Structural Change and Aggregate Fluctuations in China

Transcription

1 Srucural Change and Aggregae Flucuaions in China Wen Yao Tsinghua Universiy Xiaodong Zhu Universiy of Torono and SAIF PBOC-SAIF Conference on Macroeconomic Analysis and Predicions December 5, / 21

2 Aggregae employmen flucuaions in China Volailiy of employmen is low Employmen and oupu are barely correlaed. China US σ(l)/σ(y) ρ(l, y) Cyclical componens of oupu and employmen l y / 21

3 Employmen puzzle The low employmen-oupu correlaion in China is conradicory o sandard business cycle models Mos sudies of he Chinese business cycle simply ignore his fac 3 / 21

4 Poenial explanaions Wha explains he differences beween he US and China s labor marke? Daa issue? Discree jump in he official series in 1990 Revised daa by Brand and Zhu (2010) shows he same paern Insiuional consrain on employmen? China s labour marke is quie flexible, a leas for he non-sae secor This paper: Difference in economic srucure! Agriculure s share of employmen is much larger in China 4 / 21

5 Wha do we do? We documen new facs a secor level We argue difference in economic srucure is he main reason for he observed differences in employmen flucuaions beween China and he US We show ha a sandard wo-secor business cycle model wih srucural change can accoun for labor marke dynamics in boh economies 5 / 21

6 New fac: labor flucuaions a secor level Non-agriculure: China US σ(l n )/σ(y n ) ρ(l n, y n ) Agriculure: China US σ(l a )/σ(y a ) ρ(l a, y a ) Similar behavior a secor level 6 / 21

7 New fac: labor reallocaion across secors Table: ρ(l a, l na ) ρ(l a, z a ) ρ(l a, z na ) ρ(l na, z a ) ρ(l na, z na ) China US Similar behavior in labor reallocaion across secors 7 / 21

8 New facs and implicaions Employmen flucuaions in China and he US are similar a secor level Labor reallocaion across secors is also similar for he wo economies Differences in aggregae employmen flucuaions mus be a resul of differen economic srucure Agriculural secor is much larger in China han in he US The relaive size of he agriculural secor in China has been shrinking over ime Any explanaion also has o be consisen wih he srucural change in China 8 / 21

9 Ouline Basic model srucure Two-secor (agriculure and non-agriculure) growh model Business cycles are driven by produciviy shocks Non-homoheic preference Calibraion Use he parameers o mach he long-run rend in aggregae employmen and srucural change Use he calibraed parameers o simulae he shor-run flucuaions Presence of agriculure secor and he non-homoheic preference help he model o explain he employmen behavior in he daa. 9 / 21

10 Lieraure Srucural change: Kongsamu, Rebelo and Xie (2001) Ngai and Pissarides (2007) Herrendorf, Rogerson and Valeninyi (2013) Comin, Lashkari and Mesieri (2015) Agriculure and business cycles: Da-Rocha and Resuccia (2006) Soresleern and Zhao (2016) 10 / 21

11 The modei Planner chooses c a, c na, L a, L na o solve [ max {N C B ]} 1 + σ (L a + L na ) 1+σ s.. φ 1 ɛ a C µa 1 ɛ c ɛ 1 ɛ a + φ 1 ɛ na C µna 1 ɛ c ɛ 1 ɛ na = 1 N c a = A a N L a N c na = A na N L na N : oal populaion L i =: employmen o populaion raio in each secor ɛ : elasiciy of subsiuion µ i : income elasiciy parameer for good i (CES if µ i = 2 ε) 11 / 21

12 More on preferences Non-homoheic CES preference (Comin e. al. 2015): φ 1 ɛ a C µa 1 ɛ c ɛ 1 ɛ a + φ 1 ɛ na C µna 1 ɛ c ɛ 1 ɛ na = 1 Two properies: The elasiciy of he relaive demand for wo differen goods wih respec o aggregae consumpion is consan log (c na /c a ) log C = µ na µ a The elasiciy of subsiuion beween wo differen goods is consan log(c na /c a ) log(p na /p a ) = ɛ 12 / 21

13 Solving he model Secor employmen: L a = ( ϕ a A ε 1 a C µa 1 ϕ a A ε 1 a C µa 1 + ϕ na Ana ε 1 C µna 1 ) ε, (1) ε 1 L na = ( ϕ a A ε 1 a ϕ na Ana ε 1 C µna 1 C µa 1 Equilibrium level of consumpion: ( (1 ε) ϕ aa ε 1 a C µa 1 ( B (µ a 1)ϕ aa ε 1 a C µa 2 + ϕ naa ε 1 na + (µ na 1)ϕ naa ε 1 na + ϕ na Ana ε 1 C µna 1 C µna 1 ) ε 1 ε 1 σ C µna 2 ) ( = ϕ aa ε 1 a ) ε, (2) ε 1 C µa 1 + ϕ naa ε 1 na C µna 1 ) 1 1 ε. (3) 13 / 21

14 Calibraing preference parameers Because of non-homoheiciy, rend and cycle canno be easily separaed Le x be he rend componen of a variable x. We use he HP filer o decompose produciviies in he wo secors ino rend and cyclical componen: A a = A a e za, A n = A n e zn We choose he values of a, ε, µ a and µ na so ha when z a = z n = 0, he model implied rends of he secor employmen shares mach he rends in he daa We choose he values of B o mach he rend of aggregae employmen 14 / 21

15 oal employmen daa model Calibraion resuls Parameer Benchmark a ɛ µ a µ na σ ag employmen daa model non ag employmen daa model / 21

16 Cyclical componens of employmen in he wo secors ag employmen daa model non ag employmen daa model / 21

17 Cyclical componen of aggregae employmen and oupu 0.1 daa oal employmen y 0.1 model oal employmen y / 21

18 Simulaion Resuls for China σ(l) σ(l a) σ(l na) σ(l) σ(y) σ(y) ρ(l, y) ρ(l, zl) China Daa Benchmark σ(l a) σ(y a) σ(l na) σ(y na) ρ(l a, y a) ρ(l na, y na) ρ(l a, l na) China Daa Benchmark ρ(l a, zl a) ρ(l a, zl na) ρ(l na, zl a) ρ(l na, zl na) China Daa Benchmark / 21

19 Simulaion Resuls for he US We use he same parameer values when simulaing he US economy ag employmen daa model non ag employmen daa model / 21

20 Simulaion Resuls for US σ(l) σ(l a) σ(l na) σ(l) σ(y) σ(y) ρ(l, y) ρ(l, zl) China Daa Benchmark US Daa US Model σ(l a) σ(y a) σ(l na) σ(y na) ρ(l a, y a) ρ(l na, y na) ρ(l a, l na) China Daa Benchmark US Daa US Model ρ(l a, zl a) ρ(l a, zl na) ρ(l na, zl a) ρ(l na, zl na) China Daa Benchmark US Daa US Model / 21

21 Conclusion The cyclical behavior of aggregae employmen differs significanly beween China and he US This sharp difference a he aggregae level conceals similar behavior of cyclical employmen a secor level We argue ha he main difference beween China and he US is he size of he agriculure secor We show ha a wo-secor growh model wih echnology shocks can accoun for he srucural change and employmen flucuaions in boh China and US Our resuls highligh he imporance of srucural change for undersanding business cycle in China 21 / 21