The Im­pact of Ed­u­ca­tion In­put on La­bor Mi­gra­tion and In­equal­ity in China

LiXin(李昕)andGuanHui­juan(关会娟)

China Economist - - Contents - 1 2 Li Xin ( ) and Guan Hui­juan ( )李昕 关会娟 1 School of Sta­tis­tics, Beijing Nor­mal Univer­sity, Beijing, China 2 School of Eco­nom­ics and Man­age­ment, Ts­inghua Univer­sity; China Data Cen­ter, Ts­inghua Univer­sity, Beijing, China

Ab­stract: By in­tro­duc­ing a general equi­lib­rium frame­work to China’s dual eco­nomic struc­ture, this paper stud­ies the mi­cro­scopic mech­a­nism of ed­u­ca­tion in­put to nar­row the ur­ban-ru­ral in­come gap and how to im­prove the al­lo­ca­tion ef­fi­ciency of ed­u­ca­tion funds in China’s “new nor­mal” econ­omy. The em­pir­i­cal anal­y­sis re­sults show that ed­u­ca­tion in­put is ef­fec­tive in nar­row­ing the ur­ban-ru­ral in­come gap and achiev­ing a Pareto im­prove­ment state in both di­rect and in­di­rect ways. How­ever, the ef­fect of ed­u­ca­tion in­put at dif­fer­ent stages varies. In par­tic­u­lar, the im­pact of com­pul­sory ed­u­ca­tion on im­prov­ing in­equal­ity is more sig­nif­i­cant than the other ed­u­ca­tion lev­els. With ad­just­ment for the macroe­co­nomic slow­down, rais­ing fis­cal ex­pen­di­tures on ed­u­ca­tion can im­prove po­ten­tial eco­nomic growth by pro­mot­ing hu­man cap­i­tal ac­cu­mu­la­tion and la­bor pro­duc­tiv­ity in the long run. In all, ed­u­ca­tion in­put can pro­mote la­bor mi­gra­tion and nar­row the ur­ban-ru­ral in­come gap, which is con­ducive to al­le­vi­at­ing the con­tra­dic­tion of the struc­tural trans­for­ma­tion lag in em­ploy­ment and achiev­ing in­clu­sive growth tar­gets.

Key­words: ed­u­ca­tion in­put, the ur­ban-ru­ral In­come gap, la­bor mi­gra­tion JEL Clas­si­fi­ca­tion Codes: J31; J41

DOI:1 0.19602/j .chi­nae­conomist.2018.11.07

1. In­tro­duc­tion

Dur­ing the Twelfth Five- Year Plan pe­riod in China ( 2011- 2015), the ur­ban- ru­ral in­come gap con­tin­ued to nar­row. The ra­tio of per capita dis­pos­able in­come be­tween ur­ban and ru­ral ar­eas de­creased from 3.13 to 1 in 2011 to 2.73 to 1 in 2015. Al­though the rel­a­tive rate of the ur­ban-ru­ral in­come gap nar­rowed year by year, its value con­tin­ued to ex­pand. On one hand, the ur­ban-ru­ral in­come gap was 14833 yuan in 2011, and it ex­panded to 19773 yuan in 2015. The wage in­come gap, which con­trib­utes 75% to the ur­ban-ru­ral in­come gap, is key. On the other hand, a large num­ber of em­pir­i­cal stud­ies show that China’s in­come gap dur­ing that pe­riod, es­pe­cially the ur­ban-ru­ral in­come gap, did not fall but con­tin­ued to ex­pand after con­sid­er­ing non-mon­e­tary fac­tors such as hous­ing price, ed­u­ca­tion, health care, and so­cial se­cu­rity. Based on Chi­nese House­hold In­come Pro­ject Sur­vey data, Li (2003) finds that China’s ur­ban-ru­ral gap has been the largest in the world if in­come-in-kind and sub­si­dies are in­cluded. Han and Li (2011) show that the ur­ban-ru­ral in­come gap con­trib­utes to more than 50% of the gross

na­tional in­come gap and is the main fac­tor in China’s in­equal­ity. If pub­lic health care and un­em­ploy­ment in­sur­ance are con­sid­ered, China’s ur­ban-ru­ral in­come gap may be the largest in the world. The ex­ces­sive ur­ban-ru­ral in­come gap is not con­ducive to the de­vel­op­ment of do­mes­tic con­sump­tion, the ad­just­ment and op­ti­miza­tion of eco­nomic struc­ture, or the sta­bil­ity of so­ci­ety.

A large num­ber of schol­ars have stud­ied the in­flu­en­tial fac­tors of China’s ur­ban-ru­ral in­come gap, and have put for­ward dif­fer­ent sug­ges­tions from mul­ti­ple per­spec­tives. Some of them ex­plore the ef­fect of house­hold reg­is­tra­tion dis­crim­i­na­tion on the ur­ban-ru­ral in­come gap from China’s dual eco­nomic struc­ture. They claim that house­hold reg­is­tra­tion dis­crim­i­na­tion sup­presses la­bor mo­bil­ity, which ex­pands and so­lid­i­fies the ur­ban-ru­ral in­come gap. They ar­gue that the house­hold reg­is­tra­tion sys­tem must be re­formed to nar­row the gap (Sic­u­lar et al. 2007; Wan and Li, 2013). Some schol­ars agree that the keys to im­prov­ing the in­equal­ity be­tween China’s ur­ban and ru­ral ar­eas are the mod­i­fi­ca­tion of in­ter­re­gional house­hold reg­is­tra­tion, eco­nomic open­ness, and the im­ple­men­ta­tion of lo­cal gov­ern­ments’ eco­nomic poli­cies ( Lu and Chen, 2004). Oth­ers be­lieve that the ur­ban- ru­ral in­come gap has a nega­tive re­la­tion­ship to fi­nan­cial ef­fi­ciency and there­fore, re­duc­ing the non-equi­lib­rium of fi­nan­cial de­vel­op­ment can nar­row the gap (Wang and Qiu, 2011). In ad­di­tion to the macro­scopic per­spec­tive, a large branch of re­search uses mi­cro-sur­vey data to find that ed­u­ca­tion may be one of the most im­por­tant fac­tors in af­fect­ing China’s in­equal­ity (Sic­u­lar et al., 2007; Chen et al. 2010). The Coun­try Di­ag­nos­tics Re­port, pub­lished by The World Bank in 2016, points out that the con­tri­bu­tion of China’s ur­ban-ru­ral in­come gap to over­all in­equal­ity in­creased from 37% in 1988 to 54% in 2007, and the most se­ri­ous in­equal­ity in China is ur­ban-ru­ral in­come in­equal­ity. Ed­u­ca­tion is one of the most in­flu­en­tial fac­tors of that in­equal­ity.

Ed­u­ca­tion in­put al­le­vi­ates the ur­ban-ru­ral in­come gap both in di­rect and in­di­rect ways. In­creas­ing ed­u­ca­tion in­put can not only in­crease the mar­ginal pro­duc­tion and in­come of la­bor di­rectly, but also pro­mote la­bor mi­gra­tion in ru­ral ar­eas by re­duc­ing the mi­gra­tion costs and al­le­vi­ate the in­come gap in­di­rectly. Cao and Zhang (2015) prove that if the pro­por­tion of agri­cul­tural em­ploy­ment de­creased by 1 unit, then the ur­ban-ru­ral in­come gap would de­crease by about 1.03 units. In 2015, China’s ru­ral pop­u­la­tion ac­counted for 43.90% of the to­tal pop­u­la­tion, and agri­cul­tural em­ploy­ment ac­counted for 28.3% of to­tal em­ploy­ment, while agri­cul­tural out­put only ac­counted for 9.13% of GDP, and la­bor pro­duc­tiv­ity in the agri­cul­tural sec­tor only ac­counted for 22.34% of that in the in­dus­trial sec­tor. The dif­fer­ence of rel­a­tive pro­por­tions be­tween the ru­ral pop­u­la­tion, em­ploy­ment, out­put, and pro­duc­tiv­ity in the agri­cul­tural sec­tor shows that la­bor mi­gra­tion is still an ef­fec­tive way to im­prove pro­duc­tiv­ity and in­crease in­comes in the agri­cul­tural sec­tor, and ed­u­ca­tion may play an im­por­tant role.

With China’s econ­omy en­ter­ing the “new nor­mal,” na­tional ed­u­ca­tion de­vel­op­ment and ed­u­ca­tion fund in­puts are faced with pres­sures due to the de­cline of fis­cal rev­enues. It is im­por­tant to fur­ther re­search how to im­prove the al­lo­ca­tion ef­fi­ciency of ed­u­ca­tion funds at dif­fer­ent ed­u­ca­tional stages and al­le­vi­ate the ur­ban-ru­ral in­come gap ef­fec­tively, which are con­ducive to re­duc­ing de­vel­op­ment in­equal­ity, pro­mot­ing all so­cial mem­bers’ share in eco­nomic de­vel­op­ment achieve­ments, and achiev­ing in­clu­sive growth tar­gets. This paper com­ple­ments pre­vi­ous stud­ies in the fol­low­ing as­pects: It dis­cusses the mi­cro­scopic mech­a­nism of ed­u­ca­tion in­put on the ur­ban-ru­ral in­come gap, re­fer­ring to both the di­rect ef­fect and the in­di­rect ef­fect, and per­forms em­pir­i­cal anal­y­sis on how to im­prove the al­lo­ca­tion ef­fi­ciency of ed­u­ca­tion funds at dif­fer­ent ed­u­ca­tional stages.

The rest of the paper is ar­ranged as fol­lows: Part II con­structs a general equi­lib­rium model with the ur­ban-ru­ral dual eco­nomic struc­ture and dis­cusses the di­rect ef­fect and in­di­rect ef­fect of ed­u­ca­tion in­put on the ur­ban-ru­ral in­come gap. Part III tests the hy­poth­e­sis of the re­la­tion­ship be­tween ed­u­ca­tion in­put, la­bor mi­gra­tion, and the ur­ban-ru­ral in­come gap based on the dy­namic spa­tial panel model, and dis­tin­guishes the dif­fer­ent ef­fects of ed­u­ca­tion in­put at dif­fer­ent ed­u­ca­tional stages. Part IV con­cludes.

2. The­o­ret­i­cal Frame­work

This paper con­structs a tra­di­tional two-sec­tor model, namely, the house­hold sec­tor and the firm sec­tor. The house­hold sec­tor has the char­ac­ter­is­tic of in­tertem­po­ral con­sump­tion, and the firm sec­tor has the char­ac­ter­is­tic of dual struc­ture. On this ba­sis, this paper dis­cusses the mi­cro­scopic mech­a­nism of ed­u­ca­tion in­put on the ur­ban-ru­ral in­come gap.

2.1 House­holds

Ref­er­enc­ing Song et al. (2011), this paper con­structs a two-gen­er­a­tion Over­lap­ping Gen­er­a­tion Mod­els (OLG model).

As­sume that each in­di­vid­ual has the same pref­er­ence, and the util­ity func­tion is a log­a­rith­mic util­ity func­tion with con­stant-rel­a­tive-risk-aver­sion. Given the pop­u­la­tion growth is ex­oge­nous to zero, and the pop­u­la­tion size is stan­dard­ized as 1. The in­di­vid­ual max­i­mizes util­ity (1) sub­ject to the bud­get con­straint (2). That is,

2.2 Firms

In the de­vel­op­ment of China’s econ­omy, re­source con­straints dif­fer­ently im­pacted ed­u­ca­tion lev­els and la­bor qual­ity of in­di­vid­u­als in ur­ban and ru­ral ar­eas. The im­pact dif­fer­ence was en­doge­nous over a long pe­riod of time and, to a large de­gree, re­sulted in the ur­ban-ru­ral in­come gap (Chao and Shen, 2014). There­fore, this paper di­vides the na­tional econ­omy into the tra­di­tional agri­cul­tural pro­duc­tion sec­tor a and a mod­ern non-agri­cul­tural pro­duc­tion sec­tor b, with the num­ber of la­bor­ers La and Lb re­spec­tively. Here, La+ Lb =L, and L is the to­tal la­bor of the econ­omy.

Let it be given that the tra­di­tional agri­cul­tural sec­tor is la­bor-in­ten­sive and only em­ploys work­ers, and the mod­ern non-agri­cul­tural sec­tor hires work­ers and rents cap­i­tal in com­pet­i­tive fac­tor mar­kets.

To sim­plify the anal­y­sis, as­sume that there are no ad­just­ment costs or de­pre­ci­a­tion of cap­i­tal. The agri­cul­tural sec­tor em­ploys low-skilled work­ers and the pro­duc­tion func­tion has de­creas­ing re­turns to scale. The non-agri­cul­tural sec­tor hires skilled work­ers and the pro­duc­tion func­tion has con­stant re­turns. The pro­duc­tion func­tions in the two sec­tors take the form There are a large num­ber of iden­ti­cal firms in the sys­tem and each has ac­cess to the same prod­uct func­tion and takes A as given. The firms hire work­ers and rent cap­i­tal in com­pet­i­tive fac­tor mar­kets and sell their out­put in com­pet­i­tive out­put mar­kets. The firms max­i­mize prof­its and pay fac­tors into the mar­ginal prod­ucts.

The mar­ginal prod­uct of la­bor is

2.3 General Equi­lib­rium Frame­work

In de­vel­op­ing coun­tries, there is a sur­plus of low-skilled la­bor in the tra­di­tional agri­cul­tural sec­tor and a rel­a­tive short­age of high-skilled la­bor in the mod­ern sec­tor. As­sume that la­bor pro­duc­tiv­ity and the wage level in the non-agri­cul­tural sec­tor are both higher than that in the agri­cul­tural sec­tor. Ac­cord­ing to the dual eco­nomic model (Lewis, 1969; 1979), the high wage will en­tice work­ers to mi­gra­tion from the tra­di­tional sec­tor to the mod­ern one and in­crease their wage in­come. The mar­ginal prod­uct of la­bor and the wage in­come will show a re­verse trend with the change in the num­ber of la­bor­ers be­tween the two sec­tors. There­fore, the la­bor mi­gra­tion will grad­u­ally nar­row the ur­ban-ru­ral in­come gap in the­ory. How­ever, in prac­tice, the la­bor mi­gra­tion is based on the premise that the low-skilled work­ers im­prove their hu­man cap­i­tal and la­bor pro­duc­tiv­ity, which in­volves bor­row­ing costs, time costs, and op­por­tu­nity costs that hin­der la­bor mi­gra­tion (Har­ris and To­daro, 1970; Chau, 1997).

This paper as­sumes that is the cost of la­bor mi­gra­tion from the agri­cul­tural sec­tor to the nona­gri­cul­tural sec­tor. On one hand, ed­u­ca­tion in­put im­proves la­bor pro­duc­tiv­ity and in­come lev­els. On the other hand, higher la­bor pro­duc­tiv­ity re­sults in more job op­por­tu­ni­ties. There­fore, ed­u­ca­tion in­put can re­duce the cost of la­bor mi­gra­tion. That is, , where is the cost of la­bor mi­gra­tion from the agri­cul­tural sec­tor to the non-agri­cul­tural sec­tor.

Each in­di­vid­ual has two choices. One is work­ing in the tra­di­tional sec­tor as a low-skilled worker; the util­ity func­tion of . The other is in­creas­ing ed­u­ca­tion in­put and ac­cu­mu­lat­ing hu­man cap­i­tal to work in the mod­ern sec­tor; the util­ity func­tion takes the form of . Be­cause of the ex­ist­ing cost of la­bor mi­gra­tion, the choice made by an in­di­vid­ual de­pends on the util­ity of that in­di­vid­ual. When the eco­nomic sys­tem achieves its steady state, the util­ity is equal un­der the two

As men­tioned above, the in­come gap, which con­trib­utes to 75% of the ur­ban-ru­ral in­come gap, is the key to af­fect­ing the sys­tem in­equal­ity. There­fore, this paper de­fines the ur­ban-ru­ral in­come gap as the dif­fer­ence be­tween the non-agri­cul­tural sec­tor and the agri­cul­tural sec­tor.

Thus, the ur­ban-ru­ral in­come gap is pos­i­tively re­lated to the cost of the la­bor mi­gra­tion from the agri­cul­tural sec­tor to the non-agri­cul­tural sec­tor, and is neg­a­tively re­lated to the ed­u­ca­tion in­put. Ed­u­ca­tion in­put can nar­row the ur­ban-ru­ral in­come gap ef­fec­tively.

2.4 Mi­croe­co­nomic Mech­a­nism

This paper in­tro­duces two propo­si­tions to ex­plore the im­pact of ed­u­ca­tion in­put on la­bor mi­gra­tion and the ur­ban-ru­ral in­come gap.

Propo­si­tion 1: Ed­u­ca­tion in­put can pro­mote la­bor mi­gra­tion from the tra­di­tional sec­tor to the mod­ern sec­tor

Propo­si­tion 1 and propo­si­tion 2 il­lus­trate that ed­u­ca­tion in­put can pro­mote la­bor mi­gra­tion from the agri­cul­tural sec­tor to the non-agri­cul­tural sec­tor, and nar­row the ur­ban-ru­ral in­come gap. On one hand, with eco­nomic de­vel­op­ment, the ac­cu­mu­la­tion of ma­te­rial cap­i­tal in­creases, and the mar­ginal pro­duc­tion of ma­te­rial cap­i­tal de­clines in the non-agri­cul­tural sec­tor (as shown in equa­tion (7)), while the mar­ginal pro­duc­tion of la­bor in­creases rel­a­tively (as shown in equa­tion (8)). There­fore, in the non-agri­cul­tural sec­tor, the in­creas­ing de­mand of skilled la­bor re­sults in a rel­a­tive short­age of skilled la­bor.

On the other hand, the high wage will at­tract work­ers to mi­gra­tion to the mod­ern non-agri­cul­tural sec­tor from the agri­cul­tural sec­tor. With the con­stant num­ber the gross la­bor force, la­bor mi­gra­tion will de­crease La and in­crease Lb , and Wa will in­crease (as shown in equa­tion (9)) and Wb de­crease (as shown in equa­tion (10)) re­spec­tively.

As dis­cussed above, ed­u­ca­tion in­put can im­prove the ac­cu­mu­la­tion of hu­man cap­i­tal and la­bor pro­duc­tiv­ity and di­rectly in­crease the wage in­come in both the tra­di­tional and mod­ern sec­tors (as shown in equa­tions (11) and (12)). In general, the ed­u­ca­tion lev­els and la­bor pro­duc­tiv­ity of work­ers are both higher in the non-agri­cul­tural sec­tor than in the agri­cul­tural sec­tor. Thus, with the as­sump­tion of the de­creas­ing mar­ginal pro­duc­tion of ed­u­ca­tion in­put, the ef­fect of ed­u­ca­tion in­put pro­mot­ing la­bor pro­duc­tiv­ity is more sig­nif­i­cant than that in the non-agri­cul­tural sec­tor (as shown in equa­tions (13)).

There­fore, in­creas­ing ed­u­ca­tion in­put is con­duc­tive to nar­row­ing the ur­ban-ru­ral in­come gap.

ways. In In­creas­ing sum­mary, the ed­u­ca­tion ed­u­ca­tion in­put in­put al­le­vi­ates can im­prove the ur­ban-ru­ral the ac­cu­mu­la­tion in­come of gap hu­man both cap­i­tal in di­rect and the and mar­ginal in­di­rect pro­duc­tion of la­bor and di­rectly in­crease the la­bor in­come in both the non-agri­cul­tural and agri­cul­tural sec­tors. How­ever, the ef­fect of ed­u­ca­tion in­put pro­mot­ing la­bor pro­duc­tiv­ity is more sig­nif­i­cant than that in the non-agri­cul­tural sec­tor. Thus, in­creas­ing ed­u­ca­tion in­put can nar­row the ur­ban-ru­ral in­come gap. Be­sides that, ed­u­ca­tion in­put can also im­prove the ac­cu­mu­la­tion of hu­man cap­i­tal and the la­bor

pro­duc­tiv­ity of low-skilled work­ers, which al­lows them more job op­por­tu­ni­ties in the non-agri­cul­tural sec­tor and re­duces the cost of la­bor mi­gra­tion, which will de­crease Wb and in­crease Wa . Be­cause of the more sig­nif­i­cant ef­fect of in­creas­ing than de­creas­ing, Wb is still show­ing an up­ward trend. In ad­di­tion, be­cause the in­creas­ing rate of Wb is lower than that of Wa , ed­u­ca­tion in­put can nar­row the ur­ban-ru­ral in­come gap.

3. Em­pir­i­cal anal­y­sis

Based on the dy­namic spa­tial panel model and the data of 31 prov­inces from 1995–2014, this paper tests the hy­poth­e­sis of the re­la­tion­ship be­tween ed­u­ca­tion in­put, la­bor mi­gra­tion, and the ur­ban-ru­ral in­come gap. More­over, we dis­cuss the spe­cial ef­fect of ed­u­ca­tion in­put on nar­row­ing the ur­ban-ru­ral in­come gap and how to im­prove the al­lo­ca­tion ef­fi­ciency of ed­u­ca­tion funds at dif­fer­ent ed­u­ca­tional stages.

3.1 Vari­ables and Data

The main vari­ables in this paper in­clude the ur­ban-ru­ral in­come gap, ed­u­ca­tion in­put, and la­bor mi­gra­tion. The con­trol vari­ables in­clude the de­gree of open­ness, fi­nan­cial ef­fi­ciency, the pro­por­tion of state-owned en­ter­prises, and per capita GDP.

(1) The ur­ban-ru­ral in­come gap (GAP) is the ex­plained vari­able and is mea­sured by the ra­tio of per capita dis­pos­able in­come of ur­ban house­holds and per capita net in­come of ru­ral house­holds. The data are de­rived from the China Sta­tis­ti­cal Year­book. The Na­tional Bureau of Sta­tis­tics of the Peo­ple’s Repub­lic of China ob­tained the data from 1995–2012 based on the ur­ban house­hold sur­vey and the ru­ral house­hold sur­vey and es­ti­mated the data in 2013–2014 ac­cord­ing to the same sta­tis­ti­cal cal­iber based on the Ur­ban and Ru­ral House­hold Sur­vey data. There­fore, the sta­tis­ti­cal cal­iber of the data main­tains con­sis­tency be­fore and after 2012 in this paper.

(2) The ed­u­ca­tion in­put (edu) at dif­fer­ent ed­u­ca­tional stages is the main ex­plana­tory vari­able in this paper. A large num­ber of stud­ies have shown that ed­u­ca­tional in­equal­ity due to the dif­fer­ences in ed­u­ca­tion in­put is the most im­por­tant fac­tor re­sult­ing in the widen­ing ur­ban-ru­ral in­come gap (Bai, 2004; Terry et al. 2007; Chen et al. 2010). Ed­u­ca­tion in­put at dif­fer­ent ed­u­ca­tional stages has a dif­fer­ing ef­fect on the nar­row­ing of the ur­ban-ru­ral in­come gap (Becker and Tomes, 1979; Yang et al., 2015). There­fore, this paper tests the ef­fect of ed­u­ca­tion in­put at dif­fer­ent ed­u­ca­tional stages, such as pri­mary ed­u­ca­tion, ju­nior high school ed­u­ca­tion, se­nior high school ed­u­ca­tion, higher ed­u­ca­tion, and com­pul­sory ed­u­ca­tion. Data are de­rived from the China Ed­u­ca­tional Fi­nances Sta­tis­ti­cal Year­book. The com­pul­sory ed­u­ca­tion in­put is cal­cu­lated based on the ed­u­ca­tion in­put at the pri­mary and ju­nior high school lev­els with the weight on the num­ber of stu­dents. The av­er­age ed­u­ca­tion in­put is cal­cu­lated based on the ed­u­ca­tion in­put and the num­ber of stu­dents at each ed­u­ca­tional stage.

(3) La­bor mi­gra­tion (LTR). Ac­cord­ing to Liu (2015), the es­ti­mated for­mula for the la­bor mi­gra­tion rate is (the num­ber of ru­ral em­ploy­ees – the num­ber of ru­ral pri­mary in­dus­try em­ploy­ees) / the num­ber of ru­ral em­ploy­ees. The data are de­rived from the China Agri­cul­ture Com­pi­la­tion of Sta­tis­tics 1949– 2008 and the China Ru­ral Sta­tis­ti­cal Year­book, and the Sta­tis­ti­cal Year­book of 31 prov­inces.

(4) De­gree of open­ness (FDI). The Stolper-Sa­muel­son The­o­rem holds that a coun­try’s goods with com­par­a­tive ad­van­tages will be­come more ex­pen­sive after open­ing to the out­side world, which re­sults in a higher re­turn of the fac­tor in­ten­sively used on this kind of goods. Dur­ing the eco­nomic take-off pe­riod, most de­vel­op­ing coun­tries adapt an ex­port-ori­ented strat­egy and de­velop rapidly by im­prov­ing the la­bor-in­ten­sive in­dus­tries which in­crease the la­bor de­mand and in­come level of low-skilled work­ers (Han et al., 2012; Han et al. 2015). This al­lo­ca­tion of fac­tor and in­come is con­ducive to nar­row­ing the ur­ban-ru­ral in­come gap. In this paper, the de­gree of open­ness is mea­sured by the ra­tio of for­eign di­rect in­vest­ment to gross do­mes­tic prod­uct (GDP). Data are de­rived from the CEIC data­base.

(5) Fi­nan­cial ef­fi­ciency (Fi­nance). With the im­prove­ment of fi­nan­cial ef­fi­ciency, ru­ral res­i­dents will en­joy bet­ter fi­nan­cial ser­vices. They have more op­por­tu­ni­ties to earn cap­i­tal re­turns, which will help to nar­row the ur­ban-ru­ral in­come gap (Clarke and Zou, 2006; Beck et al., 2010). The fi­nan­cial ef­fi­ciency is mea­sured by the ra­tio of loans to de­posits in fi­nan­cial in­sti­tu­tions. Data are de­rived from the “China Com­pi­la­tion of Sta­tis­tics 1949–2008” and the Sta­tis­ti­cal Year­book of 31 prov­inces.

(6) The pro­por­tion of state-owned en­ter­prises (SOE). In a per­fect mar­ket mech­a­nism, the al­lo­ca­tion and wage lev­els of the la­bor are de­ter­mined by the mar­ket. At the present stage, China’s econ­omy shows a mixed sys­tem with var­i­ous types of own­er­ship, and the al­lo­ca­tion and wage lev­els of the la­bor are dif­fer­ent be­tween the state sec­tor and the non-state sec­tor (Dé­murger et al. 2006; Deng and Ye, 2012). It is dif­fi­cult for the la­bor in the non-state sec­tor to en­ter the state sec­tor. There­fore, the state sec­tor has mo­nop­o­lis­tic ad­van­tage and pays work­ers higher wages and so­cial se­cu­rity than that the non-state sec­tor. Thus, a big­ger pro­por­tion of state-owned en­ter­prises may re­sult in higher so­cial in­equal­ity and a greater in­come gap. In this paper, the pro­por­tion of state-owned en­ter­prises is mea­sured by the pro­duc­tion ra­tio of state-owned in­dus­trial en­ter­prises to the above-scale in­dus­trial en­ter­prises.

(7) Gross do­mes­tic pro­duc­tion per capita (gdp). Si­mon Kuznets (1955) holds that there is an in­verted U-shaped re­la­tion­ship be­tween eco­nomic de­vel­op­ment and in­come in­equal­ity. In the early stages of eco­nomic de­vel­op­ment, the non-agri­cul­tural sec­tor with higher in­come in­equal­ity de­vel­ops rapidly, and the in­come in­equal­ity in the whole so­ci­ety tends to de­te­ri­o­rate. When the eco­nomic de­vel­op­ment achieves a high level, the pro­por­tion of the non-agri­cul­tural sec­tor in­creases sig­nif­i­cantly and the in­come gap de­creases be­tween sec­tors. In ad­di­tion to the ef­fect of in­come re­dis­tri­bu­tion poli­cies, there will be more equal in­come dis­tri­bu­tion. This paper chooses per capita in­come to mea­sure the eco­nomic de­vel­op­ment level.

3.2 Space Econo­met­rics Model

This paper uses the Spa­tial Au­tore­gres­sive Model (SAR) and the Spa­tial Er­ror Model (SEM) to test the ef­fect on the ur­ban-ru­ral in­come gape of ed­u­ca­tion at dif­fer­ent ed­u­ca­tional stages. On one hand, the ur­ban-ru­ral in­come gap of a prov­ince de­pends on that of other prov­inces. The spa­tial au­tore­gres­sive model takes into ac­count the en­doge­nous in­ter­ac­tion be­tween the ex­plana­tory vari­ables. On the other hand, the spa­tial er­ror model con­sid­ers the in­ter­ac­tion of spa­tial er­ror items. The miss­ing ex­plana­tory vari­ables that af­fect the ur­ban-ru­ral in­come gap are spa­tially re­lated in the spa­tial er­ror model. When these two kinds of spa­tial panel mod­els are used at the same time, they can over­come the in­flu­ence of dif­fer­ent po­ten­tial spa­tial cor­re­la­tion fac­tors, and the re­gres­sion re­sults will be rel­a­tively ro­bust.

The func­tion of the spa­tial au­tore­gres­sive model takes the form Al­though the model con­trols a se­ries of vari­ables that af­fect the ur­ban-ru­ral in­come gap, if some im­por­tant vari­ables may be omit­ted, the re­gres­sion re­sults of the model are im­pacted. Fur­ther­more, the ur­ban-ru­ral in­come gap may af­fect the ed­u­ca­tion in­put in turn, and cre­ate en­do­gene­ity prob­lems. There­fore, this paper in­tro­duces the first or­der lagged term of the ur­ban- ru­ral in­come gap as an

ex­plana­tory vari­able. It can not only mea­sure the po­ten­tial fac­tors such as his­tor­i­cal back­ground, hu­man en­vi­ron­ment, cus­toms, and po­lit­i­cal in­flu­ences, but also solve the en­do­gene­ity prob­lem as an in­stru­men­tal vari­able to a cer­tain ex­tent.

The func­tion of the spa­tial er­ror term model takes the form

where, is the spa­tial au­to­cor­re­la­tion co­ef­fi­cient of the er­ror term, and it mea­sures the ef­fect of sto­chas­tic dis­tur­bance of other prov­inces’ ur­ban-ru­ral in­come gaps. The other vari­ables are de­fined as the spa­tial au­to­cor­re­la­tion model.

Be­cause both the spa­tial au­to­cor­re­la­tion model and the spa­tial er­ror model use the en­tire spa­tial sys­tem to cal­cu­late the spa­tial cor­re­la­tion, the re­gres­sion model can­not rule out the ex­is­tence of en­do­gene­ity prob­lems. If the or­di­nary least squares method is used, the pa­ram­e­ters of the spa­tial er­ror model are not valid, and the pa­ram­e­ters of the spa­tial au­to­cor­re­la­tion model are nei­ther un­bi­ased nor con­sis­tent (Lesage, Pace, 2010). There­fore, this paper uses the Max­i­mum Like­li­hood (ML) model

pro­posed by El­horst (2003) to es­ti­mate the pa­ram­e­ters. Been­stock and Felsen­stein (2007) ar­gue that if the spa­tial sam­ple of the re­gres­sion model is a ran­domly se­lected sub­sam­ple of a pop­u­la­tion, it is ap­pro­pri­ate to choose a ran­dom ef­fect panel model. If the re­gres­sion sam­ple is ex­actly the pop­u­la­tion, it is ap­pro­pri­ate to choose a fixed ef­fect panel model be­cause each space unit can­not be ran­domly sam­pled. The sam­ple in this paper is 31 fixed-space units di­vided by the pro­vin­cial level of main­land China, so the fixed ef­fect model is cho­sen.

3.3 Em­pir­i­cal Anal­y­sis

Ta­bles 1 and ta­ble 2 re­port the re­sults of the SAR model and SEM model. At the same time, the model re­sults in­clud­ing the in­ter­ac­tion be­tween ed­u­ca­tion in­put and la­bor mi­gra­tion are given. The in­ter­ac­tion tests whether there is in­ter­ac­tion be­tween them, that is to say, can in­creas­ing ed­u­ca­tion in­put pro­mote la­bor mi­gra­tion and can la­bor mi­gra­tion strengthen the ef­fect of ed­u­ca­tion in­put on the ur­ban­rural in­come gap.

In general, the SAR model and the SEM model are passed through the tra­di­tional La­grange Mul­ti­plier Test and Ro­bust La­grange Mul­ti­plier Test. From re­sults of R2 and Log-L, all mod­els have sig­nif­i­cant good­ness of fit, which means the spa­tial dy­namic panel data frame­work con­structed in this paper de­scribe the mi­cro­scopic mech­a­nism of ed­u­ca­tion in­put on the ur­ban-ru­ral in­come gap very well.

The spa­tial cor­re­la­tion co­ef­fi­cients of the model are greater than zero and pass the 1% sig­nif­i­cance test, in­di­cat­ing that there is a sig­nif­i­cant pos­i­tive spa­tial cor­re­la­tion re­la­tion be­tween the ur­ban-ru­ral in­come gap of dif­fer­ent prov­inces, and the es­ti­mated re­sults of the pa­ram­e­ters of each ex­plana­tory vari­able are sig­nif­i­cant.

First, ed­u­ca­tion in­put at dif­fer­ent ed­u­ca­tional stages and the ur­ban- ru­ral in­come gap show a sig­nif­i­cant nega­tive cor­re­la­tion re­la­tion­ship, in­di­cat­ing that ed­u­ca­tion in­put at each ed­u­ca­tional stage can nar­row the ur­ban-ru­ral in­come gap. Se­cond, the co­ef­fi­cient of the spa­tial dy­namic panel model with in­ter­ac­tive vari­ables is sig­nif­i­cantly pos­i­tive and more than that with­out in­ter­ac­tive vari­ables. There­fore, ed­u­ca­tion in­put at dif­fer­ent ed­u­ca­tional stages can not only im­prove la­bor pro­duc­tiv­ity and di­rectly nar­row the ur­ban-ru­ral in­come gap, but also in­di­rectly nar­row the ur­ban-ru­ral in­come gap through la­bor trans­fer, which is con­sis­tent with the pre­vi­ous the­o­ret­i­cal anal­y­sis. Third, ed­u­ca­tion in­put at the com­pul­sory ed­u­ca­tional stage, es­pe­cially at ju­nior mid­dle ed­u­ca­tional stage, has a more sig­nif­i­cant ef­fect on nar­row­ing the ur­ban-ru­ral in­come gap than that at other ed­u­ca­tional stages. There­fore, it is more ef­fec­tive to nar­row the ur­ban-ru­ral in­come gap by in­creas­ing ed­u­ca­tion in­put at the com­pul­sory ed­u­ca­tional stage, es­pe­cially at the ju­nior mid­dle ed­u­ca­tional stages. Al­though the to­tal amount of ed­u­ca­tion in­put at the com­pul­sory ed­u­ca­tional stage has in­creased sig­nif­i­cantly in China, there are still large gaps be­tween re­gions, es­pe­cially at the ju­nior mid­dle ed­u­ca­tional stage. Fig­ure 2 in­di­cates that the

pub­lic bud­get ed­u­ca­tion ex­penses per student in the eastern de­vel­oped prov­inces are five times that of the cen­tral and western prov­inces. In sum­mary, in the fu­ture, the al­lo­ca­tion of China’s ed­u­ca­tion funds should be tilted to com­pul­sory ed­u­ca­tion, es­pe­cially to ju­nior high ed­u­ca­tion.

The im­pact of each con­trol vari­able on the ur­ban-ru­ral in­come gap is also con­sis­tent with the the­o­ret­i­cal anal­y­sis in this paper. The co­ef­fi­cient of the open­ness is nega­tive, in­di­cat­ing that open­ing to the out­side world can nar­row the ur­ban-ru­ral in­come gap, con­sis­tent with the Stolpa-Sa­muel­son the­o­rem. As a de­vel­op­ing coun­try, China should be ac­tively par­tic­i­pat­ing in in­ter­na­tional trade and in­tro­duc­ing for­eign cap­i­tal to en­cour­age in­dus­tries with com­par­a­tive ad­van­tages to de­velop rapidly, which will in­crease la­bor de­mand and in­come lev­els of low-skilled work­ers and nar­row the in­come gap. The co­ef­fi­cient of fi­nan­cial ef­fi­ciency is sig­nif­i­cantly less than zero, in­di­cat­ing that fi­nan­cial ef­fi­ciency is neg­a­tively cor­re­lated with the ur­ban-ru­ral in­come gap. De­vel­op­ing an in­clu­sive fi­nan­cial sys­tem and im­prov­ing fi­nan­cial ef­fi­ciency can ef­fec­tively nar­row the ur­ban-ru­ral in­come gap in China. The im­pact of the pro­por­tion of state-owned en­ter­prises is sig­nif­i­cantly pos­i­tive, in­di­cat­ing that a big­ger pro­por­tion of state-owned en­ter­prises re­sults in a more se­ri­ous di­vi­sion of la­bor al­lo­ca­tion which will lead to a greater in­come gap be­tween the non-state sec­tor and the state sec­tor; The es­ti­mated co­ef­fi­cient of per capita GDP is also sig­nif­i­cantly greater than zero, in­di­cat­ing that China’s eco­nomic de­vel­op­ment is still in the ris­ing stage of Kuznets’ in­verted “U-shaped” curve. The co­ef­fi­cient of the lagged vari­able of ur­ban-ru­ral in­come gap also passes the sig­nif­i­cance test, in­di­cat­ing that the ur­ban-ru­ral in­come gap is im­pacted by the his­tor­i­cal back­ground, hu­man en­vi­ron­ment, po­lit­i­cal sys­tem, and other po­ten­tial fac­tors.

Ta­ble 3 re­ports the re­gres­sion re­sults of ed­u­ca­tion in­put on the per capita dis­pos­able in­come growth rate of ur­ban house­holds and the per capita net in­come growth rate of ru­ral house­holds. The spa­tial au­tore­gres­sive model and the spa­tial er­ror model both il­lus­trate the same re­sult. First, the re­gres­sion co­ef­fi­cient of ed­u­ca­tion in­put is sig­nif­i­cantly greater than zero, in­di­cat­ing that in­creas­ing ed­u­ca­tion in­put can im­prove la­bor pro­duc­tiv­ity and in­crease the in­come lev­els of ru­ral res­i­dents and ur­ban res­i­dents. Se­cond, the ef­fect of ed­u­ca­tion in­put in­creas­ing la­bor in­come is more sig­nif­i­cant for ru­ral res­i­dents than ur­ban res­i­dents. There­fore, ed­u­ca­tion in­put can nar­row the ur­ban-ru­ral in­come gap.

In sum­mary, in­creas­ing ed­u­ca­tion in­put can not only in­crease the in­comes of ru­ral res­i­dents but also in­crease the in­comes of ur­ban res­i­dents, which re­flect the Pareto im­prove­ment. Fur­ther­more, ed­u­ca­tion in­put at each ed­u­ca­tional stage can not only im­prove la­bor pro­duc­tiv­ity and po­ten­tial eco­nomic growth, but also pro­mote la­bor trans­fer and nar­row the ur­ban-ru­ral in­come gap, which is con­duc­tive to achiev­ing eco­nomic and so­cial in­clu­sive growth goals.

3.4 Ro­bust­ness Test

This paper uses dif­fer­ent sam­ple in­ter­vals to test the ro­bust­ness of re­gres­sion re­sults of the dy­namic spa­tial panel model. Con­sid­er­ing that the re­form of state-owned en­ter­prises may im­pact the in­come dis­tri­bu­tion, I used the sub­sam­ple in­ter­val 1998–2014 to test the ro­bust­ness. The re­sults show that co­ef­fi­cients of ed­u­ca­tional in­put and most con­trol vari­ables are sig­nif­i­cant, and there is no ob­vi­ous change in the val­ues and sign of co­ef­fi­cient. There is still a strong nega­tive cor­re­la­tion re­la­tion­ship be­tween ed­u­ca­tion in­put and the ur­ban-ru­ral in­come gap, that is, ed­u­ca­tion in­put can nar­row the ur­ban­rural in­come gap. The con­clu­sion is ro­bust.

4. Con­clu­sion

From the per­spec­tive of la­bor mi­gra­tion, this paper con­structs a general equi­lib­rium model to the ur­ban-ru­ral dual eco­nomic struc­ture, stud­ies the mi­cro­scopic mech­a­nism of ed­u­ca­tion in­put on nar­row­ing the ur­ban-ru­ral in­come gap, and tests the hy­poth­e­sis of the re­la­tion­ship be­tween ed­u­ca­tion in­put, la­bor

mi­gra­tion, and the ur­ban-ru­ral in­come gap based on the dy­namic spa­tial panel model.

The general equi­lib­rium the­ory shows that ed­u­ca­tion in­put al­le­vi­ates the ur­ban-ru­ral in­come gap both in di­rect and in­di­rect ways. On one hand, in­creas­ing ed­u­ca­tion in­put can im­prove the ac­cu­mu­la­tion of hu­man cap­i­tal and the mar­ginal pro­duc­tion of the la­bor, and di­rectly in­crease the la­bor in­come in both the non-agri­cul­tural and agri­cul­tural sec­tors. How­ever, the ef­fect of ed­u­ca­tion in­put pro­mot­ing la­bor pro­duc­tiv­ity is more sig­nif­i­cant than that in the non-agri­cul­tural sec­tor. Thus, in­creas­ing ed­u­ca­tion in­put can nar­row the ur­ban-ru­ral in­come gap. On the other hand, ed­u­ca­tion in­put im­proves the ac­cu­mu­la­tion of hu­man cap­i­tal and the la­bor pro­duc­tiv­ity of low-skilled work­ers, which pro­vides more job op­por­tu­ni­ties in the non- agri­cul­tural sec­tor and re­duces the cost of la­bor mi­gra­tion. The im­prove­ment of la­bor al­lo­ca­tion ef­fi­ciency is con­duc­tive to nar­row­ing the ur­ban-ru­ral in­come gap.

The em­pir­i­cal re­sults are con­sis­tent with the­o­ret­i­cal anal­y­sis. First, in creas­ing ed­u­ca­tion in­put at each ed­u­ca­tional stage can ef­fec­tively nar­row the ur­ban-ru­ral in­come gap in China, and the ef­fect of ed­u­ca­tion in­put in­creas­ing la­bor in­come is more sig­nif­i­cant for ru­ral res­i­dents than ur­ban res­i­dents, which is a Pareto im­prove­ment. Ed­u­ca­tion in­put at the com­pul­sory ed­u­ca­tional stage, es­pe­cially at the ju­nior mid­dle ed­u­ca­tional stage, has a more sig­nif­i­cant ef­fect on nar­row­ing the ur­ban-ru­ral in­come gap than that at the other ed­u­ca­tional stages. Al­though the to­tal amount of ed­u­ca­tion in­put at the com­pul­sory ed­u­ca­tional stage has in­creased sig­nif­i­cantly in China, there are still large gaps be­tween re­gions, es­pe­cially at the ju­nior mid­dle ed­u­ca­tional stage. Dur­ing the “Twelfth FiveYear Plan” pe­riod, at the ju­nior mid­dle ed­u­ca­tional stage, the pub­lic bud­get ed­u­ca­tion ex­penses per student in the eastern de­vel­oped prov­inces were five times that of the cen­tral and western prov­inces. With the ad­just­ment for the macroe­co­nomic slow­down, and the de­creas­ing growth rate of fis­cal rev­enue, fur­ther op­tional al­lo­ca­tion of ed­u­ca­tion funds at dif­fer­ent ed­u­ca­tional stages is an im­por­tant guar­an­tee for achiev­ing the strate­gic ob­jec­tives of “Ed­u­ca­tion Plan” and build­ing a com­pre­hen­sive well-off so­ci­ety in 2020. There­fore, China should fur­ther in­crease ed­u­ca­tion in­put, and the al­lo­ca­tion of China’s ed­u­ca­tion funds should be tilted to com­pul­sory ed­u­ca­tion, es­pe­cially to ju­nior high ed­u­ca­tion.

Se­cond, with the grad­ual de­creas­ing of the “de­mo­graphic div­i­dend,” the cost of China’s fac­tors is ris­ing, and the in­dus­trial struc­ture trans­for­ma­tion from a man­u­fac­tur­ing to a ser­vice in­dus­try has re­sulted in the la­bor sur­plus of low- skilled work­ers in the tra­di­tional agri­cul­tural sec­tor and a rel­a­tive la­bor short­age of skilled work­ers in the mod­ern non- agri­cul­tural sec­tor ( Chao and Shen, 2014). The lag­ging em­ploy­ment struc­ture trans­for­ma­tion has caused great de­vi­a­tion be­tween in­dus­trial struc­ture and em­ploy­ment struc­ture. In­creas­ing ed­u­ca­tion in­put can pro­mote la­bor pro­duc­tiv­ity and in­come lev­els of low- skilled work­ers in the tra­di­tional sec­tor and raise the pro­por­tion of the skilled work­ers in the la­bor mar­ket. The “mi­gra­tion div­i­dend” brought by ed­u­ca­tional in­put both can nar­row the ur­ban-ru­ral in­come gap and al­le­vi­ate the con­tra­dic­tion of the lag­ging em­ploy­ment struc­ture trans­for­ma­tion.

Third, with ad­just­ment for the slow­down of China’s macro-econ­omy, rais­ing fis­cal ex­pen­di­tures on ed­u­ca­tion can not only smooth eco­nomic growth by com­pen­sat­ing for the de­crease of pri­vate in­vest­ment in the short term, but also im­prove po­ten­tial eco­nomic growth by pro­mot­ing hu­man cap­i­tal ac­cu­mu­la­tion in the long run. In other words, in­creas­ing ed­u­ca­tion in­put can pro­mote la­bor mi­gra­tion and nar­row the ur­ban-ru­ral in­come gap, which is an im­por­tant means to achiev­ing China’s in­clu­sive growth tar­gets.

Fig­ure 1: The Di­rect and In­di­rect Ef­fects of Ed­u­ca­tion In­put Nar­row­ing Ur­ban-ru­ral In­come Gap

De­scrip­tion: The Mo­ran I is the re­sult of the spa­tial cor­re­la­tion test. R2 and Log-L il­lus­trate the good­ness of the model. λ is the Spa­tial Cor­re­la­tion Co­ef­fi­cient for Spa­tial Au­tore­gres­sive Model. ***, **, * were sig­nif­i­cant at 1%, 5%, and 10% sig­nif­i­cance lev­els, re­spec­tively. The value in brack­ets are t val­ues of the es­ti­mated pa­ram­e­ters.

Source: China Ed­u­ca­tional Fi­nances Sta­tis­ti­cal Year­book

Fig­ure 2: Per Capita Ed­u­ca­tion Ex­pen­di­ture for Ju­nior Mid­dle Ed­u­ca­tion in 2015 (yuan)

Newspapers in English

Newspapers from China

© PressReader. All rights reserved.