal reform wih he rise of consumpion ax. Therefore, we can confirm ha immigraion promoes expansion of he Japanese economy, because immigraion increases he labor force, which fuels economic growh. The immigraion scenarios increase GNP due o hree effecs: 1) he coninuous immigran flows imply less populaion aging han under no immigraion, 2) immigraion increases he labor force, and 3) he children and he grandchildren of immigrans also conribue o decreased populaion aging and labor force increase in he long run. The impac of a larger labor force is especially noiceable comparing GNP per employee under immigraion scenarios wih he no immigraion scenario. This is shown in Table 3. In Scenarios 6 o 8 (consumpion ax reform or policy-mix), GNP per employee subsanially increases each decade from 2020 o 2050, and by a leas 20 percenage poins beween 2030 and 2040 and beween 2040 and Bu in Scenarios 2 o 5, he increase is smaller because of a larger labor force. The increase in Scenario 2 (permanen immigraion) is % over 2005 GNP, smaller han oher immigraion scenarios. Scenario 4 (delayed immigraion), Scenario 3 (emporal immigraion) and Scenario 5 (half immigraion) have a comparable effec, raising GNP per employee jus over 154% in Therefore, we can confirm ha immigraion promoes quaniaive expansion of he Japanese economy, because immigraion increases he labor force, which fuels economic growh. Now, we briefly evaluae facor prices. In each scenario, due o he capial marke equilibrium, he ineres rae (wage rae) flucuae wihin a narrow range over he cenury. Saring from 3.31% in 2005, he ineres rae decreases unil 2030, increases in 2040 and declines, reaching 3.23% in he year This is rue parly because capial mobiliy equalizes world ineres raes, despie changes in he Japanese naional savings rae due o he aging populaion. The resuls of our simulaion show ha, surprisingly, facor prices are almos he same in all Scenarios. (2) Fiscal and pension effecs Generally, immigraion can be expeced o give he fiscal balance and he public pension budges ambiguous effecs hrough several channels. An inflow of working age immigrans iniially increases he ax base, and hen changes he rend of governmen revenues and pension expendiures. As a resul, governmen deb will be reduced iniially, bu dynamic long-run effecs depend on he permanence and size of immigraion. The aging of he baby boomers increases he reiree raio unil 2050 in all scenarios. In order o balance he budge, he pension premium o wages rae mus be increased. Table 3 shows ha, in Scenario 2 (permanen immigraion), compared o Scenario 1, he long run pension ax on wages is reduced by 0.31 percenage poins in In Scenario 3 (emporal immigraion), he long-run ax 12

13 rae is also reduced by 0.23%, in Scenario 4 (delayed immigraion) by 0.13%, and in Scenario 5 (half immigraion) by 0.15%. Consequenly, he inflow of immigrans reduces, o a minor exen, he ax burden effecs of he aging populaion in Japan. In addiion, hough he public deb-gdp raio gradually increases in each scenario over his cenury because of a fairly low consumpion ax rae in Japan, he raio is reduced by percenage poins in compared o no immigraion. In Scenario 3 (emporary immigraion), he public deb-gdp raio is reduced by percenage poins, in Scenario 4 (delayed immigraion) by 11.45, in Scenario 5 (half immigraion) by 12.52, in Scenario 6 (20% consumpion ax reform) by , Scenario 7 (30% consumpion ax reform) by , and Scenario 8 (policy-mix) by percenage poins. Scenario 3 shows he smalles reducion in he rae of pension ax on wages and public deb o GDP is because immigraion increases are emporary. Hence, he impac is smaller relaive o he ohers. Governmen and pension budges and he pension ax improve in he shor run because of expansion of ax base. In he long run immigrans reire and leave he ax base, decreasing revenues and increasing expendiures. Reired immigrans do no conribue o revenues, and hey do no replace hemselves in he workforce (we assume immigrans have he same feriliy rae as naives, below replacemen). Bu, like he naive populaion, reired immigrans collec pension and increase pension expendiures. Hence, he fiscal and pension balance will deeriorae in he long run. Our conclusions here are no pleasan. Firs, immigraion alone canno fully alleviae he fiscal and burden of he ageing populaion in he shor run. Secondly, reired immigrans become a burden on he Japanese public pension in he long run. (3) Welfare Figure 4 shows generaional welfare of Japanese cohors under Scenarios 1 o 8. These are welfares of subsequen five year age cohors measured in erms of lifeime uiliy relaive o he 1930 cohor under he baseline scenario. The long run increase in he pension ax on wages caused by he progress of aging reduces he amoun of resources available wihin heir lifeime. The long run increase in he public deb o GDP raio also reduces privae capial sock available and possibly decreases fuure growh. Curren and fuure generaions suffer a severe welfare loss under all scenarios. Younger generaions realize large welfare gains from he immigraion scenarios, bu no from fiscal reconsrucion scenarios (Scenarios 6 o 8) when compared wih no immigraion (Scenario 1). They gain from he reducion in he pension ax on wages. The welfares under Scenarios 1, 6 and 7 gradually decline and do no reach a minimum by 1950, bu welfares under Scenarios 2 o 5 and 8 do reach a minimum for he cohor born in In addiion, under full immigraion (Scenario 2), he generaions of Japanese born afer

14 obain a welfare gain because heir share of he pension and public deb is reduced. In paricular, he welfare of he generaion born in 2050 dramaically increases by 14.2 percenage poins. In Scenario 3, he welfare of he generaion born in 2050 also increases by 0.5 percenage poins, in Scenario 4 by 12.4, in Scenario 5 by 8.3, and in Scenario 8 by 0.8 percenage poins. On he oher hand, in fiscal reconsrucion scenarios (Scenarios 6, 7, and 8), he generaions of Japanese born afer 1940 suffer a welfare loss because he higher consumpion ax reduces heir lifeime disposable resources. Therefore, based on he comparison beween he immigraion scenarios, he consumpion ax reform scenario and he policy-mix scenario, we conclude ha immigrans are expeced o make large ne conribuions o he Japanese economy. Immigraion is a good policy for Japan based on no only he macroeconomic, fiscal, and pension impacs bu also welfare improvemens. This policy (Scenarios 2-5) enlarges lifeime resources, which offer gains o he curren and fuure generaions, and reduces he fiscal burden on o he public pension sysem. In conras, he policy ha only increases he consumpion ax reduces he fiscal pressure on he pension sysem much more, bu also reduces ciizen welfare hrough increases in he consumpion ax. Therefore, we recommend a policy-mix such as Scenario 8 which will provide a large benefi o he pension sysem and sill increase welfare. 5. Concluding remarks In his paper, we presened an dynamic general equilibrium OLG simulaion model using 16 counries and regions in order o quanify he economic impacs of immigraion o Japan. Our simulaion resuls show ha immigraion will improve he Japanese economy compared o a no immigraion scenario. Specifically, annual immigran flows of 150,000 will dramaically increase he welfare of curren and fuure generaions. On he oher hand, we can expec a significan long-run welfare improvemen solely from increasing he consumpion ax. If boh fiscal susainabiliy and welfare improvemens for curren and fuure generaions are requiremens, we will need o promoe a policy which includes boh immigraion and addiional fiscal reform, i.e. increasing he consumpion ax. Our sudy does no consider hree poins: 1) he addiional coss of immigraion such as he public educaion cos for immigrans children, 2) he difference beween he impac of high skill and low skill immigrans, and 3) he scenario in which he birhrae of immigrans does no quickly decline o he level of he naive populaion. As indicaed earlier, oher sudies concluded he firs wo are no subsanial. The las poin would make immigraion more favorable. These poins require more research. Finally, increasing immigraion is no a simple maer. Even when he populaion is ageing and growh is sagnan, Japan will face enormous difficulies implemening immigraion policy o 14

15 suppor susained economic growh. From he Japan s social perspecive, can immigraion flows of his order of magniude be socially oleraed over an exended duraion? From he immigran s perspecive, is he Japanese immigraion opporuniy aracive enough o accep he language and culural barriers? And, finally, would such an immigraion policy exacerbae he brain drain ha is already harming ou-migraion counries, and herefore be me wih srong inernaional resisance? Japan, like oher indusrialized counries, mus confron hese difficul obsacles and solve he relaed issues in order o char a producive and viable fuure for is immigran and naive-born populaion. 15

16 References Auerbach, A.J. and L.J. Kolikoff. Dynamic Fiscal Policy. Cambridge Universiy Press: Cambridge, Cabine Office. Kokumin Keizai Keisan Nenpo. Economic and Social Research Insiue, Cabine Office, Governmen of Japan, Tokyo, Dekle, R. (2004). Financing Consumpion In An Aging Japan: The Role Of Foreign Capial Inflows And Immigraion, Journal of he Japanese and Inernaional Economies, vol. 18, pp Fehr, Hans, Sabine Jokisch, and Laurence J. Kolikoff (2004). The Role of Immigraion in Dealing wih he Developed World s Demographic Transiion. FinanzArchiv 60(3): pp Homma, M., N. Aoda, Y. Iwamoo and F. Oake (1987). Nenkin : KoureikaShakai o Nennkin Seido, In Hamada, Horiuchi and Kuroda eds., Nihon Keizai no Macro Bunseki, Tokyo Daigaku Shuppankai, Jun, pp [in Japanese] Ihori, T., Ryua Ray Kao, M Kawade and Shun-ichiro Bessho (2006). Public deb and economic growh in an aging Japan, In K. Kaizuka and Ann O. Krueger, eds., Tackling Japan's Fiscal Challenges: Sraegies o Cope wih High Public Deb and Populaion Aging. [in Japanese] Kao, Ryua R. (1998). Transiion o an Aging Japan: Public Pension, Savings, and Capial Taxaion, Journal of he Japanese and Inernaional Economies vol. 12, pp Kao, Ryua R. (2002). Governmen Defici, Public Invesmen, and Public Capial in he Transiion o an Aging Japan, Journal of he Japanese and Inernaional Economies vol. 16, pp Okamoo, A. (2005). Simulaing Progressive Expendiure Taxaion in an Aging Japan, Journal of Policy Modeling 27 (3), OECD. Tax Daabase, Paris, Sadahiro, A., and M. Shimasawa (2001). Fiscal Susainabiliy and he Primary Surplus: A Simulaion Analysis wih OLG Model, JCER Economic Journal. No. 43. pp [in Japanese] and (2003). The compuable overlapping generaions model wih an endogenous growh mechanism, Economic Modeling vol. 20 No. 1 pp Soresleen, K. (2000). Susaining Fiscal Policy Through Immigraion, Journal of Poliical Economy, vol. 108 (2), pp , (2003). Fiscal Implicaions of Immigraion A Ne Presen Value Calculaion, Scandinavian Journal of Economics., vol. 105 (3), pp The Inernaional Social Securiy Associaion (ISSA). Social Securiy Programs Throughou he World: Asia and he Pacific, Social Securiy Programs Throughou he World: The Americas,

17 . Social Securiy Programs Throughou he World: Europe, Uemura, S. (2002). Lifecycle General Equilibrium Analysis on Social Securiy, Keizaironshu vol. 28, No. 1. [in Japanese] Unied Naions. World Populaion Prospecs: The 2006 Revision, New York, Whiehouse, Edward. Pensions Panorama: Reiremen-Income Sysems in 53 Counries, The World Bank,

18 Figure 1 The Toal Number of Immigrans (regisered foreign naionals) ( ) 18

19 Figure 2 Raio of Immigrans o Toal Populaion Figure 3 Reired Populaion Raio 19

20 Figure 4 Simulaion Resuls WELFARE WITH EQUIVALENT VARIATION Birh Year 20

21 Table 1 Parameer Values of he Model JAPAN USA CHINA S. KOREA INDIA INDONESIA BANGLADESH PHILIPPINES Uiliy funcion Time preference rae Ineremporal elasiciy of subsiuion 1 / Beques moive Producion funcion Technology progress Capial share in producion Physical capial depreciaion Tax policy parameers Wage ax Capial ax Consumpion ax Inheriance ax Pension policy parameers w 11.3% 11.3% 12.5% 12.8% 10.0% 10.0% 6.3% 9.3% r 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% c 5.0% 5.0% 17.0% 10.0% 12.5% 10.0% 9.5% 6.3% b 25.0% 23.5% 0.0% 25.0% 0.0% 0.0% 0.0% 10.0% Coverage of pension High High Low High Low Low Low High Naional subsidy o pension sp 50.0% 0.0% 0.0% 50.0% 0.0% 0.0% 0.0% 0.0% Replacemen raio 50.3% 38.6% 0.0% 58.5% 0.0% 0.0% 0.0% 0.0% Oher parameers Age of reiremen Q Average life expecancy Z

22 Table 1 Parameer Values of he Model (coninued) THAILAND VIETNAM BRAZIL PERU AUSTRALIA CANADA EUROPE REST OF WORLD Uiliy funcion Time preference rae Ineremporal elasiciy of subsiuion 1 / Beques moive Producion funcion Technology progress Capial share in producion Physical capial depreciaion Tax policy parameers Wage ax Capial ax Consumpion ax Inheriance ax Pension policy parameers w 10.5% 10.0% 6.9% 10.5% 11.3% 8.2% 11.5% 9.5% r 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% 20.0% c 7.0% 10.0% 16.8% 10.5% 10.0% 10.0% 17.0% 13.0% b 0.0% 0.0% 2.0% 0.0% 0.0% 0.0% 10.0% 9.0% Coverage of pension Low Low Low Low High High High Low Public subsidy o pension sp 0.0% 0.0% 0.0% 0.0% 50.0% 50.0% 50.0% 0.0% Replacemen raio 0.0% 0.0% 0.0% 0.0% 40.0% 42.5% 60.2% 0.0% Oher parameers Age of reiremen Q Average life expecancy Z

23 Table 2 Year 2005 of he Baseline Scenario JAPAN USA CHINA S. KOREA INDIA INDONESIA BANGLADESH PHILIPPINES Prediced Naional Income (% of GNP) Privae consumpion 81% 84% 68% 76% 81% 69% 92% 78% Governmen purchases of goods and services 16% 13% 18% 16% 16% 13% 13% 12% Curren accoun 6% 6% 17% 10% 5% 20% -3% 12% Ne naional saving 10% 14% 24% 19% 15% 28% 8% 23% Governmen Indicaors Pension premium o wage 11% 9% 0% 10% 0% 0% 0% 0% Gross public deb (% of GDP) 180% 64% 12% 25% 58% 34% 37% 56% Primary balance (% of GDP) -2.71% 0.00% 0.50% -1.86% 0.09% 0.18% -0.02% -0.10% Tax revenues (% of GNP) 13% 13% 18% 14% 17% 14% 13% 12% Wage ax 7% 7% 9% 7% 8% 7% 5% 7% Consumpion ax 3% 3% 8% 5% 7% 4% 6% 3% Capial ax 4% 3% 2% 3% 2% 2% 2% 2% Oher Indicaors Capial oupu raio 254% 381% 543% 377% 357% 540% 258% 499% Ineres rae 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% Acual Naional Income (% of GNP) Privae consumpion 74% 87% 50% 68% 70% 75% 78% 83% Governmen purchases of goods and services 18% 16% 14% 14% 11% 8% 5% 9% Curren accoun 3.6% -6.1% 7.1% 1.9% -1.0% 0.1% -0.3% 1.9% Ne naional saving 12.4% 0.3% 40.8% 18.6% 24.0% 16.9% 20.7% 19.4% Governmen Indicaors Pension premium o wage 14.6%* 12.4%* - 9%* Gross public deb (% of GDP) 175.3% 62.4% 22.8%** 24.7% 78.1%** 41.5%** 37.4%** 53.5%** Primary balance (% of GDP) -5.9% -1.6% N.A 1.7% N.A N.A N.A N.A Tax revenues (% of GNP) 12% 11% 9% 16% 10% 11% 8% 12% * ISSA ( ) "Social Securiy Programs Throughou he World" ** S&P (2009) "Sovereign Risk Indicaors" 23

24 Table 2 Year 2005 of he Baseline Scenario (coninued) Prediced Naional Income (% of GNP) THAILAND VIETNAM BRAZIL PERU AUSTRALIA CANADA EUROPE REST OF WORLD Privae consumpion 69% 70% 76% 75% 78% 81% 77% 79% Governmen purchases of goods and services 12% 14% 15% 15% 14% 11% 18% 16% Curren accoun 21% 19% 12% 12% 11% 11% 9% 8% Ne naional saving 28% 29% 21% 22% 21% 20% 14% 19% Governmen Indicaors Pension premium o wage 0% 0% 0% 0% 11% 12% 19% 0% Gross public deb (% of GDP) 38% 42% 45% 29% 17% 78% 77% 44% Primary balance (% of GDP) 0.35% 0.03% 0.10% 0.05% 0.41% 2.57% 0.51% -0.09% Tax revenues (% of GNP) 13% 14% 15% 15% 14% 13% 18% 16% Wage ax 7% 7% 5% 8% 7% 5% 7% 7% Consumpion ax 3% 4% 8% 5% 5% 5% 8% 7% Capial ax 2% 2% 2% 2% 2% 3% 3% 2% Oher Indicaors Capial oupu raio 567% 572% 494% 473% 414% 426% 180% 403% Ineres rae 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% 3.3% Acual Naional Income (% of GNP) Privae consumpion 72% 70% 84% 81% 78% 76% N.A N.A Governmen purchases of goods and services 12% 6% 21% 11% 19% 20% N.A N.A Curren accoun -4.6% 0.4% 1.7% 1.5% -5.8% 2.1% N.A N.A Ne naional saving 19.0% 27.7% 4.6% 8.4% 7.1% 9.5% N.A N.A Governmen Indicaors Pension premium o wage %* 9.9%* N.A N.A Gross public deb (% of GDP) 26.3%** 39%** 56%** 23.3%** 16.7% 70.3% N.A N.A Primary balance (% of GDP) N.A N.A N.A N.A 2.7% 2.6% N.A N.A Tax revenues (% of GNP) 18% 17% 13% 14% 25% 15% * ISSA ( ) "Social Securiy Programs Throughou he World" ** S&P (2009) "Sovereign Risk Indicaors" N.A N.A 24

25 Table 3 Simulaion Resuls JAPAN GNP GNP per employee Ne naional saving-gnp raio Capial-labor raio Ineres rae Wage rae Deb-GDP raio Pension premium o wage Pension burden-gdp raio Scenario % % 10.23% % 3.31% % % 11.11% 15.55% Baseline % % 16.43% % 3.25% % % 11.71% 16.40% % % 10.43% % 3.22% % % 12.39% 17.35% % % 9.67% % 3.20% % % 12.70% 17.78% % % 7.88% % 3.20% % % 12.46% 17.44% % % 4.13% % 3.25% % % 13.68% 19.15% % % -2.70% % 3.23% % % 14.80% 20.72% Scenario % % 10.27% % 3.31% % % 11.11% 15.55% High (150,000) % % 16.46% % 3.25% % % 11.71% 16.40% % % 10.46% % 3.22% % % 12.39% 17.35% % % 9.68% % 3.20% % % 12.69% 17.77% % % 7.88% % 3.20% % % 12.43% 17.40% % % 4.17% % 3.25% % % 13.57% 19.00% % % -2.32% % 3.23% % % 14.49% 20.29% Scenario % % 10.27% % 3.31% % % 11.11% 15.55% Temporary % % 16.47% % 3.25% % % 11.71% 16.40% % % 10.46% % 3.22% % % 12.39% 17.35% % % 9.69% % 3.20% % % 12.69% 17.77% % % 7.92% % 3.20% % % 12.43% 17.40% % % 4.25% % 3.25% % % 13.59% 19.02% % % -2.27% % 3.23% % % 14.57% 20.40% Scenario % % 10.27% % 3.31% % % 11.11% 15.55% Delayed % % 16.48% % 3.25% % % 11.71% 16.40% % % 10.48% % 3.22% % % 12.39% 17.35% % % 9.71% % 3.20% % % 12.70% 17.78% % % 7.90% % 3.20% % % 12.46% 17.44% % % 4.15% % 3.25% % % 13.65% 19.11% % % -2.55% % 3.23% % % 14.67% 20.54% 25

26 Table 3 Simulaion Resuls JAPAN (coninued) GNP GNP per employee Ne naional saving-gnp raio Capial-labor raio Ineres rae Wage rae Deb-GDP raio Pension premium o wage Pension burden-gdp raio Scenario % % 10.27% % 3.31% % % 11.11% 15.55% Low (75,000) % % 16.47% % 3.25% % % 11.71% 16.40% % % 10.48% % 3.22% % % 12.39% 17.35% % % 9.71% % 3.20% % % 12.70% 17.77% % % 7.92% % 3.20% % % 12.44% 17.42% % % 4.21% % 3.25% % % 13.63% 19.08% % % -2.43% % 3.23% % % 14.65% 20.51% Scenario % % 7.43% % 3.31% % % 11.12% 15.56% 20% ax % % 12.92% % 3.26% % % 11.72% 16.41% % % 19.28% % 3.23% % % 12.41% 17.38% % % 19.03% % 3.21% % % 12.72% 17.80% % % 18.57% % 3.21% % % 12.47% 17.46% % % 17.17% % 3.25% % % 13.69% 19.17% % % 14.40% % 3.23% % % 14.81% 20.73% Scenario % % 5.74% % 3.32% % % 11.12% 15.57% 30% ax % % 10.77% % 3.27% % % 11.73% 16.42% % % 23.77% % 3.24% % % 12.42% 17.39% % % 23.76% % 3.21% % % 12.73% 17.82% % % 23.83% % 3.21% % % 12.48% 17.48% % % 23.32% % 3.25% % % 13.70% 19.18% % % 21.98% % 3.23% % % 14.82% 20.74% Scenario % % 5.78% % 3.32% % % 11.12% 15.57% Policy Mix % % 10.80% % 3.27% % % 11.73% 16.42% High and 30% Tax % % 23.79% % 3.24% % % 12.42% 17.39% % % 23.77% % 3.21% % % 12.72% 17.81% % % 23.83% % 3.21% % % 12.46% 17.44% % % 23.32% % 3.25% % % 13.59% 19.03% % % 22.05% % 3.23% % % 14.51% 20.31% 26

27 27