Macroeconomic policy and employment flows

Macroeconomic policy and employment flowsLabour market adjustment mechanisms in macroeconomic models

Uma Amara Rani Matthieu Charpe Ekkehard ErnstInternational Institute for Labour Studies, ILO

Project LINK meeting, New YorkOct 22nd, 2010

Outline

1 Objectives and main message

2 Background considerations

3 A labour market flows macro model

4 Model estimation

5 Conclusion

Amara, Charpe, Ernst (ILO) Macro policy and employment flows Oct 2010 2 / 22

Objectives and main message Objectives

Objectives

Get a micro picture of labour dynamicsI Look at different margins of labour market flows

I Job creation vs. job destruction

I Assess their dynamics over the cycle

Overcoming Okun’s fatalityI Okun’s law is time-varying and asymetric

I How to promote job-rich growth?

I How does weak employment growth affect therecovery?


Objectives and main message Bottom line

Main message

Important insights from analysing labour flow dynamicsI Stark differences in labour flow dynamics over the

business cycleI Labour flow dynamics adds to the macro-economic

feedback loopsI Large (short-run) employment multipliers (> 1), lower

in the long-run (due to automatic stabilization)I Scenario analysis suggests huge fiscal benefits to be

had from additional stimulus


Background considerations Traditional macro models

Traditional macro models

Okun’s law as a corner stoneI Pre-determined employment elasticities of growth:

I Symmetric relation between output and employment growthI Trickle-down effects of stimulus

I Trickle down effects of stimulus:I First investment, than jobs (IMF: resilient recovery)I Improved demand side will eventually lead to more (and

better?) jobs

I But:I Strong evidence for time-varying Okun’s law before the crisisI Crisis makes elasticity approach obselete


Background considerations Labour flow macro models

Building labour flows into macro models

Labour flow macro modelsI Since the mid-1990s, labour flows introduced in DSGE models

I Merz, 1995; Andolfatto, 1996; Trigari, 2003; Walsh, 2005

I Increasingly becoming the industry standardI Most New Keynesian models now contain labour flowsI Recent fiscal policy models exclusively rely on them

But:I Problems with the business cycle behaviour of flows

I Volatility not high enough to match data (Shimer, 2005)I Unrealistic wage behaviour

I Estimation of the model difficultI Internationally comparable data on labour flows hardly availableI Strong assumptions wrt functional form and parameter distribution


A labour market flows macro model Labour market flows and matching

Stylised facts on labour flows

I Labour market flows present stark differences acrosscountries over the business cycle

−6

−5

−4

−3

−6

−5

−4

−3

−4

−2

02

−4

−2

02

1970 1980 1990 2000 2010 1970 1980 1990 2000 2010

Canada Germany

Japan USA

Unemployment outflows Unemployment inflows

Inflo

w p

roba

bilit

y

Out

flow

pro

babi

lity

Note: Flow probabilities have been logit−transformed


A labour market flows macro model Determinants of labour flows

Labour flow accounting

Decomposing unemployment dynamics

4Ut =4Lt −4ETt = INt −OUTt

Decomposing employment creation

4ETt = JobCreationt −JobDestructiont

⇓

Link unemployment in- and outflows to its determinantsI Job creationI Job destructionI Labour force growth


A labour market flows macro model Determinants of labour flows

What drives labour market flows?

Job creationI Investment growth, private consumption, external demandI Past employment rate, wagesI User cost of capital, Real share prices

Job destructionI Wages, real interest rates, tax wedge, external demandI Schumpeter effect: TFP, Import growth

Labour force dynamicsI Mainly driven by its own history plusI Tax incentivesI Discouraged worker effects


A labour market flows macro model A macro framework

A macro framework I

Financial acceleratorI Interest rates affect net present value of

vacanciesI Real share prices improve investment dynamics

Feedback loopsI In- and outflows feed into AD through

I Government spendingI Disposable incomeI Wages

I Government spending feeds intoI disposable incomeI private investment


A labour market flows macro model A macro framework

A macro framework II

Double Phillips curveI Hybrid Phillips curve for price dynamicsI Wage dynamics dependent on

unemployment flows

Taylor rule, depending onI Expected inflation and output gapI Government balances

Automatic stabilisers, related toI Unemployment in- and outflows


Model estimation Data and methodology

Data and methodology

DataI Macro data from OECD Economic Outlook 87I Unemployment flow data from Elsby et al. (2008)

I Estimated flows based on LFS information onunemployment duration

I Match job creation/destruction rates under certainassumptions

MethodologyI Start with single-equation identificationI Then estimate system of equationsI Full macro-model on the basis of GMM


Model estimation Single equation identification

Determinants of job creation

Decomposition of unemployment outflows shows that:I Demand components play an important role (>40%)I Indication for some financial accelerator effect (>30%)I Relative prices (wages) more moderate role (<20%)

−23.1

−19.0

−3.6

10.9

11.8

14.1

17.4

−40

−20

020

4060

Con

trib

utio

ns to

une

mpl

oym

ent o

utflo

ws

(in %

)

Use

r co

stof

cap

ital

Rea

l wag

egr

owth

Em

ploy

men

t rat

e(la

gged

)

Cap

ital s

tock

grow

th

Gro

wth

of r

eal

shar

e pr

ices

Ext

erna

l dem

and

Rea

l dis

posa

ble

inco

me

grow

th

Tot

al


Model estimation Single equation identification

Determinants of job destruction

Decomposition of unemployment inflowsI No Schumpeterian effect from import penetration (strong demand effect)I Job churning due to changes in interest rates and TFP growth

−22.9

−20.1

−9.0

10.9

14.0

23.0

−50

−25

025

50C

ontr

ibut

ions

to u

nem

ploy

men

t inf

low

s (in

%)

Gro

wth

of

real

impo

rts

Ext

erna

l dem

and

Cha

nges

inT

erm

s of

Tra

de

Labo

ur fo

rce

grow

th

Gro

wth

of T

otal

Fac

tor

Pro

duct

ivity

Rea

l sho

rt−

term

inte

rest

rat

e

Tot

al


Model estimation System-equation approach

Estimation

Two-step procedureI System-equations approach to understand the

impact of different policiesI Use GMM to estimate the full macro model

Steady-state analysisI Dynamic system allows for unique steady state

(details in the paper)I Allows to differentiate between short- and

long-run effects of policies


Model estimation System-equation approach

Short- and long-run employment multipliers−

200

2040

Con

trib

utio

ns (

in %

)

Governmentspending

Government non−wage spending

Government wagespending

Policy contributions to outflows (short− vs. long−term)

−20

020

40C

ontr

ibut

ions

(in

%)

Governmentspending

Government wagespending

Government non−wage spending

Policy contributions to inflows (short− vs. long−term)

−20

020

40C

ontr

ibut

ions

(in

%)

Unemploymentbenefits

Hiringincentives

Trainingexpenditures

Public employmentservices

Direct jobcreation

Short−term effecton outflows

Long−term effecton outflows

−20

020

40C

ontr

ibut

ions

(in

%)

Trainingexpenditures

Public employmentservices

Hiringincentives

Unemploymentbenefits

Direct jobcreation

Short−term effecton inflows

Long−term effecton inflows


Model estimation Dynamic model simulation

Business cycle

Constant INt−1 LFPRt−1 ∆TFPt−1 RIRSt−1 TaxIndt−1 Gapt OUTt

(1) INt -3.503*** 0.593*** 1.141*** 0.465*** 0.009*** 2.032*** -0.021*** -0.032***

(0.139) (0.019) (0.095) (0.028) (0.001) (0.237) (0.001) (0.009)

Constant OUTt−1 ETRt UCCt ∆WRt ∆INVt Gapt INt

(2) OUTt -1.311*** 0.606*** 1.790*** -0.008*** -0.868*** 3.517*** 0.031*** 0.145***

(0.145) (0.014) (0.135) (0.002) (0.108) (0.205) (0.002) (0.018)

Constant RIRSt−1 E {GAPt+1} GovConst−1 TAXt−1 E {πt+1}(3) RIRSt 1.201** 0.791*** 0.076*** 4.756** -6.247*** 5.168***

(0.539) (0.012) (0.016) (1.971) (1.322) (1.138)

Constant E {RSharet+1} GovInvt−1 ∆Prodt−1 RIRLt−1 ∆ETt−1 GAPt−1

(4) INVt 0.007 0.011*** 0.350** 0.831*** -0.001*** 0.473*** 0.001***

(0.006) (0.002) (0.154) (0.051) (0.000) (0.023) (0.000)

Constant OUTt INt−1 ∆INVt−1 GovConst−1 TAXt−1 ∆NetExportst−1

(8) GAPt -12.931*** 1.620*** -1.940*** 56.495*** 22.651*** -30.290* 0.001***

(1.918) (0.136) (0.388) (4.624) (2.083) (18.368) (0.000)

Constant πt−1 E {πt+1} ∆ToTt−1 GAPt−1

(9) πt 0.001*** 0.500*** 0.489*** -0.044*** 0.001***

(0.000) (0.012) (0.013) (0.008) (0.000)

Constant ∆WSSEt−1 ∆ETt−1

(10) ∆WSSEt 0.006*** 0.792*** 0.146***

(0.001) (0.011) (0.039)


Model estimation Application: Understanding recovery options

Alternatives to austerity I

Analysing crisis exit scenariosI Early consolidation (individually or globally)I Additional spending

−1.

5−

1.0

−0.

50.

0

Gov

ernm

ent n

et le

ndin

g (

in %

of G

DP

)

2005 2010 2015 2020

Baseline scenario Early withdrawalGlobal fiscalconsolidation

Additional spending for3 years (3% of GDP)

Government deficit



Alternatives to austerity II

Substantially different employment patternsI Additional stimulus pushes up employment growth beyond trend

within 2 yearsI With consolidation the gap will not be closed at the end of the

forecast period−

0.8

−0.

6−

0.4

−0.

20.

00.

2

Ann

ual r

ate

of e

mpl

oym

ent g

row

th (

in %

)

2005 2010 2015 2020

Baseline scenario Early withdrawalGlobal fiscalconsolidation

Additional spending for3 years (3% of GDP)

Employment growth



Alternatives to austerity III

Distinct transmission mechanisms of policiesI Additional spending works mainly through higher outflowsI Fiscal consolidation depresses outflows AND increases inflows

0.0

0.5

1.0

1.5

2.0

Rat

e of

une

mpl

oym

ent o

utflo

ws

2005 2010 2015 2020

Baseline scenario Early withdrawal

Additional spending for 3 years (3% of GDP) Global fiscal consolidation

Unemployment outflows

0.1

0.1

0.1

0.2

0.3

Rat

e of

une

mpl

oym

ent i

nflo

ws

2005 2010 2015 2020

Baseline scenario Early withdrawal

Additional spending for 3 years (3% of GDP) Global fiscal consolidation

Unemployment inflows


Conclusion Summary

Summary

How should countries adjust their exit strategies?

1 Labour market policies more effective than genericgovernment spending

2 Focus on measures that contribute to automaticstabilization

3 Additional stimulus might have positive fiscal implications4 Premature consolidation packages will be damaging on both

flow margins


Conclusion Future extensions

Way forward

Coverage of country specificities

1 Increase country coverage (developing economies)2 Consider more complex labour market flows3 Estimate country-specific models

Extend the macro framework1 Financial market interactions à la Wasmer and Weil (2004)2 Enrich policy framework to cover different policy instruments3 This includes proper treatment of monetary policy4 Calibrate the model to make it fit for forecasting


Documents

Macroeconomic policy and employment flows