Macro-Labour Lecture: Understanding labour market

0 downloads 0 Views 7MB Size Report
National accounts also often contain labour market information ...... Note: In the rst model calibration the informal sector accounts for 57%. ..... Public sector. Note: ...

Macro-Labour Lecture:

Understanding labour market dynamics Ekkehard Ernst Macroeconomic Policies and Jobs Unit Research Department International Labour Organization (ILO)

Ekkehard Ernst

ILO

November 2018

1 / 184

How do long-term averages compare?

USA vs. Germany

A little quiz Long-term average performance (1973-2016) USA

Germany

Unemployment rate

6.4

6.4

Employment growth

1.4

0.6

Total hours worked

1.3

-0.1

(in % of the labour force) (in % p.a.)

(change in % p.a.)

Ekkehard Ernst

ILO

November 2018

2 / 184

Overview

Purpose and learning objective Purpose of this lecture

I

Introduction into main labour market concepts

I

How concepts relate to each other

I

How to measure labour market concepts

I

Carry out a country analysis

What you will learn in this lecture

I

Main building blocks of labour market analysis

I

How they relate to macro-economic models

I

How to identify key labour market obstacles

I

How to carry out a country scan

Ekkehard Ernst

ILO

November 2018

3 / 184

Overview

Overview 1

Theoretical considerations: Making sense of the labour market

I I I 2

...and its extensions Mobility, barriers and segmentation

Empirical applications: Sources and trends

I I I I I 3

The standard model...

Assessing real-world labour markets : Labour market information Margins of adjustment: Measuring wages, ows and hours worked Institutions and regulation Employment projections How to do a country scan

Country diagnostics: Where are the key barriers?

I I

Introduction to country scans Preparing group work sessions

Ekkehard Ernst

ILO

November 2018

4 / 184

Labour market concepts, indicators and models

Ekkehard Ernst

ILO

November 2018

5 / 184

Part I - Overview

Overview: Labour market concepts and theory 1

Key indicators of the labour market

2

Sources of labour market information: Labour market surveys, National accounts, and Big Data

3

Search, matching and mismatch: The standard case

4

...with endogenous search

5

...with monopolistic competition

6

...with endogenous job destruction

7

...with dual labour markets

8

Okun's law, the Phillips curve and natural unemployment

9

Growth and unemployment

Ekkehard Ernst

ILO

November 2018

6 / 184

Key labour market concepts and information sources

Key indicators of the labour market

Decomposing the labour force: Decent work indicators

Ekkehard Ernst

ILO

November 2018

7 / 184

Key labour market concepts and information sources

Sources of labour market information

Labour market information: Sources I

Surveys of households (`persons')

I I I

I

Population census Income and expenditure survey

Surveys of establishments (`companies')

I I I

I

Labour force survey

Establishment survey of production Employment and earnings survey Occupational employment and vacancy survey

Administrative data (`records')

I I I

Educational enrolment data Migration records Employment services records

Ekkehard Ernst

ILO

November 2018

8 / 184

Key labour market concepts and information sources

Sources of labour market information

Labour market information: Comparison I Household surveys

Strengths

I

Comprehensive coverage

Weaknesses

I

of population

Sampling prevents reliable estimates for small groups

I

Detailed questioning

I

Lower quality of data on

permits precise

income, sensitive and

measurement of

employer-related topics

statistical concepts for short reference periods

I

Ekkehard Ernst

Labour ows can be

I

Cannot provide estimates

estimated when panel

of vacancies, training

structure

needs, etc

ILO

November 2018

9 / 184

Key labour market concepts and information sources

Sources of labour market information

Labour market information: Comparison II Strengths

Establishment surveys

Ekkehard Ernst

Weaknesses

I

Comprehensive coverage of larger businesses

I

Poor coverage of very small and unregistered businesses

I

Payroll records provide consistent and reliable data for income and employment by industry

I

Requires constant updating of registers (births and deaths)

I

Only source for data on vacancies, training needs, etc

I

High non-response rates

I

Sampling prevents reliable estimates for small groups

I

Data items are limited by the information in establishment's registers

ILO

November 2018

10 / 184

Key labour market concepts and information sources

Sources of labour market information

Labour market information: Comparison III Administrative records

Strengths

I

Total count allows

Weaknesses

I

Often poor coverage

I

Often not up to date

I

Data quality may be

maximum detail

I

Inexpensive to compile statistics

questionable

I

Series breaks when adm. denition changes (e.g. to receive unemployment benets)

Ekkehard Ernst

ILO

November 2018

11 / 184

Key labour market concepts and information sources

Sources of labour market information

Labour market information: Processing

How to best exploit household or labour force surveys:

I

Pre-process data to t

I

Cross-tabulates and generates labour market indicators using ILO

ADePT

denitions

I

Extensive documentation of tool and labour market concepts: ADePT documentation

Ekkehard Ernst

ILO

November 2018

12 / 184

Key labour market concepts and information sources

National accounts

National accounts information National accounts also often contain labour market information

I

Auxiliary variables to allow to calculation of ratios, e.g. output per employee

I

Can contain more detailed information than LFS:

I

I I

Hours worked, wages

Are based on the establishement principle Beware of the production concept:

I

Production is a domestic concept, i.e. all production taking place in a particular economy

I

Labour force is a national concept, i.e. all residents that are part of the labour force

I

Large discrepancies in small open economies with cross-border workers (e.g. Luxembourg, parts of Switzerland).

Ekkehard Ernst

ILO

November 2018

13 / 184

Key labour market concepts and information sources

Other sources, big data

Other sources of labour market information I

ILO October Inquiry:

I

I

Meta survey  request pre-processed data

School-to-work transition surveys:

I

Tailor-made surveys on 30 countries (2 waves) on education and labour market status of 15-29 years old

I

Child-labour surveys:

I

Tailor-made surveys on labour market status using 75 household surveys of less than 18/15 years old

I

ILO's Key Indicators of the labour market (KILM):

I I

18 key indicators collected from national sources Overview of denitions and concepts, including limitations:

KILM

Ekkehard Ernst

ILO

November 2018

14 / 184

Key labour market concepts and information sources

Other sources, big data

Unstructured data

Ekkehard Ernst

ILO

November 2018

15 / 184

Key labour market concepts and information sources

Overview of labour markets around the world

A snapshot on global labour markets

Ekkehard Ernst

ILO

November 2018

16 / 184

Labour market models

The standard Beveridge curve

How to think about labour markets Beveridge curve

V

Lower job rent (p − w) #

V1 V0 Rising mismatch

U0

U1

U

The Beveridge curve links vacancies to job seekers via the matching function:

m (V , U).

m =

In equilibrium, matching equals job destruction:

m (V , U) = σL = σ (1 − U) I

Rising wages, lower productivity, higher costs of opening vacancies lowers job creation ray;

I

Rising mismatch shifts Beveridge curve outward.

Ekkehard Ernst

ILO

November 2018

17 / 184

1. Labour supply is xed; job destruction happens at exogenous rate σ 2. Changes in unemployment are determined by the rate at which job seekers match with open vacancies, Vt ∆Ut = σ · Lt − m (Vt , Ut ) · Ut

where m (·)displays constant returns to scale. 3. In equilibrium when ∆Ut = 0 and normalizing the labour force to unity, i.e. unemployment rate is: U¯ =

Lt =

1 − Ut the

σ

 ¯ 1 σ + m θ,

where θ = UVtt represents the tightness of the labour market. 4. Firms open a new vacancy when they expects its net prots prior to the match to equal the cost of opening a vacancy: (p + w ) · m (θ, 1) · Vt

=

⇒p−w

=

(r + σ) · p · ζ (r + σ) · p · ζ m (1, 1/θ)

where p: sales price, ζ : instantaneous vacancy costs 5. The quasi rent of a match is shared through (Nash) wage bargaining: w = (1 − β) R + β · p (1 + θζ)

where β : workers' bargaining power and R : the reservation wage. 6. The job creation equation is determined through combining the wage equation and the labour demand schedule: (r + σ) + βm (θ, 1) (1 − β) (p − R) = p·ζ m (1, 1/θ) Source: Pissarides (2000), ch. 1

Labour market models

The standard Beveridge curve

The Beveridge curve: Country examples

3000

3500

4000

4500

4000

5000

2015 2007 2006 2008

5000

6000

8000

10000

12000

14000

16000

350

Unemployment

400

450

2007

2001 2002 2003

2006 2011 2000 2005 2012 2010 2004 1999 1998

200

2013 2014 2015

1997 1996 1995

150

Vacancies 2010

2008

2009

100

2009 2006 2004 2005

250

300

80000 60000 20000

Vacancies

40000

2014 2012 2013 2011

20082007

300

2010

France

2015

250

2011

Unemployment

Sweden

200

2013 2012 2009

Unemployment

2001 20022003

2014

2003 2002 2005 2004

3000

Vacancies

2001

2016

2000

600 500 400 200

300

Vacancies

2015 2015 2012 2014 2012 2011 2001 2014 2013 2000 2013 2001 2000 2007 2007 2011 20082008 2002 2010 2002 2006 2009 2003 2010 2006 2009 2005 2003 2004 2005 2004 2500

United States

6000

Germany 2016

1990 1989 1992 1991 2000

1994 1993 2500

3000

Unemployment

Source: Germany - Bundesbank; USA - FRED; Sweden, France - OECD Main Economic Indicators Ekkehard Ernst ILO November 2018

18 / 184

Labour market models

The standard Beveridge curve

What explains labour market mismatch? I

Outward movements of the Beveridge curve indicate rising mismatch:

I I

I

More open vacancies per number of job-seekers More job-seekers for each individual vacancy

Reasons?

I

Geographical mismatch:

Job-seekers don't move to locations where

jobs are

I

Skill mismatch:

Job-seekers don't have the right skills (horizontal,

vertical, quantitative)

I

Wage increases and sectoral reallocation (Mehrotra & Sergeyev, 2012): High wage increases lower demand for specic sectors

I

Institutions (Bonthuis et al., 2015):

Employment protection,

occupational licensing prevent quick restructuring

I

Demographic shifts (Bova et al., 2016):

Ekkehard Ernst

ILO

Strong labour force growth

November 2018

19 / 184

1. Skill mismatch is also known as occupational mismatch. 2. Three dierent forms can be distinguished:  Quantitative: The most general form: supply of skills does not correspond to demand (either too high or too low)  Horizontal: Field of studies in supply and demand do not match (e.g. there is a high demand for engineers but only psychologists are available)  Vertical: The level of education does not match demand (e.g. strong demand for low-skilled workers is matched only by tertiary educated graduates) 3. In the case of vertical mismatch, two problems arise:  When demand for skills is higher than supply, productivity growth suers.  When supply for skills is higher than demand, reservation wages and unemployment (among graduates) is high.

Labour market models

The standard Beveridge curve

Alternative matching theories... Traditional explanation: Random matching between jobseekers and rms with open vacancies:

I

Pissarides (2000); Mortensen & Pissarides (1994)

Convex adjustment costs of labour and capital: Productivity shocks have persistent employment eects:

I

Den Haan et al. (2000); Merz & Yashiv (2007)

On-the-job search and heterogeneity: Degree of mismatch depends on state of the macro-economy:

I

Albrecht & Vroman (2002); Pissarides (1994); Mortensen & Nagypál (2007); Nagypál (2008)

Directed search and queueing lead to endogenous matching function with asymmetric reactions to gov' interventions:

I

Kohlbrecher & Merkl (2016); Lagos (2000); Sattinger (2005)

Endogenous search intensity and job quality changes with workers' outside options:

I

Acemoglu (2001)

Ekkehard Ernst

ILO

November 2018

20 / 184

Labour market models

The standard Beveridge curve

...and implications for employment dynamics Dierence between traditional and search-and-matching framework:

I

Asymmetric reactions to business cycle upswings and downswings (job-less recoveries)

I

Search-and-matching allows to distinguish between job creation and destruction dynamics (see, e.g., Davis et al., 1998)

I

Skills mismatch (e.g. skill-biased technological change) increases in upswings

I

Persistence of unemployment dynamics with respect to shocks

I I

Sluggish labour productivity adjustment following a shock Common framework to understand ows between unemployment, employment and inactivity

I

Accounting for job ows in driving employment (for the UK: Elsby et al., 2011; for the US:

I

Shimer, 2012)

Fluctuations of participation over the business cycle (Krusell et al., 2017)

Ekkehard Ernst

ILO

November 2018

21 / 184

Labour market models

Extensions of the Beveridge curve

Extensions of the Beveridge curve 1

Endogenous search

I

Depending on the state of the labour market, workers will search more intensively

2

Monopolistic competition

I 3

Endogenous separation

I 4

Firms are price-setters, workers price-takers

Productivity shocks will endogenously destroy jobs

Labour market segmentation

I

Distinguishing between dierent types of workers (skilled vs unskilled)

Ekkehard Ernst

ILO

November 2018

22 / 184

Labour market models

Extensions of the Beveridge curve

The Beveridge curve with endogenous search Beveridge curve: Endogenous search

V

Lower cost of filling a vacancy (p − w) "

V1 V0

Higher search effort

U1

U0

U

I

When search eort goes up, matching rates go up and less unemployed are roaming for vacancies

I

Employers cut time in nding workers, the job rent goes up and more vacancies are being created

Ekkehard Ernst

ILO

November 2018

23 / 184

Labour market models

Extensions of the Beveridge curve

The Beveridge curve with monopsonistic competition V

Beveridge curve: Monopsonistic competition

Lower wages (p − w) " V1

V0 Lower search effort

U0 U1

U

I

Employers set wages unilaterally; maximise share of matching rent that goes to capital owners

I

Workers reduce search eort, employers increase vacancy posting: Ambiguous eect on unemployment

Ekkehard Ernst

ILO

November 2018

24 / 184

What is monopsony?

ˆ

Firms are price-setters as they dominate the (local) labour market

ˆ

Workers are price-takers and cannot negotiate their wages

The monopsonistic rm maximises:

max π = p · y m (L) − w (L) · L L

with the rst-order condition:

∂π = 0 ⇔ p · yLm − w − w 0 · L = 0 ∂L Let

εLw ≡

w L

·

∂L . Then, the rst-order condition writes as: ∂w

yLm − w 1 = ⇔ yLm = w w εLw

 1

+

1

εLw

 .

1 > 0, y m > w and hence y m < y 0 . In other words, both production As ε L L Lw and employment will be in the monopsonistic case, compared to the full competition case.

smaller

Labour market models

Extensions of the Beveridge curve

Beveridge curve: Endogenous job destruction Beveridge curve: Endogenous separation

V

Negative productivity shock lowers matching rent

(p − w) #

V1 V0 Negative productivity shock shifts Beveridge curve up

U0

U1

U

I

A (permenant) negative productivity shock reduces the job rent, decreases job vacancy positing

I

If only certain sectors or occupations are aected, mismatch goes up and the Beveridge curve moves outward

Ekkehard Ernst

ILO

November 2018

25 / 184

Labour market models

Extensions of the Beveridge curve

Labour market duality Tightness for high-skilled labour higher than for low-skilled A: Arbitrage condition of vacancies between high- and low-skilled labour, with jobseeker arrival probability:

θH

A B

¯ θ:

Maximum arrival rate for low-skilled labour VL ! 0

U

θH ∼ V H H

Steepness of A determined by productivity differential between H and L workers

U

θL ∼ VL L

B: Flow equilibrium locus (Beveridge curve): ¯ = LH + LL L

45◦ θ¯

θL

Source: Saint-Paul (1996); Zenou (2008)

Ekkehard Ernst

ILO

November 2018

26 / 184

Labour market duality with matching frictions

Both labour market segments have type-specic, constant-returns-to-scale matching functions:   i i i m = m V ,U ,

where i

∈ {H, L}

The arrival rate at which jobseekers match with vacancies is the slack in the labour market and dened as:   i

m V i , Ui

θ =

=m

Vi

1,

Ui

Vi

!

(1)

Curve A:

Value of a vacancy, V , is determined by an asset equation and depends on the value of a matched job, J , and the expected future change in value, V˙ . Hence, the value of H-vacancy writes as: (r + θH ) V

H

H H

=θ J

+ V˙H

(2)

with r : the discount rate (interest rate) and θH : the labour market tightness of the H -labour market. Similarly, the value of L-vacancies are determined as: (r + θL ) V

L

L L = θ J + V˙L

where θL : the labour market tightness of the L-labour market.

(3)

The (ow) value of a job, J , depends on the productivity of L- versus H -jobs. (r + σ) J

H

(r + σ) J

L

=

ρ + σV

H

=

1 + σV

L

+ J˙H + J˙L

(4) (5)

where ρ > 1, the productivity of H−jobs (i.e. the productivity dierential with respect to L-jobs) and σ: the exogenous rate of job destruction. In equilibrium, we have: V

and

L

=V

H

V˙H = V˙L = J˙H = J˙L =

0

Combining equations (2)-(5) and rearranging terms, we obtain: θ

H

=

θL ρ − θ L (1 − ρ) / (r + σ)

which is represented in the chart by the curve A. The curve A is upward-sloping and lies above the 45°-line. The larger the productivity dierential, ρ, is, the steeper is the curve. Curve B:

Curve B describes the ow equilibrium locus. The number of jobseekers can be recovered from equation (1) as:   i i i U =V h θ

Assume a labour force normalised to unity, L = 1, and, total employment, E , and employment in H - and L-jobs as E H and E L . Then, the number of available jobs must equal the number of vacancies:

Let's assume x : the share of H -type workers, then: U

H

=x −E

H

and U L

=L−x −E

(7)

L

Change in employment is determined by hirings and job separations. In steady state where have:     i i i E˙ i = m V , U − σE =

0 ⇔ Ei

E˙ i =

m V i , Ui

=

σ

0 we (8)

Using (7) and (8), we obtain the market-specic Beveridge curves:   h θi

i

u =

where uH

H = Ux

, uL

θ i/σ

H H = 1V−x , v H = Vx

U V

+ h θi

, vL = =



and v i

L = 1V−x

=

1 θ i/σ

+ h θi



(9)

and the accounting equations:

x ·u

H

+ (1 − x) · u

L

x ·v

H

+ (1 − x) · v

L

Plugging in (9) into (8) and (6), we can determine the ow equilibrium locus B in the chart as: 1 + θH/σ 1 + θL/σ x H  + (1 − x) L  =E θ /σ + h θ H θ /σ + h θ L which is downward-sloping in the

  θ H , θ L -quadrant.

(10)

Labour market models

Extensions of the Beveridge curve

Rise in the skill premium More jobs lost for low-skilled than created for high-skilled Increase in the demand for highly skilled:

I I I

θH

A

A0 B

Decreases vacancy creation for low-skilled Increases vacancy creation for high-skilled

θ¯0 < θ¯ Maximum tightness for low-skilled labour declines

Increase in the skill premium

Rise in aggregate unemployment A rise in H-productivity makes vacancy creation for L-workers over-proportionally more costly θ¯0

θ¯

θL

Source: Saint-Paul (1996)

Ekkehard Ernst

ILO

November 2018

27 / 184

Using (9), the aggregate unemployment rate is given by: U = xu

H

    h θH h θL L + (1 − x) u = x H  + (1 − x) L . H θ /σ + h θ θ /σ + h θ L

Full dierentiation yields the change in the unemployment rate as a function of changing labour market tightness on H - and L-labour markets:     H H L L dU = xµ θ dθ + (1 − x) µ θ dθ (11) where µ

  θi =

     θ i h0 θ i −h θ i /σ   2 θ i /σ+h θ i

>

0 as h (·) is convex, i.e.

h0 > 0, h00 >

0.

Changes along the B -curve are determined by dierentiating (10): 0 = xϕ where ϕ

  θi =

    H H L L θ dθ + (1 − x) ϕ θ θ

         h θ i −θ i h0 θ i /σ− h0 θ i +1/σ   2 θ i /σ+h θ i

Plugging (12) into (11) yields:

(12)

.

       µ θL · ϕ θH H  H dU = xµ θ 1−    dθ µ θH · ϕ θL

The impact of a change in H -tightness, θH , on the unemployment rate will depend on:     µ θL · ϕ θH   < H µ θ · ϕ θL

Given the convexity of

  h θi ,

1⇔

        θ L h0 θ L − h θ L θ H h0 θ H − h θ H < .   0 L 0 H σh θ + 1 σh θ +1

this condition will be satised as long as

lies above the 45°-line. Given that µ

  θi >

0 and

dθ H dρ

>

θH < θL ,

0, this establishes that

which is the case as the A-curve dU dρ

>

0.

Labour market models

Extensions of the Beveridge curve

Rise in the supply of college graduates Increased labour supply makes both segments more uid θH

I

B

B0

A0

A

Further increase in vacancy creation for high-skilled

I

But also more vacancies for low-skilled

θ¯0

θ¯

θL

Source: Saint-Paul (1996)

Ekkehard Ernst

ILO

November 2018

28 / 184

Labour market models

Wages and rigidities

Theories on wages Wage bargaining

Sharing

Ecient

matching rents

bargaining

Benchmark theory:

Internal

Fully exible wages

labour market

Tournament

Wage

Eciency wages

posting

Prevent

Limit

Worker

Gift

Wage

shirking

turnover

selection

exchange

dispersion

Ekkehard Ernst

ILO

Monopsony

November 2018

29 / 184

Labour market models

Wages and rigidities

Why do wages not adjust during downturns? I

Administered wages and prices:

I I

I

Unemployment benets

Wage bargaining:

I I

I

Minimum wages

Wages adjust only partially to shocks Wages react to nation-wide, sectoral shocks

(Downward) wage rigidities (see Bewley, 1999):

I

Nominal wage rigidities:

I I I

I

Menu costs: xed costs of changing price tags Information rigidities: xed costs of acquiring information Social norms; fairness considerations

Real wage rigidities:

I I

Employment protection Eciency wages

Matching labour demand and supply is costly, not instantaneous! Ekkehard Ernst

ILO

November 2018

30 / 184

Labour market models

Labour supply: Extensions

When wages fall too much: The reserve army eect w LD

LS

wC

wI

LC

LI

L

Source: Dasgupta & Goldar (2006)

Ekkehard Ernst

ILO

November 2018

31 / 184

Labour market models

Labour supply: Extensions

Household income and labour supply decisions w wH wH , LH : High-income households Satiation level wC , LC : Competitive equilibrium wI , LI : Informal employment

wC

Reservation wage

LS

wI

LD LH LC

LI

L

Source: Sharif (2000)

Ekkehard Ernst

ILO

November 2018

32 / 184

Labour market models

Skills and technology

Skills and education Human capital an essential investment for economic development and labour market success:

I

General education increases access to new technologies, enhances productivity

I

Job-relevant skills are needed to access particular labour market

Skills can be:

I

general (reading, writing, math);

I

occupation-specic (e.g. how to assemble a car) or

I

company-specic (company procedures, production techniques)

Returns to schooling,

α,

typically measured using Mincer equations:

Yit = Y0 +α·YearsOfSchoolingit +β·ObservedIndividualCharacteristicit +εit

Ekkehard Ernst

ILO

November 2018

33 / 184

Labour market models

Skills and technology

Skill premia in historical perspective

1.0

Wage premium: Craftsmen to labourers 1.2 1.4 1.6 1.8

2.0

Wage premium in England: 1220−2000

1200

1400

1600

1800

Source: https://ourworldindata.org/skill-premium-income-by-education/; Clark (2005) Ekkehard Ernst ILO

2000

November 2018

34 / 184

Labour market models

Skills and technology

Returns to schooling across the world Returns to schooling by educational level and region (latest available year between 2000-2011) Region

Primary

Secondary

Tertiary

GDP per capita

World

10.3

6.9

16.8

6,719

Middle East and North Africa

9.4

3.5

8.9

3,645

South Asia

9.6

6.3

18.4

2,626

Eastern and Central Europe

8.3

4.0

10.1

6,630

High-income economies

4.8

5.3

11.0

31,748

(2005 PPP)

East Asia and the Pacic Latin America and the Caribbean Sub-Saharan Africa

11.0

6.3

15.4

5,980

9.3

6.6

17.6

7,269

13.4

10.8

21.9

2,531

Source: Montenegro & Patrinos (2014)

Ekkehard Ernst

ILO

November 2018

35 / 184

Labour market models

Skills and technology

20000

GDP per person employed (in US$) 25000 30000 35000

40000

Skills and economic development

2019

2015 2011 2007

2003 1999 1995 1991 12

14

16 High−skilled workforce (in %)

18

20

Note: Data after 2014 are forecasts.

Ekkehard Ernst

ILO

November 2018

36 / 184

Labour market models

Skills and technology

Skill-biased technological change

Note: TFP = total factor productivity. Source: De Michelis

Ekkehard Ernst

et al.

(2013)

ILO

November 2018

37 / 184

Labour market models

Skills and technology

Job polarization

Source: ILO (2015)

Ekkehard Ernst

ILO

November 2018

38 / 184

Labour market models

Skills and technology

Explaining job polarization Skill complementarity Ease of automation

High

Low

High

Routine-cognitive tasks

Routine manual tasks

Low

Non-routine-cognitive tasks

Non-routine manual tasks

What explains job polarization? (see, e.g., Goos et al., 2014)

I

Routine tasks will be automized

I

Non-routine tasks at both ends of the income distribution

But: I

Rogo (2011): Can skill premia persistent in a world with labour mobility?

I

Absolute vs. comparative advantages

Ekkehard Ernst

ILO

November 2018

39 / 184

Labour market models

Macro-labour dynamics

Macro-economic owchart

Ekkehard Ernst

ILO

November 2018

40 / 184

Labour market models

Macro-labour dynamics

Okun's law I I

Arthur Okun (1962) discovered an empirical relationship:

I

A negative correlation between the rate of unemployment and the real output

I

Okun's Law is a simple equation in which it has been used as rules of thumbs

I

Many studies have been done on conrming this relationship⇒Okun's Law has become a feature in standard macroeconomic models

I

To produce more goods and services in an economy more labour input is required. Examples:

I I

I

Hiring more workers making employees work longer hours.

Arthur Okun presumed that the rate of unemployment reected the amount of labour used in the economy

Ekkehard Ernst

ILO

November 2018

41 / 184

Okun's law describes the demand side and is a reduced form relationship between aggregate demand and labour demand: 1. Shifts in aggregate demand will cause output to uctuate around potential causes rms to hire or re workers which changes the employment level. 2. It describes deviations of employment and output from their potential when the economy is working under full employment. 3. Potential output is determined by the economy's productive capacity and increases over time as a result of technological change and factor accumulation. 4. The level of employment over the long run (the natural rate of unemployment) is determined by the size of the labour force and by frictions in the labour market 5. When output is at its long run, both employment and unemployment are at their long run levels. Moreover, in the log-run (without demand shocks), unemployment and output will be unrelated.

Labour market models

Macro-labour dynamics

Okun's law II Okun's law can be formulated as an empirical relationship between changes in unemployment,

∆u ,

and GDP growth,

g

(see Okun, 1962; Ball et al., 2013;

Cazes et al., 2013):

g = A − c · ∆u When labour market ows are in equilibrium:

∆u = 0 ⇒ ⇒

m (V , U) = σE = σ (1 − U) U = f (1/V ) , ∆u = f (1/∆V )

Growth and vacancy creation are therefore positively linked as

f 0 < 0:

g = A − c · ∆u = A − c · f (1/∆V ) In order to understand the impact on prices and wages, we need to link growth to ination. This is done with the

Phillips curve via the Modied output gap

(see Sell (2016)):

e e MOG : π = π+ 1 + βg = π+1 + β [A − c · f (1/∆V )]

Ekkehard Ernst

ILO

November 2018

42 / 184

Empirical estimates of Okun's coecients Estimated Okun's coecients vary across countries and over time:

country Australia Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands New Zealand Norway Portugal Spain Sweden Switzerland United Kingdom United States

Estimates pre 1995 post 1995 -0.552 -0.433 -0.134 -0.137 -0.634 -0.310 -0.500 -0.287 -0.490 -0.369 -0.610 -0.297 -0.400 -0.335 -0.427 -0.270 -0.462 -0.382 -0.142 -0.358 -0.109 -0.209 -0.713 -0.336 -0.317 -0.426 -0.319 -0.247 -0.221 -0.463 -0.793 -0.923 -0.648 -0.362 -0.211 -0.274 -0.419 -0.215 -0.447 -0.464

Source: Ball et al. (2013)

Labour market models

Macro-labour dynamics

The Phillips curve I π PC

π0

U

U0

Ekkehard Ernst

ILO

November 2018

43 / 184

Labour market models

Macro-labour dynamics

The Phillips curve II: Demand shock π PC

π1

π0

U1

Ekkehard Ernst

U

U0

ILO

November 2018

44 / 184

1. Phillips (1958) documented for UK data from 1861 to 1957 a downward-sloping relationship between the growth rate of nominal wages and unemployment. 2. This became an integral part of policy making in the 1960s:

πt =

X

γi πt−1 − β · ut

i≥1 where

γi

the autoregressive terms of past ination rates and

slope of the trade-o between unemployment and ination. 3. Policy makers expected to be able to exploit this trade-o permanently.

 Increase in the money wage via a monetary stimulus...  ... would lead to an increase in aggregate demand...  ...which would decrease unemployment.

β

the

Labour market models

Macro-labour dynamics

The Phillips curve III: The NAIRU π PC

NAIRU

π1

π0

U1

Ekkehard Ernst

U

U0

ILO

November 2018

45 / 184

Labour market models

Macro-labour dynamics

The Phillips curve IV: Rational expectations π PC

NAIRU

π1

π0

U

U0

Ekkehard Ernst

ILO

November 2018

46 / 184

1. Both Friedman (1968) and Phelps (1968) noted - independently - that with adaptive expectations, the Phillips curve trade-o breaks down, once ination is constant:

πt = πte − β · ut Whenever

πt = πte = πt−i

then necessarily

β = 0.

2. This creates a ratcheting eect:

 Increase in money wages leads to an increase in prices

⇒Purchasing

power and the real wage are not increasing

 Production has not increased, consumption will fall back to initial level.  Unemployment will return to previous level, i.e. the NAIRU 

But:

Ination is permanently higher now.

 This became the majority view after the hyperination episode in the 1970s. 3. With rational expectations, agents do not make any mistakes on average: πte = πt . There will be no trade-o, even in the short-run (Sargent & Wallace, 1975). 4. Expansionary policy without money illusion: Purchasing power is kept constant. Ination will jump immediately to higher level without aecting unemployment.

Labour market models

Macro-labour dynamics

Linking Beveridge and Phillips curves I

U

PC , and curve, BC ,

Phillips curve, Beveridge

P C1

BC0

BC1

P C0

are connected through the Modied Output Gap,

I

MOG ; N AIRU1

Move along the (short-term) Phillips curve aects job creation,

I

N AIRU0 JC0

JC ;

Shift of the Phillips

JC1

curve shifts the Beveridge curve, 0

I NAIRU

→ 1;

π

π1

V0

π0

V1

V

is the angle of

the job creation line compatible with stable ination

I

An increase in labour market mismatch (dashed lines), increases M OG0

th NAIRU M OG1

See Dickens (2009); Sell (2016)

π

Ekkehard Ernst

ILO

November 2018

47 / 184

Labour market models

Macro-labour dynamics

Growth and unemplomyent: Overview Disembodied technological change I

Productivity increases for all jobs

I

No replacement of capital needed

I

Ex.: Economies of scale from trade

Embodied technological change I

Productivity increases only for new jobs

I

Capital investment needed

I

Old jobs become obsolete

I

Ex.: Robotization

Ekkehard Ernst

ILO

November 2018

48 / 184

Labour market models

Macro-labour dynamics

Growth and unemployment: The capitalization eect I Introducing growth in the previous framework requires to consider capital accumulation (e.g. 1.

Eriksson, 1997):

Firms choose investment,

I,

desired employment,

L

and vacancies,

V,

to maximize their prots:

ˆ



max

I ,L,V

[F (K , p · L) − w · L − pcV − I ] e −rt dt

0

K˙ = I − δ · K where p : labour-augmenting productivity, c : vacancy creation cost, σ : exogenous job destruction, and δ : capital depreciation.

subject to

2.

L˙ = q (θ) V − σ · L

and

Workers reap part of the quasi rent of the match through wage bargaining:



ζ w = βp f (k) − (r + δ) k + cθ 1−ζ where

f (k) =

Ekkehard Ernst

1

p·L F



(K , p · L) . ILO

November 2018

49 / 184

Labour market models

Macro-labour dynamics

Growth and unemployment: The capitalization eect II 3.

Labour market equilibrium using standard optimization procedures:

  (1 − β) f (k) − kf 0 (k) − 4.

Labour market tightness,

⇒Unemployment

5.

The

θ,

r +σ−g ζ βcθ − c =0 1−ζ q (θ)

increases with (exogenous) growth,

g.

and growth are negatively related.

capitalization eect:

Firms that expect higher growth have an

interest to bring forward vacancy creation as the opportunity cost of an open vacancy falls with higher growth.

Ekkehard Ernst

ILO

November 2018

50 / 184

Labour market models

Macro-labour dynamics

Growth and unemployment: Endogenous tenure Tenure T JD

I

JC

With endogenous separation rms chose both vacancy creation and tenure

I

T0

Higher growth and higher wages lead to lower tenure

θ

θ0

Ekkehard Ernst

ILO

November 2018

51 / 184

Labour market models

Macro-labour dynamics

Growth and unemployment: Creative job destruction I

With technological progress, endogenous separation increases:

Tenure T JD0

JC1

JC0

JD1

JD0 → JD1 T0

I

Technological progress also raises wages, which lowers vacancy creation

T1

JC0 → JC1 θ1

Ekkehard Ernst

ILO

θ

θ0

November 2018

52 / 184

The discussion follows closely Merz (1995) and Ernst (2016). 1.

Households acccumulate wealth through deposits D = Y − c with income generated from asset returns and gainful employment Y = rW W + w · N , with rW the average ex post return on wealth and N the level of employment: W˙

D = rW W + w · N − c

=

(13)

In addition, households face an employment constraint, which depends on the number of jobs created by rms and determines the evolution of their disposable income: N˙

(14)

L

m (θ) − σ · N

=

where σ: the rate of job destruction assumed to be exogenously given. Households, then, maximise their interetemporal utility by determining their optimal path of consumption, c , subject to the job dynamics N according to the optimal household program: ˆ∞ max c

u (c, N) e

−ρt

dt

t=0

s.t. (13) and (14)  1−η c·N −γ

with the contemporaneous utility function u (c) = where η: the (constant) relative 1−η risk aversion, ρ: the intertemporal time preference rate and −γ (1 − η): the disutility from labour. 2.

In steady state, households optimal consumption evolve according to: g =

c˙ c

=

1 η

(rW − ρ)

(15)

3.

Firms post vacancies on the labour market and invest in new capital. They transform inputs according to a constant-returns-to-scale production function with labour-saving technological progress, A: F = F (K , AN) , FK > 0, FN > 0, FKK < 0, FNN < 0. Keeping a vacancy open is costly and proportional to wage costs ζ · w V . Existing jobs are destroyed at an exogenous rate σ. Installed capital depreciates at rate δ. The rm then maximises prots selecting vacancies, V ≥ 0, and its capital stock according to the following optimal program: ˆ ∞

max

N,K ,V

0

ν1 (t) (F (K , AN) − wN − rK K − ζ · w V − rK K ) e

−ρt

(16)

dt

where ρ stands for the intertemporal preference rate and ν1 (t) for the household's shadow variable related to its accumulation constraint, which rms maximise against the constraints of employment and capital accumulation:

4.



=

q (θ) V − σ · N



=

I − δ · K.

(17) (18)

Following Eriksson (1997), bargaining power, β , describes the bounds between which wages and interest rates have to fall: Workers have a fall-back option provided by the social security system, which guarantees a replacement income in form of unemployment benets, R0 . Hence, wages will be determined as a weighted average of the sum of marginal productivity, FN , the gain of lling a vacancy, θζ0 , and unemployment benets, R0 : w = β (FN + θζ0 ) + (1 − β) R0

Both unemployment benets and vacancy costs are assumed to be adjusted in line with real wages in order to capture increases in productivity:R0 = R · w (0 ≤ R < 1), ζ0 = ζ · w . Hence: w = β (FN + θζ · w ) + (1 − β) R · w



w =

β

1 − βθζ − (1 − β) R

FN

(19)

5.

The program set up by (16) subject to (17) yields an equilibrium on capital and labour markets summarised in the following proposition. Firms select the optimal stream of jobs and investment according to: LL ≡ (1 − β) (1 − R) − βθζ −

ζβ



q (θ)

b+η σ+ρ−w

c˙ c

 =

0

(20)

Fully dierentiating (20) allows to assess that: dθ dg

=−

∂LL/∂g ∂LL/∂θ

(

< >

0 if η > 1 0 if η < 1.

Hence, whether unemployment rises or falls with growth depends on the relative risk aversion, η, which is an empirical question. The table on the next slide provides an overview of existing estimates.

Relative risk aversion - Country estimates

Zimbabwe 0.04 Sri Lanka Belarus 0.09 Bosnia and Herzegovina Netherlands 0.1 Madagascar Albania 0.14 Germany Bolivia 0.16 Mexico Panama 0.18 Cameroon Benin 0.21 Canada Korea 0.27 Slovenia Serbia 0.27 Georgia Croatia 0.31 Uruguay Dominican Republic 0.32 Honduras Ireland 0.35 India Poland 0.38 Botswana Lao PDR 0.39 Myanmar Japan 0.44 United Kingdom Ukraine 0.44 Kosovo Paraguay 0.47 Bulgaria Estonia 0.51 Portugal El Slavador 0.54 Austria Finland 0.57 Greece Armenia 0.57 Mozambique Brazil 0.63 Chile Ghana 0.63 New Zeland Russia 0.65 Vietnam Uganda 0.67 Norway Source: Gandelman & Hernandez-Murillo (2014)

0.68 0.72 0.72 0.77 0.78 0.82 0.83 0.83 0.88 0.9 0.91 0.92 0.94 1.01 1.03 1.03 1.06 1.07 1.08 1.08 1.11 1.13 1.15 1.15 1.16

Australia Moldova Tajikistan Argentina Switzerland Lithuania Indonesia Tanzania South Africa Bangladesh Macedonia, FYR United States Ecuador France Peru Belgium Kyrgyz Republic Azerbaijan Senegal Malaysia Venezuela, Rep Bol Montenegro Burundi Taiwan Uzbekistan

1.17 1.19 1.19 1.2 1.21 1.23 1.24 1.26 1.29 1.3 1.34 1.39 1.39 1.43 1.44 1.55 1.81 1.85 1.89 1.93 2.08 2.1 2.17 2.45 2.96

Labour market models

Labour market segmentation and macro dynamics

Informality and business cycles Cyclical properties: Model comparisons with and without informality

Real business

Search and

Informality I

Informality II

cycles

matching

(high switching costs)

(low switching costs)

σ (y ) = 0.5

σ (y ) = 0.41

σ (y ) = 0.21

σ (y ) = 2.59

σ (x)

σ(x) σ(y )

corr (x, y )

σ (x)

σ(x) σ(y )

corr (x, y )

σ (x)

σ(x) σ(y )

corr (x, y )

σ (x)

σ(x) σ(y )

Consumption

0.32

0.63

0.96

0.22

0.55

0.96

0.13

0.58

0.96

1.11

0.43

0.92

Investment

0.21

0.43

0.91

0.16

0.38

0.92

0.08

0.39

0.91

1.58

0.61

0.96

Formal employment

0.03

0.06

0.71

0.05

0.13

0.97

0.05

0.24

0.96

0.03

0.01

0.56

-

-

-

0.05

0.13

-0.97

0.01

0.07

-0.01

0.02

0.01

-0.40 -0.34

Unemployment Informal employment Formal wages Informal wages

corr (x, y )

-

-

-

-

-

-

0.05

0.23

-0.99

0.04

0.01

0.35

0.70

0.99

0.22

0.54

0.99

0.13

0.62

0.96

1.10

0.43

0.97

-

-

-

-

-

-

0.09

0.44

0.99

1.15

0.45

0.98

Note: In the rst model calibration the informal sector accounts for 57%. We assume a capital elasticity in the production function of 0.26; the unemployment rate is 13%. In the second calibration, the informal sector is smaller at 50% as switching costs are smaller. Unemployment is now only 5% and the capital elasticity is calibrated at 0.54. For comparison, the results of a Real Business Cycle model and a standard Search and Matching model have also been displayed. Source: Bridji & Charpe (2011)

Ekkehard Ernst

ILO

November 2018

53 / 184

Labour market models

Information and expectations

Why are networks important to understand? Production networks I

Input-output linkages between sectors: Firms in dierent sectors buy from and sell to each other

I

A shock to a rm in one sector can impact rms in other sectors (I-O analysis)

I

Depending on the strength of sectoral linkages the shocks will be felt more or less in dierent sectors

Labour market networks I

Jobseekers receive information about vacancies from friends and relatives

I I

Pockets of unemployment become persistent Information about labour market success eects one's education decision (e.g. gang violence)

Ekkehard Ernst

ILO

November 2018

54 / 184

Labour market models

Information and expectations

Why are networks important to understand?

Financial market networks I

Banks are related to each other through credit linkages

I

If one bank goes bankrupt, these risks can spread from one bank to another

I

Are some banks systemic? Could they trigger a banking crisis?

I

This has become a major eld of research since the global nancial crisis!

Ekkehard Ernst

ILO

November 2018

55 / 184

Labour market models

Information and expectations

Labour market networks and information Micro-structure of (labour) markets Unemployment duration depends on network structure:

I

Calvò-Armengol & Jackson (2004)

Emergence of social institutions can be fostered in tightly-knit interactions:

I

Social interactions are segmented, not fully random: Ernst (2004)

Shock transmission in production networks, use sectoral data:

I

Acemoglu et al. (2012); Castagna et al. (2017)

Social networks and labour market outcomes

I

Montgomery (1991, 1992, 1994)

Social-network structure and information contagion:

I

Jackson et al. (2017); Shiller (2017)

Ekkehard Ernst

ILO

November 2018

56 / 184

Labour market models

Information and expectations

Shock transmission in networks I Understanding the employment impact of a nancial market shock: The Italian case:

I

Li,j

Employment

j = 1, .., n = 18 I

for

i = 1, ..., N = 8091

municipalities in

sectors

Construct a minimum-spanning tree (MST; connected graph) using sectoral similarities

I

The distance

δij

between two sectors

i

and

j

is calculated as:

δi,j = αi + βi · Li,j where

αi =

Ekkehard Ernst

1

maxj Li,j

and

βi =

ILO

minj Li,j maxj Li,j − minj Li,j

November 2018

57 / 184

Labour market models

Information and expectations

Shock transmission in networks II Panel A: Network structure

Panel B: Shock transmission 10.0

R 4.4

6

A

5.1

B

4.7

D

5.6

E

Variation in employment (in percent) 4.0 6.0 8.0

P 4

L 4.7

N

Q

4.6 9.2

S

8.9 12.9

10

5.5

J

K

H

M

F

G 40.9

0.0

I 28.1 29.4

2.0

22.1

C

A

B

C

D

E

F

G

H

I

J

K

L

M N

P

Q

R

S

Note: Letters designate the 18 dierent sectors as follows: A: Agriculture, forestry and shing; B: Mining and quarrying; C: Manufacturing; D: Electricity, gas, steam and air conditioning supply; E: Water supply sewerage, waste management and remediation activities; F: Construction; G: Wholesale and retail trade repair of motor vehicles and motorcycles; H: Transportation and storage; I: Accommodation and food service activities; J: Information and communication activities; K: Financial and insurance activities; L: Real estate activities; M: Professional, scientic and tech. activities; N: Administrative and support service activities; P: Education; Q: Human health and social work act.; R: Arts, entertainment and recreation; S: Other service activities. Source: Castagna et al. (2017) Ekkehard Ernst

ILO

November 2018

58 / 184

Labour market models

Information and expectations

How do networks evolve?

Ekkehard Ernst

ILO

November 2018

59 / 184

Labour market models

Information and expectations

How do expectations aect labour markets? I

Employers' perspectives:

I

Asset prices, investment and jobs: Phelps curves (Zoega, 2009; Hall, 2014; Kilic & Wachter, 2015)

I

I

Hiring uncertainty and nancial market turbulence

Workers' expectations:

I

Earnings prospects and investment in skills (e.g. Popov & Laeven, 2016; Wiswall & Zafar, 2016)

I I

I

Taxes, trade and informal employment Income ows and labour market participation

Social consequences of labour market outcomes:

I

Labour prospects and social unrest (see, e.g., ILO, 2015)

Ekkehard Ernst

ILO

November 2018

60 / 184

Labour market models

Information and expectations

Asset prices, investment and unemployment Share price index deflated by wage index) 60 80 100 120

Canada

2007 1999 2000 2006 1998 2001 2005 1997 2004 2002 2003 1996

20

1995 1994 1993 1991 1992 1989 1972 1990 1987 1971 1973 1970 1988 1986 19851976 1984 1974 1977 1980 1981 1975 1978 1979

2

4

6 Unemployment rate

1983 1982

8

2007 2006 2000 2005

1973 1972 19701971 1974

40

Share price index deflated by wage index) 40 60 80 100 120

United States

10

4

6

2004 2001 1999 1998 1997 2003 2002 1996 1987 1989 1995 1994 1980 1986 1988 1981 1990 1985 1993 1983 1991 1979 1992 1984 1975 1976 1978 1982 1977

8 Unemployment rate

1970 1971 1974

1988 1987

1991

1992

1

Ekkehard Ernst

1986 1993 1994 1996 19951997 1985

20072006

1984 1983 1981 1982 1979 1980 1978 1976 1975 1977

2

3 4 Unemployment rate

2000 20051999 2004 2001 1998 2002 2003

2007 2006

1970

2005 1972

2000 2001 1999 1987 2004 2002 1998 1997 2003 1996 1995 1989 1988 1986 1990 1980 1985 1981

1971 1973

1974

1979 1976 1975 19771978 1982

40

1990

12

Australia

1989

Share price index deflated by wage index) 60 80 100 120 140

Share price index deflated by wage index) 100 150 200

1973 1972

50

Japan

10

5

2

ILO

4

6 8 Unemployment rate

1994 1993 1991

1984

1992

1983

10

November 2018

61 / 184

Labour market diagnosis

Overview

Overview of empirical methodologies Objectives of growth and employment diagnostics Diagnose the bottlenecks to job creation and growth Propose solutions

Empirical tools:

I

Presentation of basic trends

I

Principal component analysis (PCA)

I

Regression analysis; establish the relationship between dierent variables of interest

I

Forecasting and modelling - Policy scenarios

I

Input-output matrices: Sectoral and occupational breakdown of policy scenarios

Ekkehard Ernst

ILO

November 2018

62 / 184

Labour market diagnosis

Growth diagnostics

A selection tree for growth analysis Low levels of private investment and entrepreneurship High cost of nance

Low return to economic activity

Low social returns

Low appropriability Balance of payments constraints Government failures

Market failures Lack of domestic nance

Low human capital Isolation/Lack of access to markets

Macro risks; nancial, monetary, scal instability

Low domestic savings

Poor intermediation

Micro risks: Property rights, corruption, taxes

Bad infrastructure Information externalities Coordination failures

Ekkehard Ernst

ILO

November 2018

63 / 184

Labour market diagnosis

Labour market diagnostics

Where are barriers for labour markets?

Ekkehard Ernst

ILO

November 2018

64 / 184

Labour market diagnosis

Labour market diagnostics

Linking macro- and labour-analysis I I

Linking the labour market decomposition:

Lt = Ntf + Ut + NtI1 + NtI2 I

...with the production function...

  Ytf = F At , Kt , Ntf , NtI1 I

...to determine employment in the informal economy:

MPLIt1 = wtI I

...and the formal economy:

f Ntf = (1 − σ) Nt− 1 + m (Vt , Ut )

Ekkehard Ernst

ILO

November 2018

65 / 184

Labour market diagnosis

Labour market diagnostics

Linking macro- and labour-analysis II I

Switching from formal to informal status and back is possible but costly:

Vtu = I

wtI + ce rt

...and depends on the average growth of national income:

gt = (1 − nt ) {σ [rt (1 − τ ) − ρ]} + nt · gtw I

...and the rate at which vacancies are created in the formal economy: 1

g LP 1+rt t

n

Ekkehard Ernst

1 βζ(1+τw )

h io + w (Rt+1 ) (1 − x) f (θt+1 ) − ζ1 = w (Rt ) · f (θt )

ILO

November 2018

66 / 184

Labour market policies and institutions

Ekkehard Ernst

ILO

November 2018

67 / 184

Part II - Overview

Overview: Labour market policies and institutions

1

International labour standards

2

Employment protection legislation

3

Minimum wages

4

Public employment

5

Trade unions

Ekkehard Ernst

ILO

November 2018

68 / 184

Labour market policies and institutions

Labour standards and employment protection legisation

International labour standards

0

Per cent of all ILO conventions 50

100

Ratification index (2016)

Central and Western Asia

East Europe+CIS

OECD

World

Source: ILO calculations

Ekkehard Ernst

ILO

November 2018

69 / 184

Labour market policies and institutions

Labour standards and employment protection legisation

Employment protection legislation: Elements 1

Permanent contracts

I I I 2

Advance notice periods Unfair dismissal: Possibility of re-instatement

Temporary contracts

I I 3

Severance payments: linked to seniority, capped?

Maximum length of temporary contract Renewability, waiting periods

Collective dismissals

I

Additional costs and procedures when more than one individual is being dismissed

4

Enforcement

I

How strictly are regulations enforced? (Number of labour inspectors, quality of legal system)

I

Currently no indicator available for this dimensions

Ekkehard Ernst

ILO

November 2018

70 / 184

Labour market policies and institutions

Labour standards and employment protection legisation

0

1

2

3

4

Employment protection legislation: Evolution

1985

1990

1995

2000

OECD average Australia Germany

Ekkehard Ernst

2005

2010

2015

Legislative bandwidth United Kingdom France

ILO

November 2018

71 / 184

Labour market policies and institutions

Labour standards and employment protection legisation

Employment protection legislation: Mechanisms V

Beveridge curve: Employment protection

EPL lowers matching rent

(p − w) # V0

V1 EPL lowers job destruction

U0 U1 Impact on vacancies: Unambiguously

I

Impact on unemployment:

Ekkehard Ernst

U

negative Ambiguous; positive only if job creation declines faster than job destruction

I

ILO

November 2018

72 / 184

Labour market policies and institutions

Labour standards and employment protection legisation

Employment protection legislation: Issues Empirically mixed outcomes:

I

Employment protection legislation does reduce labour market ows...

I

...and aects the job-content of growth (see discussion above)

Little evidence for its eect on aggregate outcomes:

I

Aggregate unemployment barely aected

I

Some studies suggests that marginal groups are particularly aected (youth, women)....

I

...but study results are context-specic

I

More robust results on its eects on reallocation and productivity growth

Kicking the can down the road: States impose upon rms what they cannot or do not want to regulate themselves

Ekkehard Ernst

ILO

November 2018

73 / 184

Labour market policies and institutions

Minimum wages and administered prices

Minimum wages: International comparison Minimum wage

0

KGZ UZB

KAZ RUS AZE

Per cent 50

UKR CZE USA TJK EST HRV MDA ROM ESP ARM HUN BGR SVK MKD LTV CAN POL SRB GBR FRA SVN DEU ALB

100

(in % of average wage)

Note: Data refers to 2013 or earlier, except for Germany: 2015. North America and Western Europe: red bars; Eastern Europe: green bars; Central Asia: blue bars Source: Global Wage Report

Ekkehard Ernst

ILO

November 2018

74 / 184

Labour market policies and institutions

Minimum wages and administered prices

Minimum wages: Standard case V

Beveridge curve: Minimum wages

Lower profits with MW (p − w) #

V0

V MW

Higher search effort with a minimum wage

U0

U MW

U

In fully competitive markets, minimum wages:

I I I

Lower job creation rates; Increase search eorts by job seekers; Lower open vacancies and raise unemployment. Ekkehard Ernst

ILO

November 2018

75 / 184

Labour market policies and institutions

Minimum wages and administered prices

Minimum wages: Monopsonistic competition I Beveridge curve: Minimum wages with monopsony

V

V0

1 U0 Let's start with the competitive equilibrium

I

Firms create vacancies

I

...to match with

Ekkehard Ernst

U0

U

1 :

V0 ...

job-seekers.

ILO

November 2018

76 / 184

Labour market policies and institutions

Minimum wages and administered prices

Minimum wages: Monopsonistic competition II Beveridge curve: Minimum wages with monopsony

V

Lower search effort

Lower wages (p − w) "

V1

V0

2

1 U0 In a monopsony we move from

1

to

U1

U

2 :

I

Firms have the power to set lower wages, which helps create more vacancies

I

...but job-seekers will lower search eort, leading to more unemployed

Ekkehard Ernst

ILO

V1 ...

U1 .

November 2018

77 / 184

Labour market policies and institutions

Minimum wages and administered prices

Minimum wages: Monopsonistic competition III V

Beveridge curve: Minimum wages with monopsony Lower search effort

Lower wages (p − w) "

Lower profits with MW (p − w) #

V1

V MW 3

V0

2 Higher search effort with a minimum wage

1 U0 Compared to the competitive equilibrium

U

1 , in a monopsony:

higher but search eort lower, increasing both open vacancies and unemployment

I

Job creation is

I

A minimum wage will

I

A minimum wage can therefore

Ekkehard Ernst

U1 U MW

reduce vacancy creation but enhance search eort reduce unemployment: ILO

U MW < U1

2

3

November 2018

78 / 184

Labour market policies and institutions

Minimum wages and administered prices

0

Public employment (in per cent of total employment) 2 4 6 8 10

Public employment around the world

D frica frica Asia Asia bean acific OEC rth A ntral ran A Carib South t Asia & P & No & Ce Saha ca & Eas East rope meri Sub− A le d n rn Eu Mid te Lati s a E SOE

General government

Public sector

Note: Partial country coverage per region. Sub-Saharan Africa: 9 countries; East Asia & Pacic: 7 countries; Eastern Europe & Central Asia: 18 countries; Latin America & Caribbean: 15 countries; Middle East & North Africa: 8 countries; OECD: 29 countries; South Asia: 3 countries. Source: ILO Stat; Rodrik (2000) Ekkehard Ernst

ILO

November 2018

79 / 184

Labour market policies and institutions

Trade unions and wage bargaining

Trade unions: Monopoly unions w

LD

Union bargaining curve

LS

wU nion

wcomp

LU nion

Lcomp

LS;max

L

Source: McDonald & Solow (1981)

Ekkehard Ernst

ILO

November 2018

80 / 184

Labour market policies and institutions

Trade unions and wage bargaining

Trade unions: Ecient bargaining w

LD

LS

π min

wU nion Efficient bargaining curve π U nion wcomp π∗

LU nion

Ekkehard Ernst

Lcomp

ILO

LS;max

L

November 2018

81 / 184

Labour market policies and institutions

Labour market segmentation

Segmented labour markets wage LdN

LdU

LdU 0

E w∗ E0 wN 0 OU

L∗U

L∗U 0

ON

Separate labour markets but full sectoral mobility:

I

Similar wages on formal and informal labour markets

I

Relative size of formal employment depends on the strength of labour demand in the formal economy.

Ekkehard Ernst

ILO

November 2018

82 / 184

Labour market policies and institutions

Labour market segmentation

Labour market segmentation with rigid wages wage LdN

LdU

LdU 0

A

C wU

EN

wN wN 0

0 EN

OU

LC U0

LC U

ON

Labour market rigidity (e.g. unionized wages):

I

Some part of the labour market receives a wage premium

I

Shocks to the unionized sector will increase the wage premium

Ekkehard Ernst

ILO

November 2018

83 / 184

Labour market policies and institutions

Labour market segmentation

Trade unions and an (informal) minimum wage wage LdN

LdU

LdU 0

A

C wU

EN M inimum wN wN Unemployment

OU

LC U0

LC U

¯U L

ON

Rigid wages in the non-unionised lead to unemployment:

I

Example: Minimum wages, administered prices

I

Shocks to the unionized sector will increase both the wage premium and unemployment

Ekkehard Ernst

ILO

November 2018

84 / 184

Labour market policies and institutions

Labour market segmentation

Minimum wage in the monopsonistic formal sector wage M CL

LdF

LdN

A

wFM inimum wFM onop

OF

C 0 wN wN

B

onop M inimum LF LM F

ON

Minimum wage when the formal sector is monopsonistic:

I

Monopsonistic wage is below equilibrium wage, welfare loss

I

An increase in the (formal sector) minimum wage, lowers informality and increases welfare.

Ekkehard Ernst

ILO

November 2018

85 / 184

Wages and labour market activity

Ekkehard Ernst

ILO

November 2018

86 / 184

Part III - Overview

Overview: Wage dynamics

1

Wage trends

2

Evolution of bargaining power

3

Demographic changes

4

Wages, employment and activity

5

An activity indicator

6

Determinants of wage growth

Ekkehard Ernst

ILO

November 2018

87 / 184

Wages, employment and hours

Wages and rigidities

Theories on wages Wage bargaining

Sharing

Ecient

matching rents

bargaining

Benchmark theory:

Internal

Fully exible wages

labour market

Tournament

Wage

Eciency wages

posting

Prevent

Limit

Worker

Gift

Wage

shirking

turnover

selection

exchange

dispersion

Ekkehard Ernst

ILO

Monopsony

November 2018

88 / 184

Wages, employment and hours

Wages and rigidities

Why do wages not adjust during downturns? I

Administered wages and prices:

I I

I

Unemployment benets

Wage bargaining:

I I

I

Minimum wages

Wages adjust only partially to shocks Wages react to nation-wide, sectoral shocks

(Downward) wage rigidities (see Bewley, 1999):

I

Nominal wage rigidities:

I I I

I

Menu costs: xed costs of changing price tags Information rigidities: xed costs of acquiring information Social norms; fairness considerations

Real wage rigidities:

I I

Employment protection Eciency wages

Matching labour demand and supply is costly, not instantaneous! Ekkehard Ernst

ILO

November 2018

89 / 184

Wages, employment and hours

Wage trends

Exploring wages and trends Existing cross-country predictions:

I

PriceWaterhouseCoopers (PwC) produced Global Wage Projections to 2030

I

(Hawksworth & Lambe, 2013)

The Conference Board provides labour cost projections for OECD countries (Levanon et al., 2016)

I

ILO Inwork, Global Wage Database (ILO, 2014)

Wages and macro conditions:

I

The Wage-Phillips curve: Wage growth as a function of unemployment

I

Wage curve: Level of wages as a function of unemployment (Blanchower & Oswald, 1994)

I

Dierence: Shocks to unemployment have permanent level eects with wage curve, compatible with empirical evidence

Ekkehard Ernst

ILO

November 2018

90 / 184

Sources of incomes and their breakdown

Three major categories can be distinguished in household earnings: 1.

Wages remunerate labour and are related to formal or informal (dependent) employment relationships.

2.

Mixed incomes are generated by self-employed or employers of un-incorporated entreprises where a distinction between capital and labour remuneration is dicult.

Components of labour costs

Depending on the tax-benet system, a signicant dierence exist between total labour costs and net take-home pay: 1.

Gross wages include taxes and workers' contributions to social security.

2.

In addtion, employers also have to contribute their part to social security, adding to total labour costs.

3.

Together this creates a signicant tax wedge between net wages and gross labour costs (more than 50% in some countries).

Wages, employment and hours

Wage trends

What explains slow wage growth? Wages have developed slowly, even in those countries where unemployment is low. Why?

I

Unemployment higher than actually measured?

I

Labour market slack is still large but hidden: part-time, temporary employment, lower participation rates

I

Fall in worker's bargaining power

I I I

I

Trade unions less powerful (globalization, service sector employment) Reduction in public sector wages Rising labour market concentration

Political and hiring uncertainty?

I I

Hiring uncertainty still higher than before the crisis Despite the low increase in living standards only limited social movements

I

Demographic eects?

I

Older workers have typically less rapid wage growth, composition eects

Ekkehard Ernst

ILO

November 2018

91 / 184

Wages, employment and hours

Wage trends

Fall in bargaining power I: Trade unionization Evolution of trade unionisation (in %) mid 2010s 100 Iceland

80 Denmark Sweden Finland

60 Belgium Norway Malta Cyprus

40

Bolivia Italy Luxembourg Russian Federation Argentina Canada Austria Ireland Romania Hong Kong, UnitedChina Kingdom Singapore Greece New Zealand Brazil Netherlands Chile Germany Namibia Japan Portugal SpainSwitzerland Australia Latvia Poland Czech Republic Slovakia UnitedRepublic States of Korea, Ethiopia Hungary Malaysia Philippines Lithuania France

20

Malawi

Estonia Guatemala

0 0

Source: ILOStat, Industrial Relations Data. Ekkehard Ernst

20

40

ILO

60 mid 2000s

80

100

November 2018

92 / 184

Wages, employment and hours

Wage trends

Fall in bargaining power II: Bargaining coverage Evolution of bargaining coverage(in %) mid 2010s 100

France Austria Belgium Finland Sweden Denmark Netherlands Italy Spain

80

Portugal Norway

60

Germany Luxembourg Switzerland Australia Czech Republic Cyprus

40

China Romania

Ireland Canada UnitedSlovakia Kingdom Hungary

20

Greece

Malawi Estonia Japan Latvia Korea, UnitedRepublic States of Lithuania Philippines

0 0

Source: ILOStat, Industrial Relations Data. Ekkehard Ernst

20

40

ILO

60 mid 2000s

80

100

November 2018

93 / 184

Wages, employment and hours

Wage trends

Fall in bargaining power III: Public sector wages

Ekkehard Ernst

ILO

November 2018

94 / 184

Wages, employment and hours

Wage trends

Demographic decomposition of wages

Wage growth tracker Ekkehard Ernst

ILO

November 2018

95 / 184

Wages, employment and hours

Wages, employment and activity

Should wages have been growing faster? Wage elasticity with respect to unemployment

−1.0

−0.5

0.0

G20 countries average

Pre−crisis

Ekkehard Ernst

Post crisis

ILO

November 2018

96 / 184

Wages, employment and hours

Wages, employment and activity

How to measure labour market slack

Labour market slack

Inactives

Adults > 15 years not in the labour force

Ekkehard Ernst

Unemployed

Out of a job and looking for one

ILO

Short hours

Maximum annual hours minus actual hours worked

November 2018

97 / 184

Wages, employment and hours

Wages, employment and activity

An activity indicator How to measure all margins on the labour market together?

I

A jobs activity index:

JAt = Hourst × (1 − Unemploymentt ) × LabourForcet I

Dene a Maximum Potential Activity Level of:

Potentialt = 3000 × WorkingAgePopulationt , I

i.e. an absence of unemployment where all people at working age (aged 15 and over) are working 3000 hours a year

I

This corresponds to the maximum amount of hours worked observed in the database.

I

The inactivity rate calculates then as (measured in per cent):



JAt InactivityRatet = 100 × 1 − Potentialt Ekkehard Ernst

ILO

 .

November 2018

98 / 184

Wages, employment and hours

Wages, employment and activity

Inactivity versus unemployment: USA United States Contribution 100 to inactivity

10 66 Inactivity/ Unemployment rate

80

6 64

60

40 3 62 20

0

0 60 1960

Ekkehard Ernst

1980

2000

Hours

Employment

Unemployment rate (RHS)

Inactivity rate (RHS)

ILO

2020 Participation

November 2018

99 / 184

Wages, employment and hours

Wages, employment and activity

Inactivity versus unemployment: France France Contribution 100 to inactivity

75 Inactivity/ Unemployment rate 12 74

80

60

73

40

7 72

20

71 2 70

0 1980

Ekkehard Ernst

1990

2000

2010

Hours

Employment

Unemployment rate (RHS)

Inactivity rate (RHS)

ILO

2020 Participation

November 2018

100 / 184

Wages, employment and hours

Wages, employment and activity

Wage growth: Unemployment vs. total hours

−0.5 −1.0

−1.0

−0.5

0.0

Panel B: OECD countries Wage elasticity with respect to unemployment

0.0

Panel A: G20 countries Wage elasticity with respect to unemployment

Pre−crisis

Post crisis

Pre−crisis

Post crisis

−0.5 −1.0

−1.0

−0.5

0.0

Panel D: OECD countries Wage elasticity with respect to inactivity

0.0

Panel C: G20 countries Wage elasticity with respect to inactivity

Pre−crisis

Post crisis

Pre−crisis

Post crisis

Note: Nominal wage inflation regressions include controls for GDP deflator, old−age employment and year trend. Inactivity is measured as the difference between total hours worked and a hypothetical maxi− mum number of potential hours worked. Crisis year=2008 Source: OECD, ILO Global Wage Database, own calculations

Ekkehard Ernst

ILO

November 2018

101 / 184

Understanding labour market dynamics

Ekkehard Ernst

ILO

November 2018

102 / 184

Part IV - Overview

Overview: Understanding labour market dynamics

1

Measuring labour market ows

2

Labour market mobility

3

The term structure of unemployment

4

Job-to-job transitions

5

Gross unemployment risk

6

Trade and jobs

7

Remittances and labour markets

Ekkehard Ernst

ILO

November 2018

103 / 184

Labour market dynamics

Measuring labour market ows

Labour market ows I Most labour market information is on stocks only: I

Labour force

I

Employment

I

Unemployment

I

Informal jobs

Information on labour market ows is lacking: I

Job creation, destruction and transitions

I

Worker dis- and encouragement, retirement

I

Labour market turnover:

I

Gross risk of unemployment

Ekkehard Ernst

ILO

Job Creation + Job Destruction Total employment

November 2018

104 / 184

Labour market dynamics

Measuring labour market ows

Labour market ows II EI: Retirement, maternity, sickness, disability IE: Job return after inactivity

Job stayers

Inactives

EI

E

I IE UI

EU

IU UE

UE: Unemployment outows EU: Unemployment inows

U

IU: Job search after inactivity UI: Discouragement

Long-term unemployed

Ekkehard Ernst

ILO

November 2018

105 / 184

Labour market dynamics

Measuring labour market ows

Labour market ows III I

Labour force surveys only oer observations at one point in time:

I I

Typically, individuals cannot be followed from one survey to another Few countries have panel surveys where individuals are followed over at least two consecutive surves

I

I

Some surves have recall questions, which are often unreliable

But: I I I

Reported information can be used to construct proxies:

Unemployment by duration Average tenure for current employees A growing literature has made use of these procedures, including for forecasting: Barnichon & Nekarda (2012); Shimer (2012); Elsby et al. (2013)

Ekkehard Ernst

ILO

November 2018

106 / 184

Labour market dynamics

Measuring labour market ows

Measuring unemployment ows I I

Measuring the gross dynamics on the labour market:

∆Ut = ∆Lt − ∆Et = Inflowst − Outflowst I

Changes in employment can be broken down:

∆Et = JobCreationt − JobDestructiont I

Assuming constant labour force, in- and outows of unemployment serve as proxies for job matching and job destruction:

∆Ut = st (1 − Ut ) − ft · Ut where:

Ekkehard Ernst

st

:

Job destruction rate

ft

:

Job creation rate

ILO

November 2018

107 / 184

Labour market dynamics

Measuring labour market ows

Measuring unemployment ows II I

Outows are inversely correlated to unemployment duration (see Shimer, 2012):

ft ∼ I

1

UDt

Consider the change in unemployment from month

t −d

to month

t:

ut − ut−d = utNew − Ft · ut−d where

utNew :

new job-seekers between

t −d

and

t

unemployment duration of less than 1 month. I

Hence:

Ft = 1 − Ekkehard Ernst

ut − utNew ut−d ILO

and ft

=

with an

− ln (1 − Ft ) d November 2018

108 / 184

Labour market dynamics

Measuring labour market ows

Measuring unemployment ows III: Examples Inflows (right scale)

−3

2.0

Outflows (left scale)

1980

1990

2000

2010

−6

−5

−4

1.0 −4.0 −3.0 −2.0 −1.0 0.0

−6

−5

−4.0 −3.0 −2.0 −1.0 0.0

1.0

−3

Italy Inflows (right scale)

−4

2.0

Canada Outflows (left scale)

1985

1990

2000

2005

2010

−3

2.0

1980

1990

2000

2010

Outflows (left scale)

1970

1980

Inflows (right scale)

1990

2000

−6

−5

−4

1.0 −4.0 −3.0 −2.0 −1.0 0.0

−6

−5

−4.0 −3.0 −2.0 −1.0 0.0

1.0

−3

United States Inflows (right scale)

−4

2.0

Japan Outflows (left scale)

1995

2010

Source: ILO (2013) Ekkehard Ernst

ILO

November 2018

109 / 184

Labour market dynamics

Measuring labour market ows

Measuring unemployment ows IV: More examples Germany

France

UR

350

14

United Kingdom

UR

350

14

UR

350

14

GFC 12

12

12

300 10

250

300

10

8

6

300

10

8

250

6

8

250

6

200 4

150

2 2004

2006

2008

2010

2012

2014

2016

United States

4

200

2018

2 2002

2004

2006

2008

2012

2014

2016

Japan

UR

3000

2010

4

200

2018

2 2000

2002

2004

2006

2008

2012

2014

2016

Italy

UR 14

14

2010

2018

UR 14

200 500 12

2800

12

12 180

450 10

10

10 160

2600 400 8

8

8 140

2400

350 6

6

6 120

2200 4

300

4

4 100

2000

2 2000

2002

2004

2006

2008

2010

2012

2014

2016

2018

250

2 2002

Total unemployment rate (right axis)

2004

2006

2008

2010

2012

2014

2016

Inflows in excess of outflows

2018

2 2000

2002

2004

2006

2008

2010

2012

2014

2016

2018

Outflows in excess of inflows

Note: Labour market flows (in 1000s) for selected countries - outflows from and inflows into unemployment vs. the unemployment rate. UR : Unemployment Rate. GFC : Global Financial Crisis Source: ILO Labour Flow Estimates 2018, Key Labour Market Indicators

Ekkehard Ernst

ILO

November 2018

110 / 184

Labour market dynamics

Measuring labour market ows

Measuring labour market mobility I Shorrocks Mobility Index Et−1 Ut−1 It−1

Et ptEE

Ut

It

.

.

.

ptUU

.

.

.

M Shorrocks =

P n− ni p i n−1

ptII

Mobility is measured by all those that change labour market status within one year.

I ptEE : I ptUU : I

Job stayers Long-term unemployed

ptII : Discouraged workers

Source: Formby

Ekkehard Ernst

et al.

(2004)

ILO

November 2018

111 / 184

0.0 un ga Po ry rt G uga er m l a Be ny lg iu G m re ec C ze F e ch ra R nce ep ub li Sp c ai Lu Ire n xe lan m d bo ur N Po g e l a U ni the nd te rla d Ki nds ng do Au m st U Fi ria ni nl te an d St d D ate en s m ar k

H

Shorrocks mobility index 0.2 0.4 0.6

Labour market dynamics

Ekkehard Ernst

Measuring labour market ows

Measuring labour market mobility II Shorrocks mobility index, 2009 Scale: 0 − no mobility; 1 − full mobility

ILO November 2018 112 / 184

Labour market dynamics

Measuring labour market ows

Measuring labour market mobility III How to measure mobility costs? I

Assume costly labour mobility across

Cij > 0 I

for

Wi Pi

= Marginal

productivity in sector

Flow utility of being employed in sector

Vi = RealWage I

sectors at cost

Each sector is assumed to be fully competitive, so that Real Wages

I

i ∈ {1, . . . , N}

i 6= j .

Gross labour ows from sector

i

+Cij

i : Fi0 (Li )

i:

× SectoralShock

to sector

j:

Vj mij ∼ PN

k=1 Vk

I

From the empirical observation of gross labour ows one can then deduct that (implicit) mobility costs (see Artuc et al., 2010, 2015).

Ekkehard Ernst

ILO

November 2018

113 / 184

Labour market dynamics

Measuring labour market ows

Measuring labour market mobility IV Labour mobility costs across the world

Note: Estimates of labour mobility costs based on UNIDO data. Source: Artuc

et al.

Ekkehard Ernst

(2015)

ILO

November 2018

114 / 184

Labour market dynamics

Measuring labour market ows

The term structure of unemployment How does the length of unemployment duration evolve over time?

I

Most labour force surveys only have duration data at specic period lengths:

I

I

e.g. 3, 6, 12, 24 months

Often no cross-country comparability, interval length depends on country surveys

I

But: similar distribution patterns:

I I

I

High incidence at low unemployment duration Incidence declines with duration

Estimate a

I

β -distribution

on the support [0, 1]:

Continuous distribution function characterised by two parameters,

α > 0, β > 0 I

Mean and variance:

α α+β Ekkehard Ernst

and

αβ 2

(α + β) (α + β + 1)

ILO

November 2018

115 / 184

Labour market dynamics

Measuring labour market ows

The term structure of unemployment: France

Term structure of unemployment

Ekkehard Ernst

ILO

November 2018

116 / 184

Labour market dynamics

Gross unemployment risk

Job tenure and job-to-job transitions I I

How to measure job-to-job transition rates? Available information:

I I I I

I

Incidence of tenure rates for 7 tenure brackets:

Incidencet Lengtht

Average tenure (in years) for these tenure brackets: Unemployment out- and in-ows: Employment:

UEt , EUt

Et

Calculate average tenure rates (in months):

AverageTenuret = where

i

i

12 100

·

X

Incidencei,t · Lengthi,t

i

are dierent tenure categories:

= {LessThan1Month,

1MonthTo6Months, 6MonthsTo12Months,

12MonthsTo3years, 3YearsTo5Years, 5YearsTo10Years, MoreThan10Years}

Ekkehard Ernst

ILO

November 2018

117 / 184

Labour market dynamics

Gross unemployment risk

Gross unemployment risk I

Employment survival rates:

ct = e I

1 − AverageTenure

t

Surviving jobs:

EEt = ct · Et I

Job-to-job transitions:

JJt = Et − EEt − UEt

Gross unemployment risk GURt

= EUt + LTUt + UEt =

Ekkehard Ernst

Job loss + Long-term U + Short-term U

ILO

November 2018

118 / 184

Labour market dynamics

Gross unemployment risk

Gross unemployment risk: Germany vs United States Germany − August 2018

United States − September 2018

At risk: Not at risk: 90.7%

At risk:

Not at risk: 84.4%

Job loss

Job loss

Short−term U

Short−term U

Long−term U

Long−term U

Unemployment rates: Germany - 3.4%, USA: 3.7% Gross unemployment risk: USA = 66% higher than in Germany

Ekkehard Ernst

ILO

November 2018

119 / 184

Growth and jobs

Job transitions and policies

17.4

14.1

11.8

10.9

−3.6

−19.0

I I I

Total

Real disposable income growth

External demand

Growth of real share prices

Capital stock growth

Employment rate (lagged)

Real wage growth

−23.1

User cost of capital

Contributions to unemployment outflows (in %) −40 −20 0 20 40

60

What aects unemployment ows? - Outows

Demand components play an important role (>40%) Indication for some nancial accelerator eect (>30%) Relative prices (wages) more moderate role ( 0, UL < 0

0

p·c =w ·L L = U (c, L) − λ [p · c − w · L]

yields the rst-order conditions: ∂L = ∂c ∂L = ∂L

0 ⇔ Uc0

0⇔

UL0

=λ·p = −λ · w

) ⇒

Uc0 UL0

3. The competitive labour market equilibrium therefore writes as: U0 0 y =− c UL0

=

p w

The standard labour market model

Monopsonistic competition

When employers are in charge w

M CL

LD

LS

M P Lmonopson

wcomp

wmonopson

Lmonopson

Lcomp

LS;max

L

Source: Manning (2003)

Ekkehard Ernst

ILO

November 2018

164 / 184

The monopsonistic rm maximises:

max π = p · y m (L) − w (L) · L L

with the rst-order condition:

∂π = 0 ⇔ p · yLm − w − w 0 · L = 0 ∂L Let

εLw ≡

w L

·

∂L . Then, the rst-order condition writes as: ∂w

yLm − w 1 = ⇔ yLm = w w εLw

 1

+

1

εLw

 .

1 > 0, y m > w and hence y m < y 0 . In other words, both production As ε L L Lw and employment will be in the monopsonistic case, compared to the full competition case.

smaller

The standard labour market model

Monopsonistic competition

When employers are in charge: Welfare loss w

M CL

LD

LS

Harberger triangle: Welfare loss

M P Lmonopson

wcomp

wmonopson

Lmonopson

Lcomp

LS;max

L

Source: Manning (2003)

Ekkehard Ernst

ILO

November 2018

165 / 184

The standard labour market model

Monopsonistic competition

When employers are in charge: A minimum wage w

M CL

LD

LS

Harberger triangle: Welfare loss

M P Lmonopson

wcomp wM inimum wmonopson

Lcomp

Lmonopson L

LS;max

L

M inimum

Source: Manning (2003)

Ekkehard Ernst

ILO

November 2018

166 / 184

When rms can set wages unilaterally, the introduction of a minimum wage can increase both wages and employment as long as

w Minimum < w Comp .

Some essential reading on minimum wages:

ˆ

Recent overview: Neumark & Wascher (2008)

ˆ

The 2015-16 Seattle minimum wage increase: Jardim et al. (2017)

ˆ

Minimum wages and rm employment in China: Huang et al. (2014)

ˆ

Minimum wages in Latin America: Maloney & Mendez (2003)

ˆ

Earlier evidence in low-wage services: Card & Krüger (1995)

Annex 2: Further reading

Ekkehard Ernst

ILO

November 2018

167 / 184

Further reading

Bibliography I Acemoglu, D. 2001. Good jobs vs. bad jobs.

Journal of Labor Economics, 19(1),

121.

Acemoglu, D., Carvalho, V. M., Ozdaglar, A., & Tahbaz-Salehi, A. 2012. The network origins of aggregate uctuations.

Econometrica, 80(5),

19772016.

Albrecht, J. W., & Vroman, S. 2002. A matching model with endogenous skill requirements.

International Economic Review, 43(1),

283305.

Artuc, E., Chaudhuri, S.., & McLaren, J. 2010. Trade shocks and labor adjustment: A structural empirical approach.

American Economic Reivew, 100,

10081045.

Artuc, E., Lederman, D., & Porto, G. 2015. A mapping of labor mobility costs in the developing world.

Journal of International Economics, 95, 2841. Okun's Law: Fit at fty?

Ball, L.B., Leigh, D., & Loungani, P. 2013.

Barnichon, R., & Nekarda, C. J. 2012. The ins and outs of forecasting unemployment: Using labor force ows to forecast the labor market.

Brookings Papers on Economic Activity, Fall,

83117. Bewley, T. F. 1999.

Why wages don't fall during a recession.

Cambridge, MA.: Harvard

University Press. Blanchower, D. G., & Oswald, A. J. 1994. Bonthuis, B., Jarvis, V., & Valhala, J. 2015.

determinants.

Ekkehard Ernst

The wage curve. Cambridge, MA.: MIT Press. Shifts in euro area Beveridge curves and their ILO

November 2018

168 / 184

Further reading

Bibliography II Shifting the Beveridge Curve. Dual labour markets with search costs.

Bova, E., Jalles, J. T., & Kolerus, C. 2016. Bridji, S., & Charpe, M. 2011.

Calvò-Armengol, A., & Jackson, M. O. 2004. The eects of social networks on employment and inequality.

American Economic Reivew, 94(3), 426454. Myth and measurement: The new economics of the minimum

Card, D., & Krüger, A. 1995.

wage.

Princeton, NJ.: Princeton University Press.

Economic vulnerabilities in Italy: A network analysis using similarities in sectoral employment.

Castagna, A., Chentouf, L., & Ernst, E. 2017.

Cazes, S., Verick, S., & Al Hussami, F. 2013. Why did unemployment respond so dierently to the global nancial crisis across countries? Insights from Okun's Law.

Policy, 2(1),

IZA Journal of Labor

118.

Chami, R., Ernst, E., Fullenkamp, C., & Oeking, A. Forthcoming. Do remittances benet workers in poor countries? Evidence from cross-country data.

IMF Working Paper. Journal of Political

Clark, G. 2005. The condition of the working class in England, 1209-2004.

Economy, 113(6),

13071340.

Dasgupta, P., & Goldar, B. 2006. Female labour supply in rural India: An econometric analysis.

Indian Journal of Labor Economics, 49(2), Ekkehard Ernst

293310.

ILO

November 2018

169 / 184

Further reading

Bibliography III Davis, S. J., Schuh, S., & Haltiwanger, J. C. 1998.

Job creation and destruction.

Cambridge,

MA.: MIT Press. De Michelis, A., ao, Estev & Wilson, B. A. 2013. Productivity and employment: Is it a choice?

International Productivity Monitor, 25,

4160.

Den Haan, W. J., Ramey, G., & Watson, J. 2000. Job destruction and propagation of shocks.

American Economic Reivew, 90(3).

Pages 205242 of: Fuhrer, J., & Kodrzycki, Y. K. (eds), Understanding ination and the implications for monetary policy: A Phillips curve perspective. MIT Press.

Dickens, W. T. 2009. A new method for estimating time variation in the NAIRU.

Elsby, M., Smith, J. C., & Wadsworth, J. 2011. The role of worker ows in the dynamics and distribution of UK unemployment.

Oxford Review of Economic Policy, 27(2),

Elsby, M., Hobijn, B., & Sahin, A. 2013. Unemployment dynamics in the OECD.

Economics and Statistics, 95(2),

Ernst, E. 2004.

Review of

530548.

Eriksson, C. 1997. Is there a trade-o between employment and growth?

Papers, 49,

338363.

Oxford Economic

7788.

The evolution of time horizons for economic develpoment.

Ernst, E. 2015. Supporting jobseekers. How benets can help unemploymed workers and strengthen job creation?

Ekkehard Ernst

International Social Security Review, 68(3), ILO

4367.

November 2018

170 / 184

Further reading

Bibliography IV Ernst, E. 2016. Might Tobin be right? The role of market frictions in policy-induced portfolio

Pages 261287 of: Bernard, L., & Nyambuu, U. (eds), Dynamic modeling, empirical macroeconomics, and nance. New York: Springer. Ernst, E., & Rani, U. 2011. Understanding unemployment ows. Oxford Review of Economic Policy, 27(2), 268294. Ernst, E., & Viegelahn, C. 2014. Hiring uncertainty: A new labour market indicator. shifts for growth.

Formby, J. P., Smith, W. J., & Zheng, B. 2004. Mobility measurement, transition matrices and

Journal of Econometrics, 120, 181205. American Economic Reivew, 68(1), Hernandez-Murillo, R. 2014. Risk aversion at the country level.

statistical inference.

Friedman, M. 1968. The role of monetary policy. Gandelman, N., &

117.

Goos, M., Manning, A., & Salomons, A. 2014. Explaining job polarization: Routine-biased

American Economic Reivew, 104(8), 25092526. High discounts and high unemployment. Tech. rept. 19871. National Bureau

technological change and oshoring. Hall, R. E. 2014.

of Economic Research, Cambridge, MA. Hawksworth, J., & Lambe, C. 2013.

Global wage projections to 2030.

Lnodon:

PriceWaterhouseCoopers. Huang, Y., Loungani, P., & Wang, G. 2014.

from China.

Ekkehard Ernst

Minimum wages and rm employment: Evidence ILO

November 2018

171 / 184

Further reading

Bibliography V ILO. 2013. The ILO Estimates on Unemployment Flows: Introducing Flows to Assess the Variations in Unemployment.

In: Key Indicators of the Labour Market, 8th edition.

Geneva:

International Labour Oce. ILO. 2014.

Global wage report. Wages and income inequality.

Geneva: International Labour

Organization (ILO). ILO. 2015.

World employment and social outlook. Trends.

Geneva: International Labour

Organization (ILO). ILO, & WTO. 2009.

Globalization and informal employment.

ILO, WTO.

Jackson, M. O., Rogers, B., & Zenou, Y. 2017. The economic consequences of social-network structure.

Journal of Economic Literature, 55(1),

147.

Jardim, E., Long, M. C., Plotnick, R., van Inwegen, E., Vigdor, J., & Wething, H. 2017.

Minimum wage iincrease, wages, and low-wage employment: Evidence from Seattle. The Impact of Globalization on the Informal Economy - A macro-level cross country study. Kilic, M., & Wachter, J. A. 2015. Risk, unemployment, and the stock market: A rare-event-based explanation of labor market volatility. Kohlbrecher, B., & Merkl, C. 2016. Business cycle asymmetries and the labor market. Kapoor, Rakesh. 2005.

Ekkehard Ernst

ILO

November 2018

172 / 184

Further reading

Bibliography VI Krusell, P., Mukoyama, T., Rogerson, R., & Sahin, A. 2017. Gross worker ows over the business cycle.

American Economic Reivew, 107(11),

34473476.

Lagos, R. 2000. An alternative approach to search frictions.

108(5), 851873.

Levanon, G., Colijn, B., Paterra, M., & Rust, E. 2016.

Journal of Political Economy,

Labor Cost Projections.

Maloney, William F., & Mendez, Jairo Nunez. 2003. Measuring the Impact of Minimum Wages: Evidence from Latin America.

No. 9800, .

National Bureau of Economic Research Working Paper Series,

Monopsony in motion. Imperfect competition in labor markets.

Manning, A. 2003.

Princeton

University Press. McDonald, I. M., & Solow, R. M. 1981. Wage bargaining and employment.

Reivew, 71(5),

McDonald, R., & Siegel, D. 1986. The Value of Waiting to Invest.

Economics, 101(4),

Quarterly Journal of

707728.

Mehrotra, N. R., & Sergeyev, D. 2012.

Policy.

Sectoral Shocks, The Beveridge Curve and Monetary

Merz, M. 1995. Search in the labor market and the real business cycle.

Economics, 36,

American Economic

896908.

Journal of Monetary

269300.

Ekkehard Ernst

ILO

November 2018

173 / 184

Further reading

Bibliography VII Merz, M., & Yashiv, E. 2007. Labor and the market value of the rm.

Reivew, 97(4),

Montenegro, C. E., & Patrinos, H. A. 2014.

around the world.

Comparable estimates of returns to schooling

Montgomery, J. D. 1991. Social networks and labor market outcomes.

Review, 81(5),

American Economic

14191431.

American Economic

14081418.

Montgomery, J. D. 1992. Job search and network composition: Implications of the strength-of-weak-ties hypothesis.

American Sociological Review, 57(5),

586596.

Montgomery, J. D. 1994. Weak ties, employment and inequality: An equilibrium analysis.

American Journal of Sociology, 99(5),

12121236.

Mortensen, D., & Pissarides, C. 1994. Job creation and job destruction in the theory of unemployment.

Review of Economic Studies, 61,

397415.

Mortensen, D. T., & Nagypál, E. 2007. More on unemployment and vacancy uctuations.

Review of Economic Dynamics, 10, 327347. Labor-market uctuations and on-the-job search. Neumark, D., & Wascher, W. L. 2008. Minimum wages. Cambridge, MA.: MIT Press. Okun, A. M. 1962. Potential GNP, its measurement and signicance. Tech. rept. Cowles Nagypál, E. 2008.

Foundation.

Ekkehard Ernst

ILO

November 2018

174 / 184

Further reading

Bibliography VIII Phelps, E. S. 1968. Money-wage dynamics and labor-market equilibrium.

Economy, 76(4),

Journal of Political

678711.

Phillips, A. W. 1958. The relationship between unemployment and the rate of change of money wages in the United Kingdom 1861-1957.

Economica, 25(100),

Pissarides, C. 1994. Search unemployment with on-the-job search.

61, 457475.

283299.

Review of Economic Studies,

Equilibrium unemployment theory. Cambridge, MA: The MIT Press. A lost generation? Education decisions and employment outcomes during the U.S. housing boom-bust cycle of the 2000s. Rodrik, D. 2000. What drives public employment in developing countries? Review of Development Economics, 4(3), 229243. Rogo, K. 2011. Technology and inequality. Saint-Paul, G. 1996. Dual labour markets. A macroeconomic perspective. Cambridge, MA.: Pissarides, C. 2000.

Popov, A., & Laeven, L. 2016.

MIT Press. Sargent, T., & Wallace, N. 1975. 'Rational' expectations, the optimal monetary instrument, and

Journal of Political Economy, 83(2), Labour queues.

the optimal money supply rule. Sattinger, M. 2005.

241254.

http://www.albany.edu/economics/research/workingp/2005/queues.pdf.

Ekkehard Ernst

ILO

November 2018

175 / 184

Further reading

Bibliography IX

Sell, F. L. 2016. Combining the Beveridge and the Phillips Curve into an integrative model: The Modied Output Gap.

Review of Economics and Finance, 6(2),

112.

Sharif, M. 2000. Inverted "S"-The complete neoclassical labour-supply function.

Labour Review, 139(4),

Shiller, R. J. 2017. Narrative economics.

American Economic Reivew, 107(4),

Shimer, R. 2012. Reassessing the ins and outs of unemployment.

Dynamics, 15,

International

409435. 9671004.

Review of Economic

127148.

Wiswall, M., & Zafar, B. 2016.

family.

Human capital investments and expectations about career and

Zenou, Y. 2008. Job search and mobility in developing countries: Theory and policy implications. Zoega, G. 2009.

Journal of Development Economics, 86,

336355.

Employment and asset prices.

Ekkehard Ernst

ILO

November 2018

176 / 184

List of slides

Understanding labour market dynamics

List of slides I 2

A quiz: Comparing long-term averages

3

Purpose and learning objective

4

Overview

6

Overview: Labour market concepts and theory

7

Decomposing the labour force: Decent work indicators

8

Labour market information: Sources

9

Labour market information: Comparison I

10

Labour market information: Comparison II

11

Labour market information: Comparison III

12

Labour market information: Processing

13

National accounts information

14

Other sources of labour market information

15

Unstructured data

16

A snapshot on global labour markets

17

How to think about labour markets

18

The Beveridge curve: Country examples

19

What explains labour market mismatch?

20

Alternative matching theories...

21

...and implications for employment dynamics

22

Extensions of the Beveridge curve

Labour market concepts and theories

Ekkehard Ernst

ILO

November 2018

177 / 184

List of slides

Understanding labour market dynamics

List of slides II 23

The Beveridge curve with endogenous search

24

The Beveridge curve with monopsonistic competition

25

Beveridge curve: Endogenous job destruction

26

Labour market duality

27

Rise in the skill premium

28

Rise in the supply of college graduates

29

Theories on wages

30

Why do wages not adjust during downturns?

31

When wages fall too much: The reserve army eect

32

Household income and labour supply decisions

33

Skills and education

34

Skill premia in historical perspective

35

Returns to schooling across the world

36

Skills and economic development

37

Skill-biased technological change

38

Job polarization

39

Explaining job polarization

40

Macro-economic owchart

41

Okun's law I

42

Okun's law II

43

The Phillips curve I

Ekkehard Ernst

ILO

November 2018

178 / 184

List of slides

Understanding labour market dynamics

List of slides III 44

The Phillips curve II: Demand shock

45

The Phillips curve III: The NAIRU

46

The Phillips curve IV: Rational expectations

47

Linking Beveridge and Phillips curves

48

Growth and unemplomyent: Overview

49

Growth and unemployment: The capitalization eect I

50

Growth and unemployment: The capitalization eect II

51

Growth and unemployment: Endogenous tenure

52

Growth and unemployment: Creative job destruction

53

Informality and business cycles

54

Why are networks important to understand?

55

Why are networks important to understand?

56

Labour market networks and information

57

Shock transmission in networks I

58

Shock transmission in networks II

59

How do networks evolve?

60

How do expectations aect labour markets?

61

Asset prices, investment and unemployment

62

Overview of empirical methodologies

63

A selection tree for growth analysis

64

Where are barriers for labour markets?

Ekkehard Ernst

ILO

November 2018

179 / 184

List of slides

Understanding labour market dynamics

List of slides IV 65

Linking macro- and labour-analysis I

66

Linking macro- and labour-analysis II

68

Overview: Labour market policies and institutions

69

International labour standards

70

Employment protection legislation: Elements

71

Employment protection legislation: Evolution

72

Employment protection legislation: Mechanisms

73

Employment protection legislation: Issues

74

Minimum wages: International comparison

75

Minimum wages: Standard case

76

Minimum wages: Monopsonistic competition I

77

Minimum wages: Monopsonistic competition II

78

Minimum wages: Monopsonistic competition III

79

Public employment around the world

80

Trade unions: Monopoly unions

81

Trade unions: Ecient bargaining

82

Segmented labour markets

83

Labour market segmentation with rigid wages

84

Trade unions and an (informal) minimum wage

85

Minimum wage in the monopsonistic formal sector

Labour market policies and institutions

Ekkehard Ernst

ILO

November 2018

180 / 184

List of slides

Understanding labour market dynamics

List of slides V Wages and employment 87

Overview: Wage dynamics

88

Theories on wages

89

Why do wages not adjust during downturns?

90

Exploring wages and trends

91

What explains slow wage growth?

92

Fall in bargaining power I: Trade unionization

93

Fall in bargaining power II: Bargaining coverage

94

Fall in bargaining power III: Public sector wages

95

Demographic decomposition of wages

96

Should wages have been growing faster?

97

How to measure labour market slack

98

An activity indicator

99

Inactivity versus unemployment: USA

100 Inactivity versus unemployment: France 101 Wage growth: Unemployment vs. total hours

Labour market ows

103 Overview: Understanding labour market dynamics 104 Labour market ows I 105 Labour market ows II 106 Labour market ows III

Ekkehard Ernst

ILO

November 2018

181 / 184

List of slides

Understanding labour market dynamics

List of slides VI 107 Measuring unemployment ows I 108 Measuring unemployment ows II 109 Measuring unemployment ows III: Examples 110 Measuring unemployment ows IV: More examples 111 Measuring labour market mobility I 112 Measuring labour market mobility II 113 Measuring labour market mobility III 114 Measuring labour market mobility IV 115 The term structure of unemployment 116 The term structure of unemployment: France 117 Job tenure and job-to-job transitions 118 Gross unemployment risk 119 Gross unemployment risk: Germany vs United States 120 What aects unemployment ows? - Outows 121 What aects unemployment ows? - Inows 122 Labour ows and Okun's law I 123 Labour ows and Okun's law II 124 Labour ows and Okun's law III 125 Labour ows and Okun's law IV 126 Determinants of outow elasticities 127 Determinants of inow elasticities

Ekkehard Ernst

ILO

November 2018

182 / 184

List of slides

Understanding labour market dynamics

List of slides VII 128 Alternative measures of uncertainty 129 Stochastic labour demand I 130 Stochastic labour demand II 131 Stochastic labour demand III 132 Stochastic labour demand IV 133 Hiring uncertainty: Example USA 134 Labour market outcomes under hiring uncertainty 135 Financial stress, hiring uncertainty and job creation 136 The New-Keynesian Phillips curve 137 An overview of the model ows I 138 An overview of the model ows II 139 Putting the pieces together 140 Modelling methodology I 141 Modelling methodology II 142 A double (hybrid) Phillips curve 143 A word on the data and methodology 144 Estimation results: Labour block 145 Estimation results: Macro block 146 Estimation results: Fiscal block 147 Long-term vs. short-term eects: Job creation 148 Long-term vs. short-term eects: Job destruction

Ekkehard Ernst

ILO

November 2018

183 / 184

List of slides

Understanding labour market dynamics

List of slides VIII 149 Exit scenarios from the crisis: Government decit 150 Exit scenarios from the crisis: Job growth 151 Trade and informal employment 152 Trade and informal employment: Determinants 153 Remittances and labour markets I 154 Remittances and labour markets II 155 Remittances and labour markets III 156 Capital/income ows and labour markets

Group-work: Country study

158 Going forward: Exercise this week 159 A checklist for the country scan

Annex

161 Some useful labour market concepts I 162 Some useful labour market concepts II 163 Many workers and rms 164 When employers are in charge 165 When employers are in charge: Welfare loss 166 When employers are in charge: A minimum wage 167 Bibliography

Ekkehard Ernst

ILO

November 2018

184 / 184