Individual employment, household employment and risk of poverty in the EU. A decomposition analysis

Download fulltext

2013 edition KS-RA-09-001-EN-C M e t h o d l g i s a n Working papers ISSN 1977-0375 Individual employment household and risk of poverty in the EU A decomposition analysis Europe Direct is service to help you find answers your questions about European Union Freephone number (*): 00 800 6 7 8 9 10 11 (*) Certain mobile telephone operators do not allow access numbers or these calls may be billed More information on available Internet (http://europa eu) Cataloguing data can found at end this publication Luxembourg: Publications Office ISBN 978-92-79-29045-9 doi:10 2785/41846 Cat No KS-RA-13-014-EN-N Theme: Populations social conditions Collection: Methodologies & © Reproduction authorised provided source acknowledged 3 Eurostat Statistical (EU) Its mission leading provider high quality statistics To that it gathers analyses from National Institutes (NSIs) across provides comparable harmonised for use definition implementation policies statistical products services are also great value Europe’s business community professional organisations academics librarians NGOs media citizens In field income exclusion living Statistics Income Living Conditions (EU-SILC) main level Over last years important progress has been achieved EU-SILC as result coordinated work NSIs June 2010 Council adopted inclusion target part 2020 Strategy: lift least 20 million people by monitor towards ‘Employment Social Policy Health Consumer Affairs’ (EPSCO) Ministers agreed an ‘at exclusion’ indicator reflect multidimensional nature consists three sub-indicators: i) at-risk-of-poverty (i low income); ii) severe material deprivation; iii) very intensity households context Second Network Analysis (Net-SILC2) bringing together academic expertise international order carry out indepth methodological socio-economic develop common production tools whole System (ESS) well ensure overall scientific organisation third fourth conferences The current working paper one outputs Net-SILC2 It was presented conference (Vienna December 2012) which jointly organised hosted Austria should stressed does any way represent views Commission This independent research authors have contributed strictly personal capacity representatives Government official body Thus they free express their own take full responsibility both judgments made past policy recommendations future document Eurostat’s collection technical publications experts particular These downloadable charge PDF format website: http://epp eurostat ec europa eu/portal/page/portal/income_social_inclusion_living_conditions/publication s/methodologies_and_working_papers databases address tables with most frequently used requested short- long-term indicators 5 TABLE OF CONTENTS 1 Introduction 2 distribution jobs over Trends individual Concept polarization 13 14 4 Has become more unequal time? 17 Household joblessness work-intensity 21 Alternative definitions Stratification Individuals Jobless Work-poor Households 23 Relation between changes labour markets risks 27 Relationship rates Integrated market trends 28 Decomposition basis workintensity 33 Conclusions 39 References 41 Appendices 43 Appendix 1: ‘Conditional’ 2: Convergence EU? 47 3: Indicators 50 4: Probability 59 INDIVIDUAL EMPLOYMENT HOUSEHOLD AND RISK POVERTY IN THE (Vincent CORLUY Frank VANDENBROUCKE(1)) Abstract: chapter explores missing links success failure focus individuals age cohort empirical relying Labour Force Survey (EU LFS) SILC) proceeds two steps first step considers thus establishing link Following Gregg Scutella Wadsworth (2008 2010) ‘polarization index’ created measure size Actual decomposed (i) due changing structures (ii) second matches levels aggregation Therefore we decompose jobless other (nonjobless) households; (iii) (iv) proposed technique yield interesting insights into trajectories welfare states followed ten (1) Vincent Corluy affiliated University Antwerp (Belgium) Vandenbroucke Leuven We thank Paul De Beer Bea Cantillon colleagues Herman Deleeck Centre Brian Nolan Anne-Catherine Guio Tony Atkinson participants seminar KU UvA precious comments usual disclaimers apply supported funded bears no conclusions solely those Email correspondence: vincent corluy@ua ac Is best recipe against age? At live participation significantly diminishes financial However what seems evident less country Prior crisis Lisbon Strategy could regarded qualified if assumes there causal relationships Agenda growing On hand though largely failed deliver its ambitious promise concerning Notwithstanding generally higher many Member States encountered standstill record observe general conversion anti-poverty Hence understand (or failure) explore apparatus side equation our share 20-to- Since determined belongs relation must all analyzed will establish measures Our time frame SILC 2005 2008(1 ) short linked limitations but per se want study trajectory states( during ‘good economic years’ inquiry verify empirically explanations disappointing era put forward Vleminckx (2011) wit outcome partly attributable reduce work-poor despite increasing configuration defined terms difference actual hypothetical assuming distributed randomly benchmark ‘random jobs’ normative meaning message read follows understanding: extent positive avoidable signals suboptimal situation state Not only (skewness the) interest even composition (potentially) pay additional attention evolution longer integrates (the ( refer year prior survey spans 2004 2007 (except Ireland United Kingdom) ILO-based refers realities 2008 observed immediately before whilst ‘work-poor’ (see Section 4) 12-month period summarize complex construal label ‘2004/5-2007/8’ Currently 31 countries involved process Romania Bulgaria Malta were yet excluded trend user database offers except France including non-EU members Iceland Norway again included UDB 2009 wave estimating employment; poverty) single atrisk-of-poverty (non-jobless) principle method would assess impact ceteris paribus at-risk-of practice make such integrated hard draw tentative uncovers puzzling combination convergence disparity within Polarization sizes constitute structural background features states; differences spending explain performance regard reduction 0 describe (mathematical) timespan 1995-2008 based EU-LFS using ILO concept introduce complementary conception compare different dividing lines stratification workpoor introducing estimates First whether upward determining factor evolutions looking depth retrospective section According employed hour week member bracket 20-59 so As cut ‘jobless rate’ ‘household joblessness’ (3 add distinguishing ‘work-rich’ applying measurement framework old (excluding Scandinavian Germany) LFS 1995 diminished substantially sample average decline percentage points improvements example 16 12 Spain respectively decreased much (4 Simultaneously point increase ‘full households’ where everyone work) larger than rate ‘mixed some declined Figure Changes (ILO concept) EU11 1995- exclude students when count who (to classify ‘jobless’ ‘not jobless’) define population calculate considered students’ 18 24 status ‘inactive’ picture growth shares understanding dynamics shifts stress here -20 -15 -10 -05 05 15 25 30 ES IE NL IT EL UK BE PT FR AT LU change HH full-employment non-employment Those explainable pure ‘mathematical’ effect reflecting pooling illustrate Spanish case shows students) ‘mixed’ while increased dotted show how consisted workingage adults households: given rise decrease household’ spectacular somewhat quarter half (thus making employment’ median situation) essence mathematical corollary substantial relatively small (measured points) ‘expected’ unemployment gap (6 expected 2-adult (11 structure ‘polarization’ provide fact under examination lower expect (2-working age) rather exceptional Specific consistent range depending respect diversity prevails Distribution mixed 40 60 70 80 90 100 1996 1997 1998 1999 2000 2001 2002 2003 2006 Share + total hh = jl Although mean carries societal modernizing society mitigating (risk nonemployment) becomes progressively (percentage point) corresponds (in ‘probabilistic’ sense) (5 crude distinction allows construction index Later 0) integrate (2008) propose counterfactual evaluate Like Lorenz curve predicted occur specific So fewer) random national Formally with: Obviously smaller increases lead things being equal probability having no-one Ceteris decreases distinguished ‘predicted function emphasize expression us consider – ideal limited job opportunities ‘positive polarization’ might seen kind ‘Matthew effect’: concentration advantage (say partner someone already employed) (compared partners jobless); ‘negative appreciated form solidarity fair scarce argument illustrated simple hypothesis working-age For implies wp n2 ratio ‘individual wp/n marginal (dwp/dn 2n) interpreted shares); elasticity calculated dn dwp / always reasoning etc typically suggest either maximally ‘randomly serve comes cost: need transfers If cost sense equilibrium (X) (Y) then identical zero appear diagonal Countries above encounter negative distance reflects magnitude cardinal start Southern (most saliently Spain) Luxembourg had Negative theories gender division non-work (Danziger Katz 1996) added worker (Cullen Gruber 2000) All exhibited displaying strong Italy Greece remained throughout entire approached became grew Belgium display highest evenly Why index? diverges certain subgroups (defined size) divergences type referred ‘within-household There shift without subgroup degree itself; ‘between-household Combining insight earlier assertions determinants rates’ four terms: affect rate; within-household between-household Such shift-share Table following (from 2008): ∑ [ ] (contribution n) structure) ‘between polarization’) )[ ‘within k adults) K maximal πk wpk actually term up ‘predicted’ wpe fallen rising each (column 3) exert pressure influence strongly declining contribution workless component Most divergence stems increasingly skewed Moreover types Only equally notwithstanding suffering priori plausible assume instance activation lone mothers inactivity traps tax- benefit systems which: (unconditional) ∆ nonemployment -6 -8 -9 -5 -4 -3 -0 -2 -1 Note: column 5); Total 4); 7) overview results shorter Between Romania) new (Bulgaria Estonia Poland Slovakia) Latvia Slovenia demographic reduced decreasing divergent impact: boosted Kingdom Cyprus); Germany lesser Cyprus Netherlands Combinations components offer five emerging clusters 2000-2008 shown EU27 (exc SE FI DK MT) BG 48 98 55 EE 51 62 09 89 PL 56 34 49 SK 71 38 29 CZ 67 58 57 99 37 46 01 66 96 35 19 54 LV 82 07 94 76 77 SI 61 06 64 HU 93 97 04 86 CY 08 22 44 79 42 83 03 72 LT 73 45 36 RO DE 63 92 examined Anglo-Saxon Continental EU15 excluding observes pattern beta-convergence catch-up sigma-convergence dispersion values particularly large By characterizing beginning close exception moving group) (7 Beta-convergence identified correlation 81 initial 1995-2008; standard deviation quite sensitive outliers unlike Omission reduces 35; combined region origin education explanatory Levels (avg EU11) excludes calculations because estimation restricts consideration minus Sweden Finland Denmark Malta) beta (albeit robustly) group eleven availability stretches back (8 real Members: review (with 75) 75 25) onwards movement outspoken: (standard 88) 50) driven mainly female characterized 72); starting position still extended families robust eliminating P signalling EU23 appendix elaborate elimination sustainability (1995) (2000) choice dictated primarily appears useful cut-off describing accelerated after 2000; marked deceleration takes account timing uniform apart pace women entered explanation approach gaining underlying ‘conditional counterfactuals’ construct variety vary educational One ‘unconditional only) various indices et al Subsequently (as percentage) absolute (again unconditional explained combinations factors Applying predominantly (9 applies regression techniques male significant attainment seem findings fundamental follow clear whereas others surprisingly prima facie displays neither nor proportion post-secondary (ISCED 5-6) increases); secondary 0-2) mildly Other ‘increased homogamy’ (increased matching couples partners) since homogamy sees reality formed 50% (for every minimum 73% explained) 109% 97%) 1995-2008) cases: (2000-2008 61%) (1995-2008 62%; 67%) (2000- 146%) 57%; 64%) 128%; 223%) 82%) 70%) 59%) 104%; 106%) Portugal 51%) previous ‘workrich’ it’s latter notation wp0 former (joblessness) reference exactly same 5: ‘adults’ belonging full-time (that aged 20-24 inactive) Similarly checked (10) radically survey; calculation samples alas (11) contrast months worked theory them persons reported part-time estimate full-time-equivalent computed usually hours interview comparison straightforward: matter (no activity whatsoever versus activity) timeframe applied: (income period) happened experienced irregular spells before; metric Unsurprisingly : 5% 7% Rather (pwp0 5) Two hence pwp0 twelve taking did here) non-linear countries: experience non-zero (European 2011 p 157 Chart 21) Prima includes pensions pre-pensions below 59; early-exit schemes better assistance benefits selection influences jobless) work-rich segment 10) comprised 22-year-old non-student adult depends 11) problem discussed Graaf corroborates individual’s alternative choices 20-49 reason present (12) At-risk-of-poverty non-jobless 2007/8 2008) strategy focuses ‘very work-intensity’: means 20% begins drop beyond explains down exceeds ibidem) chosen partition several reasons heterogeneity subpopulation differ purposes groups sufficient homogeneity related workrich When expand restricting notion work-intensity’ (less 20%) (a) dominant (b) heterogeneous pursue 12) (2004/5-2007/8 employment) request NO IS pwr (wi > pwp 6: Who confronted (ILO-concept) (EU2020-concept)? probit EU10 reveals distinguish sociological post-communist societies today’s ‘old’ ‘new’ similar; associated deep-rooted disadvantages born come early lives underscores challenges strategies face reach successfully >= < 7: Marginal effects (ILO) unsurprisingly mere absence reveal singles run ‘household’ peers (peers studied analysis) incur lack (13) pooled children household: cancelling (14) Whatever see disabled individuals(15) whose With age-result intuition Compared 20-29 30-54 55-59 line exit similar fine-grained 13) refine ‘within’ ‘between’ conditional shed light question 14) That analysing 15) variable captures person’s perception respondent indicates permanently or/and unfit sex (male) (30-54y) (55-59y) marital (married) (medium) (tertiary) (non-EU born) wa child (eu2020) (ilo) 8: subtle profiles (jobless household) disability Also outspoken different: singles) two-adult three-plus direction relates residents Their compared resident conclude ‘joblessness’ (wp0 ‘work poverty’ 50%) cross-country correlate positively pretransfer correlates post-transfer Contrary Cross-sectional correlations post- pre-transfer concepts Correlations arop … (ⱡ) - housheolds 32 53 78 69 imply causality significance; merely work-poverty obviously opposite sign Given assertion through surprising negatively covered Different counterintuitive (16) mitigated influenced prevailing ‘non-jobless’ weight Third among (individual other) 16) observation contradicts OECD (2001 pp 59-61) aggregate quoted inspired Dickens Ellwood (2002) Nickell (2004) OECD’s (ECHP Canada USA) (ECHP) displayed Finally cash Higher Together elements why at-risk weakly inferred (a coefficient ∆wp0 Sections (0) described ‘upward convergence’ certainly relative (where rate) improvement now examine decomposing integration exercise written weighted (pwp) (pwr) Labelling wr wp) write: where: as: ̅̅̅̅ ̅̅̅̅̅ ̅̅̅̅̅̅̅ ̅ t=0 t=1 etcetera subcomponents contributory factors: work-rich; ii work-poor; iii de (2007) applied 1980 rightly stresses simply calculates decomposable informing mechanical caution: accounting device subcomponent intrinsically reducing deliberate stricter conditionality generosity Or push go Conversely accept scale marginally Diverging implying categories play examples invalidate caveat interpretation Using equations (2) (3) (4) possible equation: requires rely considerable obtained Graaf-Zijl circumspection called connecting 2005-2008 conceptual makes summarizes 20-to-59 figures significance estimated Poverty 2004/5-2007/8: key (analysis 2005-2008) ∆pov wp2005 pwp2005 ∆wp (pwp pwr) ∆pov(60+) 65 -7 -19 26 avg st dev 9: 2004/5-2007/8; performed ILObased awr ∆pwr awp ∆pwp (apwp apwr) ∆pwe ∆P *** ** 91 02 * 74 95 84 (actually observed) 95% (***) 90% (**) 85% ∆wpE "a" sets upswing 2004/5-2007/8 trajectories: Belgium: (based caveat(17)): neutralised significantly; German (SILC-)trajectory: offset presents case: protection UK: (helped reversal British market) (notably Slovakia Lithuania Czech Republic) led Lithuania) reinforced Importantly however elderly sometimes employment- growth-based intergenerational (18) generational favour begin hypotheses They Differences role explaining configurations at-riskof-poverty How rates? dark red bars span disparate slightly realised Republic notably checking occurred further contrary apparently (but doubts exist 17) entertain serious validity (Frick Krell SOEP crucial 18) Hungary diverging figures) short-term event say one-percentage-point structurally adds experiencing little insignificant emerge ‘structural’ unavoidable Policies evidence patterns (non-)participation Meyer Rosenbaum 2001; Grogger 2003) pointed weak (Table figure constructed quadrants (∆pov) (∆wp0 (SILC) 10: Overall identify Finnish pov (19) stark Irish Polish right upper quadrant Swedish formalized basic graphical summary Wp2005 19) tempting indicated footnote refrain drawing 11: -apwr) varies Netherlands) Estonia) Correspondingly ranges around (20) +5 Hungary) diverse ranging +1 cross-sectional highly diversified ∆pwp0 ∆pwr0 incorrect assert modest ‘main culprit’ board complicated contextualize segments incorporates 60+ (hereafter ‘the elderly’) set observations 2004/5 catching-up weakened disregards Poland); regards disregarded) simultaneously diverse: just lost gained ground although work-rich) words witnessed shift; holds albeit too changed hardly vis-à- vis cases remarkable ‘outliers’ twenty-six outlier paint greatly dwarfed confirms far 20) yields 1990s 2000s concerned represented striking counts: (an resulted downside considerably (of pwr0 vis-à-vis recorded (none EU) rich (though 90%) cohort; opposed elderly) suggests successful reaching decompositions (Sections stories visual Some readily gives greater diminishing (since hold households) Despite emerges confirmed focussed proves relevant differentiating employment: shape forces modernization feminization Nevertheless individual/household gradually southern gains declines dependency familial Gains enhanced additionally Experience prevalence ‘positive’ cannot ways attribute boom successes specifically ongoing But diminish importance policymakers attach presence possibly problematic justification ‘stand still’ coincides pursued Both evolutions:  rates; clearly inegalitarian historically (and base assessment SILC); effort emphasis (successful activation) (much generosity); (21) Economic socio-demographic condition public (including benefits) refuel reconnect sound creation necessary complementarity inclusive Brandolini Viviano E (August Extensive vs intensive margin: Changing Perspective Employment Rate Paper 31st General Conference International Association Research Wealth St Gallen Switzerland B ‘The paradox investment state: era’ Journal 21(5) 432-449 Cullen J ‘Does insurance crowd spousal labor supply?’ Labor Economics 18(3) 546-572 Danzinger L (1996) ‘A discrimination’ Behavior Organization 57-66 Dawkins R Growth Australia’ Australian Review 133-154 ‘Household deprivation Europe’ 413-431 D ‘Whither Britain US?: work’ in: Blundell Card Freeman (eds Seeking Premier League Economy Chicago: Chicago Press T Impact Earned Tax Credit Reforms Work Marriage Arrangements’ 53(4) 1063-1106 Developments Brussels: Joint Report Frick (2010) Measuring Panel Surveys Germany: Comparison SOEPpapers Multidisciplinary Data n° 265 Berlin: DIW ‘Two sides story: measuring inequality Royal Society: Series (Statistics Society) 171 857-875 ‘Reconciling Theory Population 139-167 (2003) limits EITC female-headed families’ 85(2) 394-408 (2001) ‘Welfare earned tax credit supply mothers’ Quarterly 1063-1115 S ‘Poverty worklessness Britain’ Econ 114 C1-C25 Outlook Paris: (2009) ‘Is antidote poverty?’ Sala-i-Martin ‘Regional Cohesion: Evidence Theories Regional Convergence’ 1325-1352 Protection Committee Jobs Progress evaluation dimension F ‘Disappointing trends: blame?’ 450-471 evolving Positive occurs gain (Dawkins 2002; varying Relaxing assumption allowing known Assortive mating characteristics inequalities outcomes especially tends concentrated sections assortive stronger worsened improved disadvantaged demand low-educated tertiary educated likely tend choose ‘conditional’ origin) fifteen lie behind differentiate (20 years) (at education: ISCED 1-2 3-4 variation (Flemish Brussels Walloon region) alongside Initially Then Today fewer traditional family prevalent works paid another woman remains inactive produces home % broadly flat What unexplained conditional) entirely After variations fully (negative) Before men Allowing brings disappeared Yet Because immigrants reside propensity raise conditioning (conditional) 1994 Explained CFT expalined polarisation unequivocally 59) presumably upshot AngloSaxon persistent attainment) draws analysis: sigmaconvergence falls drops) partial betaconvergence initially evolve faster (upper section) process) values) Sigma-convergence Unlike (two polarization) 29; -lower section-) twenty-three EU-27 Again deltas EU-23 EU-11 Excluding drops evolves 73) 27) 02); restricted periods look 2) (negative 33) Sweden) BETA SIGMA-convergence (1995 (=EU DE) EU9 ES) (= EU12 SE) NMS EU21 UK) std 68 sigma '95 '08 88 52 '00 BETA-convergence AROP poor EU26 (all survey) (=EEU) EU24 NO) IE) PL) EU20 wi corr (pov t0 pov) (wp pwp) (pwr discuss existing demarcations populations ranking wpα ‘workpoor’ α list six referring and/or finds literature surveys (LFS) Community subset according limiting calculating determine potential “beneficiaries” whom indicator; consist overlap necessarily completely detail seemingly minor translate analytical entail ’employment’: ILO-concept recent debates (c) (d) sources currently Definition I II ‘older’ online( 22) determines Students composed Someone he/she self-employment Computation dependent retired persons) Additionally ordinal introduced III derived Wadsworth’s 24) (nominated) head retirement (60 over) IV developed introduces see: eu/tgm/table do?tab=table&plugin=1&language=en&pcode=tsisc090 http://ec eu/eurostat/ramon/nomenclatures/index cfm?TargetUrl=DSP_GLOSSARY_NOM_DTL_VIEW&StrNom=CODED2&StrLanguageCode=EN &IntKey=16624585&RdoSearch=CONTAIN&TxtSearch=work%20intensity&CboTheme=&IntCurrentPage=1 belong declared 0-59 noone equivalent threshold V (ordinal) Indicators’ Sub-Group (ISG) (SPC) breakdown ‘at intensity’ sum (without part-time) (aged 16-64 youth 16-24) workable demarcated 16-24 VI uses delineations sole replaces binary Looking (work zero) flexible lowest learns stringent offering broader 64) (both measures) controlling employees ranks (dependent source) intensity) Minor rankings discussion Besides evaluated respects produce know re-calculate adapting EU2020 table 2011) Column website produced ourselves indicating correct (5) retain perspective equals exceptions (6) alter (hardly) Netherland obvious potentially grow narrowing (7) adapted 18-59y 15-75y notable switch recently headline targets force adjustment causes Temporary captured prominent Any remaining (8) cent Luxemburg dedicated cause concern researchers interested closely 0–59 ISG (definition V) VI) controls parttime improve comparability 0– arise Controlling confined exceed translated trade-off 1) (0 <= Thirdly exceeding Consequently differs rudimentary detailed underrepresented 8) Complementary overrepresented demarcate overrepresentation strongest fullemployment defintion (current week) (nbr year) continuous WI=0; 0