Impacts of Population Aging on Economic Growth and Structure Change in China

Transcription

1 Impacts of Populaton Agng on Economc Growth and Structure Change n Chna The feature of Chnese demographc structure s changng from a hgh fertlty rate, hgh death rate and low lfe expectancy to low fertlty rate, low death rate and hgh lfe expectancy, and the phenomena of ageng populaton n comng future wll become more serous. The data of Sxth Natonal Populaton Census show that the ageng rate of the populaton s hgher than expectatons, the share of populaton wth age 60 years and above s 13.26%, the share of populaton wth age 65 years and above s 8.87%; average number of members of each household s 3.10 persons, ths fgure s 0.34 person less than 3.44 persons of Ffth Natonal Populaton census n Ths demographc change has not only ncreased the burden of socal securty penson and reduced the actve labor force; t wll also nfluence the savng rate and consumpton structure and further affect the economc structure and sustanablty of Chna s economc development. CGE has been appled to many research areas, but the papers consderng the populaton age structure factors n the CGE model are not a lot. Wth the agng of the populaton receved extensve attenton, some foregn scholars began to use the CGE model studyng the agng of the populaton, such as Sang Gyoo's Yoon, Geoffrey J.D.Hewngs (2006), Eujune Km, Geoffrey J.D. HewngsHeedeok Cho (2011), Seryoung Park, Geoffrey J.D. Hewngs (2010). But CGE model study consders more about the sze of the labor supply, the homogenety of the representatve consumer assumpton, but does not take nto consderng that the agng wll affect the economy from consumpton aspect. In fact, wth the economc development, the growth of lfe, and the hgher levels of educaton, t results n the extenson of the retrement age, whch may make the labor force dd not declned as magned, especally as a country wth a large populaton. At the same tme, changes n consumer behavor of people of dfferent ages may be larger and more mportant, Hewng (1982, 1989) ponted out that the household sector and consumer behavor n the CGE model s very mportant, the famly of dfferent age structure generally have dfferent consumpton patterns. Therefore, by mprovng the DRC-CGE model and ntroducng varables to characterze the structure of household consumpton patterns n the demand sde, the paper study dfferent populaton polcy and the famly demographc changes on economc growth and ndustral structure, to provde polcy recommendatons on populaton polcy orentaton. Ths paper starts the study to analyze change of consumpton structure of household of varous age structure based upon survey data ( ) of CHIPS (Chnese Household Income Project) to explore the changng relatonshp between Chna s demographc structure and consumpton structure. Then accordng to the head of household age and famly sze of the household, we dvded the households n the CGE model nto 12 groups (sx groups n rural and urban areas respectvely) to capture the relaton between consumer behavor and demographc structure, and analyze the mpact of demographc change under three dfferent

2 populaton polces scenaros on the Chna s economc growth and structure change usng DRC dynamc recursve CGE model. I. Relatonshp between demographc structure and consumpton structure It s a common thought that agng populaton wll not only effect the relatve rato between consumpton and savng, but also t may affect drectly the consumpton structure, for example, the demand of medcal and health care servce of the elders wll far exceed that of the youth. But there s nearly no domestc quanttatve data whch reflects drectly the relatonshp between populaton structure and consumpton structure. The Natonal Bureau of Statstcs of Chna generally announces only the data of consumpton structure based upon ncome groups, there s no nformaton related to populaton structure n correspondence. Currently, there s only one avalable database to be substtute,.e. survey data of CHIPS (Chnese Household Income Project). Ths database ncludes data of household consumpton and age of head of household, and the varaton of age of head of household of dfferent household can be used to reflect the change of populaton age structure. Ths study used the survey data of the year 1995, 2002 and 2007 of CHIPS (Chnese Household Income Project) to analyze the changng relatonshp between household wth head of household of dfferent ages and consumpton structure. The followng content lsts the statstcal results of groupng of urban Result of analyss of rural household s bascally smlar to urban household. Concrete results are shown n appendx of ths paper. Ths study classfes the households wth age structure of house holder n years nto 9 groups wth nterval of age of fve years; n addton, two groups of household wth age of house holder below 25 years (25 years ncluded) and age of head of household above 70 years (70 years ncluded),there are 11 groups all together. In order to compare the change of consumpton structure n the year 1995, 2002 and 2007, dagrams showng the change of share of consumpton of 8 categores to total consumpton wth the change of age of house holders s prepared n ths paper (Please see the attached appendx). Table 1: The Consumpton Structure of urban household accordng the age of the head of household n 2007 (%) Educaton, Age of Head Food clothng resdence Equpment and daly necesstes Health care and medcal servces Transportaton and communcaton culture and recreaton servce Others total

3 Wth reference to the data of consumpton structure of 2007, the followng facts are shown. (1) Average share of food consumpton s above 40%; there s ncrease of share of food consumpton post retrement because of decrease of ther ncome. (2) The share of consumpton of clothng, resdence, transport and communcatons shows a declnng trend wth the growth of age. (3) The share of expenses on health care and medcal servces ncreases wth the ncrease of age, and t rases relatvely faster post the age of sxty. (4) The share of expenses on educaton, culture and recreaton servce vares wth changng age n the shape of nverted U, peak appears n households wth the age of the head around years. Therefore, the consumpton structure of the old people s dfferent explctly wth the young people, the age of the head of households has mportant mpact on consumpton structure, wth the consumpton of health care and medcal servce wthn the eght large categores of consumpton of urban household, the share of health care and medcal servce s around 5.41% for the group wth age of the head of households around years, whle ths fgure s 13.31% for the group wth age of the householder around years. The major features of change of consumpton structure from 1995 to 2007 are shown n the followng: (1) There are nearly no sgnfcant changes of share of expendture on food and clothng of urban household, the relatonshp wth the age of house holder s bascally smlar from 1995 to 2007; the share of expendture on clothng has a declnng trend wth ncreasng of age, whle the expendture on food s n rsng wth the ncreasng of age. (2) There are relatvely large changes of share on expendture of resdence, health care and medcal servce, transport and communcaton from 1995 to The share of expendture on resdence s around 2% n 1995, t s rsen to above 8% n 2002 and The share of expendture on health care s around 3% n 1995, and there are no sgnfcant changes n accompanyng the change of age. It s ncreased to above 6% n 2007, t shows a rsng trend n accompanyng the ncreasng of ages. The share of expendture of transport and

4 communcaton s around 2% n 1995, and t has nearly no change wth the change of ages, t s rsen above 9% n 2002 and 2007, and t has a declnng trend n accompanyng the ncreasng of ages. (3)There are relatvely large changes of share of expendture on educaton, culture and recreaton n the year of 1995, 2002 and 2007, the share of expendture on educaton, culture and recreaton s the lowest n the year 1995 whle t s the hghest n It may be questoned why t s declned n 2007? Ths may be related to exempton of educaton cost around that year. In addton, the share of expendture on educaton, culture and recreaton has an nverted U structure wth the age of head of household, the peak value appeared n the same age group,.e. n the household wth age of house holder from years old. It can be seen that there s sgnfcant relatonshp between the age structure of house holder and consumpton structure of household from prevous analyss, or to express the fact n dfferent way, there are sgnfcant dfferences of consumpton structure of household wth dfferent age groups. In short, the followng rules are exsted ether from the cross secton data or panel data: consumpton of food ncreases wth the ncrease of age of house holder; share of consumpton of clothng, transport and communcaton decreases wth the ncrease of age of house holder; the share of consumpton of health care and medcal servces ncreases rapdly n accompanyng the ncrease of age of the house holder; the share of consumpton of educaton, culture and recreaton has nverted U shape wth the change of age. II. CGE model descrpton and scenaro desgn 2.1 Man Assumpton of Module of CGE In the orgnal DRC-CGE model the household demand sector only dvdes nto urban households and rural households, so ths could not reflect the mpact of populaton agng from the consumpton demand. In order to reflect the mpact of populaton agng on consumpton, we dvde the household sector nto 12 groups accordng the age of the head and sze of the famly. The followng content manly descrbe the household demand sector n CGE, other module referred to the related lterature ( L Shan tong and He Janwu (2010)). Table 2. Household sectors n urban or rural area n CGE sze of the The age of the head of The age of the head of household The age of the head of famly household below 40 years old between 40 and 60 years old household above 60 years old 1-3 persons (agel40,small) (age4060, small) (ageg60,small) Above 3 persons (agel40,large) (age4060,large) (ageg60,large)

5 Through the detaled descrpton of household sector n CGE model, the changes of the rato of each household group to the whole populaton, reflected demographc changes. Ths paper use ELES consumpton functon to descrbe the demands of each group. µ µ s U = ( C θ ) ( Save ) n s.. t P C + Save = Y n = 1 µ + µ = 1 s P :Consumer commodty prces; Y:dsposable ncome; U :utlty functon; C :quantty of consumpton goods; Save :household savng; θ :survval consumpton of goods; µ :the margnal consumpton tendency. Usng CHIPS date, we estmate budgets share parameter and ncome elastcty of dfferent famly group n basc year However, to mantan the balance of SAM table wth a base perod of consstent data, we use least squares method to calbrate budgets share parameter. 2.2 Scenaros desgn The domestc academc hold dfferent vews on whether the exstng fertlty polcy should be adjusted, and how to adjust the fertlty polcy. They Can be roughly dvded nto three types: to mantan the exstng polcy(to mantan the exstng low level of fertlty),to mpose the strct "one-chld polcy" (the fertlty polcy enforcement should be strengthened, the fertlty rate can be further reduced, such as L Xaopng (2007), Ye Wenzhen(2010), Zhang Yong (2006)) and the relaxaton of exstng polces (advocates adjustng or relaxng the exstng fertlty polcy, such as L Janxn (2005), Lang Zhongtang(2006)).So how the Dfferent populaton polcy wll mpact on Chna's future economc growth and structure, and what s Chna's populaton polcy orentaton? Researchng the problem s sgnfcant. In order to study the mpact of structural change of the populaton, t s necessary to desgn the possble scenaros of structural change of the populaton. By the mean tme, n order to combne better ths study wth Chna s realty, the scenaro of Chna s Populaton Projecton under condtons of varous brth rate polces done by Zeng Y and Jang Lawe n 2009 s used to be the bass of desgn of ths study. It s necessary to pont out that the projecton done by Zeng Y and Jang Lawen (2009) s projecton based on households,.e. projecton of

6 change of future number of households and average number of persons per household wth house holder of dfferent age groups. Accordng to populaton forecast of Zeng Y,Jang lawen (2009), we set up 3 scenaro, wth low fertlty scenaro, medum fertlty scenaro(baselne scenaro), and hgh fertlty scenaro. Table 3 Scenaro assumptons Scenaros Assumptons 1. The total populaton trends and age structure s exogenous, whch s from the medum populaton forecast n Zeng Y Jang Lewen(2009) -Promotng Two chldren for couples ether one s only chldren and Two chldren wth late chldbearng. Medum Scenaro 2. The supply of labor s exogenous, The supply of agrcultural land s exogenous. 3. The tax rate and transfer s exogenous. 4. Balance of payments wll gradually acheve balance from 2010 to Government consumpton s exogenous. 6. TFP s exogenous, the growth rate of TFP s about 2% 1. The total populaton trends and age structure s exogenous, whch s from the LOW scenaro of populaton forecast n Zeng Y Jang Lewen(2009) LOW Scenaro --Mantan the exstng famly plannng polcy,only one chld per famly except both couple are only chld 2.Others assumpton s the same as that of Medum scenaro. 1. The total populaton trends and age structure s exogenous, whch s from HIGH Scenaro the HIGH scenaro of populaton forecast n Zeng Y Jang Lewen(2009)-- Completely release the two-chldren polcy, the fertlty rate surged) 2. Others assumpton s the same as that of Medum scenaro. Furthermore, followng content lsts the key varables change n the demographc changes under the three scenaros. 2,100,000 1,900,000 1,700,000 1,500,000 1,300,000 1,100, , , , Hgh medum low Fgure 1: Total populaton forecast, Unt:thousand

7 Table 4 Labor Supply of 3 Scenaros Unt:10 thousand Scenaros Baselne LOW HIGH LOW relatve to the baselne(%) HIGH relatve to the baselne(%) Table 5 Percentage changes of urban households of the three scenaros (%) Type of househould Rato of the age of head less 40 and scale of 1-3 persons Rato of the age of head above 60 and scale of 1-3 persons Scenaro Baselne Hgh Low Baselne Hgh Low Smulaton results Smulaton results of baselne scenaro Economc growth rate Smulaton results n baselne scenaro show that: the economc growth rate gradually declnes. The average growth rate from 2008 to 2010 s about 9.47%. The economc growth rate s expected to be around 6.83% from 2026 to The economc growth rate n the perod from 2046 to 2050 s expected to be around 2.95%. From perspectve of the source of economc growth, the man drvng force of Chna's economc growth stll les n the accumulaton of captal. About 1.45 percentage ponts, accountng for 49.25% of the correspondng GDP growth rate s due to accumulaton of captal. At the same tme, labor contrbuton to economc growth s small before 2020, and becomes negatve after The contrbuton to GDP growth from 2046 to 2050 s negatve 0.33 percent. In short, the accumulaton of captal s stll the engne of economc growth, but the

8 contrbuton of labor gradually becomes gradually negatve. Table 6 Economc growth rate from 2008 to 2050 (%,baselne scenaro) Baselne 2008~ ~ ~ ~ ~ ~ ~ ~ ~20 50 Economc growth rate Labor growth rate Captal growth rate Growth rate of TFP Contrbuton to economc growth Labor captal TFP Data source : Smulaton result from CGE model. Level of GDP per capta From economes of scale, at 2007 prces, Chna's total GDP n 2030 s lkely to acheve bllon dollar and exceed the total U.S. economy, the world's largest economy countres. As for the GDP per capta, t reaches $ 35,683 n 2050, but t s stll less than that of the U.S.A n Table 7 GDP and GDP per capta n baselne scenaro Indcator GDP(trllon yuan) GDP per capta(10 thousand yuan) GDP(trllon U.S. dollar) GDP per capta (dollar ) Notes: constant prces at year 2007, Exchange rate: 1 dollar =6.6883yuan. Economc structure Due to the agng of the populaton and declnng savng rate, household consumpton share of GDP gradually rse from percent n 2010 rose to 42.66% n 2030 to 59.32% n

9 2050. At the same tme, nvestment rate decreased from 47.78% n 2015 to %. Table 8 Structure of GDP from 2007 to 2050(% baselne scenaro) Structure of GDP Household consumpton Governmental consumpton Captal formaton Net export The proporton of prmary ndustry gradually reduces down to about 5.95% by 2030, a further declne to 0.74% by The proporton of secondary ndustry ncreased slghtly from 2007 to Due to serous populaton agng, the proporton of secondary ndustry has been slowly declne from 50.10% n 2010 to 41.28% n At contrast, the proporton of the tertary ndustry rse from percent n 2010 to 57.98% n Table 9 ndustry structure from 2007 to 2050 (%,baselne scenaro, current prce) Baselne scenaro Prmary ndustry Secondary ndustry Tertary ndustry Comparatve analyss wth the control scenaro In order to analyze the mpact of populaton agng on the economy, we have also desgned a control scenaro that dd not have agng problem. Total populaton growth n control scenaro s the same as that n the baselne scenaro. However, the populaton age structure n control scenaro keeps wth ageng structure n 2010, Therefore, there s no agng problem after 2010 n the control scenaro. Impact of populaton agng on economc growth rate Wth the deepenng of populaton, the negatve mpact of populaton agng on the rate of economc growth ncreases. GDP growth rate of the control scenaro n 2015 s 8.15%, 0.25 percentage ponts hgher than that of baselne scenaro (7.89%). GDP growth rate of the control scenaros n 2030 s 6.76%, 0.57 percentage ponts hgher than that of baselne scenaro. GDP growth rate of the control scenaro n 2050 s 3.86%, 1.26 percentage ponts hgher than that of the baselne scenaro. Therefore, populaton agng has a greater negatve mpact on the chna s GDP growth rate.

10 Table 10 Supply of Labor and GDP growth rate Unt:10000 person Labor supply of baselne scenaro Labor supply of control scenaro Labor supply margn(%) Dfference of GDP growth rate(%) Table 11 rato of household groups n the baselne and the control scenaro Household groups scenaro Below 40 year old and 1-3 persons baselne control above 60 year old and 1-3 persons baselne control Impact of populaton agng on level of GDP Wth the deepenng of populaton agng, the adverse mpact of populaton agng on the level of GDP s ncreasng. Chna's GDP n the control scenaro s trllon yuan n 2030, 7.13% hgher than the baselne scenaro s GDP ( trllon yuan). Moreover, GDP n the control scenaro s trllon yuan n 2050, 24.81% hgher than that of baselne scenaro. Because the populaton of the control scenaro s the same as that of the baselne scenaro, the changes of GDP per capta s the same as that of level of GDP. Table 12 GDP of baseln scenaro and control scenaro Unt: trllon renmnb yuan Scenaro Baselne Control Dfference Note: constant prce at Smulaton results of dfferent populaton polcy scenaro Economc growth rate The result shows that In 2030 GDP growth rate of LOW scenaro s 5.75%, 0.44 percentage ponts lower than the baselne scenaro, and n 2050 LOW scenaro of GDP growth rate of LOW scenaro s 1.32%, 1.28 percentage ponts lower than the baselne scenaro, low-ampltude up to 49.1%. In other words, f we mantan strngent populaton polcy unchanged, n 20 or 30 years, the total populaton and changes n the structure of Chna's economc growth rate wll have a greater adverse mpact on GDP growth rate. Level of GDP per capta Judgng from the sze of the economy, due to the largest populaton and the least degree

11 of agng of Hgh scenaro, total GDP of the hgh scenaro s the most among the 3 scenaros. In 2050, the total GDP of the Hgh scenaro s trllon yuan, percent hgher than that of baselne scenaro. On the contrary, the LOW scenaro's GDP s less than the baselne scenaro, by 2050, Total GDP of the low scenaro s trllon yuan, 20.49% lower than that of the baselne scenaro. Table 13 :GDP of the 3 scenaros Unt: trllon renmnb yuan Scenaros Baselne Low Hgh Low scenaro relatve to baselne (%) Hgh scenaro relatve to baselne(%) From the perspectve of GDP per capta, per capta GDP of hgh scenaro s less than the baselne due to the large populaton n hgh scenaro. In 2030, GDP per capta of hgh scenaro s yuan, 2300 yuan less than that of baselne scenaro. In 2050,Per capta GDP of Hgh scenaro s yuan, 8000 yuan less than that baselne. For the Low scenaro, due the less populaton n the Low scenaro before 2035, per capta GDP of low scenaro s yuan, a lttle more than that of baselne. However, owng to the negatve mpacts of agng, the per capta GDP s 8300 yuan less than that of baselne. All n all, f the populaton polcy of baselne s mplemented, t s conducve for Chna to cope wth the agng of the populaton agng, and makes the hghest level of GDP per capta. Table 14 : GDP Per capta of the 3 scenaro Unt:10 thousand yuan Scenaros Baselne Low Hgh Low relatve to baselne(%) Hghr elatve to baselne (%) GDP expendture structure and ndustral structure The more serous of agng, the hgher the share of household consumpton to GDP, and the lower the share of nvestment to GDP, n 2050, the share of household consumpton to GDP n Low scenaro s 60.56%, 1.25 percentage ponts hgher than the baselne scenaro, 1.84 percentage ponts hgher than n the HIGH scenaro; the share of nvestment to GDP n the LOW scenaro s 18.36%, 5.11 percentage ponts lower than the baselne scenaro, 6.95 percentage ponts lower than n the HIGH scenaro. Table 15:GDP expendture structure of 3 scenaros (%)

12 GDP expendture structure baselne Rato of Household consumpton Low Hgh baselne Rato of Captal formaton Low Hgh From the perspectve of ndustral structure, the more serous of the degree of agng of scenaro, the share of f the tertary ndustry s hgher. In 2050, rato of the tertary ndustry of Low scenaro accounted for 60.38%, 2.4 percentage ponts more than the baselne, 2.96 percentage ponts hgher than the Hgh scenaro,. On the contrary, the proporton of the added value of the secondary ndustry showed the opposte trend wth tertary ndustry. In 2050, the share of the secondary ndustry of the low scenaro accounted for 38.9%, 2.38 percentage ponts lower than that of the baselne scenaro, 2.95 percentage ponts lower than that of HIGH scenaro. Table 16 : Industral structure of 3 scenaros (%) Industral structure Scenaros Rato of the tertary ndustry Rato of the secondary ndustry baselne Low Hgh baselne Low Hgh III. man conclusons Based on the above analyss, there are the followng conclusons n ths paper: (1) Populaton agng wll have greatly negatve mpact on the economc growth. As the agng become serous, GDP growth rate reduces greatly. Smulaton results show that n 2030, GDP growth rate of control scenaro s 6.76%, 0.57 percentage ponts hgher than the baselne scenaro; n 2050 GDP growth rate of control scenaro s 3.86%, 1.26 percentage ponts hgher than the baselne scenaro. (2) In the baselne scenaro when Promotng Two chldren for couples ether one s only chldren and Two chldren wth late chldbearng, the level of GDP per capta s hghest. Smulaton results show that In 2050, the GDP per capta of the Hgh scenaro s 230,700 yuan, 8,000 yuan lower than the baselne medum,

13 the ampltude s 3.34%; the GDP per capta level of LOW scenaro s 230,400 yuan, 8,300 yuan lower than the baselne.