The Demographic Window, Human Capital Accumulation and Economic Growth in China -- An Applied General Equilibrium Analysis

Transcription

1 The Demographic Window, Human Capial Accumulaion and Economic Growh in China -- An Applied General Equilibrium Analysis Xiujian Peng Ausralian Insiue for Social Research The Universiy of Adelaide Absrac: This paper invesigaes he conribuion of human capial o economic growh and is poenial scope o neuralize he effecs of populaion ageing in China during he firs half of he 21 s cenury by using an applied general equilibrium model. In paricular, i examines wheher he demographic window which opened during he 1990s and is projeced o say open unil he 2030s can be exploied o injec a growh boos ino he economy. Such a boos could be delivered hrough policies of energeic human capial formaion and significan governmen spending on social infrasrucure. Policy simulaions show ha acceleraing human capial formaion may dramaically raise boh oal oupu and per capia real income. Public expendiure on educaion is, herefore, welfare enhancing, miigaing he adverse effecs of populaion ageing while providing a respecable environmen for China s rapidly increasing elderly populaion. Key words: Populaion ageing, Human capial, Demographic Window, Economic growh, China The auhor wishes o hank Peer Dixon, Yinhua Mai, Dierich Fausen and wo anonymous referees from Asian Populaion Sudies for heir very useful commens. xiujian.peng@adelaide.edu.au. 1

2 1. Inroducion Along wih many developed counries, he dramaic feriliy decline since he 1970s and increasing life expecancy in China are giving rise o a serious ageing problem. Even hough his problem will aain is full scope a couple of decades laer han in mos of he advanced economies, i is likely o be significanly more severe. China s subsanially lower income and significanly faser ageing rae, relaive o he ypical advanced economy scenario, will give rise o subsanially larger resource requiremens o cope wih populaion ageing problem. Scholars are increasingly concerned abou he fallou from he rapid progress of populaion ageing on economic growh and living sandards, and hey sugges ha an acive policy response is urgenly indicaed. However, some scholars argue ha he above skech exaggerae he negaive effec of populaion ageing. Day and Dowrick (2004) argue ha declining feriliy, increasing female labour force paricipaion and increasing human capial formaion may offse he negaive effecs of populaion ageing. Fouègre and Merée (1999) observe ha by raising he presen value of fuure real wages, 1 ageing could creae new incenives for fuure generaions o inves in human capial. This, in urn, could compensae for he poenially negaive effecs of ageing. The curren debae abou he poenial economic effecs of populaion ageing concenraes on he conribuion of human capial o economic growh. Growh of he human capial sock migh miigaes he adverse effecs of populaion ageing, since whils he working age populaion is becoming smaller relaive o he aged populaion, i is also becoming smarer and more producive and is increasing is supply of labour (Day and Dowrick, 2004). Human capial formaion and he ensuing produciviy improvemen, herefore, provide a venue hrough which he auhoriies could seek o alleviae he adverse economic effecs of populaion ageing. If he slower growh in he labour force were o be accompanied by a faser rae of change in he labour-augmening echnical change, he slowdown in he growh of effecive labour migh be subsanially smaller han he slowdown in he working age populaion (Bosworh, Bryan and Burless, 2004). 1 The declining increase in he labour supply as a resul of populaion ageing will drive up he real wage. 2

3 If he above argumens are correc, hen he adverse effec of populaion ageing in China will be reduced o some exen or may even be enirely offse by an appropriae expansion of he human capial sock in China. This paper invesigaes he conribuion of human capial o economic growh and is poenial scope o neuralize he effecs of populaion ageing in China during he firs half of he 21 s cenury. China s populaion ageing process will creae a demographic window consisen wih he experience of he new indusrialized Souh-Eas Asian Economies. Accordingly, one aim of his paper is o explore wheher he demographic dividend can be harvesed by boosing produciviy improvemen in China hrough he accumulaion of human capial. This produciviy growh would lay a solid foundaion for supporing he rapidly rising proporion of he elderly afer he demographic bonus. The paper conains six secions. The inroducion is followed by he modelling framework and idenificaion of key assumpions of baseline and policy scenarios in secion wo. The hird secion explores he conribuion of human capial o China s economic growh by over he period 2000 o The fourh secion evaluaes he macroeconomic effecs of acceleraing he rae of human capial formaion by means of increased public expendiure on educaion. The penulimae secion presens conclusions and policy implicaions, and he final secion discusses he limiaions of he model. 2. The modelling framework and key assumpions The analysis is based on he ageing profile associaed wih he medium varian populaion projecion conduced by he Unied Naions (2004). The growh rends of oal populaion, working age populaion and elderly populaion from 2000 o 2050 are summarized in Table A1. 3

4 2.1 The modelling framework The model I used in his paper is a modified version of PRCGEM, an applied general equilibrium model of China. 2 In his analysis, PRCGEM is adaped o run in a sequenial long-run simulaion. Following Mankiw, Romer and Weil (1992), human capial is inroduced ino he producion funcion as one of he inpu facors. 3 The producion funcion in PRCGEM is accordingly modified o become α β η = A K ( L H LAND (1) Y ) where Y denoes real GDP a year, A, K and L are oal facor produciviy (TFP), he physical capial sock and oal employmen, respecively, a year. H is he human capial sock a year represened by average years of schooling of he populaion aged 15 o 64. In his funcion, L H is a skill-adjused measure of labour inpu. LAND represens naural resource land a year. α, β, and η are he shares of he respecive inpus in facor cos. 4 Taking logs and differeniaing oally yields he growh equaion g = α k + β ( l + h ) + ηland + a (2) 2 See Zheng and Fan (1999) for deails of he original PRCGEM and Peng (2005) for deails of he modified version of PRCGEM. 3 There exiss a wide variey of models and heories o explore he imporance of human capial for economic growh. Basically, he differen models can be grouped ino wo heoreical perspecives on he relaionship beween human capial and economic growh (Aghion and Howi, 1998; Benhabib and Spiegel, 1994; and WöБmann, 2000). The firs is he neoclassical view which is pioneered by Mankiw, Romer and Wail (1992). The accumulaion of human capial as a facor of producion drives economic growh, so ha differences in levels of human capial are relaed o differences in oupu levels across counries. This implies ha differences in growh raes across counries are relaed o differences in he raes of human capial accumulaion. The effec of human capial on economic growh in his neoclassical view is called level effec. The second, echnical-progress view, is he cenral par of many endogenous growh models. Here oal facor produciviy (TFP) growh depends on he sock of human capial hrough eiher he domesic producion of echnological innovaion (Romer, 1990) or hrough he adopion and implemenaion of new echnology from abroad (Nelson and Phelps, 1966). In eher case, he growh of TFP is posiively relaed o he average level of human capial. The effec of human capial on economic growh in his echnical-progress view is called growh effec. 4 α, β, and η are 0.49, 049 and 0.02,respecively, in 2000 in China. 4

5 where g is he growh rae of real GDP a year. k, l, h are he growh raes of physical capial, employmen and human capial, respecively, a year. a is he growh of TFP. Since he inpu facor land is assumed o be in fixed supply, is growh rae land is zero. Equaion (2) decomposes he growh rae of oupu ino growh of TFP, and a weighed average of he growh raes of physical capial sock and skill-augmened labour. The long-run perspecive is achieved by assuming as exogenous he rae of reurn on capial and he employmen level. The rae of reurn on capial is deermined by he exogenously given world rae of reurn. Capial mobiliy, boh inernally beween secors and exernally beween China and he res of he world, mainains he domesic rae of reurn a he world level. Given flexible wages, changes in aggregae employmen in he long run are deermined by he growh of working age populaion and by labour force paricipaion raes. Wages adjus o ensure ha he labour marke mainains equilibrium. The pah of oal facor produciviy is given exogenously. Is hypohesized increase is based on empirical findings and on a hisorical simulaion of PRCGEM. The growh rae of human capial is reaed as exogenous, and is effec on TFP is ignored in he following discussion. Naional saving raes are exogenous. The paern of aggregae naional savings is assumed o be consisen wih he life-cycle hypohesis and o display a hump shape following he evoluion of he populaion age srucure in China. Financing addiional human capial formaion requires public expendiure ha reduces naional savings. To he exen ha increased invesmen in educaion is funded by foreign invesors, he resuling deb service liabiliies o foreigners widen he gap beween GDP and GNP. This pus pressure on he curren accoun, and may affec he erms of rade and household consumpion. Accordingly, in order o analysis he welfare effecs of addiional human capial formaion, ne foreign asses are endogenously deermined by he model. The addiional human capial formaion alers he aggregae income - expendiure balance. 5

6 These changes are refleced in ne cross-border movemens of goods (ne expors) and funds (capial flows). 2.2 Key assumpions of he baseline and policy scenarios Some imporan srucural characerisics of he economy are esablished in he baseline simulaion and reained in he policy simulaion Growh of Toal Facor of Produciviy: The growh rae of Toal Facor Produciviy (TFP) is reaed as exogenous and se a 2.5 per cen per annum. This assumpion is informed by he following consideraions: Exising empirical research: Young (2000) esimaes ha he annual growh rae of TFP in China from is 1.4 per cen afer a few adjusmens o he official GDP growh rae. Wang and Yao (2003) find ha he average annual growh rae of TFP from 1978 o 1999 is 2.41 per cen. Hisorical CGE simulaion from 1993 o 2000: I have used he PRCGEM o deermine he implici rae of TFP improvemen in China for he period 1993 o This hisorical simulaion yields an annual growh rae of TFP over ha period of 3.5 per cen. The macroeconomic environmen in China: Wih enry ino he WTO and wih he coninued progress of globalizaion, economic aciviy in China is exposed increasingly o global markes. Increasing openness and he associaed simulus o compeiion and innovaion are likely o raise he average produciviy growh rae beyond hisorical levels. These hree sources of informaion yield quie varied esimaes of TFP growh ha range from 1.4 per cen o 3.5 per cen. The sysemaic economeric esimaes by Young (2000) and Wang and Yao (2003) place he produciviy growh rae below 2.5 per cen. A he same ime, he changes in China s economic environmen esablish a presumpion in favor of a shif o a higher growh rae. Consequenly, a rae of he order of 2.5 per cen would seem o be a plausible working hypohesis. 6

7 2.2.2 Labor force paricipaion rae: The aggregae labour force paricipaion rae (ALFPR) depends on he legal reiremen age, demographic age srucure and a hos of oher socioeconomic facors. Following Dugan and Robidoux (1999), I use a simple accouning framework o deermine he rend of he aggregae paricipaion rae from 2000 o PR = j i= 1 s i. PRi, where (3) s / i, = WPi, WP (4) where PR denoes ALFPR in year, and i PR i, is he paricipaion rae of cohor i in year, s. is he share of cohor i in he oal working age populaion aged 15 o 64 ( WP ) in year. Equaion (3) shows ha changes in he ALFPR reflec eiher changes in cohor paricipaion raes or in cohor shares. Tha is, changes in labour force paricipaion may be sympomaic of changes in he composiion of he working age populaion for given cohor paricipaion raes - he demographic composiion effec. Reiremen policy affecs cohor paricipaion raes. In China, he reiremen age is 55 for women and 60 for men. I assume ha he reiremen age will no change unil The developmen of educaion affecs all cohor paricipaion raes. Increases in he school enrolmen rae, especially in upper secondary and eriary enrolmens, ends o reduce he paricipaion rae, especially for he populaion aged 15 o 29. However, higher educaional aainmen of he young generaion will increase heir paricipaion rae as hey reach age 40 years and older. Evidence from Ausralia and Canada shows ha here exiss a posiive relaionship beween educaion level and labour force paricipaion, especially for women (Day and Dowrick, 2004; Dugan and Robidoux, 1999). Given he fac ha a age 45 o 65 women s paricipaion rae in China is subsanially lower han men s (Table A2), he educaional aainmen of women will increase heir paricipaion rae when hey reach 45 years. In his case, he ne effecs of educaion developmen on he ALFPR of all cohors aged 15 o 64 are no quie clear. So I ignore he effecs of human capial accumulaion on he cohor paricipaion raes during he simulaion period and only calculae he demographic composiion effec. 7

8 Daa from China s one-percen-populaion-sample-survey of 1995 shows ha he ALFPR was 85 per cen. The deailed cohor and gender specific paricipaion raes are shown in Table A2. I esimae he rend of ALFPR during 2000 o 2050 by assuming ha he cohor paricipaion raes remain a heir 1995 level. I is convenien o define his effec wih he following equaion: PR = j i= 1 s i. PRi,95 (5) PR is he aggregae paricipaion rae ha would have been observed a ime if all cohor paricipaion raes remained a heir 1995 levels. The esimaion resul is shown in Table A2. The evoluion of he demographic age srucure reduces he ALFPR from 85 percen in 1995 o approximaely 80 percen in Accordingly, he demographic composiion effec over he period of 2000 o 2050 amouns o five percenage poins The unemploymen rae: According o he China Saisical Yearbook he unemploymen rae has flucuaed around 3 per cen since I is assumed ha his level of he unemploymen rae will be mainained unil Baseline scenario: he conribuion of human capial o economic growh Human capial has come o be regarded as a primary source of long-erm economic growh along wih physical capial (Azariadis and Drazen, 1990; Barro and Lee, 1993; Mankiw, Romer, and Weil, 1992). Exising empirical research has demonsraed ha human capial has played an imporan role in he process of China s rapid economic growh. Wang and Yao (2003) esimae ha human capial has conribued 11 per cen o China s economic growh during he reform period ( ). 3.1 The key assumpions in he baseline scenario 3.1.1The growh pah of human capial sock during he period 2000 o 2050 In he baseline scenario, he assumed growh pah of human capial is based on he hisorical evidence, he curren educaion siuaion, and on China s low expendiure on 8

9 educaion. 5 This informaion maps a conjecural growh pah of human capial ha declines from an iniial level of one per cen per annum ( ), o a level of 0.5 per cen ( ), and furher o 0.25 per cen ( ). Under hese assumpions, he average years of schooling of he working age populaion will be 6.90 in 2010, 7.63 years in By 2050, i reaches 8.02, represening an increase of 2.1 years (approximaely 30 per cen) over he saring level (5.96) in Exising empirical resuls Table 1 repors he resuls of wo major empirical sudies of human capial formaion in China: Wang and Yao (2003) and World Educaion Indicaor Programme (2002 ediion). 6 Noe in paricular ha he average growh rae of human capial was lower in he 1990s han in he 1980s. One poenial explanaion for his declining growh rae is ha i is easier o increase he exen of schooling if one sars from a relaively low base. Anoher relevan facor is ha he uni cos of secondary educaion is significanly higher han ha of primary educaion. According o esimaes of he World Educaion Indicaor Programme (WEIP), annual expendiure on primary educaion per suden in China in 1999 was PPP$372 (Purchasing Power Pariy), PPP$467 on junior secondary educaion. Expendiure on senior secondary school per suden increased sharply o PPP$1768, and o PPP$5798 on eriary educaion. In oher words, he cos of one year senior secondary schooling corresponded o almos five years primary schooling and more han hree years junior secondary educaion. The average educaional aainmen of he working age populaion in year 2000 was 5.96, which corresponded approximaely o he level of compleed primary educaion. Since furher increases in average schooling exend ino he secondary ier, and given he proporionaely higher cos of secondary educaion, I assume ha he growh rae 5 The lieraure on long-run forecass of China s human capial sock is quie limied. The very imporan work in his field is Cao and Luz (2004). Since heir research does no forecas he graduaion raes of differen educaion levels and, furhermore, and does no disinguish beween junior and senior secondary educaion, i is difficul o calculae he average years of schooling based on heir projeced resul. 6 Two oher sudies (Barro and Lee, 2000; Ahuja and Filmer,1995) come up wih reasonable findings, bu I did no repor he resul because of boh sudies use differen age ranges. 9

10 of average years of schooling will coninue o decline if he Chinese governmen coninues o mainain he curren low share of GDP spen on educaion. Table 1: The esimaion of human capial sock in China (Average years of schooling of working age populaion aged 15-64) Average annual growh rae (per cen) Wang and Yao * WEIP** * This average annual growh rae covers he year 1990 o 1999 ** WEIP is World Educaion Indicaor Programme Sources: Wang and Yao (2003) and WEIP (2002 ediion). Curren educaion siuaion and low educaion spending in China China has developed a wide base of primary educaion. The ne enrolmen rae of primary school reached 98.6 per cen in 2002 (Minisry of educaion of China, 2003). However, developmen of secondary and eriary educaion is relaively low. Gross enrolmen rae of secondary school was 64 per cen while he gross enrolmen rae of upper secondary school was 38 per cen in 2002 (Minisry of educaion of China, 2003). The graduaion rae from upper secondary was only 16 per cen in The gross enrolmen rae of eriary educaion developed fas in recen years. I reached o 11 per cen in 2000, which is sill lower han he average level of world oal (16.2 per cen in 1996). The sagnan raio of public expendiure on educaion o GDP is one major reason for he slow developmen of upper secondary and eriary educaion (Figure 2). Since he 1980s, public expendiure on educaion in China has flucuaed around 2.5 per cen of GDP, which is even lower han he average level of lowincome counries (around 3 per cen). Considering China s huge populaion size, low income, and low enrolmen and graduaion raes of upper secondary and eriary school, given he proporionaely higher cos of he secondary educaion and subsanially higher cos of eriary educaion, a conjecural growh pah of human capial ha increase from a saring level of 5.96 years in 2000 o 8.02 years in 2050 would seem o be a plausible working hypohesis if he Chinese governmen does no change he curren low share of GDP on educaion. 10

11 3.1.2 The naional saving rae In he baseline scenario I assume ha he governmen will no change he proporionae expendiure of naional oupu on educaion. Hence, I ignore he effecs of public educaion expendiure on naional saving raes and only consider he demographic effecs. The projeced change of he naional saving rae during he simulaion period (2000 o 2050) is consisen wih he life-cycle hypohesis and based on he empirical research conduced by Heller and Symansky (1998). 7 I assume ha China s naional saving rae will increase o 40.8 percen in Subsequenly i will fall o 23 per cen by To reach his arge, I shock he saving rae from 2000 o 2010 o rise by 0.2 percen annually, and o fall by approximaely 1.4 per cen annually from 2010 o Shocks in he baseline scenario Given he foregoing assumpions, annual percenage changes of he supply of labour and populaion derived from populaion projecion (refer o Table A1) are fed ino he model as well as he growh rae of TFP, naional saving rae, and human capial sock. Table 2 summaries he shocks in he baseline scenario. Table 2: Shocks in he baseline scenario (p.a., per cen)* Year Populaion Labour force Toal facor produciviy Saving rae Human capial sock The growh rae refers o he average annual growh rae over each decade Baseline simulaion resul Table 3 repors he growh pahs of real GDP and primary facor inpus during he simulaion period. The average annual growh rae of real GDP is 7.7 per cen during he firs decade, and hen decreases gradually o 3.3 per cen in he fifh decade. The annual 7 See Peng (2005) for he deained discussion on he change of naional saving rae from 2000 o

12 growh rae of GDP is he resul of he change in employmen, physical capial sock, human capial sock and produciviy improvemen. Table 3: Average annual growh rae of real GDP and facor inpus (per cen) Year Real GDP Physical Capial sock Employmen Toal facor produciviy Human capial sock As Table 3 shows, afer he iniial increase during he firs wo decades, employmen falls coninually. This reflecs he progress of populaion ageing. As a resul, China has o rely on increases in he physical capial sock and produciviy improvemen o susain her rapid growh, given he fixed land resource. However, he rae of growh of he capial sock deerioraes sharply afer he firs weny years o approximaely wo per cen a he end of The reason for he low increase of he physical capial sock is ha he demand for capial falls when he rae of growh of he labour force falls. 8 The rapid decline in he labour force and he low growh of capial indicae ha he growh rae of oupu mus be suppored by produciviy improvemen which may be achieved hrough increased oal facor produciviy or hrough expansion of he human capial sock. Table 4 compares he economic growh pahs wih and wihou incorporaing human capial ino he model. Noe ha wih consan growh of TFP (ignore he effec of human capial on TFP) of 2.5 per cen per annum, human capial formaion will miigae he adverse effec of rapid populaion ageing on economic growh by raising he growh raes of all he relevan macroeconomic indicaors. On average, human capial lifs he growh rae of GDP by 17 per cen, invesmen by 8 per cen and households consumpion by 12 per cen during he observaion period. 8 See Back of he Envelope analysis for deailed explanaion for he low growh of physical capial sock in Peng (2005). 12

13 year Table 4: Baseline scenario: he conribuion of human capial o China s economic growh (Annual growh rae, per cen) Real GDP wihou HP* wih HP Invesmen wihou HP wih HP Households consumpion Expor Impor wihou Wih Wihou wih wihou wih HP HP HP HP HP HP * HP: Human capial The simulaion also shows ha China accumulaes ne foreign asses (Figure 1). The projeced pah of ne foreign asses implies a ne surplus in he saving-invesmen balance. Even hough he naional saving rae is almos cu in half during he observaion period (from 40 o 23 percen), China coninues o run capial ouflows. This implies ha domesic invesmen declines even more 9. This simulaion resul is consisen wih sandard neoclassical growh heory and he empirical findings in he lieraure 10 : slower labour force growh associaed wih populaion ageing will reduce invesmen demand. The fall in invesmen demand will end o exceed he reducion in saving a some sage of he populaion ageing process (Bosworh, Bryan and Burless, 2004). Figure 1: The sock of Ne foreign asses in he baseline scenario 8000 year US$ billion The assumpion of a fixed capial-invesmen raio ies he growh pah of invesmen rigidly o physical capial formaion (Table 4). 10 Bloom and Williamson (1998), Higgins and Williamson (1997), and Higgins (1998). 13

14 4. Policy shock: acceleraing human capial accumulaion aking advanage of he demographic window 4.1 The growh poenial of human capial sock The imporan role of human capial in he process of economic growh is very clear from he baseline simulaion. Now I will explore he macroeconomic implicaions of a higher growh rae of human capial beyond he levels specified in he baseline. The experience of oher recenly indusrializing counries is examined o provide some coarse indicaion of a plausible arge level for educaional aainmen ha Chinese policy-makers migh srive for. Singapore, Hong Kong, Taiwan and Korea have achieved remarkable economic growh performance during he pas 50 years -- a performance ha has been regarded as he Asian economic miracle. The dramaic progress in educaion (Table 5) was one imporan reason (Tilak, 2002). No unexpecedly, his achievemen was suppored by a subsanial commimen of resources. In Korea, he average raio of he public expendiure on educaion o GDP during 1980 o 2000 was around 3.8 per cen per annum (Figure 2). Table 5: Educaional aainmen of he oal populaion aged 15 and over (Average years of schooling) Year Hong Kong Korea Taiwan Singapore Average annual growh rae (%) Source: Barro and Lee (2000) Anoher imporan reason for he dramaic increase in human capial is ha hese counries have fully aken advanage of he demographic dividend. The declining school 14

15 age populaion and low oal dependency raio driven by he feriliy ransiion during he las half-cenury have creaed he demographic windows in hese counries. The demographic window combining wih he subsanial invesmen in educaion and effecive educaion reform has boosed he human capial formaion. This common experience provides an insrucive example ha China could profiably follow. I involves wo basic elemens: Firs, a significan increase in public expendiure on educaion Compared wih he Asian igers, China s public spending on educaion is quie low (Figure 2). I has sagnaed around 2.5 per cen of GDP since he 1980s. A he same ime, roughly 30 percen of GDP is devoed o physical invesmen. In he U.S. hese figures are 5.4 and 17 percen, respecively. In Souh Korea, hey are 3.7 and 30 percen. China s raio of annual invesmen in physical capial o human capial is much higher han in mos counries around he world (Heckman, 2002). The goal of China s governmen is o increase he share of public expendiure on educaion o four per cen of GDP (Minisry of Educaion, 1996). This goal has been se up in However, he share coninues o sagnae a 2.4 percen in If he Chinese governmen can manage o fulfill is goal, hen he growh rae of human capial will be significanly acceleraed. 15

16 Figure 2: Public expendiure on educaion (as per cen of GDP) % year China Hong Kong Singapore Souh korea Low income counries * Daa for Souh Korea in 1996 and 1997 were unavailable Sources: World Bank (2004) and UNESCO (2003). Secondly, exploi he opporuniy of demographic window The dramaic feriliy rae decline since he early 1970s has creaed a demographic window for China (Figure 3). The oal dependency raio has been declining since In year 1990, he oal dependency raio declined o he level below 0.5. This golden age srucure characerized by oal dependency raio below 0.5 percen will las approximaely 40 years (from 1990 o around 2030) in China. Afer 2030, he oal dependency raio will once again rise above 0.5 since populaion ageing drives up he old dependency raio. The rapidly increasing old populaion will impose a heavy burden on China s public expendiure. This implies ha China may no be able o afford o increase he share of public expendiure on educaion beyond However, before oal dependency raio rises up 0.5 again, China can ake advanage of his demographic window o expand her educaion rapidly. This argumen is based on he following facs: 16

17 Figure 3: Trends of dependency raios and he "Demographic Window" in China year Youh dependency raio Old dependency raio" Toal dependency raio Source: Unied Naions (2004) Declining School age populaion: he fall in youh dependency raio permis schooling per child o rise, adding furher o fuure economic growh (Bloom, Canning and Malaney, 1999). In China he rapid feriliy decline has caused a noable reducion of he school-aged populaion (Table A3). The percenage of school age populaion aged 5-14 (including primary and lower secondary school) has declined since he early 1980s. From 2000 o 2050, he school age populaion aged 5-14 is expeced o decline by 33 per cen and upper secondary educaion populaion (aged 15-19) is expeced o decline by 27 per cen. Togeher, he oal school age populaion (aged 5-24) is expeced o decline by 29 per cen. Changing Composiion: As feriliy raes fall, demand for primary educaion coninues o drop and, over ime, his effec will be repeaed a he secondary level. However, demand for higher educaion will simulaneously be rising (a leas for a period of ime), as hose leaving secondary educaion consider aemping o gain more advanced qualificaions (Bloom, Canning and Sevilla, 2003). The dramaic decline in primary school age sudens permis he ransfer of educaion expendiure from primary o lower and upper secondary school or o eriary educaion. Empirical research suggess ha human capial may play a sronger role in he growh process once i reaches a cerain hreshold. Upper secondary 17

18 and eriary achievemen is imporan for economic growh o ranslae ino seady growh (WEIP, 2002 Ediion). Policy-makers can use he demographic dividend, i.e. he saving from he reducion in number of primary level, o shif resources oward broadening access o more advanced forms of educaion. The ne resul of hese demographic changes in an environmen of sable spending on educaion is o raise he level of educaional aainmen. The reiring populaion is characerized by low educaional aainmen. The young school-age cohor is he beneficiary of addiional educaion. The ineracion of hese processes increases he average educaional aainmen of labour force. However, he poenial benefi of he demographic window can no be reaped auomaically. I occurs only when a counry has social, economic and poliical insiuions and policies ha allow i o realize and capure his growh poenial (Bloom and Williamson, 1998; Asian Developmen Bank, 1997; Bloom, Canning and Sevilla, 2003). The demographic ransiion creaes condiions where people will end o inves in heir and heir children s healh and educaion, offering grea economic benefis, especially in he modern world s increasingly sophisicaed economies. Bu governmen invariably plays a vial role in creaing an environmen where high-qualiy healh and educaion provision is possible (Bloom, Canning and Sevilla, 2003). The demographic dividend is ime-limied. If he Chinese governmen akes advanage of his demographic window of opporuniy and increases educaion spending significanly, hen i may accelerae he growh rae of human capial formaion, possibly achieving by 2050 average educaion levels ypical of developed counries in he early 1990s of more han en years. In order o achieve his goal, I assume ha he Chinese governmen subsanially increases he spending on educaion o he level ha ensures he average years of schooling of working populaion increase by 1.5 per cen per annum during he firs hree decades, and by 0.5 per cen during he las wo decades of he firs half cenury. Based on his assumpion, he average years of schooling of he working age populaion will be years by year

19 4.1.2 Naional saving raes in policy scenario The significan increase in educaion expendiure by he governmen will reduce he public saving rae. Since deailed daa of he public educaion cos per suden a differen school levels is no available for China, I canno esimae he effecs of addiional human capial formaion on naional saving raes. I will use he findings of Heller and Symansky (1998) o roughly represen he effecs. Their esimaion of he change in he naional saving rae akes accoun of he addiional effecs of governmen policy change such as educaion, medical oulays and pension coverage besides he effecs of populaion ageing. According o heir esimaion, China s naional saving rae will decline from 40 percen in 2000 o 38.3 in 2010, and furher o 20 percen by 2050 (hree percenage poins lower han he baseline scenario). Based on he foregoing discussions, he deviaions from baseline of he growh raes of human capial sock and naional saving raes in he policy scenario are pu ino he model 4.2 Policy simulaion resul The policy simulaion shows ha rapid human capial accumulaion is expeced o boos all macroeconomic variables (Table 6). For example, he annual average growh rae of real GDP is likely o be raised by 0.66 percenage poins during he firs half cenury. By 2050, real GDP will be 24 per cen larger han baseline. The rapid expansion of China s economy provides a relaively solid foundaion for supporing he rapid growh of he elderly populaion. Households living sandards also improve. The annual growh rae of per capial real consumpion increases by approximaely 0.6 percenage poins during 2000 o By he end of 2050, per capial real consumpion exceeds he baseline scenario by 22.3 per cen. Noice ha he resource requiremens of addiional human capial formaion reduce consumpion growh below GDP growh (Table 6). Furhermore, he induced increase in physical capial formaion increases invesmen demand. Boh facors narrow he imbalance beween saving and invesmen. The cumulaive resul is by 2050 ne foreign 19

20 asses will be US$1600 billion lower in policy simulaion han in he baseline scenario (Figure 4). 11 This reards he growh of GNP compared o real GDP (Table 6) and, herefore, he growh of per capia real consumpion. 12 Table 6: Macroeconomic effecs of higher growh rae of human capial relaive o baseline Changes in average annual growh rae (percenage poin) Cumulaive deviaion from baseline by 2050 (per cen) real GDP Consumpion Capial sock Invesmen expor impor Real wage rae Per capia real GDP Per capia real GNP per capia real consumpion In sum, he rapid increase in human capial miigaes he adverse effecs of populaion ageing by raising no only he growh rae of oal oupu, bu also he growh in households living sandards. These resuls suppor he finding ha expanding educaion is likely o offse he erosion of living sandards of an ageing populaion by promoing faser produciviy growh (Fouègre and Merée, 1999, and Day and Dowrick, 2004). 11 Policy simulaion shows ha wih he exra invesmen demand induced by addiional human capial formaion, China sill mainains capial ouflows and domesic saving is sill sufficien o mee domesic invesmen demand. 12 Anoher reason for he lower growh of per capia real consumpion is ha faser growh of expors causes he erms of rade o decline. 20

21 Figure 4: Cumulaive deviaions from baseline of he sock of ne foreign asses US$ billion year 5. Conclusions and policy implicaions This paper uses a CGE model o explore he macroeconomic effecs of acceleraing human capial accumulaion by increasing public expendiure on educaion. The policy shock dramaically raises oupu and real consumpion in boh oal and per capia erms o miigae he erosion of living sandards of an ageing populaion., In general erms, he basic policy messages of he simulaion exercise are: Invesmen in educaion is welfare enhancing. The Chinese governmen should significanly increase spending on educaion. This invesmen should be underaken wihou delay because of he long implemenaion lags of such policies in alering he sock of human capial. Imporanly, his increase in expendiure should be imed o uilize he demographic window. Demographic ransiion has significan effecs on he process of human capial formaion (Bloom, Canning and Sevilla, 2003). However a congenial policy environmen is required o ake full advanage of his demographic opporuniy. Such an environmen should be designed o encourage and faciliae access o educaion, especially in rural areas, and o promoe higher reenion raes. 21

22 6. Limiaions and exensions The presen paper ignores he effecs on oal facor of produciviy of changes in he size of he human capial sock. This may underesimae he acual conribuion of human capial accumulaion o economic growh. Secondly, he assumpions in he presen paper abou he growh pah of human capial over he sudy period are very coarse esimaes. The inenion of his simulaion is o provide some qualiaive indicaion of he poenial conribuion ha human capial may make o economic growh in he conex of rapidly changing age srucure of populaion in China. Thirdly, changes in he enrolmen and compleion raes, especially a he upper secondary and eriary educaion affec he oal labour force paricipaion rae. I ignore hese effecs and only ake accoun of he impacs of demographic srucure change. This simplificaion may overesimae he conribuion of human capial. Deailed modelling of he relaionship beween addiional educaion expendiure, changes in human capial and changes in labour force paricipaion rae will be included in an exended version of his paper. Fourhly, Populaion size is assumed no o respond o changes in human capial sock. Empirical evidence shows ha educaion aainmen affecs feriliy rae adversely. Higher educaion is expeced o have lower feriliy level. Wih he increase in human capial sock, he feriliy rae is expeced o be reduced furher in China. 22

23 References: Aghion, P. and Howi P. (1998) Endogenous Growh Theory. Cambridge, Mass. and London: MIT Press. Ahuja, V. and Filmer, D. (1995) Educaional Aainmen in developing Counries New Esimae and projecions Disaggregaed by Gender, Policy Research Working Paper, 1489, World Bank. Azariadis, C. and Draxen, A. (1990) Threshold Exernaliies in Economic Developmen, The Quarerly Journal of Economics, Vol. 105(2), pp Barro, R. J. and Lee, J-W (1993) Inernaional comparison of Educaional Aainmen, Journal of Moneary Economics, 32, pp (2000) Inernaional Daa on Educaional Aainmen: Updaes and Implicaions, Cener for inernaional Developmen a Harvard Universiy Working Paper no. 42. Benhabib, J. and Spiegel M. M. (1994) The Role of Human Capial in Economic Developmen: Evidence from Aggregae Cross-Counry and Regional U.S. Daa, Journal of Moneary Economics, Vol. 34, pp Bloom, E. D, Canning, D. and Malaney, P. N. (1999) Demographic Change and Economic Growh in Asia, CID Working paper No.15 and Williamson, G. J. (1998) Demographic Transiion and Economic Miracles in Emerging Asia, The World Bank Economic Review, Vol.12 (3), pp , Canning, D. and Sevilla, J. (2003) The Demographic Dividend A New Perspecive on he Economic Consequences of Populaion Change, RAND, Sana Monica. Bosworh, B. P., Bryan, R. C. and Burless, G. (2004) The Impac of Ageing on Financial Markes and he Economy: a Survey, Working Paper, Brookings Insiuion Washingon, D. C. Cao, G-Y and Luz, W. (2004) China s Fuure Urban and Rural Populaion by Educaion Level in Luz, W., Sanderson, W. C. and Scherbov, S. (eds.), The End of World Populaion Growh in he 21 s Cenury: New Challenges for Human Capial Formaion and Susainable Developmen. Earhscan, London, pp

24 Day, C. and Dowrick, S. (2004) Ageing Economics: Human Capial, Produciviy and Feriliy, Agenda, Vol.119 (1), pp Dugan, B. and Robidoux, B. (1999) Demographic Shifs and Labour Force paricipaion Raes in Canada, Canadian Business Economics, Summer 1999, pp Fouègre, M. and Merée, M. (1999) Populaion ageing, Inergeneraional Equiy, and Growh: Analysis wih an Endogenous Growh Overlapping Generaion Model, Working paper of he Deparmen of Finance, Canada. Heckman (2002) China s Invesmen in Human Capial, NBER working paper, No Heller, P. S. and Symansky, S. (1998) Implicaions for Savings of ageing in he Asian Tigers, Asian Economic Journal, Vol.12 (3), pp Higgins, M. (1998) Demography, Naional Savings, and Inernaional Capial Flows, Inernaional Economic Review, Vol.23, and Williamson, J. G. (1997) Age Srucure Dynamics in Asia and Dependence on Foreign Capial, Populaion and Developmen Review, Vol.23 (2), pp Mankiw, N.G., Romer, D. and Wei, D. (1992) A Conribuion o he Empirics of Economic Growh, The Quarerly Journal of Economics, Vol. 107(2), pp Minisry of Educaion (1996) The Ninh Five-Year Plan and Developmen Plan 2010 of Naional Educaion, (2003) Saisical Repor of Educaion Developmen in China, Nelson, R. R. and Phelps, E. S. (1966) Invesmen in Humans, Technological Diffusion, and Economic Growh, American Economic Review, Vol.56 (2), pp Peng, X. J. (2005) Populaion Ageing, economic Growh and Populaion Policy opions in China A compuable General Equilibrium Analysis, PhD Disseraion, Monash Universiy. Romer, P. (1990) Human Capial and Growh: Theory and Evidence, Carnegie- Rocheser Conference Series on Public Policy, Vol.32, pp Tilak, B. and Jandhyala, G. (2002) Building Human Capial in Eas Asia: Wha Ohers Can learn, World Bank. UNESCO (2003) UNESCO on line Educaion daa, UNESCO Insiue for Saisics. 24

25 Unied Naions (2004) World Populaion Prospecs: The 2004 Revision online daa. Wang, Y and Yao, Y. (2003) Sources of China s Economic Growh : Incorporaing Human Capial Accumulaion, China Economic Review, Vol.14, pp WöБmann, L. (2000) Specifying Human capial: A Review, Some Exensions, and Developmen Effecs, Kiel Working Paper No.1007, Kiel Insiue of World Economics, Germany. World Bank s World Table, DX for Windows, version 3.0, 2004 World Educaion Indicaors Programme (2002) Financing Educaion Invesmens and Reurns: Analysis of he World Educaion Indicaors, UNESCO Insiue for Saisics and OECD. Young, A. (2000) Gold ino Base Meals: Produciviy Growh in he People s Republic of China during he Reform Period, NBER Working papers Zheng, Y. and Fan, M. (1999) A Chinese CGE Model and is Applicaion o Policy Analysis, Social Sciences Documen Publishing House, Beijing. 25

26 Annex: Year Size (million) Toal populaion Table A1: Populaion projecion in China Average annual growh rae (%) Working age populaion (15-64) Size (million) Average annual growh rae (%) Toal (million) Elderly populaion Proporion of elderly over oal populaion (%) The growh rae refers o he average annual growh rae over each decade. Source: Unied Naions (2004) 26

27 Table A2: Deailed Demographic composiion effec on labour force paricipaion rae in China from 1995 o 2050 (%) Age group PA* (1) cohor shares (2) conribuion o APR* (2)*(1)/100 cohor shares (3) conribuion o APR wih 1995 cohor raes (3)*(1)/100 cohor shares (4) conribuion o APR wih 1995 cohor raes (4)*(1)/100 cohor shares (5) conribuion o APR wih 1995 cohor raes (5)*(1)/ Toal

28 Table A2: Deailed Demographic composiion effec on labour force paricipaion rae in China from 1995 o 2050 (coninued) Age group cohor shares (6) conribuion o APR wih 1995 cohor raes (6)*(1)/100 cohor shares (7) conribuion o APR wih 1995 cohor raes (7)*(1)/100 cohor shares (8) conribuion o APR wih 1995 cohor raes (8)*(1)/ Toal * PA denoes paricipaion rae and APR denoes aggregae paricipaion rae. Source: Daa in column one is calculaed by auhor based on China s one-percen-populaion-sample survey of 1995 and daa in columns wo o eigh is based on Unied Naions medium varian populaion projecion (2004). 28

29 Table A3: school-age populaions in China Year Populaion aged 5-14 Populaion aged Populaion aged Percenage* Toal size (million) Percenage Toal size (million) Percenage Toal size (million) * % of oal populaion Source: Unied Naions (2004). 29