“job quality, labour market performance and well-being”__barbieri

41
ATYPICAL EMPLOYMENT AND WELFARE REGIMES A SOCIOLOGICAL INSIGHT Paolo Barbieri Department of Sociology and Social Research, Trento University and FamIne ERC2010 Project EC-OECD seminar on Job Quality, Labour Market Performance and Well-Being 27-28 November 2014, Brussels Funded under the 7° FP Call: ERC-2010

Upload: statscommunications

Post on 27-Jul-2015

137 views

Category:

Data & Analytics


2 download

TRANSCRIPT

ATYPICAL EMPLOYMENT AND WELFARE REGIMES A SOCIOLOGICAL INSIGHT

Paolo Barbieri

Department of Sociology and Social Research, Trento University

and FamIne ERC2010 Project

EC-OECD seminar on Job Quality, Labour Market Performance and Well-Being

27-28 November 2014, Brussels

Funded under the 7° FP

Call: ERC-2010

Consequences of atypical (flexible) employment relations?

I. Labour market career and flexible employment (LM)

II. The effects of Lm precarity on (first) motherhood (DEMO) Is there a common, negative, effect or is it just a problem for the ‘mediterranean’ sub-protective countries?

Consequences at the micro level of atyp jobs and how they are shaped by institutions

The ATYPWR Project: Topics

ITALY GERMANY HUNGARY NL FRANCE SWEDEN G.B. DK

Period : 1970-2005

ILFI

1985-2007;

(1995-07)

SOEP

Youth 2000 1980-2000 GEN.98 +

FQP 1985-

1993-2003

LNU 2000 1994-2006

BHPS

LFS

definition of

atypical

employment

FTC, TWA,

Apprent.,

Solo-self

FTC, TWA FTC, TWA,

Solo-Self

FTC, TWA,

Apprent.,

Solo-self

FTC, TWA,

Voc., Solo-

self

FTC + Voc

FTC

FTC (FT/PT)

The ATYPWR project

Plus: ECHP long.; EU-SILC long.; EU-LFS

Structure of the Presentation

Macro theoretical framework… and some micro

I. LM: labour market (trap/springboard?)

II. DEMO: fertility (delay?)

macro dimensions and micro consequences of flexible employment

Results from the 7 country study

…and then? Crisis : FTC and LTU risks

Definitions - Italy

“Labour Market Flexibilisation and its

Consequences in Italy” European Sociological

Review (2009) 25 (6): 677-692 (with Stefani

Scherer)

Three main “macro” dimensions, strictly interconnected, structuring EU LMs and the consequences of LM flexibilization at the micro level:

- labour market institutions: EPL, ETS

- the characteristics of the process of labour market flexibilisation (institutions of LM deregulation):

at the margins Vs universalistic

- welfare: “Beveridge Vs Bismark”

LM MACRO Dimensions (the literature)

- School to Work lit.

- E.P.L. / P.M.R. / Flexicurity & T.L.Ms.

- Globalife, Flexcareer, TransEurope…

V.O.C. (Hall and Soskice, 2001)

OECD; Blossfeld; Soskice-Hall; Esping-

Andersen-Regini; Streeck; Blanchard

(“Insider disease”); Blau and Kahn; Di Prete

et al.; Maurin, Postel-Vinay “the

wage/security trade off ” for LM

adjustments; Scarpetta

Titmuss; Korpi; Esping-Andersen; Palier;

Ferrera…

Flexicurity

Citizenship-based welfare systems

Beveridge

Low EPL => “universalistic” LMflex

Higher WS

support

Vocational

ETS =>

Nordic W.R.

DK

NL SWE

Liberal W.R.

UK

Scarce WS

support

Generalistic

ETS =>

Skill divide

OLMs

GER

Conservative W.R.

FR IT

Conservative /

Sub-protective W.R.

Age divide

ILMs

High EPL => Partial & Targeted LMflex

(INS/OUTS)

Work-based (labouristic) welfare systems

Bismarck

HU

LM Institutions + Deregulation + Welfare

Job Search theory

screening hypothesis => speed up the entry process for the highly productive workers;

effort eliciting mechanism => for the best qualified and productive workers = bridges to secure job = EVERYWHERE, BUT MORE IN HIGH EPL COUNTRIES

On the job hunter side, it’s a s.c. “multi-armed bandit” problem that is a problem of search from unknown distribution of job offers: the searcher does not even know the distribution of offers, so by searching longer, more is learned about the range of job offers available.

=> better search: delays entry = ITALY!

Signalling theories

Individual characteristics: Gender, Age, Ethnicity, as well as Family duties, Employment history or Education & Training, work as “proxies” for workers’ productivity. = EVERYWHERE, BUT MORE IN HIGH EPL COUNTRIES and/or ILMs assets

Reservation Wage Th.

R.W. is expected to be higher for FTC… but should decrease with job search duration. => duration dependence

-W.S., Individual and family characteristics are expected to affect job seeker’s reservation wage…

-Economic and LM conditions, affecting the job offers arrival rate, can influence individual’s reservation wage…

Job Shopping Theory: but can be also a ‘negative signalling’!

Production theories “Rhein model of capitalisms”

FIRM’s characteristics and market competition strategies (Labour costs reduction VS Diversified Quality Production) are key mechanisms orienting employers’ incentives as regard workforce policies, and therefore the payoff of Flexible forms of employment on the labour market

LM MICRO Mechanisms (literature)

Micro consequences regarding atypical employment ? Citizenship-based welfare systems

Beveridge

Low EPL => “universalistic” LMflex

Vocational

ETS =>

DK, SWE? (NL?)

Low trap;

No gender segregation;

No/Limited cohort

divide;

High exit rates;

Unskilled at risk

UK, HU?

No trap;

mid gender

segregation;

Cohort divide;

High exit rates;

Generalistic

ETS =>

Skill divide

GER, NL?

Some trap;

gender segregation;

No/Limited cohort

divide;

Mid exit rates

Unskilled at risk

FR, IT

High entrapment;

High cohort divide;

Youngsters mainly at

risk;

Gender segregation;

Low exit rates

Age divide

High EPL => Partial & Targeted LMflex

(INS/OUTS)

Work-based (labouristic) welfare systems

Macro Contexts and Micro Consequences

Institutions intervening in the motherhood decisions:

- Welfare State

- Market (services and labour Mks) the “welfare triangle”

- Family

Within these institutional boundaries (macro level),

different theories suggest various research hypothesis (micro level)

DEMO MACRO Dimensions

Consequences at the micro level, differently influenced by various “welfare equilibria”

THEORIES

1) New Home Economics: (opportunity costs matter)

- intra-family allocation of paid / unpaid work (male/female roles: functional advantages entailed in Becker’s specialization theory)

=> HYP. : + Effects of partner’s income/class/LM position/mobility/ (Intra-family dependency)

- the independence hypothesis: increased female education increased female LM participation increased incompatibility between female employment and family roles (Becker, 1973, 1981; Sweeney, 2002)

=> HYP. : Effects of woman’s education/income/class/LM position/mobility/ Delay and subsequent catch-up for highly educ. women

DEMO MICRO Theories

2) Career entry hypothesis: focus on the increasing difficulties to access stable, secure employment and to achieve economic and work stability for young people (Oppenheimer, 1988, 1994; Billari, 2008)

“scarring effect”

=> HYP. : Effects of woman’s work insecurity/ income/class /LM position/ Delay of motherhood

=> HYP. : Combined effects of both partners’ precariousness

=> HYP. : Effect of early LM instability (individual or coupled) on motherhood (Blossfeld et al. 2005; Bernardi et al. 2005; Oppenheimer 1988)

3) Theory of the “second demographic transition” : no reasonable way to deal with it, no data on values, attitudes, cultural orientations, perceptions, preferences…

DEMO MICRO Theories II

Given the macro scenarios, what micro consequences regarding motherhood ?

Citizenship-based welfare systems

Beveridge

Low EPL => “universalistic” LMflex

LM adjustment based on (higher) Wage inequality

Vocational

ETS =>

Nordic W.R.

DK, SWE, NL(?)

WState support to motherhood;

No Intra-family dependency;

No Delay;

No effects of atyp.job

Liberal W.R.

UK

Market ‘support’;

Delay for educ women;

Intra-family dependency

(+ Partner’s effect)

Low effects of atyp.job

Generalistic

ETS =>

Skill divide

Conservative W.R.

GER

WState support to Family care; /

Intra-Family dependency;

Some effects of atyp.job

Cons./Sub-protective W.R.

FR, IT

FR: WState support;

IT: Family care;

Intra-Family dependency;

Strong effect of atyp.job

Age divide

High EPL => Partial & Targeted LMflex

LM adjustment based on Job insecurity (INS/OUTS)

HU

DEMO Macro Context and Micro Consequences

“7 countries * 8 models”

The ATYPWR Project: research design

1 ) LM CONSEQUENCES

Q1. WHO ENTER S LM ATYPICALLY ?

Table 1: Type of FIRST employment Multinomial logit (typical employment; atypical employment ; seasonal+off - the - books)

Q2. WHO GET S A SECURE JOB?

Table 2: Entry in the first secure job , Event History Analysis (EHA) piecewise constant exp.

Q3. WHO MAKES THE TRANSIT ION TO A STABLE JOB? / WHO REMAINS TRAPPED?

Table 3: Exit from atypical employmen t EHA piecewise const . exp. competing risk model ( permanent job ; repeated atyp . job; Not - working)

Q4. LONG TERM SCARRING EFFECT ?

Table 4: Employment situation at age 35 A ) NON - EMPLOYMENT risks at 35 (logistic regression) B ) ENTRAPMENT IN ATYP.EMPLOYMENT at 35 ( versus being in secure employment – log . regr . )

2) DEMOGRAPHIC CONSEQUENCES

Q5. BIRTH OF FIRST CHILD

Table 5 : Event History Analysis, piecewise constant

Q6. LONG TERM SCARRING EFFECT ?

Table 6 : Family situation at a ge 35 2 - A: living in a couple (logistic regression) 2 - B: being parents (logistic regression)

Results

Some key results…

the Outsiders: who are they?

EU-SILC LONG 04-06

AMEs on the probability of having a fixed term contract, controlling for sex, age, isced, and change of job between

t-1 and t, being unemployed or inactive during the previous year.

Age effect: being young 15-24 versus over55 having a fixed term contract (FTC) versus Permanent employment

-10%

0%

10%

20%

30%

40%

50%

60%

HU UK IE IT NL SE FR DE

DE 25-34: 8,2% (6-10.5)

Gender effect: being women vs men having a fixed term contract (FTC) versus Permanent employment,

-1%

0%

1%

2%

3%

4%

5%

6%

HU UK NL SE DE FR IT IE

EU-SILC LONG 04-06

AMEs on the probability of having a fixed term contract, controlling for sex, age, isced, and change of job between

t-1 and t, being unemployed or inactive during the previous year.

the Outsiders: who are they?

the Outsiders: who are they?

EU-SILC LONG 04-06

AMEs on the probability of having a fixed term contract, controlling for sex, age, isced, and change of job between

t-1 and t, being unemployed or inactive during the previous year.

Education effect: being Isced 0-2 versus Isced 3+ having a fixed term contract (FTC) versus Permanent employment,

-4%

-2%

0%

2%

4%

6%

8%

10%

UK SE NL IT IE HU FR DE

ECHP 1995-2001 DK

GB (exit rates: males 80%, females 70%)

DE (exit rates: males 60%, females 50%)

IT (exit rates: males > 40%, females < 40%)

12

34

56

1 2 3 4 5 6order

12

34

56

1 2 3 4 5 6order

12

34

56

1 2 3 4 5 6order

12

34

56

1 2 3 4 5 6order

1) permanent

2) FTC

6) non-employment

Labour Market Trajectories: EXIT from FTC

Micro consequences of atypical employment: entrapment? Citizenship-based welfare systems

Beveridge

Low EPL => “universalistic” LMflex

Wage inequality LM adjustment

Vocational

ETS =>

DK, SWE:

Low/NO trap; √ No gender segregation; √

No/Limited cohort divide; √ High exit rates; √ Unskilled at risk√

UK:

No/LOW trap; √ mid gender segregation; √

Cohort divide; √ High exit rates; √

Generalistic

ETS =>

Skill divide

DE:

Some trap; √

gender segregation√

Limited cohort divide; √

Mid exit rates√

Unskilled at risk √

IT, HU / FR

High entrapment; √

High cohort divide; √

Youngsters mainly at risk; √

Gender segregation; √

Low exit rates √

Age divide

High EPL => Partial & Targeted LMflex

Job insecurity LM adjustment

(INS/OUTS)

Work-based (labouristic) welfare systems

NL

Results from country studies: Labour Market

Results: DEMO – Fertility

Consequences of atypical employment on transition to first maternity:

• UK: very weak, negative effect…

• De: some negative effect of unemployment, no effect of atypical

employment,

• Swe: no effect,

• Fr: no effect, (but: strong class effect!)

• Italy, Hu, (Spain): strong, negative, effect of atypical employment!

More in detail:

0.0

00.2

50.5

00.7

51.0

0

0 50 100 150 200 250analysis time (month)

typical empl atypical

seasonal/irregular never entered

0.0

00.2

50.5

00.7

51.0

0

0 50 100 150 200 250analysis time (month)

typical empl atypical

seasonal/irregular never entered

0.0

00.2

50.5

00.7

51.0

0

0 50 100 150 200 250analsis time (month)

typical empl atypical

seasonal/irregular never entered

ALL WOMEN Birth cohort 1951-65 Birth cohort 1966-80

Entry first childbirth by first job contract. Kaplan-Meier survival estimates.

Women (15-45), ITALY

Atypically employed

Paolo Barbieri and Rossella Bozzon, The family and fertility consequences of labor market deregulation ‘at the margins’ in Italy and Spain, W.P. FamIne, 2014

Transition to First Childbirth: Predicted Probabilities per Age ITALY 0

.00

2.0

04

.00

6.0

08

.01

pr(

ha

zard

)

10 20 30 40 50age

permanent self-empl

atypical other

out of work

Atypical Job

Permanent Job

Not Working

Fertility Decisions

Paolo Barbieri and Rossella Bozzon, The family and fertility consequences of labor market deregulation ‘at the margins’ in Italy and Spain, W.P. FamIne, 2014

Model 3 in table 1 - Appendix

0

.00

5.0

1.0

15

.02

Pr(

haza

rd)

10 20 30 40 50age

permanent self-empl

atypical other

out of work

Transition to First Childbirth: Predicted Probabilities per Age SPAIN

Atypical Job

Permanent Job

Not Working

Fertility Decisions

Paolo Barbieri and Rossella Bozzon, The family and fertility consequences of labor market deregulation ‘at the margins’ in Italy and Spain, W.P. FamIne, 2014

Model 3 in table 1 - Appendix

0

10

20

30

40

50

60

70

atypical career other workers never worked atypical career other workers never worked

have a partner and a child have a partner

Italy

Spain

+

only

Fertility Decisions The situation at 30 BiProbit pred. values (Partner&Child)

Paolo Barbieri and Rossella Bozzon, The family and fertility consequences of labor market deregulation ‘at the margins’ in Italy and Spain, W.P. FamIne, 2014

Model 3 in table 1 - Appendix

ATYPWR project Conclusions - the Macro Context

While welfare and market – as regulatory principles – seem to realise low levels of Insider-Outsider division, low levels of LM segmentation, and seem to bring very low consequences on individuals’ LM careers and on women’s fertility decisions…

The corporatist assets turn to be the most segmenting.

Nonetheless...

Within the “insider-outsider” countries (Italy, Hungary, Germany, and partially France) there is quite heterogeneity, suggesting the existence of a much deeper rupture of the “corporatists” cluster than many of the macro theories would suggest.

A possible line of interpretation, therefore, would stress the role of the (pre-existing) institutional arrangements (welfare state, educational system, apprenticeship and vocational educ. but also family types) which – in Germany (but also in France, as regard exits from secondary LM and maternity chances) - succeeded in differentiating within the outsiders and therefore in reducing the distance between the insiders and the upper segments of the outsiders.

As regard the demographic side of the ins-outs scenario, the German “WS & family mix” seems to cope better with the new societal uncertainties than the classical Italian (and Spanish) “mommy model”. Mediterranean countries are the very losers of the (pre-crisis!) LM deregulation process.

What happened with the crisis? Flexible forms of employment and LTU risks

FOCUS:

• THE EVOLUTION OF UNEMPLOYMENT and LTU RISKS (in 15 European

Countries) AND THE ROLE OF INSTITUTIONAL FACTORS IN MANAGING

SUCH RISKS

• MORE IN DETAILS:

• WE WANT TO TEST THE IMPACT ON (levels, composition, stickiness of)

UNEMPLOYMENT & LTU OF TWO ALTERNATIVE STRATEGIES OF

INTERNAL VS EXTERNAL FLEXIBILITY, DURING the 2008-2011 CRISIS

• AS WELL AS THE IMPACT ON UNEMPLOYMENT & LTU OF DIFFERENT

LABOR MARKET & PRODUCT MARKET ADJUSTMENTS POLICIES (EPL,

PMR, A/PLMPs INDEXES), IN A LONGER TERM PERSPECTIVE (2003-2011),

ACROSS EU COUNTRIES

European Central Bank, 2012

A snapshot of the economic crisis

Unemployment Trend, EU 12 Unemployment Dispersion: SD of EU12

unemployment rates

European Central Bank, 2012

Employment breakdown by educational level

Employment breakdown by age group

Employment breakdown by professional status and type of contract

Descriptives: Occupational impact

The «loosers»:

- Less educated

- Temporary employment

Crisis? Data and Methods

The European Labour Force Survey (Eu-LFS in a pseudo-panel form) covering the

entire time span between 2003 and 2011 and coping with the need for a unified, cross

country and longitudinal dataset of European countries with a quite extended

observational window

Countries

-DK, FI, NO, SE (clustered as “Northern Countries”)

-AT, BE, DE, NL (clustered as “Continental Countries”)

-GR, IT, FR, PT, ES (clustered as “Southern Countries”)

-UK, IE (clustered as “Liberal Countries”)

-We propose the construction of an

unbalanced pseudo-panel using 2003-2011

(LTU RISKS) EU-LFS cross-sections,

collapsing individuals sharing relevant time

invariant characteristics into “-individuals”

univocally identified by each possible

combination of such variables

-Grouping homogeneous individuals into

“meta-individuals” we are able to “follow”

them along the entire time span (up to 8

waves), and to treating the averages within

these “meta-individuals” as conventional

longitudinal observations.

(The mean number of individuals within each

meta-individual exceeds 390 per year)

Crisis? Data and Methods

Main advantages of this approach:

-The possibility to transform a cross-sectional dataset into a longitudinal

(new) one, thus reconstructing retrospective information and including dynamics

(lagged values) in regression models at the “meta-individual” level

- The implementation of proper panel estimations and the provision of a reliable

within variation of dependent and independent variables along time (within

meta-individual and within country) effects of macro variables estimations based

on variation over time and not across countries

Sample

Males and Females aged between 25 and 60; Excluded those out of LM.

-12 millions individuals, aggregated in a final estimation sample of about 32 thousand of observations

(i.e. about 4200 meta-individuals followed on average for 8 points in time)

Regression Models

We rely on :

Dynamic RE models: (RE models augmented with y at t-1 both to reduce UH role at the

micro level and look to inertia in the outcome variable. Country dummy included)

(Quantile) FE regressions: (time invariant unobserved heterogeneity is accounted for)

where the analysis is not meant to provide a description of employment determinants, but to

approximate a causal interpretation of the dynamics of unemployment trends in response to

the changing levels of “internal / external flexibility”, delta in national PLMP-ALMP

expenditure, PMR Index

Dependent variables: 1) Unemployment risks at meta-individual level

2) Long-Term Unemployment risks at meta-individual level

(Both continuous outcomes ranging between 0-1)

Data and Methods

Micro (meta-individual) level independent variables:

Sex

Young (< 30 years)

Education level (compulsory, secondary, tertiary)

LM status at t-1 (unemployment in previous year; metaindividual shares of FTC)

(Country, year of leaving ETS, sex and education used in the definition of the unit of analysis)

Macro level independent variables:

Time trend

Crisis dummy variable (2008-2011)

Time varying macro (national specific) variables:

EPL(gap)

Active/Passive LMPs (gap)

PMR Index

-Variables with no cross-section within individual variation: interpretation of the coefficient will be identical to that one

given in a common micro level panel setting

-Variables with cross-section within variation: the “meta-individual” variable computed as mean/rate of individuals

values; the regression coefficients will represent the impact on response variable, given a unit change in meta-individual

exposure to x

Results: diffs in risks of LTU – pseudopanel descriptives

Deltas in LTU risks for metaindividuals at relevant

points (p10/p90) of the unemployment distribution,

before and after crisis

Widening LTU risks in concomitance with economic

crisis, more pronounced in SC and LC :

«Metaindividual» time constant variables: Country of residence, Gender. Year of leaving education, Educational level (3 dummy variables)

Results: distributional effects of LTU – educational groups

Dynamic RE models on the probability of LTU: 3way Interaction Crisis-Country-Cluster-Education

Results: distributional effects of LTU - Gender and Age

Dynamic RE models on the probability of LTU

3 way Interaction Crisis-Country-Cluster-Gender dummy

Dynamic RE models on the probability of LTU

3 way Interaction Crisis-Country-Cluster-Youth dummy

Gender effect on LTU

Youth (<=30) effect on LTU

Results: Adjustment levers – previous FTC effects on LTU risks

(is there a scarring effect of LM segmentation? S.C.!!)

Adjustment levers models: Panel FE models on the probability of long term unemployment.

Interaction: metaindividual’s FTC exposure at t-1 * Crisis dummy variable

Provisional Conclusions

- As regard the consequences of the 2008-2011 crisis, on a macro level, significant

differences arise between the four countries clusters, with moderate effects of the crisis in

Continental and Northern countries, while major concerns apply to Southern and Liberal

clusters that experienced a stronger employment retrenchment (even if a different pace of

recovery can be expected)

- Moreover, micro level consequences in terms of polarization of unemployment risks

follow the same pattern, with Southern countries and Liberal countries displaying the higher

impact on young and/or less qualified workforce, and with Southern countries experiencing

a further increase in LTU rates

- The ‘scarring effect’ of the interaction between having being previously employed in a

secondary LM (FTC) and the crisis, is particularly significant and relevant in Southern

Countries: ITALY, SPAIN, FRANCE, GREECE, PORTUGAL: the losers of the EU process

of LM flexibilization “at the margins”…

Thank you!