United but divided: Policy responses and people’s perceptions in the EU during the COVID-19 outbreak

(1)

ContentslistsavailableatScienceDirect

Health

Policy

jo u rn al h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / h e a l t h p o l

United

but

divided:

Policy

responses

and

people’s

perceptions

in

the

EU

during

the

COVID-19

outbreak

Iryna

Sabat

a,∗

_,

_Sebastian

_{Neuman-Böhme}

b

_,

_Nirosha

_Elsem

_Varghese

c

_,

_Pedro

_Pita

_Barros

a

_,

Werner

Brouwer

b,d

_,

_Job

_van

_Exel

b,d

_,

_Jonas

_Schreyögg

e

_,

_Tom

_Stargardt

e a_Nova_School_of_Business_and_Economics,_R._Holanda_1,_2775-405,_Carcavelos,_Portugal

b_Erasmus_School_of_Health_Policy_and_Management,_Erasmus_University_Rotterdam,_P.O._Box_1738,₃₀₀₀_DR,_Rotterdam,_the_Netherlands c_Centre_for_Research_on_Health_and_Social_Care_Management,_CERGAS,_Bocconi_University,_Via_Röntgen_n._1,_20136,_Milano,_Italy d_Erasmus_School_of_Economics,_Erasmus_University_Rotterdam,_P.O._Box_1738,₃₀₀₀_DR,_Rotterdam,_the_Netherlands

e_Hamburg_Center_for_Health_Economics,_University_of_Hamburg,_Esplanade_36,_20354,_Hamburg,_Germany

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received27May2020

Receivedinrevisedform16June2020 Accepted17June2020 Keywords: COVID-19 Publicpolicies People’sperceptions Europe

a

b

s

t

r

a

c

t

TounderstandthepublicsentimenttowardthemeasuresusedbypolicymakersforCOVID-19 contain-ment,asurveyamongrepresentativesamplesofthepopulationinsevenEuropeancountrieswascarried outinthefirsttwoweeksofApril2020.Thestudyaddressedpeople’ssupportforcontainment poli-cies,worriesaboutCOVID-19consequences,andtrustinsourcesofinformation.Citizenswereoverall satisfiedwiththeirgovernment’sresponsetothepandemic;however,theextentofapprovaldiffered acrosscountriesandpolicymeasures.Anorth-southdivideinpublicopinionwasnoticeableacrossthe Europeanstates.Itwasparticularlypronouncedforintrusivepolicymeasures,suchasmobiledatause formovementtracking,economicconcerns,andtrustintheinformationfromthenationalgovernment. Considerabledifferencesinpeople’sattitudeswerenoticedwithincountries,especiallyacrossindividual regionsandagegroups.Thefindingssuggestthattheepidemicactsasastressor,causinghealthand economicanxietieseveninhouseholdsthatwerenotdirectlyaffectedbythevirus.Atthesametime,the burdenofstresswasunequallydistributedacrossregionsandagegroups.Basedonthedatacollected, wedrawlessonsfromthecontainmentstageandidentifyseveralinsightsthatcanfacilitatethedesign oflockdownexitstrategiesandfuturecontainmentpoliciessothatahighlevelofcompliancecanbe expected.

1. Introduction

TheoutbreakofCOVID-19triggeredawiderangeofresponses fromgovernmentsintheEuropeanUnion.Giventhatthedisease wasnewandeffectivemedicalcountermeasuresdidnotexistin early2020,governmentshadtoadoptnon-medicalmeasures aim-ingatthecontainmentandmitigationofCOVID-19.Withtheaimof ¨flatteningthecurve,¨thesepoliciesincludedbansonpublic gather-ings,closuresofacademicinstitutionsandpublicplaces,national andinternationalmobilityrestrictions,conﬁnement,andseveral others[1].

ItalywastheﬁrstcountryinEuropetoapplyintervention mea-suresfromthebeginningofMarch2020inresponsetotheseverity oftheCOVID-19outbreak.OtherEUcountriesfollowedsoon

after-∗ Correspondingauthor.

E-mailaddress:iryna.sabat@novasbe.pt(I.Sabat).

ward,usingsimilarcountermeasuresaroundmid-March2020[2]. Theadoptionofthesepoliciesvariedintheirscale,stringency,and paceacrosscountries.WhilemostEuropeanstatesimplemented confinementmeasures,theextentoflimitationsofpeople’s free-domsdifferedacrossindividualcountries.Lockdownswereusually strictest where the pandemic was deadliest (Italy, Spain, and France), imposingsevere limitationsonpopulationmovements. Somegovernments choselessstringentversionsofconfinement ornolockdownatall,forinstance, änintelligentlockdownïnthe Netherlandsor ¨freedomunderresponsibilityïnSweden[3].

Forcedtoreactswiftlytotheunfoldingepidemicsituation, pol-icymakersineverycountrytriedtobalancetheimplementationof containmentpoliciesagainstnumerousimportantfactorswiththe prioritymostlygiventotheprotectionofthepopulation’shealth. Consequently,therehasbeenalotofdebateineverysocietyabout whethermeasurestakenbythegovernmentwereappropriateor not.Somepartsofthepopulationhavebeenvoicingsupportfor moreseverecontainmentpoliciestominimizethespreadofthe https://doi.org/10.1016/j.healthpol.2020.06.009

(2)

virus.Suchattitudeswerelikelyfueledbypeople’sworriesabout theirhealthandthepotentialoftheirnationalhealthcaresystem towithstandtheepidemic.Meanwhile,othersexpressedtheir con-cernsaboutthesocialandeconomicconsequencesofsuchpolicies, therebyadvocatingforlessseverecontainmentmeasures[4].

Asthepandemicbegantoabate,governmentsstarted design-ingthelockdownexitstrategies andrestartingtheireconomies. However,theriskthatthenewwaveoftheepidemicmayhappen didnotdisappear,especiallygiventhatthevaccinedevelopment takesa longtime, andherdimmunity wasnot achieved[5]. In thislight,theissueofliftinglockdownshasbecomeanew sub-jectofpublicdebateacrossandwithinEuropeancountriesraising discussionsabouttheappropriatenessoftiming,risks,and poten-tialconsequencesofendingtheconfinement[6].Liftinglockdown restrictionscreatesacutedilemmastothepolicymakerssincethe economicandhumancostsofanyexitstrategyseemtobeclosely linkedtogether.Takingautilitarianapproachinthissituationcould backfireifthesociety’sunderstandingisnotpreliminarilysecured orexpectationsarenotfulfilled.

Policymakersandpublichealthexpertshavetopersuadetheir citizenstomakebehaviorchangesandrespectfuturecontainment interventionswhilefacingthedifficultyofenforcingsuch regula-tions.Therefore,itbecomescrucialtounderstandpeople’sworries aboutthepandemicandtheirperceptionsoftheeffectsof contain-mentpolicies,sothatthedesignoffurtherpoliciesandcontingency measuresiswell-informed,andahighlevelofcompliancecanbe expectedfromthepopulation.Moreover,trustinthegovernment andsocialinstitutionsmaybecomecentraltoachievinga success-fulimplementation offuture measures,whereas lack ofit may turndetrimentaltothefightagainstthepandemic.Henceitisof paramountimportancetounderstandwhopeopletrustmostso thatpublichealthmessagescanbeamplifiedusingcorrectmeans ofcommunication.

Weprovideatimelydescriptionofthecurrentsituationand drawlessonsfromthecontainmentstagetoinformthedesignand implementationofthelockdownexitpolicies.

2. Materialsandmethods

Inordertounderstandthepublicsentimenttowardsthe COVID-19containmentmeasuresandtoinformfuturepolicydevelopment, wecollectedinformation onpeople’s supportfor thesepolicies, theirworriesinrelationtotheunfoldingepidemic,andtheirtrust indifferentsourcesofinformation.Wesurveyedover7000people representativeoftheadultpopulationinsevenEuropeancountries: Denmark,France,Germany,Italy,Portugal,theNetherlands,and theUK.The ﬁeldworkwasconductedonlineduringApril2–15, 2020,using multi-sourced online panels provided by the mar-ketresearchcompanyDynata.Toensurethatthesamplingframe wasrepresentativegiventheonlinenatureofthestudy,the com-panyapplieddiverserecruitingprocedures toreachthegeneral population(throughopenrecruitment,loyaltyprograms,afﬁliate networks,mobileapps).Itthenusedquotastomatchthenational censussharesineachcountry.

The questionnaire was designed by the authors of the study except for the worryitems that wereadopted from the WorldHealthOrganization(WHO)COVID-19Snapshot Monitor-ingproject[7].Thequestionnairewascarefullytranslatedintosix otherlanguagesbynativespeakersandthenimplementedusing theQualtricsplatformﬁrstasapilot(10%ofthesampleinevery country)andnextasalarge-scalesurvey.Thedatafromthepilot studywereincludedinthetotalsample.

Ineachcountry,wecollecteddatafromasampleof1000 respon-dentsrepresentativeofthenationalpopulationintermsofregion, age,gender,andeducation.GiventhattheItalianregionLombardy

wasthemostseverelyhitbytheCOVID-19outbreak,wecollected 500additionalresponsesinthisregionrepresentativeintermsof ageandgender.Learningaboutperceptionsandattitudesof peo-plewhoresidetherecouldprovideessentialinsightstoresearchers andpolicymakers.TheextradatacollectedfromLombardywere notincludedintherepresentativesampleofItaly.Thus,no weight-ingwasused astheadditionalLombardy samplewasanalyzed separatelyanddenotedasLombardyintheresultssection.

3. Results

3.1. Policysupport

Weassessed people’sapprovalofpolicymeasuresthatwere taken(orwerelikelytobetaken)bytheirnationalgovernmentin responsetotheCOVID-19outbreak.Inparticular,wecoveredsuch issuesasschoolclosures,bansonpublicgatherings,border clo-sures,bansimposedontheexportofmedicalequipment,ﬁnesfor quarantineviolations,randomtemperaturechecks,curfews, pub-lictransportsuspensionsandutilizationofmobilephonedatafor trackingCOVID-19casesandtheircontacts.

Onaverage, 68 %ofpeoplein thesevenEuropeancountries approvedofthepoliciestakenintheircountryinresponsetothe pandemic,implyingconsiderablepublicsupport.Nevertheless,the extentofapprovaldifferedbycountryandbypolicymeasure.

Themostapprovedmeasureswereﬁning14-dayquarantine violations, ban of publicgatherings, and border closures (each supportedby83%of respondents).By thetimeof thesurvey’s ﬁeldwork,restrictionsonpublicgatheringshadbeenadoptedinall countriescoveredbythestudy,whereasinternationaltravel con-trolshadbeenimposedtoacertainextenteverywhere,exceptthe UK[8].

Priortocomplete border closuresin mid-March2020, some countries(forexample,Italy,France,Germany,Denmark)hadbeen requiringscreeningand14-dayquarantineforarrivalsfrom high-riskregionsalready sinceFebruary.Incontrast,othercountries, suchas Portugaland the Netherlands,started later and turned directlytostrictmeasures,suchasbanningarrivalsfrom high-riskareasandimposingpartialborderclosures.Thelattertypically impliedeitherlimitationonentriesofnonresidentsorclosureof onlycertaintypesofborders(land,sea,air),whileensuring ¨green lanes¨forfreightvehicles transportinggoods.However,complete borderclosuresoccurredhaphazardlyandledtodisrupted com-merce and stranding citizens.Among countries covered in our study,Denmarkwastheﬁrsttocloseallbordersinmid-March, whereas the UK did so only in the second half of May 2020. Moreover,atthetimeofﬁeldwork,theUKdidnothaveroutine screeningsatitsairportsorquarantinerequirementfortravelers [8,9]. Thus, theresultsfor theUK showedthe extentof public supportthatthesemeasureswouldhavereceived,hadtheybeen implementedearlier.

Meanwhile,themostopposedcontainmentpolicieswere pub-lictransportsuspension(37%ofrespondentsagainstit),banof medical export,useof mobilephone datafortracking, andthe impositionofacurfew(eachdisapprovedbyapproximately23%of respondents).

These trends might reﬂect within-country regional and age structure of thepopulation. For example,older individuals and thoselivinginremoteareastendedtobethemoststronglyopposed topublictransportsuspension.Infact,amongcountriescoveredby thesurvey,publictransportsuspensionwasimplementedonlyin Italy,whereasitsvolumewasreducedinallotherstatesexcept forGermany[8].Thestay-at-homeordersweremostsigniﬁcantly opposedbytheyoungestrespondentsagedbelow25.This

(3)

mea-surewasenforcedinallcountriescoveredbythesurveyexceptfor Denmark,whereitwasintroducedasarecommendation[8].

Overall,anorth-southgradientcouldoftenbenoticedinthe EUregardingpolicysupport:peoplelivinginthesouthernstates (Portugal,Italy,andFrance)tendedtoapproveofthecontainment policiesmorethanresidentsinthenortherncountries(Denmark, GermanyandtheNetherlands).Noteworthy,thelargestshareof supportersforeverycontainmentmeasurewasnoticedamongthe residentsofItalyandparticularlyinLombardy.Here,onaverage, 79%ofthepopulationapprovedofthegovernment’sresponseto thepandemic.

Fig. 1 illustrates the average degree of approval of several selectedcountermeasuresacrosssevenEuropeancountries (mea-suredonaLikertscalefrom1-stronglydisapproveto5-strongly approve),whichhighlightshowdiverseEuropeisinthe percep-tionsofCOVID-19policyresponses.Higherintensityofthecolor reﬂectsahigherlevelofapprovalofaspeciﬁcpolicybythe popu-lationineachcountry.

Interestingly,themostsigniﬁcantshareofthepopulationwho explicitlyopposedeachofthecontainmentpoliciestakenbytheir governmentwasidentiﬁedinDenmark.Here,forexample,22%of respondentsdisapprovedofschoolclosuresand48%disapproved oftheimpositionofacurfew.Incomparison,theaverage disap-proval ofthesemeasuresin othercountrieswasaround 8%for schoolsand20%forcurfews.

Themostpolarizingopinionswereobservedconcerningtheuse ofmobiledatafortrackingCOVID-19casesandtheircontacts.The mostsigniﬁcantshareofpeopleexplicitlyopposingsuchpolicywas identiﬁedinDenmark(34%),theNetherlands(31%),andGermany (25%).Itwasparticularlydisfavoredbytheyoungestagegroup(33 %ofrespondentsagedbelow25againstit).

Thispolicyreceivedsigniﬁcantmediaattentionassome coun-triesandtheEuropeanCommissionstartedthecollaborationwith telecomproviderstoaccessindividualgeolocationdatafor pre-dictionandsurveillanceofCOVID-19spread[10,11].AsofMarch 2020,Deutsche Telekomprovided German authoritieswiththe anonymizeddataonthemovementofitsusers.InItaly,Vodafone, WindTreandTelecom Italiaofferedaggregateduserdata provi-siontothegovernmentforthesamepurpose.Authoritiesinthe Lombardyregionused mobilephone data tocheckcompliance withthelockdownrestrictions[10–12].Othercountrieseither ini-tiatedthedevelopmentoftheirownmobilephonetrackingappsor cooperatedonthecreationofcommonsoftware,suchasthe Pan-EuropeanPrivacy-PreservingProximityTracing(PEPP-PT)project ledbyGermany.However,thelaunchofthePEPP-PTwasdelayed attheendofAprilduetothedataprotectionconcernsvoicedby expertsandevensomeoftheprojectparticipants[13].

Whileproponentsofthecontact-tracingmeasuresclaimthat usingmobiledataisofparamountimportanceinresponsetothe COVID-19pandemic,manypeopleworryaboutthegovernment’s useoftechnologyduetopossibleprivacyviolations,therebyraising debates about the appropriatenessof such social control mea-sures[10,13,14].Accordingtoourdata,peopleinsomeEuropean countriesexpressedconsiderablereluctanceaboutsupportingsuch policy, which therefore makes future compliance questionable. Moreover,suchprivacy disputes, asin thecase of thePEPP-PT projectlaunch,mighttriggerhigherreluctanceamongthepotential userstouseanycontact-tracingappinthefuture,whichcouldbe detrimentalfortheimplementationofaviabletracingtechnology [13].

Tobetterunderstandpublicopinion oncertain policies,it is essentialtolookatthebigpictureandplaceobtainedresultsinto thenationalcontexts.People’sattitudeswerelikelybasedontheir perceptionsofthegeneralstateofaffairsintheircountry, particu-larlyintermsoftheepidemicsituationandrestrictionstheywere subjecttoatthatmoment.

Inviewofthat,Table1summarizesthescaleofthepandemic andthestringencyofgovernment’sresponseinsevenEuropean countriesatfourpointsoftimespacedaroundApril12(whenthe survey’sﬁeldworkwas99%completeineverycountry).The pub-lichealthsituationineachstateisdescribedusingtotalconﬁrmed casesofCOVID-19andtotaldeathsattributedtoCOVID-19,both measuredper1millionpeopleandreportedbytheEuropeanCentre for DiseasePrevention andControl[15]. Thestringencyof gov-ernment’sresponseismeasuredwiththeCOVID-19Government ResponseStringencyIndex,acompositemeasureofcontainment policiesrangingfrom1to100,where ahigher valuedenotesa stricterresponse[8].

Atthetimeofthesurvey’sﬁeldwork,theepidemicsituationwas worst,andthestringencyindexwashighestinItalyand France [8,15].Clearly,therewasanorth-southgradientinthestringency ofgovernmentresponse:Italy,FranceandPortugalimposedmore demandingpoliciesthanDenmark,Germany,theNetherlandsand theUK.Nevertheless,althoughpeopleinsoutherncountrieswere exposedtomoreseverecontainmentmeasures,theyapprovedof themmorethanpeopleresidinginnorthernstates,who experi-encedlessstringentrestrictions.

Turningnowtowithin-countryvariations,weobserved consid-erableheterogeneityofattitudestowardsmanypolicyresponses within individual countries with particularly marked differ-encesbetweenregions andagegroups inItaly, France,and the Netherlands.

Hereinafter,wegroupedregionsbasedontheseverityofthe COVID-19outbreakdistinguishingbetweenthemostandtheleast affectedareas.Noteworthy,Lombardydenotestheextrasample collectedinItalyandwasanalyzedseparatelyfromthe represen-tativeItaliansample.Overall,wedidnotﬁndsigniﬁcantdifferences inpolicysupportbetweenLombardyandtherestofItaly.

Toillustratewithin-countrydifferences,Fig.2reﬂectsregional andage-relatedheterogeneityofpublicopinionsinFranceandItaly towardbanningtheexportofmedicalequipment,suchasmasks.In fact,thismeasurewasbrieﬂyundertakenbyGermanyandFranceat theonsetofthepandemicinearlyMarch2020,leadingtopolitical tensionsbetweentheEUmemberstates.Germanydeclaredthatthe reasonwastoavoidshortagesofmasks,glovesandsafetyglasses withinthecountry,whereasFrancearguedthatthebanwasneeded fortheassessmentofinventoryandstoragecapacity[16].Following thecallforsolidarity,bothcountriesliftedthewithin-EUexportban onequipmentinmid-March[17].

Whilesupportforthispolicytendedtobesimilarinthemostand theleastseverelyaffectedpartsofItalyandFrance,theapprovalof theexportbanconspicuouslydifferedacrossagegroups.Older indi-vidualsapprovedmoreofthispolicythanyoungerpeople,which, besidesotherfactors,mayberelatedtothelevelsofworrypeople intheseagecategorieshaveabouttherisksthatCOVID-19posesto theirhealth.Wefoundthat51%ofFrenchand46%ofItalian respon-dentsagedabove65perceivedriskstotheirhealthfromCOVID-19 ashighorveryhigh,whilethecorrespondingshareamongpeople agedbelow25equaled30%inFranceand17%inItaly.

3.2. Worriesabouthealthandtheeconomy

ToaddressthementalhealthimplicationsoftheCOVID-19 out-breakandsubsequentcontainmentmeasures,weassessedlevels of worryprevailingin Europeansocietiesover severaldomains (health,economic,emotional,work,andfuture).Morespeciﬁcally, we addressed concerns about losing a close person, becoming unemployed,health systemgettingoverloaded, schoolclosures, small companies running out of business, recession, restricted accesstofoodsupplies,blackouts,andsocietygettingmore ego-istic.TheseitemswereadoptedfromtheWHOCOVID-19Snapshot Monitoringproject,whichwillallowfuturecomparisonswith

(4)

sim-Fig.1.Meansupportofgovernmentpolicies.

Table1

TotalconﬁrmedCOVID-19casesanddeaths(per1millionpeople)andgovernmentresponsestringencyindex.

Country Date

March12,2020 April12,2020 May12,2020 June12,2020a

Denmark Cases 89 1035 1815 2078 Deaths 0 45 91 102 Stringencyindex 37.96 72.22 65.74 62.96 France Cases 35 1437 2138 2383 Deaths 0.74 212 408 450 Stringencyindex 28.7 90.74 76.85 60.19 Germany Cases 19 1438 2035 2216 Deaths 0.04 32 90 105 Stringencyindex 32.87 73.15 64.35 50 Italy Cases 206 2519 3636 3906 Deaths 14 322 508 565 Stringencyindex 85.19 93.52 62.96 48.15 Netherlands Cases 29 1425 2497 2816 Deaths 0.29 154 318 353 Stringencyindex 41.67 79.63 68.52 62.96 Portugal Cases 6 1568 2715 3522 Deaths 0 46 112 148 Stringencyindex 32.41 87.96 75 71.3 UnitedKingdom Cases 7 1164 3286 4293 Deaths 0.1 171 472 608 Stringencyindex 11.11 75.93 75.93 70.37

(5)

Fig.2.HeterogeneityofpublicattitudeswithincountriesbyregionandagecategoryinFranceandItaly.

ilardatacollectedforothercountriesandatdifferentpointsintime [7].

We foundthat themean trendwas similarin allcountries: peopleworriedmostofallaboutthehealthsystemgetting over-loadedsothatthecapacitiescouldbecomeinsufﬁcienttocopewith thesurgein COVID-19cases.We observedthateven incase of householdsthathadnotbeendirectlyhitbythenovel coronavi-rus(above75%ofrespondentsinthetotalsample),thepandemic mighthaveactedasastressorcausinghealthandeconomic anxi-eties.

Fig.3presentspeople’sworryaboutselectedissuesacrossseven EUcountries(measuredonaLikertscalefrom1-notworryatallto 5-worryalot),wherethehigherintensityofcolorreflectsalarger shareofthepopulationwhoworry ¨quiteabit ör ä_lot¨.Cross-country differenceslooksubstantial,andanorth-southdivideintheworry causedbytheCOVID-19outbreakisconspicuous.

For instance, 84 % of respondents in Portugal and 81 % in Italy mentioned that they worried ¨quite a bit ör älot äbout the national health system becoming overloaded, while the corre-spondingsharesinDenmarkandGermanywere54%and62 %, respectively.Thesehealthconcernsmighthavereflectedthe devel-opmentof thepandemic.AsshowedinTable1,theprogressof theepidemichadanorth-southpatternwithmoreCOVID-19cases anddeathspermillionofthepopulationinsouthernstatesthan innorthern.TheexceptionwastheUK,wheretheepidemicwas thirddeadliestafterItalyandFrance,butgovernmentresponsewas lessstrictthanincountrieswithabetterepidemiologicalsituation [8,15].

Similarly,more peopleinPortugal andItaly wereconcerned withtheeconomicconsequencesofthepandemicthaninother Europeancountries. For example, 68% ofPortuguese and 56 % ofItalianswereworriedaboutlosingtheirjobs,whilerespective sharesintheNetherlandsandDenmarkwere27%and16%, corre-spondingly.

Thesecross-countrydifferencesineconomicanxietiesmaybe relatedtopeople’sperceptionsoftheeconomicandfinancial coun-termeasurestakenbytheirnationalgovernmentandtheEU.During thepandemic,Europeancountriesimplementedseveralfiscaland monetarymeasurestomitigatetheeconomicimpactofthe COVID-19 outbreak.Thesepoliciestypicallyincludedsupportofwages underthereduced-hourscheme,postponementoftaxpayments forcompanies,directfinancialsupportsandgrantstosmall enter-prisesandself-employed,theextensionofunemploymentbenefits, provisionofcapitalbufferstobanks,etc.[1].Nevertheless,there weresubstantialvariationsinthetimingandspecificcontentof thesecountermeasuresacrossthestates.

Tobrieﬂyoverviewthescaleofeconomicsupportprovidedby thegovernmentineachofthesevencountries, Table2 summa-rizesvaluesoftheeconomicsupportindex,acompositemeasure reﬂectingincomesupportanddebt/contractreliefprovidedbythe nationalgovernmenttohouseholds[8].Itismeasuredona0–100 scale,whereahighervaluereferstoamoresubstantialeconomic assistance.

Atthetime of thesurvey’sﬁeldwork,allcountriesprovided sometypeofeconomicrelieftotheirresidents.Nevertheless,the extentofsuchsupportwasconspicuouslydifferent:Franceandthe

(6)

Fig.3. Theproportionofrespondentswhoworry ¨quiteabit ¨or ¨a_lot¨.

Table2

Economicsupportindex.

Country Date

March12,2020 April12,2020 May12,2020 June12,2020a

Denmark 37.5 37.5 87.5 87.5 France 0 100 100 75 Germany 0 87.5 87.5 62.5 Italy 0 50 50 75 Netherlands 0 62.5 62.5 62.5 Portugal 25 75 75 50 UnitedKingdom 0 100 100 100

a_Or_closest_available_date.

UKrankedhighest,whileDenmark,Germany,andItalyranked low-est[8].Hence,itmaybepossiblethathigherlevelsofeconomic concernsinsomecountriesindicatedpeople’sbeliefsinthe insuf-ﬁciencyofthegovernment’sresponse,whichwillbesubjecttothe analysisinthenextwavesofthesurvey.

Moreover,thecompositionofemploymentvariesacrosstheEU, especiallyintermsofinformalandtemporaryemployment. Tem-porarycontractsprovidelowerlevelsofsocialprotectionandjob securitytoemployees,buttheirprevalencehasincreasedoverthe lastyears,particularlyintheNetherlands,Italy,andFrance.Asof 2019,theshareoftemporaryemployeesinthetotalnumber of employedwashighestinsouthernEuropeancountries:Portugal (17.4%),France(13.3%),andItaly(13.1%).Incontrast,itwas sig-niﬁcantlylowerinnorthernstates:theUK(3.8%),Denmark(8.3 %),andGermany(9.3%).TheonlyexceptionwastheNetherlands,

wheretemporaryworkersconstituted13.6%ofallemployees[18]. Thus,suchdifferencesintheemploymentcompositionmaybein partresponsibleforthecross-countrydissimilaritiesineconomic concerns.

Wealsoobserveddifferencesinthelevelsofconcernwithin individualcountries. Fig.4showstheextentofworryaboutthe healthsystemandarecessioninItaly.Wegroupedregionsbased ontheseverityoftheCOVID-19outbreakanddistinguishedthe levelsofanxietyacrossagecategories.Higherintensityofthecolor reﬂectsagreaterextentofworry.

Overall,thelevelofworryinthehighlyaffectedregionsofthe countrywasnot higher than elsewherein Italy, except forthe youngestagegroup.However,economic concernstended tobe unequallydistributedacrosstheagegroups.Forinstance,worries abouttherecessionandsmallcompaniesrunningoutofbusiness

(7)

Fig.4. Heterogeneityofpeople’sworriesinItalybyregionandagecategory.

werehigheramongolderindividualsthanyoungeragecohort.This patternwassimilarinallcountriescoveredbythesurvey. 3.3. Trustinsourcesofinformation

Weaskedpeopleaboutthemainsourcesofinformationfrom wheretheyreceivednewsaboutCOVID-19.Thedatashowthat overall 94 % of respondents closely followed the news onthe situationwithCOVID-19,implyingahighlevelofpublic aware-ness.Regardingthesourcesofinformation,86%ofrespondents mentioned receiving updates from the TV and 50 % addition-allysearchedforinformationontheInternet.Presumably,reliable informationpresentedthroughthetelevisionemergedasthebest channeltoreachthepopulationatlarge.

Next,weassessedtheextentofpeople’strustintheinformation receivedfromvarioussourcesinthecontextoftheCOVID-19 situa-tion.Thetrustinthefollowinginformationsourceswasaddressed: nationalgovernment,theEU,theWHO,hospitalsandGPs,national newschannelsandnewspapers,socialmedia,relativesandfriends. Fig. 5 shows mean values of trust in information from six selected sources across seven Europeanstates (measured on a Likertscalefrom1-notrust atallto5-trustverymuch).Higher intensityofthecolorreﬂectsahigherleveloftrustinthe informa-tionfromaspeciﬁcsource.

Thedata showthat overall peoplehad thehighest levelsof trustininformationfromhospitals,familydoctors,andtheWHO, followedbyinformationfromthenationalgovernmentandmain nationalnewschannels.Thisrankingofsourcesbytrustwas simi-larinallcountriescoveredbythesurvey,exceptforFrance,where citizenshadahighlevelofconﬁdenceonlyinhealthcareproviders andplacedrelativelylittletrustinallothersources.

Moreover,anorth-southdividecouldbenoticedinthelevelof trustininformationfromthenationalgovernment.Trustwas high-estinDenmarkandtheNetherlands(morethan70%ofrespondents

trusted ¨much ¨or ¨verymuch¨),whereasitwaslowestinFrance(27% ofrespondentshadahighleveloftrust).

Furthermore,asimilarnorth-southgradientwasobserved con-cerningthetrustintheEU:trustwashighestinDenmark(45%), Germany(40%),theNetherlands(39%)andtheUK(35%),whereas itwaslowestinItaly(24%) andFrance(21%).Portugalwasan exceptiontothiscasesincethecorrespondingvaluehere consti-tuted46%.

Finally,wealsoobservedconsiderableregionalheterogeneities inlevelsoftrustwithincountrieswithparticularlynoticeable dif-ferencesacrossindividualregionsinItaly,France,andGermany. Fig.6showspeople’strustininformationfromthenational gov-ernmentinthecontextofCOVID-19inGermanyandFranceasan example,wherethehigherintensityofthecolorindicatesagreater extentof trust. Whiletrust didnot differsigniﬁcantly between regionsgroupedwithrespecttotheCOVID-19severity,itwas het-erogenousacrosstheagegroups.

Althoughthesurveyaskedabouttheleveloftrustininformation fromdifferentsourcesinthecontextoftheCOVID-19situationand notabouttheoveralltrustininstitutions,thesetwoarelikelytobe related.Generally,trustreﬂectspeople’sperceptionsofwhether institutionsaredoingwhatisright.Thus,trustintheinformation theyprovidecanbeconsideredanindicatoroftheconﬁdencethat citizenshaveintheseinstitutions[19].

4. Policyimplications

TheCOVID-19pandemicraisednewchallengesfor policymak-ersacrosstheEU.Theimminentthreattopublichealthattheonset ofthepandemicledmostgovernmentstoimposealockdownon society.However,asthepeakofthepandemicabated,thefocusof attentionturnedtothesocialandeconomicconsequencesofthe containmentmeasures.Giventhatwithoutacquiredherd immu-nitytheriskofanewwaveoftheepidemicremainshigh,andthe

(8)

Fig.5.MeantrustininformationsourcesinthecontextofCOVID-19situation.

(9)

productionanddistributionofvaccinesmaytake12–18months [20],governments musttrytostrike therightbalancebetween effectsonpublichealth,sociallifeandtheeconomywhen consid-eringpossibleexit-strategiesfromthecurrentlockdownsituation. Intheabsenceofmedicalintervention,policymakersandpublic healthofﬁcials mustresorttonon-medicalbehavioral interven-tions. Lifting the lockdown requires that citizens support and adheretothepolicymeasuresthataimtocontainthespreadof thevirusassocialandeconomicactivitygraduallyrestarts.Given thedifﬁcultyofenforcingsuchregulations,futuremeasuresneed tobebothwell-designedandwell-communicatedtothepublic. Themorepeoplearewillingtocomplyvoluntarilywiththenew measures,theless enforcementand supervisionwillbeneeded to achieve high compliance. For this, people’s perceptions and attitudesneedtobefactoredin atthepolicy-designand imple-mentationstages.

Oursurveysoughttocapturethepublicsentimenttoward mea-surespreviouslytakenbypolicymakerstocontainCOVID-19and addressedpeople’ssupportforpolicies,worriesaboutthe conse-quencesofCOVID-19,andtrustindifferentsourcesofinformation. Theﬁrstinsightsobtainedfromthedatashowedthatcontainment and mitigating policiesundertaken bynational governments in responsetotheinitialstagesoftheCOVID-19pandemicwere gen-erallywell-receivedbythepopulationinallcountriescoveredby thesurvey.Nevertheless,theextentofapprovalvariedacrossstates andspeciﬁcpolicymeasures.

Severallessonscanbedrawnforthedesignandimplementation ofpoliciesfortheprolongationorgradualremovaloflockdown restrictions.

First, we observed a north-south divide in people’s percep-tions,worriesandtrustacrosstheEuropeancountries.Thisfinding suggeststhat furthercontainment measuresand lockdownexit strategiesneedtobebalancedagainstthefactorsthatworry peo-pleineachspecificcountry.Onenoteworthyexampleisthelevel ofimportancethatpeopleinEuropeancountriesattributetothe conceptsofindividualfreedomandprivacy.Usingmobiledatafor trackingCOVID-19casesandtheircontactsmaybeacontroversial decisiontotakeeventhoughitisbelievedbymanyexpertstobea usefultooltomanagetheCOVID-19outbreak.Theeffectivenessof thispolicycriticallydependsonasufficientlevelofadoptionofthe technologybythepopulation[8].Ourdatasuggestthatthismay notbeachievedeasilyinsomeEuropeancountries.

Acleartakeawayisthatanopendialoguewithsocietyonthis matterisneeded.Explainingtheneedforandtheadvantagesof suchintrusivepoliciesthroughtrustedmeansofcommunication, whileaddressingpeople’sconcernsexplicitlyandbeingopenabout therisksofusingsuchpolicymeasuresmayhelpraisethesupport andcomplianceinsocietytoasufﬁcientdegree.

Anothercriticalissueisthebalancebetweensavinglivesand savinglivelihoods.According tothesurvey, peoplein southern Europeancountriesare substantiallymore concernedaboutthe economicaspectsoftheCOVID-19outbreakthanpeoplein north-ernEuropeancountries. Economicanxieties,ifleftunaddressed, mayhaveadverseeffectsonthementalhealthandwellbeingof thepopulation,aswellascausedownwardadjustmentsin con-sumptionbehavior,therebyexacerbatingtheeconomicsituation inacountryiftherecessionindeedhappens.

Second, we found considerable heterogeneities in people’s approvalofpolicieswithinindividualcountries.Thistendencywas particularlynoticeableinFranceandItaly.Onepossible determi-nantofregionaldifferencesinpublicsupportcouldbetheextentof thedevolutionofdecision-makinginthecountry.Ontheonehand, devolutioncouldenableregionalorlocalauthoritiestomakebetter decisionsduetotheirbetterawarenessofregion-speciﬁc circum-stances.Ontheotherhand,itcouldharmthecoordinationofpolicy responsesbetweenthecentralandregionalauthoritieswithin

indi-vidualcountries.Thus,itiscrucialtounderstandthedeterminants ofsuchdifferencesandaddressthemtosecurepublicsupportof futurepoliciesandensurehighcompliancewithgovernment mea-sures.

Furthermore, our results showed that the burden of stress tendedtobeunequallydistributedacrossandwithincountries. Evenincaseofhouseholdsthatwerenotdirectlyhitby COVID-19,thepandemicmayhaveactedasastressorcausinghealthand economicanxieties.Suchworriesmaybedetrimentaltoindividual mentalhealthandwellbeing,andtheymaybecomefurther exac-erbatedbytheimpositionofself-isolationpolicies.Thus,itmaybe reasonabletoconsideranasymmetricapproachtothedesignof exitstrategiestakingregion-speciﬁclevelsofsupportandworry intoaccount.Thisincludestheidentiﬁcationofvulnerable cate-goriesofthepopulationnotonlyintermsofhealthrisksbutalso withrespecttosocialandeconomicactivities,andaddressingtheir concernssatisfactorily.

Third,duringapandemic,publictrustinthegovernmentand theinformationitprovidesisofparamountimportance.Toexpect highcomplianceoverextendedperiodsoftime,policymakersneed toadopteffectivestrategiesandmeansofcommunicationwhereby securingasufﬁcientleveloftrustandconﬁdencefromthesociety. Asourresultssuggest,somecountriesweremoresuccessfulinthis respectthanothers.

Societyneedstobewell-informedaboutthedilemmasfacedby policymakers,andforthis,thecommunicationbetweenthe gov-ernmentandthecitizensmustbeclearandtransparent.Thedata showedthat94%ofrespondentscloselyfollowedthenewsonthe situationwithCOVID-19mainlyusingtelevisiontokeep them-selvesupdated.Thus,televisionemergedasthebestchannelto reachthepopulationatlarge,suggestingthatpresentingreliable informationthroughthismeansisaneffectivestrategytofollow.

Nevertheless,giventhatthedatashowregionalandage-related heterogeneitiesintrustandpolicysupport,itmaybeworth tai-loringmessagesandmeansofcommunicationtospeciﬁcgroups ofthesociety.For example,cooperationwithpublicﬁguresand well-knownexpertscanbeusedtodelivergovernmentandpublic healthmessagesinasimplelanguage,orlocalvoicescouldbeused toamplifysuchmessagesinindividualregionsofthecountry.

Overall,informationprovision,publiceducationandeffective communicationstrategiesshouldbeamongthekeyguidelinesfor policymakerswhen implementing exitstrategies and designing future containmentmeasuressothat thesepolicieshavepublic supportandhighcompliance.

AdditionalwavesofthesurveyarescheduledinJuneandAugust 2020.Thiswillallowustoinvestigateinmoredetailhowthe pop-ulationcopeswiththehealth,socialandeconomicconsequences oftheCOVID-19pandemicasthesituationevolves.

Authornote

Allauthors contributed tothe manuscript. IS, SNB and NEV equallycontributedtothedesignandimplementationofthe sur-veywithexpertiseandfeedbackfromPPB,WB,JE,JSandTS.IS,SNB, NEV,PPB,WB,JE,JSandTSanalyzedthedata.ISwrotetheoriginal draft,PPB,WB,JE,JS,TS,SNBandNEVreviewedandedited.

DeclarationofCompetingInterest

None.

Acknowledgements

ThisprojecthasreceivedfundingfromtheEuropeanUnion’s Horizon2020researchandinnovationprogrammeundertheMarie

(10)

Skłodowska-CuriegrantagreementNo721402,theworkwas sup-portedbyfundingundertheExcellenceStrategybytheGerman federalandstategovernments,aswellasbytheUniversityof Ham-burg,ErasmusUniversityRotterdam,andNovaSchoolofBusiness &EconomicsLisbon–ChairBPI|“Fundac¸ãoLaCaixa”onHealth Economics.

Wethankourcolleaguesfortheirfeedbackandworkonthe adoptionofthesurveytonationalcontexts:HelenBanks,Joana Pestana,MaartenHusen,LaurieRachetJacquet,NicolaiFink Simon-sen.

References

[1]IMFpolicytracker;2020.Retrievedfrom https://www.imf.org/en/Topics/imf-and-covid19/Policy-Responses-to-COVID-19AccessedonApril24,2020. [2]FlaxmanS,MishraS,GandyA,UnwinHJT,CouplandH,MellanTA,etal.

Estimatingthenumberofinfectionsandtheimpactofnon-pharmaceutical interventionsonCOVID-19in11Europeancountries.2020;2020,http://dx. doi.org/10.25561/77731.

[3]Dolow-trustsocietiesdobetterinapandemic?TheEconomist2020,2020, May2)Retrievedfrom https://www.economist.com/europe/2020/05/02/do-low-trust-societies-do-better-in-a-pandemic.AccessedonMay7,2020. [4]LockdownfatiguehitsasEuropeenforcescoronavirusrestrictions.

Finan-cialTimes;2020(2020,April1)Retrievedfromhttps://www.ft.com/content/ eb81dc96-c2bd-4c1c-ac8a-bf88653b0b8d.AssessedonApril24,2020. [5]GraedenE,CarlsonC,KatzR.Answeringtherightquestionsforpolicymakers

onCOVID-19.TheLancetGlobalHealth2020;8(6):e768–9,http://dx.doi.org/ 10.1016/s2214-109x(20)30191-1.

[6]Reopeningsmarkanewphase:global’trial-and-Error’playedoutinlives. TheNewYorkTimes;2020,2020,May)Retrievedfromhttps://www.nytimes. com/2020/05/07/world/europe/coronavirus-reopening-costs.html. Accessed onMay8,2020.

[7]WHORegionalOfﬁceForEurope.COVID-19SnapshotMOnitoring(COSMO Standard):Monitoringknowledge,riskperceptions,preventivebehaviours, andpublictrustinthecurrentcoronavirusoutbreak-WHOstandardprotocol. PsychArchives2020,http://dx.doi.org/10.23668/psycharchives.2782. [8]HaleT,WebsterS,PetherickA,PhillipsT,KiraB.OxfordCOVID-19government

responsetracker.BlavatnikSchoolofGovernment;2020.Retrievedfromwww. bsg.ox.ac.uk/covidtrackerAccessedonJune11,2020.

[9]European Commission, 2016, December 06) Retrieved from https://ec. europa.eu/home-affairs/what-we-do/policies/borders-and-visas/schengen/ reintroduction-border-controlen Accessed on June 11, 2020 Temporary reintroductionofbordercontrol;2016.

[10]OECD, Retrieved from https://www.oecd.org/coronavirus/policy-responses/ tracking-and-tracing-covid-protecting-privacy-and-data-while-using-apps-and-biometrics/ Accessed onMay 6, 2020 Tracking and tracing COVID: protectingprivacyanddatawhileusingappsandbiometrics;2020. [11]Howwillgovernmentsknowwhentoliftrestrictions?Covid-19Health

Sys-temResponseMonitor;2020(2020,May7)Retrievedfromhttps://analysis. covid19healthsystem.org/index.php/2020/05/07/how-will-governments-know-when-to-lift-and-impose-restrictions/AccessedonMay10,2020. [12]Europeanmobileoperatorssharedataforcoronavirusﬁght.Reuters;2020

(2020,March18)Retrievedfrom https://www.reuters.com/article/us-health- coronavirus-europe-telecoms/european-mobile-operators-share-data-for-coronavirus-ﬁght-idUSKBN2152C2AccessedonJune12,2020.

[13]EUpush forcoronaviruscontacttracingsufferssetback.DeutscheWelle; 2020(2020,April23)Retrievedfrom https://www.dw.com/en/eu-push-for-coronavirus-contact-tracing-suffers-setback/a-53219372AccessedonJune12, 2020.

[14]IencaM,VayenaE.Ontheresponsibleuseofdigitaldatatotacklethe COVID-19pandemic.NatureMedicine2020;26(4):463–4,http://dx.doi.org/10.1038/ s41591-020-0832-5.

[15][dataset]OurWorldinData,Retrievedfrom https://github.com/owid/covid-19-data/tree/master/public/data/ AccessedonJune11,2020Completeour worldindataCOVID-19dataset;2020.

[16]E.U.Seekssolidarityasnationsrestrictmedicalexports.TheNewYorkTimes; 2020(2020,March7)Retrievedfromhttps://www.nytimes.com/2020/03/07/ business/eu-exports-medical-equipment.htmlAccessedonJune11,2020. [17]BretonT(2020,March15)Retrievedfromhttps://twitter.com/thierrybreton/

status/1239159582821408770?lang=en Accessed on June 11, 2020 After intense discussions,I welcomethatDEandFR now allowfor exportof #COVID19protectiveequipment.IwillcontinuetofollowsupplytoITclosely. NosingleEUcountrycanwinthisbattlealone.#cooperation#solidarity;2020. [18][dataset] Eurostat (2020, April 20) Retrieved from https://ec.europa. eu/eurostat/databrowser/view/tesem110/default/table?lang=en.Accessedon June11,2020Temporaryemployeesaspercentageofthetotalnumberof employees;2020.

[19]OECD.Trustingovernment,policyeffectivenessandthegovernanceagenda. In:Governmentataglance2013.Paris:OECDPublishing;2013,http://dx.doi. org/10.1787/govglance-2013-6-en.

[20]TheraceagainstCOVID-19.NatureNanotechnology2020;15(4):239–40,http:// dx.doi.org/10.1038/s41565-020-0680-y.