Download fulltext
a362

European Sociological Review 2020 Vol 36 No 2 317–332 doi: 10 1093/esr/jcz052 Advance Access Publication Date: 12 October 2019 Original Article Solidarity between Generations in Extended Families: Old-Age Income as a Way Out of Child Poverty? Gerlinde Verbist1 * Ron Diris2 and Frank Vandenbroucke3 1Department Sociology Centre for Social Policy Herman Deleeck University Antwerp 2000 Belgium 2Department Economics Maastricht 6211 LK the Netherlands 3University Amsterdam 1000 GG *Corresponding author Email: gerlinde verbist@uantwerpen be Submitted July 2018; revised 2019; accepted September Abstract We analyse intergenerational solidarity within multigenerational households (MGHs) assess how formation these is related to poverty across countries Our aim this type household coping strategy with respect financial distress families children Using data from EU Statistics on Living Conditions (EU-SILC) we examine three specific empirical questions regards complex form notably (i) identify what extent co-residence MGHs financially beneficial young and/or old generation; (ii) income brought into by generation impacts child (iii) test sensitive impact hypotheses about way resources are shared define those where generations cohabit The results indicate that MGH operates mainly older younger Although not designed purpose pensions alleviate most prevalent Introduction In article look at sharing affects pov- erty may part house- holds Pensions key component but their role an anti-poverty device has hardly been studied Europe let alone international comparative per- spective Exploiting unique features database soli- darity 32 countries: measure contributed hypothe- ses Southern especially Eastern relatively common they far less Western Northern Different factors can explain such VC Author(s) Published Oxford Press All rights reserved For permissions please e-mail: journals permissions@oup com ranging individual preferences ex- ternal socio-economic or cultural context post- Communist combination difficult transition market economies hardship entailed many together still rela- tively low level development welfare states some provide plausible explanation high prevalence (Romania exemplary case social economic MGHs; see Preoteasa Vlase Tufva 2018) found both legacy Saraceno Keck (2010) have called ‘familialism default’ (i e neither publicly provided alternatives nor support family care) more recently fi- nancial crisis which halted long-term socio- logical downward trend [a documented Glaser et al (2018) England Wales France Greece Portugal Austria] con- trast Scandinavian highly developed long-standing ‘de-familialization’ (Saraceno 2010) obvious play previous work researchers generally focused labour supply time spent informal formal care (e g Pezzin Steinberg Schone 1999; Bertrand Mullainathan Miller 2003; Dimova Wolff 2011) However one import- ant implication left out: elderly typically bring poten- tially other substantial size As espe- cially Albertini Kohli (2012) transfers parents adult clusters (Nordic Continental Southern) do sider first analysis gains pro-child (when proportionally income) pro- (which occur through econo- mies scale) EU-SILC 2013 each scenarios (pro-child pro-elderly mutually beneficial) second relates risks logistic regression third ana- lysis contribution reduction under different cost-sharing resource-sharing A stand- ard practice study distribution as- sume fully literature however becoming increas- ingly critical assumption Such criticism hold fortiori therefore needs tested To our knowledge old-age bearing analysed studies perspective general rare dataset information intra-household allows us fill gap organized follows Background section position Data Methodology discuss meth- odology underpinning mentioned earlier Financial Multigenerational Households present final presents conclusions contributes drivers measurement putting classical Evidence genera- tions mostly refers United States seen significant increases since 1970s Studies attributed increase rises divorce rates single-parent female force participation incarceration over same period (see Baker Silverstein Putney 2008; Turney 2014) addition find among unemployment higher suggesting main motivations Cultural religion migrant status ethnicity cited another explanatory factor (Luo 2012; Pilkauskas 2012) Mutchler investigate insecurity material living grandparent-headed increased risk health three-generation no difference food housing comparison two-parent Research setting scarce inter- national An exception who share people decreasing Austria around 1981 early 2000s whereas it rising Romania [the discussed depth (2018)] They firm characterized disadvantage With well-being research South Africa indi- cates budget positive only health- clothes consumption shares (Hamoudi Thomas 2005) also cogni- tive physical (Duflo 2000) school enrolment (Case Menendez 2007) It appears shift bargaining power male head grandparent (generally grand- mother) benefits even when controlled changes any evidence indicates least extra pen- sion used benefit cannot extrapolate findings low- middle-income (such Africa) high-income coun- tries there vast tax-transfer system Ba´rcena-Martı´n Blanco-Arana Pe´rez-Moreno Salanauskaite Verbist 2013; references therein) though little known chil- dren specifically Diris Vandenbroucke (2017) estimate direct spending uncover ambiguous pension size: worsens relative thereby alleviates This performed aggregate i aims country order justice cross-country differences micro-level needed currently lacking inevitably triggers standard assumptions applies equivalence scale derive needs-adjusted metric coined ‘equivalized income’) Assigning equivalized means assumes equal shar- ing If violated misleading might drawn (Atkinson 1975; Decancq growing body equal-sharing lacks theoretical foundation sup- port Behrman Orsini Spadaro 2005; Burton Phipps Woolley Several rejected ‘classical’ model dif- ferent individuals levels often strong gender di- mension 1990; Schultz Fortin Lacroix 1997; Bennett 2013) Typically anal- yses exclusively working-age adults without likely apply number within-household re- source-sharing very limited stud- ies conducted [see Cantillon Nolan (2001) Ireland; (2007) Canada] consider will set out sections backdrop de- scriptive Therefore consists four subsections: Definitions briefly EU- SILC Measuring Direction concepts applied direction solidarity; Multivariate Analysis multivari- ate poverty; Simulation Resource-Sharing Assumption contains representative samples private (the member plus Croatia Iceland Norway Serb Republic Switzerland) defined here working age person survey than 18 64 aged 1 reveal several characteristics Supplementary Appendix Table SA1; variables controls section) First all much grandmothers grandfathers frequent large majority report suffer poor suggests need concerns major behind for- mation Another characteristic likeli- hood having substitute parent figure non-EU background Nordic Anglo-Saxon lower human capital intensity Remarkably absent reflect propen- sity operate opposite directions eld- erly could require take away activity serve facili- tators acting care- givers grandchild These SILC-based observations largely confirm (2018): ‘grandparent households’ associated (whether measured marital education level) study; grandfathers; formed head- count rate takes relevant group in- (1) Nordic: Denmark Finland Sweden; (2) Continental: Germany Luxembourg Switzerland; (3) Anglo-Saxon: Ireland Kingdom; (4) Southern: Cyprus Italy Malta Spain; (5) Eastern: Bulgaria Czech Estonia Hungary Latvia Lithuania Poland Slovak Slovenia geographical- historical certain mature rich cluster) pretend delineated clear-cut dynamics From involved elder- ly (throughout use ‘beneficial’ mean ‘fi- nancially beneficial’) ‘pro-child’ ‘pro-elder- ly’ respectively describe whose solely ‘mutual’ sim- ple framework classify distinct categories if come counterfactual formally P þ NP below line Following prac- tice Union 60 per cent median income; with: > ESMG ESCFC incomes modified OECD composition account attributes weight 0 5 individu- als 14 3 facilitate presentation clus- ter sample five groups basis geography history former communist now (candi- date) members constitute cluster] ¼ sum non-equivalized mem- ber(s) (mostly hence ‘P’ note included P) 65; parameter equivalize exclude (persons 65þ); Expression written as: incorporated scales or: – (in MGH) improves threshold (compared whereby would excluded household) Hence counter- factual expressions formalize simple insight: forma- tion ratio divided non-elderly larger generated terfactual latter words add gain similar reasoning perspec- elderly: finan- if: makes possible surpass so lives either ‘pro- child’ presume importance question cial Children Elderly People Europe: Prevalence Poverty Outcomes show predominantly poverty- alleviating effect co-residing potentially largest subsequent em- ESCFE pirical focuses its relation cated contribu- total ESCFE¼ Equation tells equiva- lence Equations simultaneously following condition satisfied: nevertheless make up SA2) being regres- sions dependent variable (Poori) whether (0) independent interest (MGHi; yes/no) old-aged (yes/no) distinction man (YOAM) woman (YOAF) (YOAFM) thus two regressions country: < Model 1: Poori b MGHi cXi Ei short label does hold; elderly’ holds; ‘pro-elderly’ 2: contradiction b1YOAFi b2YOAMi b3 YOAFMi include control (X): (yes) more; son suffers bad (yes PH010 4 [bad] [very bad]) separate woman; born); (iv) attained educa- degree; (v) household; (vi) models allow compared two-generation (observable) circumstances correct fact tend ad- verse Note behavioural causes example because grandparents act caregivers corrected tensity reflected estimated perform simulation sess were investigating compo- nents help reduce pre-post analysis: before after inclusion com- ponent Levy Lietz Sutherland 2007; 2013)? static reactions limitation method well-known (Bergh Jesuit Mahler 2010 Marx 2016) Nevertheless provides indication important lifting above expand upon assump- standardly dis- tributive analyses full-sharing probably unrealistic selection simulations sensitivity out- comes Two extreme full hand Neither realis- tic exercises upper bounds valuable (Burton kind (examples Jenkins 1991; 1995) few investigated using explicitly ask gree income-sharing Marshall 1994) self-reported approximate true degree construct additional scenario best know- ledge indeed occurs (See SA3) partial SA4) moderate centres 70 scen- ario determined weighted average reported Notes MGHs: costs due proportion- calculate change baseline cur- rent situation un- changed scenarios: ‘No unchanged’: removed changed; corresponds persons cost taken altered While realistic absence scale’: correspondingly MGH; effectively form(s) split div- ided sub-households roof not- ably consisting person(s) adapted gets value 75 (ra- ther 1) live there- fore divide ‘Part changed’: partially old- (based SILC-reported degrees) his/ her crease construction move (as baseline) depending changing dominates results: descriptive elderly; gauges Membership Explanatory Factor estimates membership Impact Poverty: Pre-Post assesses resource Figure shows grouped regions There considerable cross- variation close somewhat smaller excep- well particu- larly typical MG On amounts 19 7 slightly (20 cent) wide (Figure 2a) statis- tically (exceptions Spain Poland) statistically non-MGHs continental instance almost twice (32 (17 should noted small surprising exist particular subgroup seem es- pecially dire offers 73 versus contrast at- risk-of-poverty percentage points (18 point difference) (16% centage (15% percent- outcomes mixed; Share Notes: Within ranked Countries resp white Source: Own calculations marked 2b) reverse Especially counterparts When comparing face fers uses formulae presented Most appear MGH: half figures substantially 90 80 Slovakia 40 (dir- ect) 15–20 while consistently pattern surpris- accrue compositional effects conclude goes cross-sectional ob- servation matches trends grandchildren highlighting according 2013: (a) (b) asterisk (*) name ‘in MGH’ ‘not (at 95% confidence interval) SA5 numbers increasingly supportive rather supported Given result focus remainder enrich 2a applying control- ling represented mutual (per Number cases Pro-elderly Mutual Pro-child SE 29 20 51 31 17 22 DK 34 8 9 56 69 53 37 NO 13 6 23 63 26 55 27 FI 25 68 66 67 IS 16 39 43 35 38 24 21 DE 83 15 76 NL 81 CH 28 48 FR 11 62 57 44 BE 47 61 41 LU 77 AT 95 58 IE UK 30 87 72 CY 52 46 MT 54 92 82 IT 341 293 ES 378 322 GR 59 188 167 PT 232 209 CZ 139 108 SI 454 EE 261 182 HU 65 288 234 SK 227 178 LT 222 196 LV 380 314 HR 89 325 237 RO 281 233 BG 290 260 PL 85 79 1041 785 RS 955 754 74 Country averages unweighted marginal (ME) inde- pendent (Model originated 2; SA6) yields negative ME arrangements given characteristics) meaning linked likelihood Logistic Member Only (OAF) (OAM) Both (OMF) 127 016 (omitted) -0 125 017 044 019 023 041 043 027 061 04 123 004 078 067 057 183 116 010 050 026 059 100 160 036 073 117 020 042 066 130 051 035 185 005 033 053 134 060 152 231 054 401 088 098 068 192 112 025 137 039 107 070 045 055 062 065 031 034 142 029 195 030 136 069 022 090 121 c 153 032 040 012 037 047 089 180 015 000 Numbers bold interval put italics Statistically coefficients OAF OAM; OMF; OAM OMF statistical significance MEs implies observed explained logistics looks separately mothers jointly signifi- cant signs coming men women presence reduces overcomes Exceptions (Austria) (Norway) exhibits (conditional) correlation cover expenses poverty: trolling back- ground do; ference becomes although itself lead greater stronger receive contribute Interestingly (Malta Greece) woman-only man-only Having ‘second’ brings comparatively indicating enough compensate his comparative- compensation suggest underlying reasons already patterns lack adequate protection fac- tors driven Other choices insights determinants considerations cash carry insight next performing Pre-post Resource- Sharing compare them column ap- proach Column security tent (with incomes) 50 (without ‘no sharing’ illustrates pass non-pension Removing story ignores scale: rest suppose cover- of) own costs; including overestimates By removing [column (2)] leads drop removes [compare columns One alter- native benchmark pertains remain current recipients (income) Sharing: Full Part Equivalence Unchanged Split 71 33 86 45 49 42 Total particularly Serbia conduct continue (thus benefiting overall remains reductions (especial- countries) (0)] pointing im- portance instru- ment avoid Finally gives (a plausible) Though consequences whole reflects SA3 size- able element preventing streams partly confirmed longitudinal year strongly traditional indicators underestimate reality overestimate Conclusion la- bour (in)formal dimensions country-specific providing range MGHs—which incomes—affects established domin- antly significantly clearly rele- vance cope giving operationalization concept doing critically resource- novel constructed crucial hypothesis picture too rosy: observe sub- calls caution interpretation Unsurprisingly (notably gener- osity degrees establish conclusion policy stimulate short-term ‘coping strategy’ directly inadequate safety nets non-financial personal room extended Moreover modernizing societies presumably past future policymakers implementing reform; ‘full resources’ tangible is—for good reasons—rationalized pension- heavy must parallel develop- systems terms services optimistic dimin- ish urgency conclusion: severe definition misclassification just limit Eyeballing EUROSTAT months worked during reference theoretically has/have category ad hoc module ‘Intra-household respondents answer Tables available ESR online Acknowledgements grateful constructive input an- onymous referees Funding was Belgian Science BELSPO (Contract BR/165/A4/CIRCLE_JPIMYBL) Joint Programming Initiative ‘More Years Better Lives’ (Part Horizon Grant Agreement 643850) References M generational conflict family: transfer regimes 1–13 Atkinson B (1975) Inequality Oxford: Clarendon L J hard- ship Journal Marriage Family 947–962 N (2008) Grandparents raising States: forms stagnant policies Societal & 53–69 E S tries: pro-poor targeting targeting? 739–758 R (2003) Intrahousehold Rosenzweig Stark O (Eds ) Handbook Population 1A North Holland Amsterdam: Elsevier pp 125–187 F (2013) Researching distribution: overview developments debates methodological chal- lenges 582–597 Bergh (2005) problem state research: redistribution? 345–357 D Public families: World Bank Economic 27–50 reconsidered Micklewright Reexamined 103–125 households: measuring non-monetary Feminist 5–23 Case Does money empower elderly? agincourt demographic surveil- lance area Health 157–164 K (2014) evolution Union: Reconciling Work Reduction How Successful Are Welfare States? 60–93 -C (2011) Do enhance maternal labor supply? 911–933 G taxes pro-poorness orientation Socio-Economic 745–775 Duflo (2000) American 393–398 (1997) unitary col- lective 933–955 Trends grandchild(ren) selected Ageing 237–250 Hamoudi Pension grandchildren: new California Center 043-05 (1991) agenda action 457–483 V Comparing government redistribution second-order Quarterly 91 1390–1404 H C Swapping poli- cies: alternative tax-benefit strategies 625–647 Luo Y grandchil- dren: population-based continuity Issues 1143–1167 I (2016) optimal redistribu- Forces 1–24 Resources Household: Multi-Country Microsimulation Determinants ‘Strategic Weight’ Differentials Their Distributional EUROMOD Working Paper Series EM3/05 (1999) Intergenerational giving: approach Human 475–503 (1995) impli- cations Canada Canadian 177–204 Three-generation ferences structure birth 931–943 reaction crisis: alteration precarious Romanian Societies 111–130 Is neighbour’s grass greener? New 315–331 W Can Europe? 675–696 T (1990) Testing neoclassical fertility 599–634 Women Fiscal 1–22 Intra-household allocation: ferential 635–664 mass incarceration: implications children’s contact 93 299–327 (1994) inequality Wealth 414–432 senior researcher Her interests evaluations microsimulation modelling performance vulnerable migrants) modern post-doctoral School Business His lie economics cog- nitive non-cognitive skills educational tracking Professor He teaches he chair ‘Herman Deleeck’