You are on page 1of 12

Search in the Labor Market and the Real Business Cycle Monika Merz by Max Glischinski March 14,

2011

Max Glischinski

14.03.2011

Introduction

Neoclassical real business cycle model fails to capture many stylized labor market facts frictionless market no persistency in output wages equaling the marginal product of labour correlation of employment and productivity Solution: Model with frictions and transacting costs (search-model) quantitative implications in a general equilibrium setting

Max Glischinski

14.03.2011

Related Literature

Kydland and Prescott (1982), and Long and Plosser (1983) Diamond (1982), Mortensen (1982), and Pissarides (1988)

Max Glischinski

14.03.2011

Physical Environment
Continuum of homogenous workers (more precise households) assumed large - perfectly insure themselfes Continuum of homogenous rms Perfect goods and capital markets Labour market inherits search frictions driven by market-tightness: matches evolve accroding to matching function: M rms incurring vacancy posting cost: a workers search with certain search intensity: c (S ) exogenous seperation at rate:

Max Glischinski

14.03.2011

Planners Problem
Social planner faces objective function

E0 t U (Ct , Nt ) = U (Ct ) G (Nt ) , 0 < < 1 t =0 with U (Ct ) = log (Ct ) , G (Nt ) = N 11/ / (1 1/)

and production technology Ct + It + e t c (St ) (1 Nt ) + e t aVt e (1 )(t +zt ) Kt Nt1 with c (St ) = c0 St , c0 > 0, > 1, St 0, , > 0 subject to laws of motion Kt +1 = (1 ) Kt + It Nt +1 = (1 ) Nt + Mt 0 < , < 1 with Mt = Vt1 [St (1 Nt )] thus endogenous probabilities for job-nding and hiring: pt = Mt /St (1 Nt ) = St1 t 1 qt = Mt /Vt = St t1 where = Vt / (1 Nt )
Max Glischinski 14.03.2011

Planners Problem
Social planner chooses{Ct , Kt +1 , Nt +1 , St , Vt }=0 to maximize the t objective function

zt , Kt , Nt

I {t , Vt , St }

max

U Ct , Nt

+E

W |W

subject to
Ct = t +1 = K Nt +1 = zt +1 = I Yt t c (St ) (1 Nt ) aVt I (1 ) Kt + t (1 ) Nt + Mt zt + t +1

Max Glischinski

14.03.2011

Planners Problem
Euler Equations

Uct =
aUct MVt =

E E

UCt+1 fKt+1 + (1 ) | zt , Kt , Nt

UCt+1 fNt+1 + c (St +1 ) GNt+1 aUCt +1 + MV (1 ) + MNt 1 | zt , Kt , Nt t +1 UCt+1 fNt+1 + c (St +1 ) GNt+1 UCt +1 cSt +1 (1Nt +1 ) (1 ) + MNt 1 | zt , Kt , Nt + MSt +1

Uct cSt (1Nt ) = MSt

Max Glischinski

14.03.2011

Model Calibration
Matched parameters:
Output elasticity wrt K Discount rate Depreciation rate Convexity of

c (St )

Negative of Frisch elasticity Elasticity of job matches wrt S Seperation rate Growth rate Advertising cost Search cost level Autocorr shock St.dev shock

a c0

0.360 0.990 0.022 1.000 -1.250 0.400 0.070 0.004 0.050 0.005 0.950 0.007

Calibrational exercise: Comparing articial (100 times model simulation) and US time series (59:1-88:2, CITIBASE) outcome
Max Glischinski 14.03.2011

Results

Max Glischinski

14.03.2011

Results

Max Glischinski

14.03.2011

Results

Max Glischinski

14.03.2011

Conclusion

Incorporating labour market frictions into the neoclassical growth model inherits numerous improvements: productivity more volatile than wages productivity leads employment over the cycle countercyclical behaviour of labour share unemployment and vacancies highly volatile history dependence in wages (persistence ion un/employment and aggregate output)

Max Glischinski

14.03.2011