the rise of economic insecurity in the eu: concepts and ...the rise of economic insecurity in the...

LPS–LaboratoriodiPoliticheSociali

ThisworkingpaperishostedbyDAStU,DipartimentodiArchitetturaeStudiUrbani,www.dastu.polimi.it/pubblicazioni/working_papers

©Copyrightisretainedbytheauthor

The rise of economic insecurity in the EU: concepts and measures

CostanzoRanci(PolitecnicodiMilano)JasonBeckfield(HarwardUniversity)

LauraBernardi(UniversitédeLausanne)AndreaParma(PolitecnicodiMilano)

DAStUWorkingPapers

n.02/2019(LPS.04)ISSN2281-6283

DAStU Working Papers – LPS The rise of economic insecurity in the EU: concepts and measures| Ranci, Backfield, Bernardi, Parma

2

Abstract

Economicinstability,anarrayofsocialchanges,andwelfarestateretrenchmentplacethequestionofeconomicinsecurityhighonthescholarlyandpoliticalagenda.Wecontributetothesedebatesbydrawingconceptualdistinctionsbetweeninequalityandinsecurity.Inequalityconcernsthecross-sectionaldistributionofresourcesacrossindividuals,whileinsecurityconcernstheover-timeexposuretofinancialstrainsthatcanunsustainablethehouseholds’economicconditions.Themultiplicityanddynamismofinsecurityinformourdevelopmentofanewmeasureofeconomicinsecurity,usinglongitudinaldatafromtheEU-SILCdatabase.WethenuseournewmeasuretoanalyzethedistributionofinsecurityinEurope.OuranalysisshowsthatinsecurityiswidespreadacrossEurope,affectingcountrieswithdifferentinequalitylevelsandwelfarestructures.Second,itiswidespreadacrossincomegroupsandoccupationalclasses,reachingintothemiddleclasses. Key-words

Class;inequality;poverty;socialstructure;welfarestate;insecurity. Shortbioofauthors

CostanzoRanciisFullProfessorofEconomicSociologyandDirectorofSocialPolicyResearchUnitatthePolytechnicofMilan.Hehaspublishedonsocialpolicyandthewelfarestate,thethirdsectorandvoluntaryorganizations,socialexclusion,socialrisksandsocialmovements.HehasextensivelyworkedininternationalresearchprojectswithintheVIIEuropeanFrameworkProgramme.E-mailaddress:costanzo.ranci@polimi.itJasonBeckfieldisProfessorofSociologyandAssociateDirectoroftheCenterforPopulationandDevelopmentStudiesatHarvardUniversity.Hisresearchinterestsareintheareasofsocialstratification,politicaleconomy,andpopulationhealth.Currently,hisresearchinvestigatesthedeterminantsofcomparativeandhistoricalvariationinhealthinequalities,andheisdevelopinganinstitutionaltheoryofpopulationhealthdistribution.E-mailaddress:jbeckfie@wjh.harvard.eduLauraBernardiisProfessorofLifeCourseDemographyandSociologyattheFacultyofSocialandPoliticalSciencesattheUniversityofLausanne,Switzerland,andDeputyDirectoroftheSwissNationalCenterforCompetenceinResearchLIVES.Hermajorresearchareasarefertilityandfamilydemography,migration,andlifecourseandsocialinequalities.E-mailaddress:laura.bernardi@unil.chAndreaParmaiscurrentlyapost-docfellowatPolytechnicofMilan,wherehecollaborateswiththeSocial Policy Laboratory of DAStU. His recent research interests include school and residentialsegregation,economicinsecurityandlabormarketpolicies.E-mailaddress:[email protected]

LPS – Laboratorio di Politiche Sociali http://www.lps.polimi.it/

How to cite this working paper: Ranci, C., Beckfield, J., Bernardi, L., Parma, A. (2019) The rise of economic insecurity in the EU: concepts and measures, in DAStU Working Paper Series, n. 03/2019 (LPS.04).


3

Introduction In the postwar period, economic growth, welfare states, and social dialogue supported thedevelopment of strong and stable middle classes in many European societies. Currently, theweakeningofthe“Europeansocialmodel”(BonoliandNatali,2012),togetherwithderegulationoflabourmarketsandtheemergenceofnewsocialrisks,maybeerodingthefoundationsofthemiddleclassanddiffusingeconomicinsecurityfromthepooresttotheostensiblymoresecuremiddleclasses.TousethewordsofLeisiringandLeibfreid(1999),currenttransformationsmayhaveled“notonlytoapermanentexclusionofthe‘useless’membersofsocietybutalsoandatthesametimeto thecreationof the fluctuatingmass insociety,peoplewhoexperience insecurityratherthanexclusion”(p.251).ThispapercontributesnewdescriptiveandanalyticalevidencetothisdebateoverthepotentiallychangingfortunesofEuropeanmiddleclasses.WebuildonthepioneeringworkofGornickandJaentti (2014) andWhelan et al (2015b) in examining economic insecurity among Europeanmiddleclasses.Specifically,wetestthehypothesisthateconomicinsecurityisexperiencednotonly at the bottom of the income distribution, but also in themiddle classes, and across thevarietiesofcapitalismandwelfareregimes(Mauetal.,2012).Whilegrowinginequalitythreatensthe“intermediate”positioningofthemiddleclassinthesocialstructurebyincreasingtheriskofitsrelativedeprivation(duetoeitheritsrelativeimpoverishmentorupwardmobilityoflowerclasses),theriseofinsecurityisachallengetothestabilityofthis“majorityclass”(Mau,2015).Insecurity makes adverse events more unpredictable and weakens the system of socialguaranteestraditionallyattachedtothemiddleclass.Thespreadofeconomic insecurity into themiddleclasses is somewhatsurprising forEurope,which,incontrasttothedebateover“middle-classsqueeze”intheUnitedStates(Pressman,2007;Hacker, 2008; Scott andPressman, 2011; Frank, 2013), has tended to debate insteadpovertytrendsandsocialexclusion(FörsterandVleminckx,2004;CantillonandVanderbroucke,2014).Thismayreflecttheassumptionthatthestill-strongredistributioncapacityofEuropeanwelfarestates insulates the middle class from high financial risks or poverty (Dallinger, 2013). As aconsequence, in Europe there is still relatively little comparative research on middle-classinsecurityspecifically(FouargeandLayte,2005;Jenkinsetal.,2012;Kenworthy,2014;cf.Whelanetal.,2015b).Wecontributetothisemergingdebateintwoways.First,wedevelopnewmeasurestocapturethe substance of economic insecurity, and to clarify its distinctiveness frommore traditionalmeasuresofpovertyandmaterialdeprivation.Anoriginalprincipal-componentsanalysis(PCA)givesstrongempiricalsupporttoourhypothesis.Second,weconductalongitudinalanalysis,tocapturethedynamicandtransitorycharacterofeconomicinsecurity.Specifically,weusepaneldatafromtheEUStatisticsofIncomeandLivingConditions(EU-SILC)todescribethedistributionofshort-termeconomicinsecurity.WefindthatinsecurityvariesacrosscountriesinEurope,butis present across welfare regimes and varieties of capitalism. We then show that economicinsecuritylocateswelluptheclassscaleandtheincomedistribution.Thepaper is structured in the followingsections.After this introduction,Section2presentsamultidimensional, dynamic approach to the study of insecurity. Section 3 elaborates originalmeasures andmethods aimed to capture economic insecurity. Section 4 presents our results,startingwiththePCA,followedbylongitudinalheadcountsofcriticalsituationsexperiencedbyhouseholds,across-countryanalysisofeconomicinsecurity,andananalysisoftheincomeandclass distributions of insecurity. Section 5 concludes the paperwith a summary and broaderdiscussionoftheresults,includingthelimitationsofouranalysis.


4

2.BackgroundThe concept of economic insecurity has been recently proposed to capture increasinglywidespread situations characterised by significant threats to the financial sustainability ofhouseholds. IfStiglitz,SenandFitoussi (2009)noted that insecurityabout the futurematerialconditions“reflectstheexistenceofavarietyofrisks,inparticularassociatedwithunemployment,illness, and old age”, in a more recent OECD report economic insecurity was defined as “asignificantdownsideeconomicrisk–i.e.ahazardordanger–loomingintheindividuals’economicfuture,whichtheyareunabletoadequatelyinsureagainstoravoidorignore”(Osberg2015:5).Building on these general definitions, various empirical concepts have been proposed todistinguishtheobjectiveaspectsofeconomicinsecurityfromsubjectivefeelingssuchasasenseofuncertaintyorriskadversion(Mauetal.,2012).The debate has generated several conceptualizations. Some scholars have defined economicinsecurityasahighrisk/probabilityofeconomic losses(Westernetal.,2012),adverseshocks(BossertandD’Ambrosio,2013)ordownsidehazards(OsbergandSharpe,2014;Osberg2015).Insecurityhavebeenalsoassociatedtothedegreeofprotectionagainsteconomiclossesofferedbyrelevant“buffers”suchasprivatewealth(BossertandD’Ambrosio,2009)orstateprotection(Hackeretal.,2014).Oneinfluentiallineofworkstressesinsteadthedifficultyofhouseholdsincoveringstandardexpensesortopaytheirfinancialcommitments(Whelanetal.,2015a).We contribute to this debate by offering an encompassing, multidimensional, longitudinalconceptualization of economic insecurity that differentiates it frommaterial deprivation. Weconceptualizeeconomicinsecurityasaneconomicstrainthreateningthefinancialsustainabilityofthehousehold.Weunderscoretwoaspectsofthisconceptualization:a)financialunsustainabilitythatcanbecapturedonlybyconsideringapluralityofaspects;b)aneconomicstrainshort-termsituation,clearlydifferentiatedfrompermanent,chronichardshipconditions.

2.1Financialunsustainability

Sincetheonsetofthefinancialcrisis,increasingattentionhasbeenpaidtothehighprevalenceofhouseholds experiencing economic insecurity without necessarily falling into severe poverty.Scholarshavestartedtoelaboratenewmeasurementstocapturethespecificityofsuchsituations.While low income or poverty conditions were the focus of previous research, attention hasrecently shifted to sustainability and to the possible means/endsmismatches in households’consumptionbehavior.Theconsequencehasbeenontheonehandtheproliferationofdistinctmeasuresofeconomicinsecurityfocusingondifferentaspectslikeover-indebtednessoreconomicstrain,andontheotherhandthedevelopmentofadynamicassessmentsofpoverty(transitorypoverty).Over-indebtedness:Thelasttwodecadeshavewitnessedanincreaseinhouseholddebt,bothinWesternEuropeandintheUS(AngelandHeitzman,2015).Whilehouseholdindebtednesswasmainly driven by deregulation of the financial sector and proliferation of new financialinstruments through 2007, since the onset of the financial crisis indebtedness has beenincreasinglyassociatedwithaworseninginthefinancialconditionsofhouseholds(FligsteinandGoldstein,2015).Inrecentyears,scholarlyresearchhasincreasinglyinvestigatedtheroleofdebtin determining household financial vulnerability, and this interest has also inspired newtheoreticaladvances (AnderloniandVandone,2011). Inneo-classicaleconomic theorydebt isconsideredasawayofanticipatingspendingbasedonexpectationsofincreasedfutureincomereceipts; within this framework, indebtedness guarantees heightened economic welfare bysmoothing consumption over time. However, the growth of household debt, and especiallyunsecured debt (i.e. consumer credit), in recent years seems less related to consumptionsmoothingthantoprevailingfinancialdifficultiesofoverlyindebtedhouseholds(Jappellietal.,2013).


5

Economicstrain:Researchhasalsoincreasinglyacknowledgedtheeconomicstrainofhouseholdsexperiencingextremelylowconsumptionlevels,strongcompressionoftheirlivingstandards,andshortagesduetoilliquidityorhighfinancialvulnerability(WhelanandMaitre,2005).AconceptandrelatedmeasurementhavebeenrecentlyproposedbyWhelanetal.(2015b)tocapturetheeconomic vulnerability of households facing substantial financial difficulties that are excludedfromtraditionalindicesofpovertyormaterialdeprivation.Inbuildingupmeasuresoffinancialstress,Whelanetal.(ibidem)includenotonlyobjectiveover-indebtedness,butalsoitemsaimedatmeasuringabroadersubjectivenotionofunsustainablespendingbehavior.Transientpoverty:Inthefieldofincome-basedstudiesthefocushasshiftedalsofrompersistentpoverty to transient poverty.While in theUS context great attentionhasbeenpaid to incomevolatility,understoodasshort-term income fluctuationresulting inyear-to-year incomedrops(Westernetal.,2012;Hackeretal.,2014;Osberg,2015;NauandSoener2017),inEuropeincomevolatility isconsidereda lessrelevant issuethanktotherolestillplayedbyincomestabilizerssuch as thewelfare state. Instead, increasing attention has been paid to economic insecuritymeasures considering the impact of downside risks on individuals’ or households’well-being.Movingfromastatictoadynamicassessmentofpoverty,longitudinalmeasuresofpovertyhavecapturedtheincreasingnumberofhouseholdsfacingtransientpovertysituations,characterizedbyupsanddownsaroundthepovertyline(Whelanetal.,2003).Sandovaletal.(2009)considerthe number of spells below the poverty level, and find that in US poverty is becomingmoretransitorybutthattheriskofpovertyisbecomingmorewidespread.Astrongassociationbetweenpovertyriskandlowworkhouseholdintensity(CantillonandVandenbrouke,2014)corroboratesthe idea that transitory poverty has increased as consequence of higher unemployment, thespreadoftemporaryworkandprecariousemploymentinmanyEuropeancountries.

2.2Adynamicapproach

Thedistinctionbetweeninequalityandinsecurity(Westernetal.,2012)highlightstheover-timedynamicofmaterialwellbeing.Inequalitydescribesthedistributionofincomeandothermaterialresourcesacrossindividualsorhouseholds,whileinsecurityhastodowithexposuretorisksthatcanworsen livingconditions.Changes in inequalityproducevariation in thedistancebetweensocialgroups,inthediffusionofrelativepoverty.Changesininsecurityaffectthepredictabilityandpotentialharmsofstressfuleventssuchasincomeloss,unemployment,familybreakdowns,or financial breaks. Although strongly interweaved, these two concepts capture two distinctaspectsofongoingchanges inthesocialstructureofcontemporarysocieties:whiletheformerrefers to distance among individuals and social classes, the latter describes the exposure ofindividuals or social groups to financial unsustainability.We followWestern et al. (2012), inconceptualizinginequalityasmainlystatic(thoughwithsignificantinter-temporaleffects),andinsecurityasdynamic.Astimeisanessentialelementofinsecurity,cross-sectionalanalysiscannotcaptureitssubstance.First, incross-sectionanalysischronicsituationsoverlapwith temporarysituations,andthesetwoaspectscannotbereciprocallydetected.Second,acrudedistinctionbetweenpoorandnon-poor(ordeprivedandnotdeprived)isused,withnoattentiontointermediatesituationsaffectedby transitory hardship or latent poverty (Leisering and Leibfried, 1999). Finally, householdsaffectedbycontingentproblems,orvolatilityofbasicresources,cannotbeidentifiedandanalyzedassuchinstaticanalysis.Weadoptalongitudinalapproachusingpaneldatatoaddresstheseshortcomings.WediscussourmethodologicalapproachbelowinSection3,butherewewanttoclarifyourgeneralperspectiveandhowitdiffersfromextantlongitudinalresearchfocusedonhardshipconditions.Atraditionalapproachanalyzespovertyspellsoverlongperiodsoftimeandthereforedistinguishesbetweenpermanentandtransitorypovertythroughsequenceanalysis.TheclassicresearchofBaneandEllwood(1986)showedthat incomevolatilitywasveryhigh in theUS,andthat individuals in


6

permanentpovertywereaverysmallminorityofthetotalpopulation.Notonlydoesthisapproachadoptaverylongtimeframe,butalsoitiscoherentwitha“Beveridgeanperspective,”wherebyproblematic situations are identified by reference to a collectively fixed threshold (e.g., thepovertyline).Basedontheseassumptions,sequenceanalysesofpovertyhavemainlyfocusedontherelativelysmallproportionofhouseholdslivinginpermanentpoverty(Vandecasteele,2010;2011).Amore recent approach is focusedon income volatility, considering either short-term incomevariabilityorlargeincomelosseswithnoregardtotheirdirectimpactonthelivingconditionsofpeople. Itseemsimplicitthata largedrop(forexample,by25%ofthetotal income)producesinstability whatever is the previous income level of the household (Western et al., 2012).Consequentially, the income-volatility approach backgrounds the question of dynamics ofhouseholdlivingconditions.Webuildonboththeseapproachesintwoways.First,wehypothesizethateconomicinsecurityismainlyrelatedtoshort-termvariationsandisthereforeseparatedfrompermanentpovertyormaterialdeprivationtrajectories.Second,weadoptarevisedversionofthe“headcountapproach”proposedbyAlkireandFoster(2011)thatallowsthesimultaneousanalysisofmultiplevariationsoccurringinalimitednumberofspells.Economicinsecurityintegratesthesetwodimensionsasitfocusesondownwardfluctuationsinamultiplesetofaspectsthatareabletopushhouseholdsintofinancialunsustainability.

3.DataandmethodsTodevelopmultidimensional,dynamicmeasuresofeconomicinsecurityweusetheEU-SILCpaneldatabase,whichprovidesa4-yearbasedrotationalpanelforalltheEUcountrieshereconsidered.As our study focuses on temporary flows and short-term occurrence of different forms ofinsecurity,atimespanoffouryearsisadequatetostudysuchvariations,withsomelimitationsrelatedtothedifficultytocapturelong-termvariability.Thestudyconsidersdatafortheyears2007-2012 (justafter theonsetof the recent recessions).Wepooldata from threesuccessiverotations(startingin2007,2008,and2009).Wefocusontheworking-agepopulation:onlyhouseholdswhosemainearneratthestartoftheobservationperiodwasunder60yearsoldareincludedintheanalysis.Finally,weinclude8Europeancountriesasrepresentativeofdifferentvarietiesofcapitalismandwelfareregimes:DenmarkandSwedenasNordiccountries;theUKasthemostrepresentativecountry for the Anglo-Saxon regime; France1 as a Continental regime and Italy and Spain ascomponents of its Mediterranean version; Hungary and Poland as representative of central-easterncountries.

3.1Dataweightsandattrition

EU-SILCdoesnotprovidelongitudinalhouseholdweights.Therefore,weusethecross-sectionalhousehold weights provided in the EU-SILC for the first spell of each rotation to adjust fordifferencesintheprobabilityofahouseholdbeingsampledaccordingtodemographicdifferencesacrosscountries.Weadjustbytheeffectivesizeofeachcountryandweestimatenewweightstocontrolthevariabilityofthepanelcompositionoveritstimespan.Themainproblemtobeaddressedhereissampleattritionduetolossofinitialsamplemembers.Ina four-years timespanattritionrate is34%for theoverall sample,withcountries’samplesrangingfrom25%(Poland)to40%(Italy),andonlytwoexceptions:Denmark,wheretheuseofregister data grants no attrition over the time period (100% of households responding for 4 1TheFrenchSILCpanelfollowspartofthehouseholdsfor8years.Werandomlysampledonlyonerotationforsuchhouseholds.


7

years), and United Kingdom, with attrition up to 69%. However, the absolute number ofhouseholdsincludedintheBritishpanelisstillhigh2.Wecontrolledattritionbyestimating,foreachhouseholdincludedinthefirstwave,itsprobabilitytobeinthepanelforthewholedurationoftherotation.Variablesconcerninghouseholdtypology,social classstructure, the family incomeandhousehold’smainearnerageandeducation levelwere included in theregressionmodelrunat thecountry level tocalculatesuchprobabilities.Results were used to generate Inverse Probability Treatment Weights (IPTW), which wereappliedtoallhouseholdsinthepanel.IPTWsweretheninteractedwiththeoriginalcrosssectionhouseholdweightsforspell1togenerateanewweight.A secondary problem is related to over-time changes in the composition of SILC households(though limited in a short time-span). We addressed this problem using the shared weightsmethod (LatoucheandNaud,2001),which take individualsmoving in/outof households intoaccount,toadjustthenewweightscalculatedasexplainedinthepreviousparagraph.

3.2Lackofsynchronicityamongincomedataandotherinformation

EU-SILCincomequestionsrefertotheyearbeforetheyearofinterview,inallcountriesanalyzedhere except the UK. Economic theory on the impact of income variations on householdexpendituresorfinancialproblemsisstillinconclusive,rangingfromassumptionsthatshort-termincomechangesdonotgreatlyaffectexpendituresandconsumptionsmoothingisalwayspossibleforhouseholds,to“rule-ofthethumb”theoriesassumingthathouseholdcanspendonlywhattheyhavejustearned(JappelliandPistaferri,2006).Lessextremeassumptionsgenerallyconcludethatincomeshort-termchangesdonotsubstantiallyaltereithertheirconsumptionleveloravailableliquidity(MeghirandPistaferri,2011).Thus,weconcludethatthereislittleevidencethatpresent-yearincomeaffectseconomicinsecuritymorethanprevious-yearincome,andsoinallanalysesweretainallfourwavesofdata.

3.3Headcountmethod

Ouraimistocalculatetheoccurrenceandfluctuationsofeconomicinsecurity.Weadoptarevisedversionoftheheadcountapproach(AlkireandFoster2011,Alkireetal.,2014),alsoknownasspell,orduration,approach.Theclassicheadcountapproachmeasuresmultidimensionalpovertyby aggregating a class of different indices across multiple dimensions and over time. Morespecifically,itdefinesaunit(individualorhousehold)aspoorifitexperiencesscarcityofbasicresourcesandifsuchscarcityisrepeatedlyexperiencedovertime(numberofspells).Researchersexogenously fix the thresholds indicating themultiple dimensions of poverty and time (each“dimensionalcut-point”andthe“timecut-point”).Inspiredbytheclassicheadcountapproach,weproceededasfollows:

1. WefirstrunaPCAonmultipleitemstoidentifythedimensionsofinsecuritytobeusedintheanalysis.

2. We calculated an additive index for each dimension of insecurity found in the PCA.Building on Nolan andWhelan (2009), the value of each item has been weighted bycountryspecificprevalenceweights3tocontroldifferencesbycountryintheirdiffusion.

2Numberofhouseholdsremaininginthepanelafter4spells:Denmark1859,Spain5352,France5490,Hungary3872,Italy6742,Poland5721,Sweden2603,UnitedKingdom2368.3Prevalenceweightswerecalculatedtocontrolbythevariabilityofthedistributionofitemswithineachcountry,sothat items less prevalent in each country contribute less to the total counting. The formula used is: [1-(share ofhouseholdsreportingitemi/totalshareofhouseholdsreportingitemsincludedintheindex)].


8

3. Wethenfixeddifferentdimensionalcut-offsforeachindexbasedonthenumberofitemsincludedintheindex4.

4. Foreachdimension,wecountedthenumberofspellsinwhichthehouseholds’scoreishigherthanthecut-offpoint5.Theprocesscanbesummarizedasfollow:

! d = 1ifi > k*

+,-

𝑤ℎ𝑒𝑟𝑒𝑡 = 𝑦𝑒𝑎𝑟; 𝑖 = 𝑣𝑎𝑙𝑢𝑒𝑜𝑓𝑡ℎ𝑒𝑑𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛𝑎𝑙𝑖𝑛𝑑𝑒𝑥; 𝑘 = 𝑑𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛𝑎𝑙𝑐𝑢𝑡𝑜𝑓𝑓𝑝𝑜𝑖𝑛𝑡Thiswaywefollowedthedevelopmentovertime(4-yearperiod)ofthethreeindexesforeachhousehold.

5. Finally, following Bossert et al. (2012)we calculated a continuityweight to take intoaccount the aggravating condition of persistence in insecurity and/or deprivation forlongertime.Weweightedourheadcountstwiceeverytimethehouseholdwasbelowoneof the cut off points consecutively rather than in separated spells. As a result a newvariableof“timecontinuity”(rangingfrom0to7)wascreatedandusedinthefollowinganalysis.

4.Results4.1Thedimensionsandpeculiarityofeconomicinsecurity

To test the hypothesis that economic insecurity is a multidimensional phenomenon clearlydistinct frommaterialdeprivationorextremepoverty,weconductedanexploratoryPrincipalComponentsAnalysis6 by considering a large rangeofmultiplequalitativenot-income relateditems. FollowingWhelan et al. (2015a), we consider both subjective and objective items. AllavailableitemsincludedintheEU-SILCsurvey(withafewexceptions)7informingonthefinancialconditionsofhouseholdswereusedinthisanalysis.Inthefollowingstepweaddedalongitudinalmeasureofshorttermpovertytocapturealsothisdimensionofeconomicinsecurity.Inpreviousresearchmaterialdeprivationhasbeenidentifiedasamultidimensionalmeasureableto capture the households’ inability, due to financial constraints, to obtain a wide range ofconsumptions, facilities, and social conditions “generally regarded as acceptable in thecommunity” (Nolan and Whelan, 1999, 2011). In this analysis qualitative items have beencombinedtogetheraccordingtodifferentcriteria,butalwayswiththeaimofdefiningauniquemeasure of deprivation, on the assumption that a plurality of dimensions can be combinedtogetherinonesingleindex.ThesameapproachhasbeenadoptedbytheEuropeanUnion,whichhasincludedamaterialdeprivationindicatorintheofficialsetofstatisticalmeasuressupportingtheEurope2020strategy(Guioetal.,2009;Eurostat,2016).

4WeidentifiedthreeindexesbasedonPCAresults(seesection4forfurtherdetails).Thecut-offpointfor“economicstrain”(whichadds6items)wasestablishedat3,for“multipledeprivation”(5items)at2,for“over-indebtedness”(whichadds4items)at1.5Contrary toAlkireandFoster’sapproach,wedidnotdefinea secondmulti-dimensional cut-off tocreatea singleinsecuritymeasureastheaimof thearticle is toanalyzethe interconnectionofdifferentdimensions.Forthesamereason,wedidnotsetapre-determineddurationcut-off.Thisallowsshowinghowdimensionaltransitoryandchronicsituationsaredistributed.Finally,whiletheoriginalheadcountapproachintroducesadjustmentsbasedontheintensityofpoverty(AlkireandFoster,2011),ourheadcount-continuitymethodconsidersintensityasanadditionalaspectthatmaybeanalyzedseparatelyfromtheduration-continuityanalysis.6ThePCAwas runon spell1by considering the full sample.Weuseda tetrachoric correlation,which is apropertechniqueforestimatingcorrelationbetweendichotomousvariables.Varimaxrotationwasapplied.Factorsshowingeigenvalue>1wereretained,producinga3-factorsolution.AstandardPCAshowedconsistentresults. 7Weincludedalltheavailableitemswiththeonlyexclusionofthoserelatedtothehousingconditionsofhouseholds(notfrequentlyconsideredinsuchanalysis)andaspecificitem(telephonenotaffordable)thatdoesnotappearinsomeofthecountrieshereanalyzed.


9

OurexploratoryPCAresults(seeTable1),however,supporttheideathateconomicinsecurityincludes different dimensions that are differentiated from other aspects captured by thedeprivationindex.Given14dummyvariables,threestatisticallyindependentfactors(weusedaVarimaxrotation)explain65.6%ofthetotalvariance.ThePCAidentifiedtwolatentvariablesthatcapturemostofthemultiplicityofeconomicinsecurity,andreflectthedimensionsofinsecurityalready discussed on a theoretical basis in section 2.1. The third factor describes a moretraditionaldimensionofdeprivation.BelowwedescribethePCAsolution.Financial strain is our label for the first factor, which loadsmost heavily on items that askrespondentsabouttheirhousehold’sinabilityto(1)affordoneweekholidayonceayear,(2)affordameal withmeat, chicken, fish (or vegetarian equivalent) …. every second day, (3) keep homeadequatelywarm,(4)faceunexpectedfinancialexpenses,(5)makeendsmeet,and(6)[bearthe]financial burden of the housing costs. Through their strong association, these items showsituationsofhigheconomicpressureandconsumptioncompression,whichendangerthecapacityofhouseholdstocopewithfinancialdemandsandsatisfytheirownneeds,whichvaryaccordingto the socio-economic contexts where they live. While income-based measures show thehouseholds’ deficit in resources flows or stocks, our index of financial strain looks atconsumption/expenditure, and shows the stresses faced by many households in keeping anadequatelivingstandard.Over-indebtednessisourlabelforthesecondfactor,whichloadsmostheavilyonitemsthataskrespondentsabout(1)arrears[on]loanpayments,(2)arrears[on]utilitybills,(3)arrears[on]rent,mortgage,(4)heavyfinancialburdenoftherepaymentofdebtsfromhirepurchasesorloans.Theseitems are usually considered as good predictors of over-indebtedness (Angel and Heitzman,2015).Following therecommendationofaEuropeangroupofexperts (EuropeanCommission2008)our indexcombines togethera)at leastone financialcommitment (arrears)andb)onefinancialcommitmentperceivedas“heavy”(aburden).Whiletheover-indebtedhouseholdhasbeenconceptuallydefinedas“onewhoseexistingandforeseeableresourcesareinsufficienttomeetitsfinancialcommitmentswithoutloweringitslivingstandards”(Fondevilleetal.,2010,p.4),more practicalmeasures have been adopted in empirical research. According to previousresearch,subjectiveorobjectiveindicatorsconsideredseparatelyshowstronglimitationsduetodifferent individual judgmentofwhat “difficulty”means, andhugevariability in national legalregulations governing late payment. Therefore, a mixed strategy combining subjective andobjectiveitemsseemstobepreferable(Whelanetal.,2015a).Absolutedeprivationisourlabelforthethirdfactor,whichloadsmostheavilyonitemsthataskrespondentswhetherthey(1)can’taffordcar,(2)can’taffordPC,(3)can’taffordwashingmachine,(4)can’taffordcolorTV.Theseitemsmeasurethehouseholdresourceendowmentastheabilityto afford durable goods.When some of these durable goods aremissing due to affordabilityproblems,wehaveasituationofabsolutedeprivation.Theseitems,thoughsomeofthemarenowunderrevisionthatcanaffect theheadline indicator itself (Guioetal.,2012;GuioandMarlier,2013),arestillregularlyincludedwithincurrentlyusedmultipledeprivationindexes(Eurostat,2016).AConfirmatoryFactorAnalysiswascarriedout tochecktherobustnessof these findings(seeTable1).GoodnessofFittestsconfirmthatthePCAsolutionisadequateforthewholesample.Testswererunalsoatthecountrylevelwiththesamepositiveresults8. 8Country-levelstatisticalresultsareavailableonrequest.


10

Table1.PrincipalComponentAnalysis(PCA)factorloadings. Comp1 Comp2 Comp3 Unexplained

WashingMachine 0,5418 0,275Car 0,415 0,404PC 3417 0,417

ColourTV

0,4727

0,381

Arrearsrent 0,5271 0,288Arrearsbill 0,4337 0,314Arrearsloan 0,5753 0,221Debtburden 0,2083 -0,279 0,3479 0,370Unexpectedexpenses 0,2742 0,237Makeendsmeet 0,3874 0,206Holidays 0,3819 0,237Meat 0,3822 0,317Keephousewarm 0,4393 0,353Housingburden 0,4591 0,361ConfirmatoryFactor–GoodnessofFitTests RootMeanSquareTest 0,054ComparativeFitIndex 0,910TuckerLewisIndex 0,889

StandardizedRootMeanSquareResiduals 0,036Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Note:loadingslessthan.2inabsolutevaluearenotshown. Two theoretical assumptions are therefore confirmed by the 14-items PCA and followingconfirmatoryanalysis.First, economic insecurity isadistinctdimension in respectofmaterialdeprivation,whichismoreproperlyrelatedtothelackofbasicdurablegoods.Then,twomainindependentnot-incomebaseddimensionsofeconomicinsecurityhavebeenidentified:financialstrain(showingtheeconomicinsecurityandconsumptioncompressionofhouseholds),andover-indebtedness(i.e.difficultyofhouseholdstopaytheirfinancialcommitments).To take in account transient poverty (see Section 2.1) we integrated these results with alongitudinal income-based measure of poverty, calculated as number of fluctuations of thehouseholdequivalent incomebelow theyearlynationalpoverty line (60%ofnationalmedianincome).Finally,distinctindexesofthesedimensionsarecalculatedbyusingthelongitudinalheadcountmethodologydescribedinsection3.Table2reportsthemainstatisticalresultsofsuchanalysis.


11

Table2.Longitudinalheadcountsofindexesofinsecurityanddeprivation

Financialstrain

Overindebtedness

Incomepoverty

AbsoluteDeprivation

Perc.totalheadcount(atleastonce) 30,9 17,4 28,9 1,0Average duration (number ofspells) 2,2 1,8 2,2 1,4% average duration (over max.duration) 55,0 45,0 55,0 35,0Averagecontinuity 3,2 2,5 3,2 1,7Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Buildingontheseresults,wedevelopthenextanalysisinthreesteps.First,weanalyzethetimedynamicofourindicatorsandwelookattheintersectionsamongthem.Thisallowsustobuildupanewtypologyofeconomicinsecurity,whichiscomposedbyseveraldimensionsandisclearlydistinguishedfromabsolutedeprivationorpermanentpovertysituations.Second,weobservethecross-country differentiation of this typology to investigate whether and to what extentdifferentiated social and welfare regime contexts may differently affect economic insecurity.Third,weconsiderhowhouseholdsaffectedbyeconomicinsecurityaredistributedacrossdiverseclassgroups.

4.2Thedynamicofeconomicinsecurity

Figure 1 gives a sensitive representation of the relevance and characteristics of temporarydifficulties as opposed to more chronic (maximum four years) situations. Householdsexperiencingat leastonecriticalspell(whateverdimensionisconsidered)areconcentratedintwobiggroups:thoseonlywithonespellmostlyinvolvingonesingledimension,andthosefacingmany spells and multiple difficult conditions at the same time. While chronicity very ofteninvolvesaprogressiveaccumulationofdisadvantages,manyhouseholdsexperienceonlyaverytemporary(mostlyonlyoneyear)criticalsituationaffectingonlyonedimension.

Figure1.Shareofhouseholdsincriticalsituationbytimecontinuityandnumberofdimensionsexperienced(total=householdswithatleastonecriticalspell).

1 2 3 4 5 74 dim 0,0 0,0 0,0 0,0 0,1 0,33 dim 0,1 0,2 0,4 0,9 0,9 5,12 dim 1,2 1,5 2,4 1,8 2,4 5,71 dim 12,9 1,9 3,7 1,2 1,8 2,1

0,0

2,0

4,0

6,0

8,0

10,0

12,0

14,0

16,0


12

Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Fromtheseresults,weelaborateatypology,shownintable3wheredifferentcriticalsituationsaredynamicallydescribed.Inthefirstgroup,wehavemanyhouseholdsexperiencingshort-term,one-dimensionaleconomicinsecurity,butnotabsolutedeprivationorpermanentpoverty:9.6%experiencefinancialstrain,3.5%faceover-indebtedness.Formostofthesehouseholdssuchsituationsoccuronlytemporary.Manyother households (10.5%) experience income fluctuations sufficient to temporarybringthem under the poverty threshold, but fewer signs of financial stress (see householdscharacterizedby“transientpoverty”).Forsuchhouseholdslowincomeapparentlycoincideswithnosignificant financialorconsumptionpressure:asituationdescribedas“integratedpoverty”(Paugam,1996;Böhncke,2008).Insum,weobservethat23.6%ofhouseholdshavebeenaffectedbyashort-term,one-dimensionalformofeconomicinsecurity.Inanintermediatepositionthereisasmallergroup(4.6%)ofhouseholdsthatcumulatefinancialstrain and over-indebtedness, but not poverty spells: these are households in strong financialdifficulty due to costs that they are unable to meet even though income has no significantdownward fluctuations in the observed time span. Furthermore, very often such financialdifficulty occurs on a temporary basis,with significant long-term consequences only in a fewcases.Inthesecondgroup,alargeshareofhouseholds(17.5%)experiencesaprolongedtrajectoryofmultiple hardship, involving frequent fluctuations under the poverty line combined withcumulatedfinancialstrainorover-indebtedness.Thesemultiplecriticalsituationsdonotoftenoccur synchronically (in the same year), but, more frequently, households shift from incomepovertytoilliquidityproblemsorover-indebtednessinaviciouscirclemadeofincomeshortageandconsequentstrongconsumptioncompressionandhighfinancialvulnerability.Chronicizationandmulti-dimensionalitydescribethesehouseholds,whicharedefinedasfallingina“multiplepoverty-based hardship”. Finally, a small amount (1.0%) of households is affected byabsolutedeprivation that almost always intersects other hardship dimensions: households here arecharacterized by multiple difficult situations, where deprivation (considered here as lack ofdurablegoods)ispersistentlyassociatedwithlong-standingpoverty,financialstrainand/orover-indebtedness.ThisisanextremeandcumulativehardshipsituationwhichisveryrarethroughEurope.Table 3 shows also the over time distribution of the various categories. As expected, one-dimensional financial strain, over-indebtedness or temporary poverty are mainly conditionsaffecting households for only one year or two separate years. On the contrary, householdsexperiencingmultipleformsofhardshiporabsolutedeprivationaremorelikelytobeaffectedforthreeorfourspells.Iftheheadcountiscalculatedbyyearsincriticalsituationsratherthanbynumberofhouseholds(see table 3), theweight of chronically poor/deprivedhouseholds becomeshigher, consistentwithpreviousresultsbyBaneandEllwood(1986).Table3showsthattrajectoriescharacterizedbylong-termmultiplehardshipcountfor52.9%oftotal years even though they involve only 28.3%of households. On the other hand, 30.5%ofhouseholdsexperiencingatransitory(onlyonespell)economicinsecurityaccountsforjust8.1%of total years.Wehave thereforeahigh concentrationofdifficult spells ina relatively limitednumberofhouseholdsontheonehand,andalowdiffusionofdifficultspellsinalargenumberofhouseholdsontheother.Thissecondgroupisnotaffectedbyhighriskofsocialexclusion(Mood,2015),butexperiencesaformofeconomicinsecuritycharacterizedbydifficultsituationsforaverylimitedamountoftime,overthefour-yearperiodexaminedhere.


13

Table3.Headcountstatisticsfordifferentcriticalsituations.

Short-termone-dimensionaleconomicinsecurity IV

Financialstrain+over-

indebtedness

Long-termmultiplehardship

Total

IFinancialstrain

IIOver-

indebtedness

IIITransientpoverty

VMultiplepoverty-basedhardshi

p

VIAbsolutedeprivatio

n

Headcountratio 9,6 3,5 10,5 4,6 17,5 1,0 46,6

Continuityindex Columnpercentoverthenumberofhouseholdsincriticalsituation1 52,8 70,8 51,1 11,4 4,3 6,7 30,52 9,5 5,2 8,0 12,0 6,7 4,8 7,93 15,7 13,9 16,2 17,3 11,2 7,7 14,04 5,2 2,9 5,4 13,8 11,0 6,7 8,15 6,8 5,8 8,9 15,1 15,0 8,7 11,1

7 10,0 1,4 10,4 30,4 51,8 65,4 28,3

TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Continuityindex Columnpercentoverthetotalnumberofcriticalspells1 21,6 40,8 20,1 2,6 0,8 1,2 8,12 7,8 6,0 6,3 5,6 2,5 1,7 4,23 19,3 24,0 19,1 12,0 6,3 4,1 11,24 8,6 6,7 8,5 12,8 8,3 4,7 8,65 13,9 16,7 17,5 17,5 14,0 7,6 14,87 28,8 5,8 28,5 49,4 68,1 80,7 52,9

TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Tosumup, thenumberofhouseholdsdealingwitheconomic insecuritycapturedthroughthisapproach is very high.Over 46%of households in our eight countries underwent a period ofdifficultywithina four-year timespan(see table3).Manysufferedonly temporary insecurity,whichmainlyaffectedonlyonedimensionoftheirlivingconditions.Theseinsecurehouseholdsdonotconstitutealargepercentageofthesociallyinsecureinanyoneyear,giventhattheyarenotpermanentlyinsecure.Buttheseinsecurehouseholdsdoconstituteaverylargeshareofthepopulation,representingaround24%ofthetotalhouseholdsinoureightcountries.

4.3Cross-countrycomparison

ThedistributionofourcategoriesacrosscountriesshowsthathouseholdsaredifferentlyaffectedbytheseproblemsacrossEurope(seeTable4andFigure2).HouseholdsinCentralandEasternEuropean (CEE) countries aremore likely to be in difficult situations than households in the


14

westernpartofEurope.InHungaryandPolandcriticalsituationsareremarkablyconcentratedinmultiple poverty-based hardship: a clear sign that low income, constrained consumption andfinancial strain are often interweaved problems in these countries, and that these situationsaccumulateinagreatnumberofhouseholds.Ontheotherhand,theshareofhouseholdsaffectedby transitory poverty in CEE countries is remarkably lower than in the Western Europeancountriesincludedhere.InWesternEuropeancountries,theshareofmultiplelong-termhardshipisverylow,andlowestinSwedenandDenmark.Mediterraneancountriesshowhigherlevelsofmultiplepoverty-basedhardship(ourdatareferto2007-2012,whenItalyandSpainwerestronglyaffectedbyeconomiccrisis).Finally,one-dimensional transitorypovertypredominates in theUKandSpain,but it iscomparativelyveryhighbothinSwedenandDenmark.Thisfactmayconfirmtheideathatriskoftransitory income poverty is higher in countries with lower inequality and higher welfareprotection. In these countries, indeed, while shorter average distance from the poverty lineincreases the likelihood of poverty spells, the high diffusion of minimum income programsenables the temporarypoor to quickly recover from their situation. Finally, financial strain ishigherinMediterraneancountries,FranceandtheUK.

Figure 2. Area plots of headcounts of households in critical situation, by country. Horizontal axis: finalheadcountweightedbycontinuity;verticalaxis:shareofhouseholdsSource: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.In sum, transitory economic insecurity is a widespread phenomenon in Europe, although itsprevalencevariessystematicallyacrosscountries.Thesecross-nationaldifferences,detailed inTable4,aresummarizedasareaplotsinFigure2.Thehigherconcentrationoftransitoryeconomicinsecurity in Continental countriesmay reflect amarked dualisation in the labormarket, thespreadofunstableemploymentpositionsinthisarea,andhigherexposureofhouseholdstorisksrelatedtofinancialisation.Ontheotherhand,thegeographicaldistributionofmultiplepoverty-basedhardship(closetochronicity)ismuchmoredifferentiated,withhigherconcentrationinCEEandMediterraneancountries.Inthefigure,weobserveaclearUshapeddistributioninCEEandMediterranean countries, to show that both long-term poverty and short-term insecurity

0,02,04,06,08,0

10,012,014,016,018,020,0

1 2 3 4 5 7

ALLCOUNTRIESAbsolutedeprivation

Financialstrain+indebtedness

Multiple poverty-basedhardship

Transitory poverty

Over-indebtedness

Financialstrain0,02,04,06,08,0

10,012,014,016,018,020,0

1 2 3 4 5 7

DENMARKAbsolutedeprivation



Transitory poverty

Over-indebtedness


10,012,014,016,018,020,0

1 2 3 4 5 7

SWEDENAbsolutedeprivation



Transitory poverty

Over-indebtedness

Financialstrain

0,02,04,06,08,0

10,012,014,016,018,020,0

1 2 3 4 5 7

FRANCEAbsolutedeprivation



Transitory poverty

Over-indebtedness


10,012,014,016,018,020,0

1 2 3 4 5 7

SPAINAbsolutedeprivation



Transitory poverty

Over-indebtedness


10,012,014,016,018,020,0

1 2 3 4 5 7

ITALYAbsolutedeprivation



Transitory poverty

Over-indebtedness

Financialstrain

0,02,04,06,08,0

10,012,014,016,018,020,0

1 2 3 4 5 7

UKAbsolutedeprivation



Transitory poverty

Over-indebtedness


10,012,014,016,018,020,0

1 2 3 4 5 7

HUNGARYAbsolutedeprivation



Transitory poverty

Over-indebtedness


10,012,014,016,018,020,0

1 2 3 4 5 7

POLANDAbsolutedeprivation



Transitory poverty

Over-indebtedness

Financialstrain


15

predominate.NordiccountriescanbecharacterizedbyaJ-shapedcurve,whereonlytransitoryproblemspredominate.FranceandtheUKareinamixedposition.

Table4.Headcountsofhouseholdsincriticalsituationbycountry.

Short-termone-dimensionaleconomicinsecurity IV

Financialstrain+over-

indebtedness

Long-termmultiplehardship

0

Well-being

IFinancialstrain

IIOver-

indebtedness

IIITransientpoverty

VMultiplepoverty-basedhardship

VIAbsolutedeprivation

Tot.

Denmark 77,7 2,8 4,1 9,5 1,7 3,5 0,7 100Sweden 73,4 1,7 5,3 12,5 1,3 5,5 0,4 100Spain 46,0 9,2 3,5 13,3 5,1 22,1 0,8 100France 60,1 6,9 4,7 10,1 4,3 12,6 1,3 100Italy 49,2 12,1 2,5 8,5 6,2 20,6 0,9 100Hungary 36,2 24,9 3,2 2,8 11,4 18,5 3,1 100Poland 45,9 16,8 3,0 8,5 4,0 20,3 1,6 100UK 56,6 6,2 3,1 13,0 3,0 17,7 0,6 100Total 53,4 9,6 3,5 10,5 4,6 17,5 1,0 100Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.4.4Economicinsecurityandthemiddleclass Figure3showsthedistributionofhouseholdsincriticalsituationacrossincomedeciles9.Whilemultiple poverty-based hardship, absolute deprivation, andobviously also transient poverty areconcentrated in the lowest three deciles and drop significantly in the fourth, our indexes oftransitory financial strain and over-indebtedness are more broadly distributed across incomedeciles,with significant drops (below10%of householdswithin the decile) only at the ninthdecile.Thisisclearevidencethateconomicinsecuritymainlyrelatedtodifficultconsumptionandhighriskofindebtednessspreadstothemiddle-incomegroups,whilethepooresthouseholdsaremostlyaffectedbytransitoryorpermanentpoverty10.Moreover,suchresultshowsthateconomicinsecurityandpovertydonotoverlapforthesameincomegroups.Whilepovertyisobviouslyconcentratedinthepoorestdeciles,itislessobviousthatlower-middleincomegroups(betweenthethirdandfifthdecileoftheincomedistribution)aremorelikelytofacedifficulties intheirfinancialsustainabilitythanlow–incomegroups.Forthesegroups,itisnotthescarcityofincomeinitselftobeproblematic,butexpenseordebtlevelsthatresulttobedisproportionatetotheirdisposable income, especially since they are traditionally prone to high consumption andconsumptionsmoothingbehaviour.

9 Income deciles are calculated for the firstwave of our panel. The graph therefore shows howdifferent types ofhouseholdsincriticalsituationhavedevelopedovertimestartingfromaspecificpreliminaryincomesituation.10Country-basedanalysis (notshown in thepaperbutavailableonrequest)confirmedthesameresults forall thenationalcontextshereconsideredwithonlyminorexceptions.


16

Figure3.Distributionofhouseholdsincriticalsituationbyincomedeciles.Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Ofcourse,theanalysisabovehaslimitedutilityforsociologicalclassanalysis,sinceitisbasedontheincomedistributionratherthanclasscategories.Toaddressthis limitation,wealsousedasimplified version of the European Socio-economic Classification (ESeC)11 to analyze thedistributionofdifferenttypesofcriticalsituationsbysocialclass.Figure4showsthathouseholdsaffectedbyabsolutedeprivationordurablemultiplepoverty-basedhardshiparemoreconcentratednotonlyintheworkingclass(categories1-2)butalsointhelowermiddleclass(category3).Suchdeprivationishardlyexperiencedbyhouseholdsinthemiddle-upperanduppersocialclasses.Dimensionsofeconomicinsecurity,however,donotlineupsoneatlywithsocialclass.Forexample,transitorypovertywassignificantamongthelowermiddleclass in2007-2012,probablyowingtotheeconomicrecession12.The financialstrain typecuts

11ESECclassificationhasbeendevelopedfollowingRoseandHarrison(2007)’ssimplifiedprocedure.Inthe2009-2012rotation,asonlytheISCO-08classificationwasprovided,re-allocationtoISCO-88categorieshasbeendoneusingtheISCOcorrespondencematrix.WhereanISCO-08submajorgroupwasdividedbetweentwoISCO-88classes,caseswereallocatedaccordingtotheirmodalvalue.Thestandard9classesESECweresynthetizedinsixclassesgroupingtogether“Small employer and self employed occupations”, “Self employed occupations”, “Lower supervisory and lowertechnicianoccupations”and“Lowerservices, salesandclericaloccupations” intoa “Smallemployers, lowermiddleclass”group.12Thehighvalueofpovertyforthesmallemployers/lowertechnicalclassmayalsoreflectadifficultyofEU-SILCtocontroltheincomedataofsmallemployers,whoshowhighlevelofunderreportinginsurveyinanumberofcountries.

©

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

Financialstrain

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

Over-indebtedness

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

TransitoryPoverty

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

Multiplepoverty-basedhardship

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

Absolutedeprivation

0,0

5,0

10,0

15,0

20,0

25,0

30,0

35,0

40,0

45,0

1 2 3 4 5 6 7 8 9 10

Financialstrain+over-indebtedness


17

moststronglyacrossclasses,andonlydropssignificantlyforthehighestclass.Theintermediateclass, which combines technicians and public employees with medium to high qualification,experiencessubstantialfinancialstrainandover-indebtedness.

Figure4.Distributionofhouseholdsincriticalsituationsbysocialclass(ESECclassification).Source: EU-SILC panel database: pooling of three rotations from 2007/10 to 2009/12, authors’ ownelaboration.Lookingindetailatmiddleclassgroups(categories3-4-5),temporaryfinancialstrainandover-indebtedness are thehighestandmostencompassingaspectsofeconomic insecurity.Thehighconcentrationofsuchindexesinallthemiddle-classgroupsgivesanideaoftowhatextent, incurrentEuropeansocieties,consumptioncompressionand inability topaydebtsendangerthefinancialsustainabilityofarelevantpartofthemiddleclass,includinghigh-skilledprofessionalsandmanagersthatarepartofthe“Lowersalariat”category.Formiddle-upperclasspeoplewithhigher qualifications, sustainability problems mean a large discrepancy between highconsumptionassociatedtotheir life-styleandtheirsocialreputationontheonehand,andtherelativescarcityorinstabilityofmaterialresourcesthatarenecessarytosustainthatsocialstatus.Forthem,economicinsecuritybasicallymeansariskof“relativedeprivation”(Frank,2007),asignificantlackofresourcestosustainalife-styleinaccordancewiththeirsocialidentity.Inthelower-middleclassgroup,instead,criticalfinancialsustainabilitydoescometogetherwithhigherleveloftransitorypoverty.Thisisthesituationcharacteristicofthe“pettybourgeoisie”mostlyrepresentedinthe“Smallemployer/lowmiddleclass”class.Thissocialgroup(composedby shop-owners, small entrepreneurs, self-employed technicians, craft workers, etc.) istraditionally characterized by relatively low education and consumption levels, and relativelyhigh income. The self-employed,however,hasbeenoneof the social categoriesmostlyhit in

Figure4.Distributionofdifferenttrajectoriesofhardshipbysocialclass(ESECclassification)

.0,0

5,0

10,0

15,0

20,0

25,0

HighsalariatLowersalariatIntermediateSmallempl/low-middleclass

LowtechniciansRoutine

Financialstrain

0,0

5,0

10,0

15,0

20,0

25,0



Over-indebtedness

0,0

5,0

10,0

15,0

20,0

25,0



Transitorypoverty

0,0

5,0

10,0

15,0

20,0

25,0



Multiplepoverty-basedhardship

0,0

5,0

10,0

15,0

20,0

25,0



Absolutedeprivation

0,0

5,0

10,0

15,0

20,0

25,0



Financialstrain+over-indebtedness


18

recentyearsduetothehugecrisisofsmall-sizeenterprises.Inthiscase,thereisthereforeaclearfunctionallinkbetweenhighpovertyrisksandfinancialdifficultiescompressingconsumptionandexposingtoover-indebtedness.

5.DiscussionThis paper sheds new light on economic insecurity in Europe by incorporating multipledimensions of disadvantage simultaneously, and by incorporating a dynamic perspective thatrevealssignificantcross-countrydifferences.BuildingontheworkofGornickandJantti(2013),Westernetal.(2012),andWhelanetal.(2015a),amongothers,wedevelopasyntheticapproachtoeconomicinsecuritythatincorporatesmultipledimensionsofshort-termvulnerabilityandadynamicperspective.Todemonstratetheutilityofthisapproach,weanalysefourwavesofdatafromtheEuropeanUnionSurveyofIncomeandLivingConditions(EU-SILC),referringtotheyears2007-2012.Aprincipalcomponentsanalysisyieldsevidencesupportingourhypothesisthattherearedistinct typesof insecurity, and that theseaspects are clearly separate fromameasureofabsolutedeprivation.Usingthiscategorization,wecalculatedynamicheadcountsofhouseholdsincriticalsituation,andshowhowthesecountsvaryacrossEuropeancountries.Finally,weshowthateconomicinsecurityisbroadlydistributedacrossEuropeanhouseholds,reachinghighuptheincomeandclassscales.The primary overarching finding is the peculiar social profile of economic insecurity. First, itcrisscrosses a wide range of social class positions. Secondly, it affects not only lower classhouseholds but also white-collar intermediate workers and households whose income is inmiddledeciles.Itconfirmsthateconomicinsecurityaffectsnotonlytheworkingclassandlowerincomehouseholds,butalsoasubstantialpartofthemiddleandevenupper-middleclass.FutureEU-SILC datawill need to be analyzed to determinewhether thiswidespread vulnerability isstructural,orconfinedtotherecenteconomiccrisis.Butwhetheritenduresordissipates,thiswidespreadvulnerabilitydemonstratesthelimitedcapacityofcontemporaryEuropeanwelfarestatestosecurehouseholdsfrommarketvolatility(HuberandStephens,2014).TheevidencethateconomicinsecurityiswidelydistributedacrosstheclasshierarchyinEuropesuggests thatmany households experience insecurity on a day-to-day basis that shapes theirstandardofliving.Weemphasizethatthisinsecurityconcernsnotonlytheabsoluteamountofavailableresources,butalsothestabilityofresourceflows(asshownbyourresultsconcerningfluctuationsunder/overthepovertythreshold)andtherelationofincometoconsumption(asourindicatorsaboutfinancialstrainandover-indebtednessclearlyshow).Economicinsecuritynotonlyincreasestheriskofpovertyormaterialdeprivation.Rather,itisadiffuseconditionformanyhouseholds.Itisthisconditionofbeinginaturbulent,unstableconditionthatisunderstoodhereas “economic insecurity”. In other words, economic insecurity not only causes problems forhouseholds,butmaybeasocialprobleminandofitself.Many aspects of thiswidespread insecurity still need to be investigated. Although this paperdescribesthedistributionofeconomicinsecurityacrossEuropeannationsandhouseholds,wehavelefttheexplanationofthesepatterstofuturework.Ourresultspointtothenecessityofnewresearch on, for instance, the relationship between trigger events and specific insecuritysituations. Furthermore, structural conditions potentially responsible for the distribution ofeconomicinsecurityneedstilltobeproperlyassessed.Wealsoneedtoanalysetheconsequencesof insecurity, in termsof socialbehaviour, investments in the future,andpoliticalorientation.Short-term insecuritymay have different impacts on people depending on the role played bycollectiveaswellindividualbuffers:forsomegroupsitcouldbeatransitoryexperiencewithnolong-term consequences, while more fragile groups could end up entrapped in a sort ofpermanent,recurrentfluctuation.Inthissense,wecallfornewresearchtounderstandwhatrolemaybeplayedbybroadwelfareregimesandspecificsocialpolicies inprotectingpeople frominsecurityanditspotentialconsequences.


19

References Alkire,S.,andFoster,J.(2011)Countingandmultidimensionalpovertymeasurement,inJournalofpubliceconomics,95(7),476-487.Alkire,S.,Apablaza,M.andChakravarty,S.R.(2014)MeasuringChronicMultidimensionalPoverty:ACountingapproach,inOPHIWorkingPaperNo.75.Anderloni,L.,andVandone,D.(2011)Riskofover-indebtednessandbehavioralfactors.InLucarelli,C.,andBrighetti,G.(Eds.)Risktoleranceinfinancialdecisionmaking,113-132,PalgraveMacmillan.Angel,S.,andHeitzmann,K. (2015)Over-indebtedness inEurope:Therelevanceofcountry-levelvariables for the over-indebtedness of private households, in Journal of European Social Policy,25(3),331-351.Azpitarte,F.(2012)Measuringpovertyusingbothincomeandwealth:Across-countrycomparisonbetweentheU.S.andSpain,inReviewofIncomeandWealth,58(1),pp.24-50.Bane,M.J.andEllwood,D.(1986)Slippingintoandoutofpoverty:Thedynamicsofspells,inJournalofHumanResources21,1-23.Beck,U.(1992)Risksociety:Towardsanewmodernity.Sage.Böhnke,P.(2008)Arethepoorsociallyintegrated?Thelinkbetweenpovertyandsocialsupportindifferentwelfareregimes,inJournalofEuropeanSocialPolicy,18(2),133-150.Bonoli,G.,andNatali,D.(2012)Thepoliticsofthenewwelfarestate.Oxford,OxfordUniversityPress.Bossert,W., Chakravarty, S. R., andD’Ambrosio, C. (2012)Poverty and time, inThe Journal ofeconomicinequality,10(2),145-162.Bossert,W.andD’Ambrosio,C.(2009)MeasuringEconomicInsecurity,inECINEQWorkingPapersSeries2009-111.BossertW.andC.D’Ambrosio(2013)MeasuringEconomicInsecurity,inInternationalEconomicReview,Vol.54/3,1017-1030.Cantillon, B., and Vandenbroucke, F. (2014). Reconciling work and poverty reduction: howsuccessfulareEuropeanwelfarestates?Oxford,OxfordUniversityPress.Castel, R. (2000) The roads to disaffiliation: insecure work and vulnerable relationships, inInternationalJournalofUrbanandRegionalResearch,24(3),519-535.Dallinger,U. (2013)Theendangeredmiddle class?A comparativeanalysis of the role playedbyincomeredistribution,inJournalofEuropeanSocialPolicy,23(1),83-101.European Commission (2008), Towards a common operational European definition ofoverindebtedness,inEuropeanCommission.Brussels:EuropeanCommissionEurostat (2016), Europe 2020 indicators - poverty and social exclusion, accessed athttps://goo.gl/JMGrSRonOctober12,2016.


20

Fligstein,N.,andGoldstein,A.(2015)TheemergenceofafinancecultureinAmericanhouseholds,1989–2007,inSocio-EconomicReview,13(3),575-601.Fondeville,N.,Özdemir,E.,andWard,T.(2010)Over-indebtedness.NewevidencefromtheEU-SILCspecialmodule,inEuropeanCommissionResearchNote,4.Förster, M. F., and Vleminckx, K. (2004) International comparisons of income inequality andpoverty:findingsfromtheLuxembourgIncomeStudy,inSocio-EconomicReview,2(2),191-212.Fouarge, D. and Layte, R. (2005)Welfare regimes and poverty dynamics: The duration andrecurrenceofpovertyspellsinEurope,inJournalofSocialPolicy34:407_/26.Frank, R. (2007) Falling behind: How rising inequality harms the middle class, University ofCaliforniaPress.Gornick,J.,andJäntti,M.(2013)Introduction,inGornick,J.,andJäntti,M(eds)Incomeinequality:Economicdisparitiesandthemiddleclassinaffluentcountries,StanfordUniversityPress,1-47.Gottschalk,P.,andMoffitt,R.(2009)TherisinginstabilityofUSearnings,inTheJournalofEconomicPerspectives,3-24.Guio,A.C.,Fusco,A.,andMarlier,E.(2009)AEuropeanUnionapproachtomaterialdeprivationusingEU-SILCandEurobarometerdata,inIntegratedResearchInfrastructureintheSocio-economicSciences(IRISS)WorkingPaperSeries,19,2009.Guio, A., Gordon, D and Marlier, E. (2012) Measuring Material Deprivation in the EU, inLuxembourg.PublicationsoftheOfficeoftheEuropeanUnionGuio,A.C.,andMarlier,E.(2013)Alternativevs.currentindicatorsofmaterialdeprivationatEUlevel:Whatdifferencesdoesitmake?,inImPRovEWorkingPaperSeries13/07,HermanDeleeckCentreforSocialPolicy-UniversityofAntwerp,Antwerp.Hacker,J.S.(2008)Thegreatriskshift:TheneweconomicinsecurityandthedeclineoftheAmericandream,OxfordUniversityPressHacker,J.S.,Huber,G.A.,Nichols,A.,Rehm,P.,Schlesinger,M.,Valletta,R.,andCraig,S.(2014)TheEconomicSecurityIndex:anewmeasureforresearchandpolicyanalysis,inReviewofIncomeandWealth,60(S1),S5-S32.Huber, E., and Stephens, J. D. (2014) Income inequality and redistribution in post-industrialdemocracies:demographic,economicandpoliticaldeterminants,inSocio-EconomicReview,12(2),245-267.Jappelli, T., Pagano, M., and Di Maggio, M. (2013) ‘Households' indebtedness and financialfragility’,JournalofFinancialManagement,MarketsandInstitutions,1(1),23-46.Jappelli,T.,andPistaferri,L.(2006)Intertemporalchoiceandconsumptionmobility,inJournaloftheEuropeanEconomicAssociation,4(1),75-115.Jenkins, S.P., Brandolini, A.,Micklewright, J., andNolanB. (2012)The Great Recession and theDistributionofHouseholdIncome,Oxford,OxfordUniversityPress.


21

Kenworthy,L.(2014)Hasrisinginequalityreducedmiddle-classincomegrowth?,inGornick,J.,&Jäntti,M.(Eds.)Incomeinequality:Economicdisparitiesandthemiddleclassinaffluentcountries,StanfordUniversityPress,101-114.Latouche,M.andNaud,J.J.(2001),WeightingoftheSwissHouseholdPanel,FORSWorkingPapers8_01. Available at http://forscenter.ch/wp-content/uploads/2013/12/Weight_SHP_W2_detail_E.pdfLeisering, L., and Leibfried, S. (1999)Time and Poverty inWesternWelfare States, CambridgeUniversityPress.Mau, S., Mewes, J., and Schöneck, N. M. (2012) What determines subjective socio-economicinsecurity?Contextandclassincomparativeperspective,inSocio-EconomicReview,10(4),655-682.Mau, S. (2015) Inequality,Marketization and theMajority Class:WhyDid the EuropeanMiddleClassesAcceptNeo-liberalism?,PalgraveMacmillan.Meghir,C.,andPistaferri,L.(2011)Earnings,consumptionandlifecyclechoices,inHandbookofLaborEconomics,4,773-854.Mood,C.(2015)Thenot-very-richandtheverypoor:PovertypersistenceandpovertyconcentrationinSweden,inJournalofEuropeanSocialPolicy,25(3),316-330.Nau,M.andSoener,M.(2017),Incomeprecarityandthefinancialcrisis,inSocio-EconomicReview,2017,Vol.0,No.0,1–22,onlineNolan,B.andWhelan,C.T.(1999),LoadingtheDice?:AStudyofCumulativeDisadvantage,Dublin:OakTreePress.Nolan,B.andWhelan,C.T.(2009),UsingNon-MonetaryDeprivationIndicatorstoAnalyzePovertyandSocialExclusion inRichCountries:Lessons fromEurope?, inUCDSchoolofAppliedScienceWorkingPaperSeries,WP09/11Nolan,B.andWhelan,C.T.(2011)PovertyandDeprivationinEurope.Oxford:OxfordUniversityPress.Osberg,L.andSharpe,A.(2014),Measuringeconomicinsecurityinrichandpoornations,inReviewofIncomeandWealth,60,S1,pp.53-76Osberg, L. (2015)How Should One Measure Economic Insecurity?, in OECD Statistics WorkingPapers,2015/01,OECDPubl,accessedathttp://dx.doi.org/10.1787/5js4t78q9lq7-enonJanuary7,2017Paugam,S.(1996)Povertyandsocialdisqualification:AcomparativeanalysisofcumulativesocialdisadvantageinEurope,inJournalofEuropeanSocialPolicy,6(4),287-303.Pressman, S. (2007)TheDeclineof theMiddleClass:An InternationalPerspective, in Journal ofEconomicIssues,41,1(2007),181-200.Rose,D.,andHarrison,E. (2007)TheEuropeansocio-economicclassification:anewsocialclassschemaforcomparativeEuropeanresearch,inEuropeanSocieties,9(3),459-490


22

Sandoval,D.A.,Rank,M.R. andHirschl,T.A. (2009)The IncreasingRiskofPovertyAcross theAmericanLifeCourse,inDemography,46,717–737.Scott,R.H.,andPressman,S.(2011)ADoubleSqueezeontheMiddleClass,inJournalofEconomicIssues,45:2,333-342Stiglitz, J., Sen,A. andFitoussi, J.P. (2009),Report by theCommissionon theMeasurement ofEconomic Performances and Social Progress, accessed athttp://ec.europa.eu/eurostat/documents/118025/118123/Fitoussi+Commission+report onNovember3,2017Vandecasteele,L. (2010)Poverty trajectoriesafter risky life courseevents indifferentEuropeanwelfareregimes,inEuropeanSocieties,12(2),257-278.Vandecasteele,L. (2011)Life course risksor cumulativedisadvantage?The structuringeffectofsocial stratification determinants and life course events on poverty transitions in Europe, inEuropeanSociologicalReview,27(2),246-263.Western, B., Bloome, D., Sosnaud, B., and Tach, L. (2012) Economic Insecurity and SocialStratification,inAnnualReviewofSociology,38:341–59.Whelan, C. T. Layte, R.andMaître, B; (2003)Persistent Income Poverty and Deprivation in theEuropeanUnion,inJournalofSocialPolicy,32(1):1-18Whelan, C. T. andMaître,B; (2005)EconomicVulnerability,Multi-dimensionalDeprivationandSocialCohesioninanEnlargedEuropeanUnion,inInternationalJournalofComparativeSociology,46(3):215-293.Whelan, C., Nolan, B., andMaítre, B. (2015a)The “SqueezedMiddle” in the Great Recession: AComparativeEuropeanAnalysis of theDistributionofEconomicStress, inGearyWorkingPaper,2015/12,June.Whelan,C.T.,Russell,H.,andMaître,B.(2015b)EconomicStressandtheGreatRecessioninIreland:Polarization,Individualizationor‘MiddleClassSqueeze’?,inSocialIndicatorsResearch,1-24.

the rise of economic insecurity in the eu: concepts and ...the rise of economic insecurity in the...

Documents