multidimensional inequality and covid-19 in brazil · public policy brief, no. 153 2 contents 3...

Levy Economics Institute of Bard College

Public Policy BriefNo. 153, 2020

MULTIDIMENSIONAL INEQUALITY AND COVID-19 IN BRAZIL

LUIZA NASSIF-PIRES, LAURA CARVALHO, and EDUARDO RAWET

Public Policy Brief, No. 153 2

Contents

3 Preface DimitriB.Papadimitriou

4 Multidimensional Inequality and COVID-19 in Brazil LuizaNassif-Pires,LauraCarvalho,andEduardoRawet

18 About the Authors

The Levy Economics Institute of Bard College, founded in 1986, is an autonomous research organization. It is nonpartisan, open to the examina-tion of diverse points of view, and dedicated to public service.

The Institute is publishing this research with the conviction that it is a constructive and positive contribution to discussions and debates on relevant policy issues. Neither the Institute’s Board of Governors nor its advisers necessarily endorse any proposal made by the authors.

The Institute believes in the potential for the study of economics to improve the human condition. Through scholarship and research it generates viable, effective public policy responses to important economic problems that profoundly affect the quality of life in the United States and abroad.

The present research agenda includes such issues as financial instability, poverty, employment, gender, problems associated with the distribution of income and wealth, and international trade and competitiveness. In all its endeavors, the Institute places heavy emphasis on the values of personal freedom and justice.

Editor: Michael StephensText Editor: Elizabeth Dunn

The Public Policy Brief Series is a publication of the Levy Economics Institute of Bard College, Blithewood, PO Box 5000, Annandale-on-Hudson, NY 12504-5000.

For information about the Levy Institute, call 845-758-7700, e-mail [email protected], or visit the Levy Institute website at www.levyinstitute.org.

The Public Policy Brief Series is produced by the Bard Publications Office.

Copyright © 2020 by the Levy Economics Institute. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information-retrieval system, without permission in writing from the publisher.

ISSN 1063-5297

ISBN 978-1-936192-69-4

Levy Economics Institute of Bard College 3

not explain the observed disparities in death rates in Brazil, for instance. For a more comprehensive picture of relevant inequali-ties, the authors focus on the persistence of unequal access to healthcare and variations in the prevalence of comorbidities, both of which contribute to the severity of illness and number of deaths. The authors also reveal that, in terms of the gaps in observed rates of infection and hospitalization for COVID-19, racial inequality appears to have played a more significant role than income inequality.

Nassif-Pires, Carvalho, and Rawet note that Brazil entered the pandemic on the heels of slow GDP growth, high unemploy-ment, and rising inequality—along with ongoing implementa-tion of an austerity program spurred by a new, constitutionally enshrined ceiling on the growth of federal expenditure. With the COVID-19 emergency, those fiscal restraints were temporarily lifted: Brazil mounted a fiscal response equivalent to 6.5 percent of GDP (in health and nonhealth measures), with almost half of this response dedicated to a cash transfer program, Auxílio Emergencial. The authors find that, for the bottom half of the income distribution, this program raised incomes by more than the crisis-induced fall in wages. As a result, poverty was reduced to its lowest level ever recorded and, so far, the impact of the crisis in terms of rising income inequality has been neutralized.

Amidst uncertainty over how long the Auxílio Emergencial will be extended, rising bankruptcies among small firms, and rumblings of a return to an austerity regime, the authors stress that the current fiscal measures must be both retained and enhanced (they recommend more action on credit measures for businesses). Moreover, the effects of structural inequalities on the country’s infection rates and death count make it clear that broader policy changes are necessary for addressing other dimen-sions of inequality, particularly those rooted in structural racism. As always, I welcome your comments.

Dimitri B. Papadimitriou, PresidentSeptember 2020

As major global crises often do, COVID-19 has exposed coun-tries’ political, policy, and socioeconomic fault lines and vulner-abilities. In a previous public policy brief, Luiza Nassif-Pires led a study examining the feedback loops between the pandemic and racial, gender, and income inequalities in the United States (Public Policy Brief No. 149, “Pandemic of Inequality”). For this current policy brief, Nassif-Pires, Laura Carvalho, and Eduardo Rawet use a similar analytical framework to examine Brazil’s experience with COVID-19—a country whose high income inequality and history of slavery predispose it to share similar structural weaknesses in the face of the pandemic.

Nassif-Pires, Carvalho, and Rawet note that Brazil is suffer-ing from some of the worst per capita numbers in the world in terms of cases and deaths, and they explore here how yawning racial, regional, and class disparities can help account for why COVID-19 has had such a deleterious impact on the Brazilian population. Although they find that fiscal policy has so far been successful at mitigating the impacts of the crisis with respect to wage inequality, the existence of structural inequalities along racial lines in particular have resulted in the public health burden of this pandemic being unequally borne.

The authors construct an index to measure the social bases of vulnerability to the virus, focused mainly on risks driven by living and working conditions such as informal employment or cramped living arrangements. The index reveals significant dis-parities in the risk of infection that break down along lines of race, region, income, and education. Moreover, the overlap of racial inequalities with income or educational inequalities exacerbates these disparities—the authors find this to be especially conspicu-ous with respect to the intersection of race and low educational attainment. They note that infection, hospitalization, and death rate microdata targeting these intersections of race, gender, and class would help better guide effective public policy.

Nassif-Pires, Carvalho, and Rawet find that the disparate public health impact of the pandemic reflects the inequalities identified by their vulnerability index, particularly with respect to infection rates. Nevertheless, the authors explain that the index only reveals part of the story. Social vulnerability alone does

Preface


IntroductionSincethe2008globalfinancialcrisis,incomeandwealthinequal-itieshavegainedrenewedattention in theeconomic literatureandwiderpolicydebates.Giventheeconomicandpoliticalcostsofthebroadlyacknowledgedriseofincomeconcentrationatthetopof thedistributionsince the1980s,economistsandpoliti-ciansinthepastdecadehaveputforwardvariousinterpretationsaswellasproposalsforreducingthegapbetweentheveryrichandtherestofthepopulation.However,noneofthesediscus-sionsseemtohavepreparedoursocietytobattlethedevastat-ing consequences of inequality during the COVID-19 crisis.On the one hand, inequality aggravates the pandemic, as thewidegapbetweentherichandpoor—intermsofincome,typeof employment, living conditions, access to health, and otherdimensions—hasmajorconsequencesforthedistributionofthedeathtollwithinandbetweencountries.Ontheotherhand,thepandemicexacerbatesinequalitybywideningthisgapthroughitsdeepeconomicandsocialimpacts.

Based on data from 175 countries after five significantepidemics—SARS (2003),H1N1 (2009),MERS (2012),Ebola(2014), and Zika (2016)—a study by Furceri, Loungani, andOstry (2020) suggests that these episodes have contributedto raising income inequality by almost 1.5 percent in the fivesubsequentyears.ThiseffectmaybesubstantiallylargerintheCOVID-19pandemic,withhealthandeconomicburdensdis-proportionatelylaidonthoseatthebottomofthedistribution.First,themostvulnerablearemorepronetobeinfectedbythevirus,duebothtotheneedtocontinueworkinginpersonandto inequalities in livingconditions.Second,precarioushealth-careandtheunequaldistributionofcomorbiditiesplayaroleinexplainingwidedisparitiesintheseverityofcasesandthenum-berofdeaths.Third,thelossofincomegeneratedbythecrisisseems todisproportionatelyaffect self-employedand informalworkers,aswellaslower-skilledemployeesintheservices,retail,andconstructionsectors.

Hence,afteraninitialperiodinwhichhigh-incomecoun-trieswere theepicenterof theCOVID-19pandemic,develop-ingcountriesnowaccountformorethanhalfofglobaldeaths.AstudybyMurrayetal.(2006)suggeststhatmortalityratesduringthe1918–20flupandemicwereupto30timeshigher inpoorregionsoftheworld.Simonsenetal.(2013)showthatduringtheH1N1pandemicin2009,mortalitywas20timeshigherinSouthAmericathaninEuropeancountries.In2020,LatinAmericancountriesareattractingworldwideattentionfortheirinabilityto

fightthecoronavirus.InAugust,theLatinAmericandeathtollpassed200,000,whileBraziltopped100,000deaths,rankingsec-ondintheworldinabsolutenumberofdeaths.Ifthiswerenotenough,theIMFprojectsafallof9.3percentinLatinAmerica’sGDPin2020—anumberthatmakesthe4.9percentcontractionprojectedforglobalGDPlooklikeamildrecession.

In addition to the ineffectiveness of lockdownmeasures,wide structural inequalities, high levels of informality in thelabormarket,andthe importanceof theservicesandtourismsectors in these economiesmay help explain these disastrousresults.Moreover,theregionwasexperiencingaperiodofslowgrowthandthusfacedhighlevelsofunemploymentpriortothepandemic.Inthiscontext,manyofthesecountrieslackedthefis-calspacetoreactproportionately:asofMay2020,morethan$1trillionhadalreadybeenobtainedasloansfromtheIMFtofightthecrisisinLatinAmerica.

BythebeginningofAugust,countrieslikePeru,Chile,andBrazilhadthefiscalspacetospendmorethantheirneighborsbuthavenonethelesspresentedsomeofthehighestnumbersintheworldwith respect todeathsper100,000people.1 Incon-trast,Uruguayranked125thwhilespendinglessthan2percentofGDPtofightthepandemic.Whileotherdifferencescertainlyhaveplayedarole,Uruguayisknownforitsrelativelylowlevelofincomeinequalityintheregion:the2018GiniindexforBrazilwas53.9andonly39.7inUruguay(WorldBank2018).

High inequality may contribute to explaining why, sincemid-June 2020, Brazil has had the second-highest number ofcasesintheworld,evenafterspendingmorethan6percentofGDP in fiscalmeasures to fight theCOVID-19 crisis. Settingaside the antisciencediscourseof the federal government andtheoveralldisastrousapproachonthehealthfront,thenextsec-tionwillexaminethecountry’swideinequalitiesasariskfactorinthepandemic.ThefollowingsectionwillbuildapreliminaryanalysisoftheunequallydistributedeconomicandhealthcostsoftheCOVID-19crisisinBrazil.Thefinalsectionconcludesthepolicybrief.

Inequality as a Risk Factor for COVID-19: Measuring Social Vulnerability in BrazilSeveraldimensionsofstructuralinequalitiescanbeidentifiedaspotentiallyincreasingtheriskofinfectionanddeathduringthepandemic.First, low-incomepopulationsaremoreexposedtocontamination.Thisisduetothedifficultyofadheringtoquar-antinemeasures,as theycannot forgotheir labor income.For


thosewhocontinuetowork,theuseofpublictransportationandthehighconcentrationofjobsintheservicesectormakessocialdistancingdifficult.Moreover,theexposuretocontaminationisgreaterforlow-incomepopulationsevenwhentheyareinquar-antine,astheydonotnecessarilyhaveaccesstomodernplumb-ingandliveinmorecrowdedspaces.

Second, access to healthcare is not evenly distributed inBrazil.Thisisduebothtoregionaldisparitiesandtotheexis-tenceoftwohealthcaresystems:aprivateandauniversalpublicsystem.Althoughthetotalnumberofhospitalbedsinbothnet-worksissimilar,thelatterservesthemajorityofpopulation(71percent)anditsbedsareunequallydistributedacrossregions.

Third, low-income populations infected with the novelcoronavirustendtohaveworseoutcomes.Studieshaveshownthat the incidence of comorbidities that are associated withhigher rates of hospitalization and deaths, such as diabetesandhypertension, ishigheramong low-incomeandracializedpopulations(PrestonandTaubman1994;Margolisetal.1992;Gaskin,Thorpe,andMcGinty2014;Viacavaetal.2019;Maltaetal.2019;Leiteetal.2015).

Inthissection,wewilldiscussindetailthesethreeaspectsofinequalitiesinBrazilandconstructasocialvulnerabilityindexthatcanhelp shed lighton the rootsof theunequalobservedhealthoutcomesbyregion,income,andraceinBrazilduringtheCOVID-19pandemic.

Inequality and the Risk of InfectionTheCOVID-19epidemic spreadquickly, releasinga scientificracetounderstandtheeffectsofthevirus.Bioscientistsandmed-icalauthoritiesascertainedthatafewfactorsincreasedtheriskofcontractinganddyingofCOVID-19:age,sex,andunderly-inghealthconditions.Epidemiologistswarnedthatlargesocialgatherings,poorhygiene,andclosedspaceswouldleadtomoreinfections.SocialscientistscautionedthatstructuralinequalitiesincreasedtheriskofminoritiesandpoorpopulationsgettingsickanddyingfromCOVID-19.Thewarningsofmedicalauthori-tiesandepidemiologistswere turned intoguidelinesandpoli-cies,butthewarningsbysocialscientistsweremostlyignoredbyauthoritiesandthevirus’sarrivalintounequalsocietiesuncov-eredachallenge.Policiesthattreatedeveryoneasequalsexac-erbatedstructuralinequalitiesandrevealedthediscrepanciesinliving,working,economic,andhealthconditions.

Data on the evolution of the virus’s spread by neighbor-hoodinNewYorkCitysooncorroboratedthatsocioeconomic

characteristics were responsible for stark differences in infec-tion,hospitalization,anddeathrates.AsoftheendofJuly2020,datafromtheNYCDepartmentofHealthandMentalHygieneshowed that Latino and black populations’ infection rates arearound 1.6 times higher, with hospitalization and death ratesaround two-times higher than those of the white population.Furthermore, infection, hospitalization, and death rates arehigher among the poor.According to theCenters forDiseaseControl and Prevention (CDC 2020), “inequities in socialdeterminants of health put racial and ethnicminority groupsat increased risk of getting sick and dying from COVID-19.”Withinsuchinequities,theCDCciteseducational,income,andwealthgaps, aswell asdifferences inaccess tohealthcare, joboccupation,andhousingconditions.

Indeed, studiesonprevious respiratory infectionepidem-ics (1918–20 flu, H1N1, and SARS) have shown that socialinequalitiesareadeterminantfortherateoftransmissionandseverityofthesediseases(CordobaandAiello2016;Mamelund2017; Tricco et al. 2012; Bengtsson 2018; Bucchianeri 2010).Multidimensionalpovertyisresponsibleforthefactthatthoseliving near the poverty line did not have themeans to avoidinfection.Structuralracismplaysanimportantroleinexplain-ingwhyminoritiesarealsoparticularlyvulnerabletoinfections.They are overrepresented in essential jobs andmore likely tobedependentonpublictransportation.Minoritiesinbigcitiesareconcentrated inneighborhoods that experienceoutbreaks,and the higher population density and likelihood of sharingsmallerhouseswithmorepeoplethereforeincreasescontagionwithinthefamily.Furthermore,onceinfected,poorandminor-itypopulationsaremorelikelytohaveworseoutcomesduetoahigherprevalenceofcomorbiditiesandmoreprecariousaccesstohealthcare(Nassif-Piresetal.2020).

TheUnitedStatesandBrazil share twocrucial character-istics:highincomeinequalityandahistoryofslavery.ItisthusexpectedthatraceandlowincomeinBrazilwouldalsoberiskfactorsforCOVID-19infection.Toinvestigate,webuildanindextomeasureindividuals’socialvulnerabilitytothevirususingtheNationalHouseholdSampleSurvey(PNAD-Contínua).Wecon-structbinaryvariables to indicatetheriskof infectionaccord-ingtolivingandworkingcharacteristics.Thework-relatedriskfactorsconsidered foran individualare:employment ina jobthathasbeendeemedessentialbythefederalgovernment,beinginformallyemployed(nothavingcarteiraassinada),2notowningacaroramotorcycle,andnothavinginternetaccess.Regarding


livingconditions,weconsiderlivinginahousewithmorethanthreeoccupantsperbedroomorwithnoplumbingandsewagesystemtobearisk.Wecomputethesocialvulnerabilityindexasthesumofriskfactorsforeachindividual.

According to our social vulnerability index, the averageBrazilian has a score of 1.53, though there are large regional,racial, gender, and educational discrepancies in the values.NorthandNortheastregions,whicharethepoorestandmostunequalones,3haveanaveragesocialvulnerabilityindexhigherthan the national average (2.35 and 1.92, respectively), whilethose living in theSouth,Southeast, andMidwest regionsareless vulnerable than the average Brazilian according to ourindex.Thevalueofthesocialvulnerabilityindexalsodecreaseswitheducationalattainment,asthosewhohavelessthanahighschooldegreearefoundtobemorevulnerablethantheaverageBrazilian.Forthosewithacollegedegreeandabove,theindexisapproximatelytwiceassmall(0.98)thanforthosewithoutanyformaleducation(1.86).

Table 1 presents the averages of the social vulnerabilityindex by race and sex. The Brazilian Institute of GeographyandStatistics(IBGE)providesfiveraceclassifications:Brancos,Pretos, Amarelos, Pardos, and Indígenas, here translated aswhite,black,Asian,brown,andindigenous.Thosecorrespondrespectivelyto46percent,9percent,1percent,44percent,and0.4percentofBrazil’spopulation.ThetermPardoaggregatesthevastmajorityofAfricanandnativeBraziliandescendents.4

Forallracialgroups,womenarelessvulnerablethanmen.Althoughtheaveragewomanislessvulnerablethantheaverageperson,thisisnottrueforblack,indigenous,andbrownwomen.Whenwelookattheindexbyrace,onlywhiteandAsianpersonsarelessvulnerablethanaverage.

Lower infection ratesamongwomen,higher ratesamongminorities, and lack of intersectional data on cases poses a

difficulty in inferring the health impacts of COVID-19 onwomenofcolor.OurresultsfortheintersectionofsexandraceshedsomelightonthisdiscussionandstresstheimportanceoftakingtheunequalhealthburdenofCOVID-19onwomenofcolorintoaccount.

We now turn our discussion to the intersection betweenclassandrace.Ithasbeenestablishedthatlow-incomeandracial-izedpopulationsareatincreasedriskofbeinginfectedanddyingfromCOVID-19,butstructuralracismmakes it impossible toseparatethesefactors.Figure1presentsthemeanofthesocialvulnerability indexby incomepercentilewith information fortworacialgroupings:black,brown,andindigenous—whichintheaggregateexperienceabove-averagerisk—andwhite,whichareculturallydominantinBraziliansociety.Incomepercentileswerecalculatedfrompercapitahouseholdincomereportedinthesurvey.5Thefirsttwointerestingaspectstonoticearethat,except for the fifthdecile, theaveragevulnerability index fallssteadily with income, and that black, brown, and indigenouspopulationsareathigherriskthanwhitepopulationsineveryincomepercentile.TheaverageBrazilianinthebottom60per-centoftheincomedistributionpresentsariskabovethenationalaverage and race seems to be a determining factor. Forwhitepeople,theaveragevulnerabilityindexfallsbelowthenationalaverageatthefourthdecile;black,brown,andindigenouspeo-pleonlyescapetheabove-averageriskinthetopthreedeciles.


Anotherworryingpatternisthefactthattheracialgapinthevulnerabilityindexishigheratthebottomofthedistributionthanatthetop.ThoseattheintersectionofraceandpovertyarethereforemuchmorevulnerabletoCOVID-19.Anevenmoredramaticpatternisobservedforthe intersectionbetweenraceandloweducationalachievement,withwhitesofalleducationalgroups facing below-average risk, while for black and brownpeople,onlythosethatstartedorearnedasecondarydiplomapresentasocialvulnerabilityindexbelowthenationalaverage.

Our results highlight the importance of recording infec-tion,hospitalization,anddeathratesnotonlybyrace,sex,andincome,butalsofortheintersectionofthosegroups.Thelackofmicrodatathatwouldallowresearcherstoassesshowstructuralinequality,racism,andsexismimposesanunequaldistributionofvulnerabilitiesacrossdifferentsocialgroupsisanobstacletothedesignofeffectivepolicymeasures.

Our social vulnerability index shows thatwhenCOVID-19startedtospreadlocallyinBrazil,racializedandlow-incomepopulations were at increased risk of being exposed to thevirus.Thenextsectionprovidesoneadditionalexplanationforobserving socioeconomic inequalities in thenumbersof casesanddeathtolls:previousinequalitiesinhealthaccessandout-comescanleadtoinequalitiesintheseverityofcases.

Inequality and Access to Healthcare Brazilhadanadvantage indealingwiththepandemic:oneoftheworld’s largestuniversalpublichealthcare systems.Asdic-tatedbytheBrazilianconstitutionof1988,healthisauniversalrightandaresponsibilityofthestate.Theconstitutionalsostatesthathealthcareshouldbeequallyaccessibleandthatthesystemshouldberegionallydecentralized.TheimplementationoftheUnifiedHealthSystem(SistemaÚnicodeSaúdeorSUS)beganin 1990 and has allowed Brazil to slowly address the nation’shealth inequities,alreadyexacerbatedduringthepreceding20yearsofmilitarydictatorship.Nonetheless,thissystemstillfallsshortofachievingitsegalitariangoals.6

AnotherimportantaspectofBrazil’shealthcaresystemisthecoexistenceofalargeprivatehealthnetwork,mostlyavailabletothosewhocanaffordhealthinsurance.AccordingtoGuimarães(2020), the overall number of beds available in the privateandpublicsystemsarecomparable,thoughnumbersfromtheNationalHealthSurvey(PNS)showthat71percentofBraziliansrelyonthepublichealthcaresystemand72percentofBraziliansdonothaveprivatehealthinsurance.Thosenumbersarehigher

forblack,brown,andindigenouspopulations(80percentand81 percent, respectively) and for those that did not completemiddleschool(83percentand84percent,respectively).

Furthermore, PNS indicates that only 65 percent ofBraziliansself-evaluatetheirhealthasgoodorverygood;thatproportionintheSoutheastregionis11percentagepointshigherthanintheNorth.TheNorthernregion’spopulationisalsothemostdependentonthepublicsystem(80percent)andishometothelowestproportionofthosethathavebeentoadoctorinthelastyear(61percent).Infact,18percentofthosethatdeclarehavingpooror verypoorhealth in theNorthern regionhavenotseenadoctorinthelastyear.Thesamediscrepanciescanbeobservedwhenweevaluate raceandeducational attainments,withblack,brown,andindigenouspopulationsandthosethatdidnotcompletemiddleschoolbeingmoredependentonthepublicsystemandlesslikelytohavebeentothedoctorinthelastyear,eveniftheyevaluatetheirhealthmorepoorlyonaverage.ThisisfurtherevidencethattheSUSfailstodeliverequalanduniversalhealthcare.

Indeed,Racheetal.(2020)estimatesthatofthe316healthregionalunits,14.9percentofthepopulationthatreliesonthepublichealthsystemlivesinoneofthe142unitswithnointen-sivecareunit(ICU)beds.Theyalsostatethatin72percentofthehealthregionsthenumberofICUbedsper10,000peopleisbelowwhatisadequateforatypicalyear,withouttheinfluenceofCOVID-19.AreportbytheBrazilianIntensiveCareMedicineAssociation(AMIB2020)estimatesthatthenationalaverageofICUbedsper10,000people is 2.2, but it is 4.9 in theprivatesystemand1.4inthepublic.Tomakemattersworse,thosenum-bersvarywidelybyregion,withthelowestaveragenumberofICUbedsper10,000people inthepublicsystembeingintheNorth(0.9)andthehighestintheSoutheast(1.8).

UnequalhealthcareaccessinBrazilhadtwoconsequencesforthepandemic.First,theBrazilianpublicsystemwasunpre-pared,andparticularlysointheNorthandNortheastregions,toprovideservicestothoseinfectedbythevirus.Second,evenbeforethepandemic,manyindividuals—especiallythoseracial-izedandlivinginpoorerregions—hadcompromisedimmunesystems. This last argument is further scrutinized in the nextsubsection.


Inequality and ComorbiditiesAccordingtotheWorldHealthOrganization(2020),themajor-ityofpeopleinfectedbyCOVID-19(80percent)willexperiencemildsymptoms.Itisconsistentlyobservedacrosscountriesthatageandunderlyinghealthconditionscanexplainsevereillness,needofhospitalization,anddeath(Guanetal.2020;InstitutoSuperiorediSanitá2020).

AstudycarriedoutintheUnitedStatesamongpeoplediag-nosedwithCOVID-19(Chowetal.2020)showsthatthehospi-talizationrateforthosewhodidnothaveanyunderlyinghealthproblemwas 7 percent overall and 2 percent in ICUs. Thesenumbers increase to 30 percent and 15 percent, respectively,forpeoplewith reportedcomorbidities.7Furthermore,Gaoetal.(2020)reportthattherateofhospitalizationintheChineseprovince ofHubeiwas 1.8 times higher for patientswith onecomorbidity and2.6 timeshigher for thosewith twoormorecomorbidities.

As studies show, gaps in life expectancy (Preston andTaubman1994),prevalenceoflowerrespiratoryillness(Margolisetal.1992)anddiabetes(Gaskin,Thorpe,andMcGinty2014)are correlatedwith educational attainment, income, and race.Brazil does not escape this pattern, with regional and educa-tionalgapsamongthosediagnosedwithdiabetesandhyperten-sion(Viacavaetal.2019;Maltaetal.2019;Leiteetal.2015)andincome gaps associated with cardiovascular disease mortalityrates(Ishitanietal.2006).

Previousstudies for theUnitedStateshavepointedto theimportanceofaccountingforthedifferentratesofprevalenceofcomorbiditiestoexplaintheracialandincomegapsinCOVID-19’sobservedcasesanddeathtoll(Nassif-Piresetal.2020;KimandBostwick2020;Price-Haywoodetal.2020).ToinvestigatethisissueinBrazil,weusethe2013PNSandevaluatethecor-relationbetweenincidenceofriskfactorsandeducationalattain-ment.Weconsider risk factors to includebeingover60 yearsofageanddiagnosedwithdiabetes,hypertension,asthma,lungdisease, coronarydisease, or chronic kidneydisease.The factthat the information on diseases is self-reported and requiresa previous diagnosis by a doctor leads to racial, educational,andregionalbiases.Topartiallycorrectforthis,werestrictoursample to individualswhohaveconsultedadoctorwithin thelastyear,whichcorrespondsto72percentofourinitialsamplepopulation.8

Theproportionofoursamplewhofallintothegroupcon-sideredat-risk forCOVID-19is42percent.However, therisk

factors are not equally distributed among the population. Asshown inFigure 2, theproportionof peoplewhodeclared tohaveattendedonlyelementaryschoolandpresentoneormoreriskfactorsis60percent,comparedto32percentforthosewhoattendedhighschooland37percentforthosewhohavestartedahigherdegree.Thisdifferenceisevengreaterwhenconsider-ingthosewhohavemorethanoneriskfactor,withafrequencyamongthosewhoattendeduptomiddleschool2.5timeshigherthanamongthosewhoattendedhighschoolandtwiceashighthanforthosethatstartedagraduatedegree.

Whenweconsiderallresultspresentedinthissection,itisclearthatwhenCOVID-19reachedBrazil it foundastructur-ally unequal country, where certain social groups weremorevulnerabletoinfection,lesslikelytohaveaccesstohealthcare,andmorelikelytodevelopsevereillness.Finally,thejuxtaposi-tionofthesethreelayersofvulnerabilityweremoreprominentforthoseattheintersectionofclassandrace.Therefore,with-outastrongpolicyresponsetosupportvulnerablegroups,theCOVID-19 health burden in Brazil will be necessarily higherforracialized,poor,andless-educatedpopulations.InthenextsectionwewilldiscusstheeconomicpoliciesimplementedandstudytheobservedhealthandeconomicimpactsofCOVID-19inBrazilsofar.WethenanalyzeifthepoliciesimplementedhavebeensuccessfulinmitigatingtheexpectedunequaldistributionofthehealthandeconomiccostsofCOVID-19.


The Impact of the Pandemic in Brazil: Health, Social, and Economic Effects

Economic and Policy Context Asopposedtoseveralhigh-incomecountriesthatenteredtheirCOVID-19 crises following long periods of economic expan-sionandreductionsinunemploymentrates,Brazilexperienceda7percent fall inGDPin2015–16andhadsincebeengoingthroughtheslowesteconomicrecoveryinitshistory.Moreover,asa resultof thecombinationbetweena fiscalexpansionandaslowdownin taxreceipts, theelectoralyearof2014broughtaboutthefirstincreaseinBrazilianpublicdebtrelativetoGDPinthe21stcentury.9

Intheyearsthatfollowed,theriseinpublicdebtwasusedtojustifytheadoptionofafiscalconsolidationprogramfocusedoncuttingpublicinvestmentandapprovingstructuralreformsinthepensionsystemandothersourcesofmandatoryexpendi-tures.In2016,congresspassedanamendmenttotheconstitu-tionthatestablishedaceilingforfederalprimaryexpenditures:thebudgetwouldonlybeallowedtoexpandattherateofthepreviousyear’sinflation.Inotherwords,thecountrywasheadedtowardasubstantial reduction in thesizeof thestate,aspub-lic expenditures would not be allowed to keep upwithGDPgrowth.ItalsoattachedBrazil’scommitmenttoausteritytotheconstitutionandrestrictedfuturegovernments’abilitytoimple-mentanticyclicaleconomicpolicies.

Asunemploymentalmostdoubled—from6.5millionpeo-ple in 2014 to 13.2million people in 2017—former presidentMichelTemerapproveda labor reform that allowed formoreflexibleworkcontracts.Notsurprisingly,theagendawasunabletodeliver thepromisedsurge in investors’confidence, leadingtosuccessivefrustrationsinGDPgrowthprojectionssincetherecoverystartedin2017.Eveniftheeconomyhadcontinuedtogrowat thesamepace,Brazilwouldonlyhavereturned to itsprecrisisrealGDPlevelby2025—morethantenyearsafterthe2014peak.Tomakemattersworse,since2015,incomeinequal-itygrewanaverageof50percentfasterthanitfellinthe2000s,andhouseholdswerestilltryingtocopewithsignificantlevelsofdebtandincreasinglyprecariousjobs.BeforetheCOVID-19crisis,Brazilcounted38.4million informalworkers(41.3per-centofthelaborforce)and12.5millionunemployed(11percentunemploymentrate).

Thus,theelectionofPresidentJairBolsonaroin2018hap-penedinacontextofmountingfrustration.Thecrisiscoincided

with the largest corruption investigation in Brazilian history(knownasLavaJato,orOperationCarwash),whichfacilitatedthe simplistic yet understandable perception among the gen-eralpopulationthatcorruptionitselfwasthemaincauseoftheeconomicmeltdown.From thisperspective, it becomes easiertounderstandhow,incontrasttomultiplefar-rightnationalistcandidacies around theworld,Bolsonarowaselected throughthecombinationoftheusualmorallyconservativediscourseandanultra-liberaleconomicplatform—gettingridofacorruptstateinallareas(exceptpublicsecurity)wassoldasasolutiontoallofthecountry’sproblems.10PauloGuedes,Bolsonaro’sUniversityofChicago–educatedeconomicguru,becameBrazil’sministerof finance.With alleged success in the financial industry andmarketfundamentalistdiscourse—includingpromisesofpriva-tizingallpublicassetstopayoffpublicdebtandexplicitpraiseoftheChileaneconomicsuccessunderPinochet—Guedeshelpedgathersupportfromfinancialelites.

TheCOVID-19 shock came right after another round offrustrating GDP numbers released in earlyMarch: economicgrowth in 2019—the first year of Bolsonaro’s presidency—slowedto1.1percent, farbelowmarketexpectationsfromthebeginningoftheyearofaround2.6percent.Thereactionincon-gress to thedisappointingeconomicperformancehadalreadyrevealed growing discontent with the ability of the austerityagendaandthespendingceilingtodelivereconomicgrowth.OnMarch16,2020,lessthanamonthafterthefirstcaseofCOVID-19wasreportedandonlyafewdaysbeforelockdownmeasureswereimposedbystategovernorsandmayorsaroundthecoun-try,EconomyMinisterPauloGuedes stated that theBrazilianeconomyhad“itsowngrowthdynamics”and“couldperfectlygrow2.5percent [in2020]” (CNNBrazilBusiness2020).Thedenialphasedidnotlastlong.Lessthanaweeklater,PresidentBolsonaro decreed a state of public emergency, allowing gov-ernmentexpendituresduringthepandemictogobeyondwhatisallowedbyfiscalrules(includingthespendingceiling).TheCOVID-19 crisis thus put an end to five years of austerity inBrazil:“theone-timestarminister isbeing forcedtoreconcilehisfreemarket‘ChicagoBoy’identitywiththeneedformassivegovernmentintervention,”reportedtheFinancial TimesonApril28(HarrisandSchipa2020).


Policy Responses and the Socioeconomic Impact of COVID-19Brazilhasadoptedloose,inarticulate,andinsufficientlockdownmeasures, as the president continuously denied scientific evi-denceandblamedtheeconomiccollapseonstategovernorsandmayorswho imposedanyrestrictivemeasures.OnApril11th,Brazilhadofficiallyreported1,000deathswhileranking131stinanindex11thatclassified178countrieswithrespecttostrictnessofgovernmentresponses.

The samedisregard cannot be attributed to theBraziliangovernment’s fiscal response. According to the IMF’s “FiscalMonitorDatabaseofCountryFiscalMeasures inResponse totheCOVID-19Pandemic”(IMF2020),thetenadvancedecono-mies12oftheG20spent,onaverage,6.6percentofGDP(includ-ingdeferredtaxes)onfightingthepandemic.Theaveragehealthexpendituresinthesecountrieswere0.5percentofGDP,withthe greater part being destined for income transfers and jobmaintenanceprograms.Inthetendevelopingeconomies13oftheG20,thefiscalresponseonlyaddedupto2.8percentofGDPonaverage,fromwhichthesame0.5percentofGDPwasclassifiedashealthexpenditures.Inotherwords,thisgroupofdevelopingcountrieshasspentthreetimeslessinnonhealthareasrelativetothesizeoftheireconomies(andtentimeslessifweconsidertheabsolutedollarvalueoftheresponse)thantheadvancedecon-omiesconsideredabove.However, thesamedatabaseshowsatotalof6.5percentofGDPinadditionalspendingandforegonerevenuesinBrazil(0.9percentofGDPinthehealthsectorand5.6 percent of GDP in the nonhealth sector). Brazil has thusmatched the average fiscal responseof the groupof advancedeconomiesintheG20relativetoitsGDP.Itsfiscalresponsewassmaller than that of the United States, Japan, Germany, andAustralia,butgreaterthanthatofCanada,France,Italy,Korea,Spain,andtheUnitedKingdom.Asaconsequence,fiscalpro-jections suggest that the countrywill run aprimarydeficit ofmorethan8percentofGDPin2020andthatpublicdebtwillgobeyond100percentofGDPin2026.

Almosthalfof the total additional expendituresapprovedbyMay15,2020wereallocatedtotheemergencycashreliefpro-gramAuxílioEmergencial,whilethejobmaintenanceprogram,which guaranteed partial or full payment of unemploymentinsurancetoworkerswithreducedorsuspendedworkcontracts,accountedfor22percentofthetotalexpenditure.Spendingmorethan2percentofGDPonAuxílioEmergencialwasnotadeci-sionmadebytheexecutivebranchofthegovernment,butrathera package pushed by the national congress with the support

of numerous actors in civil society. The program, originallyapproved for three months and already extended by anothersix,transferredR$600(around$110)14peradultonamonthlybasisinthefirstfivemonths(andwilltransferR$300inthelastfourmonths)tounemployedandinformalworkers,aswellastobeneficiariesofthecashtransferprogramBolsaFamília.15TheadministrativecapacitydevelopedinBrazilformanagingBolsaFamíliaandothersocialprotectionprogramsinthepastdecadeshelpedintheimplementationofAuxílioEmergencial.Brazilianswhowere registered as potential beneficiaries for other socialprograms but were not drawing a pension or unemploymentinsurancewereautomaticallyqualifiedtoreceivetheemergencycashrelief.Otherinformalandunemployedworkerswereabletofilloutaformthroughamobileappreleasedbythepubliccom-mercialbankCaixaEconômicaFederaltoapplyforthebenefit.ByJuly,morethan60millionpeoplehaddirectlyreceivedthecashtransferandmorethanhalfthepopulationhadbenefitedfromit.

According to a special June 2020 National HouseholdSurvey(PNAD-COVID)releasedbyIBGE,cashtransfersfromAuxílioEmergencialmore thancompensated forbeneficiaries’incomelossesduringthecrisis.AsobservedinFigure3,thelossof labor income for thebottom50percentof thedistribution


wassmaller than theaveragepercapita incomegain fromtheemergencyprogram.WhilethereareimportantmethodologicaldifferencesbetweenPNAD-COVIDandpreviouslyconductedsurveys, data also suggests that the Auxílio Emergencial pro-gramwasresponsibleforreducingpovertytoitslowesthistoricallevelpreciselyduringwhatmaybecomethedeepest recessioninworld history (Duque 2020).When focusing on black andbrownrespondents,theaverage17percentlossinlaborincomehasbeenfullyneutralizedbythecashreliefprogram.

Hence, the programhas so far fully neutralized the pan-demic’s initial impactonincomeinequality:whilethePNAD-COVIDsurveysuggeststhattheGiniindexforpercapitalaborincomeincreased16from0.64to0.67duringthecrisis—repre-sentingasignificant5percentriseininequality—thisindexfallsto0.56whenaddingthepercapitavalueobtainedfromAuxílioEmergencialinthecorrespondinghousehold.

However,astheorignalbenefit(approvedforthreemonthsand thenextended foranother two)hasbeenreducedbyhalfasof September and isonly approved to lastuntilDecember,thesenumbersmaychangequickly.Inparticular,thesubstantialfiscalresponseinBrazilwasnotmatchedbyanadequateexpan-sionofcredit lines tobusinesses,17whichseems tobe leadingtoamassbankruptcyofsmallfirms.UntilthefirsthalfofJuly,anothersurvey(IBGE2020b)revealedthat716,000companieshadclosed—99.8percentofwhichhadfewerthan49employ-ees.Asthesmallbusinessesgoingbankruptareconcentratedinsectorsinwhichlow-skilledlaborpredominates(services,retail,andconstruction),thecrisismaycontinuetodisproportionatelyaffectjobsandwagesatthebottomofthedistribution.AlabormarketsurveyreleasedinAugustbyIBGE18showsthatbetweenAprilandJune2020,thegreatestjoblosseshappenedinthefoodandhousingsector,domesticservices,construction,andotherservices (a reduction of 26.1 percent, 24.7 percent, 19.4 per-cent,and17.5percent,respectively,relativetothesamequarterin2019).InthePNAD-COVIDsurvey,thereductioninlaborincomeofindividualswhohaveatmostamiddleschooleduca-tionwasalreadyat18.5percent,relativetoa13percentdecreaseforworkerswithacollegedegreeormore.

Asthecrisisisexpectedtobefarfromoverbytheendof2020,thegenerousbuttemporaryAuxílioEmergencialwillmostlikely have only postponed the pandemic’s effect on incomeinequalityinBrazil.Anabruptterminationoftheprogramandareturntoausteritymeasuresin2021couldthusamplifythesame

inequalitiesthatmadethecountrysovulnerabletothehealth,economic,andsocialeffectsofCOVID-19.

Social Vulnerability, COVID-19 Infections, and MortalityThenumberofreportedCOVID-19casesanddeathsinBrazilclearlyreflectthecountry’sdeepracialandregionalinequalities.Whenthedeathtollreached54,488people,atechnicalreport(Medeiros,Cravo,andTatsch2020)basedonofficialhealthsta-tisticsshowedthat61percentofthedeadwereblackorbrown(categories that together make up only 54 percent in Brazil’spopulation, according to the census). In theNortheast, blackandbrownpeoplemadeup82percentoftotaldeathswhileonlyaccountingfor70percentoftheregion’spopulation.DatafromPNAD-COVIDpointinthesamedirection.AmongrespondentswhodeclaredhavinghadatleastthreeCOVID-19symptomsinthepreviousweek, 62percentwereblack andbrown—apro-portionsignificantlyhigherthanthe55percentshareofblackandbrownpeopleinthefullsample.ForthosewhohadmorethansixsymptomsassociatedwithCOVID-19,thissharegoesup to 66 percent. Finally, considering only those who had tobehospitalizedforoneormoredays,60percentareblackandbrown(thissharegoesupto70percentofthosewhorequiredaventilator).


The evidence reflected in the deep racial inequalities inour social vulnerability index, the incidenceof comorbidities,andaccess tohealthcaremayhelpexplain suchracialdispari-tiesintheproportionandseverityofCOVID-19infections.Infact, racial inequalities seemtohaveplayedamore importantroleinexplainingdifferentratesofinfectionandhospitalizationthanincomeinequalitiesformostofthepopulation.TheshareofrespondentsineachsectionoftheincomedistributionwhodeclaredhavinghadatleastthreeCOVID-19symptomsinthepreviousweekcorrespondsexactlytoitsshareinthetotalpopu-lation,ascanbeobservedinFigure4.Inotherwords,beingatthebottom50percent(oratthetop10percent)oftheincomedistributioninBrazildoesnotseemtoincrease(ordecrease)thelikelihoodofinfection.Whenobservingtheshareofhospitaliza-tionsand,evenmoreso,theshareofrespondentswhohavebeenputonaventilator,disparitiesbecomeapparentatthetopoftheincomedistribution: for the top10percent, theproportionofindividualswhorequiredaventilatorisonly1.6percent.

Afewhypothesescouldexplainthispattern.First,thoseatthebottom50percentoftheincomedistributioninBrazilarehighly concentrated in rural areas in the country’sNorthandNortheast,whichpresentedamuchlowerrateofinfectionthanmetropolitanareasintheSoutheast.Additionally,theseareareaswithloweraccesstohospitals,aspreviouslyobserved.Both

thesefactscouldhelpexplainwhythebottom50percentofthedistributionseemstohavea lowerrateofhospitalizationthanthe40percentatthemiddleofthedistribution,whichincludessociallyvulnerableworkersinbigcitiesintheSouthernstates.Second,Brazil’sdeepincomeinequalitieshavetraditionallybeenassociatedwithadisproportionateconcentrationof incomeattheverytop:thetop1percentofBraziliansintheincomedistri-butionreceivemorethanonequarterofnationalincome.Asaconsequence,incomedifferencesbetweenthetopandmiddleofthedistributionaremuchhigherthanthosebetweenthemiddleandbottom,contributingtothedisparitiesintheuseofventila-torsappearingat the top.Thesedisparitiesareevenmoresig-nificantwhenonetakesintoaccountthataccesstoventilatorsislargelyconcentrated intheprivatehealthcaresystem,which isonlyavailabletotheportionofthepopulationthatseemstohavebeentheleastinneedofsuchaccess.

Finally,whenitcomestotheroleofregional inequalities,Figure5showsthecorrelationbetweenoursocialvulnerabilityindex and accumulated cases and deaths per 100,000 people.ThePearsoncorrelationcoefficientsarecalculatedperdaysincethenumberofcasesandnumberofdeathsfirstsurpassed100.PanelCdisplaystheevolutionofthecorrelationcoefficientsbyday,whilepanelsA andBpresent scatterplots for thedays inwhichweobservethestrongestcorrelationbetweentheindex


andcasesper100,000(May29th)andthestrongestcorrelationbetweentheindexanddeathsper100,000(June15th).PanelsAandBshowthatthesocialvulnerabilityindexaverageforeachBrazilianstatepositivelycorrelateswiththenumberofCOVID-19casesper100,000andCOVID-19deathsper100,000,respec-tively.InpanelA,wecanobservethatstatesintheNorthandNortheastregionswithahighersocialvulnerabilityindexratingalsohaveahigh infectionrate,even if thesestatesrepresentalargershareofruralareas(lessaffectedbythevirus);highsocialvulnerabilityinthelargestcitiesinthesetworegionscouldhelpexplainthisapparentcontradiction.Manaus,thecapitaloftheNorthernstateofAmazonas,appearedtohavethefastestcon-tagionrateandanexplosioninmortalityinthefirstmonthsofthepandemic(Orellanaetal.2020).Asaconsequence,mortal-ityratesinthestateofAmazonashavebeenfour-timeshigherthanthenationalaverageandeveninremotetownspeoplehavebeenaslikelytogetsickasinNewYorkCity.Tragically,thevirusspreadalongtheAmazonRiverandexactedanespeciallyhightoll on indigenous people (Andreoni, Londoño, and Casado2020).Figure5,panelAthereforesuggeststhatsocialvulnerabil-ityasmeasuredinthisbriefhadsubstantialexplanatorypowerforCOVID-19infectionrates.

ExaminingpanelsBandC,wenoticethatthepositivecor-relationbetweenoursocialvulnerability indexanddeathsper100,000islessrobust.Thisisevidencethatthefactorstakenintoaccount in our social vulnerability index explain the vulner-ability to infectionwell, but that other factors,not taken intoaccount inour index, are important for explaining thediffer-encesindeathrates.Thiscorroboratestheimportanceoftakingintoconsideration thedifferences inprevalenceof comorbidi-tiesandaccesstohealthwhenexplainingtheseverityofcases,aspreviouslydiscussed.

Thecorrelationbetweenoursocialvulnerabilityindexandthenumberof reportedCOVID-19 infections anddeathshasalsochangedduringthepandemic,aswecanobserveinpanelC.Itwasrelativelylowwhenthecountryfirstreached100casesinmid-Marchand100deathsattheendofMarch,itincreasedbetweenthenandtheendofMay,anditstartedtofallinmid-June.Asthefirstreportedcaseshavebeenassociatedwithtravelabroad,thevirustooktimetospreadfromelitecirclestomorevulnerable areas. One possibility is that social vulnerabilityincreaseditsroleasanexplanatoryfactorforinfectionratesasthepandemicaffectedtheentirepopulation.Asarelevantpro-portionofthesociallyvulnerablewereexposedtothevirusin

majormetropolitanareas, theinfectionrateamongthisgroupmayhavestartedtofall,reducingthiscorrelationinathirdstageofthepandemic.AstudybyGomesetal.(2020)suggeststhatCOVID-19infectionratesstarttofallafter10percentto20per-centofthepopulationhavebeenexposedtothevirus.AstudycarriedoutinSãoPaulo(Tessetal.2020)revealedthatbetween15percentand20.9percentofthepopulationinBrazil’sbiggestcityhavealreadybeenexposedtothevirus.

ConclusionAs of August 8, 2020, Brazil has reported 100,000 deaths byCOVID-19 andmay soonhave the highest total accumulateddeathsinthepandemic.Ifthedisastrousresponseonthehealthfrontby theantiscience federalgovernmentwerenotenough,this briefhasdemonstratedhow structural inequalitiesplayedanimportantrole inexplainingthis tragicoutcome.First,oursocial vulnerability index, built around several dimensions ofinequality (i.e., work, transportation, infrastructure, and liv-ing conditions) that potentially increase the risk of infection,is positively associated with the number of COVID-19 casesacrossBrazilianstatesandrepresentsdeepracial,income,andregionalinequalities.Second,thecountry’sstarkinequalitiesinaccesstohealthcareduetothedualitybetweentheprivateandpublicsystemscontributetoexplainingwhysocialvulnerabilityalonedoesnotaccountforobserveddisparitiesinthenumberofdeaths.Third,thecomorbiditiesassociatedwithmoreseverecasesofCOVID-19arealsoshowntobeunequallydistributed,thushelpingexplain theobservedgapbetween the topof thedistributionandtherestofthepopulationintermsofventilatoruseduringthepandemic.

When addressing the social and economic effects ofCOVID-19,wehaveshownthatthesubstantialfiscalresponsepushed by congress—particularly the implementation of theemergencycashreliefprogramAuxílioEmergencialduringthepandemic—hasbeenable,throughareductionofpovertylevelstoahistoricallow,toneutralizetheinitialriseinwageinequalitycausedbythecrisis.Whilethisshort-termresponsewasinsuf-ficienttocompensatefortheeffectsofstructuralinequalitiesonthecountry’sinfectionratesanddeathtoll,ithascertainlybeenabletopreventadditionalsocialandeconomiccoststothemostvulnerable.However,thefutureoftheprogramisstilluncertainandthesharpincreaseinpublicdebtduringthepandemichasalready presented an opportunity for Bolsonaro’s ultraliberaleconomicteamtopushforthereturnofanausterityagendain


2021thatwouldbeevenmoreseverethanthefiveyearsofpre-COVIDausteritypolicies.

A few policy recommendations can be derived from ouranalysis.Intheshortrun,wefindthatallocatingasizablepor-tionofthebudgettoextendingthebreadthandlengthofsocialwelfareprogramsissufficienttoovercometheunequaleconomicimpactsofaverydeepeconomiccrisis.However,whenconsid-eringtheconsequencesofthepandemicforpublichealth,ourresultssuggestthattheproblemsofmultidimensionalinequalityandstructuralracismareofparamountimportance.Deepracialandregionaldifferencesneedtobeaddressedthroughagovern-mentefforttopermanentlystrengthensocialwelfareprogramsandpublichealth,education,andinfrastructuresystems.

Unfortunately, this does not seem to be the direction inwhich Brazil is heading. In light of the evidence that socialinequalitiesincreasethebreadthandlengthofthepandemicandthepresenceofadeepeconomicrecession,thethreatofareturntoaneconomicagendacenteredoncuttingsocialexpendituresposesmajorhealthandsocialrisks.Moregenerally,pursuingthepastdecades’economicframeworkonagloballevelwillacceler-ate the same tendencies in the labormarketand in inequalitythatimposedveryhighsocial,health,andeconomiccostsdur-ingtheCOVID-19pandemic,pavingthewayforfurthertrag-edies.OuranalysissuggeststhatCOVID-19hasexacerbatedtheinequalities thatmade thepandemicworse, thus requiringanevenmoresubstantialeffortbygovernmentstocounterbalancethesetendencies.

In this context, low- and middle-income countries withhigh levels of inequality require an evenmore substantial fis-calresponsetoneutralizetheseeffects.However,thesearealsothecountriesthataremoresusceptibletoexternalrestrictions,capitalflight,andbudgetconstraintsintheeraoffinanciallib-eralization.Thewayourglobalizedworldandeconomicsystemhasbeen shaped in the recentpast can thereforebe seenasamajorcomorbidityincompoundingthesymptomsofthepan-demic in 2020. Structural changes to the global labormarket,increasing underemployment in high-income countries, andeconomicinsecurityelsewherearelikelytowidenhealthineq-uitiesinthelongertermandleaveusevenmorevulnerabletofuturepandemics.

Notes1. In thebeginningofAugust2020, theyranked5th,9th,and11th,respectively.2.Labor rights inBrazil areonly enforced for employees thatare officially registered with a worker booklet signed by theemployer(carteiraassinada).3.With average household incomes of R$895 andR$905 andGini of 0.55 and 0.57, respectively, compared to R$1600 andhighestGiniof0.52inotherregions.4. Since races are socially constructed, assigned at birth andlater on self-declared, ethnic background and skin pigmenta-tionmightnotbethedeterminantfactorsinsomeonedeclaringthemselvesPardo,black,orindigenous.ItisimportanttonotethatthetermPardoissubjecttocontestationand,accordingtoCarneiro(2000),“itlendsitselfonlytoaggregatethosewho,havetheirethnicandracialidentityshatteredbyracism,discrimina-tionandthesymbolicburdenthatblacknesscontainssocially.”5.Asiswell-documented,householdsurveystendtounderesti-matetheincomeatthetopofthedistribution.ThisisalsotrueforBrazil,asmadeclearbythecombinationofsurveyandtaxdatainSouzaandMedeiros(2017)andMorgan(2017).6.Beforethe1988constitution,publichealthcarewasonlyavail-ableforformalworkersandlocalbudgetswereproportionaltocontributionsbytaxpayers(MIS2012).TheSUSinheritsmanyofitsregionalinequalities.7.Someillnesseshaveledtohigherhospitalizationrates:chronickidneydiseaseanddiabeteshadICUadmissionrates11and8.5timeshigher,respectively;chronicpulmonarydiseaseshad3.4timesmorehospitalizationsand6.5timesmoretransferstotheICU.8.Theseproportionsarehigher forwhiteBrazilians(~74per-cent), smaller for black and brown (~68 percent), as well ashighestinSãoPaulo(79percent)andlowestinAmapá(57per-cent).Moreover, theproportionofthosewithsomeuniversityeducationormorethathavebeentothedoctorinthelastyearisapproximately9percentagepointshigherthanfortherestofthepopulation.9.Thiswasaconsequenceofthesharpfallinoilpricesandothercommodities,andofgenerous taxcutscreatedby thegovern-mentinpreviousyears;seeCarvalho(2018).10.ForanextensiveaccountofgovernmentresponsesandtheBrazilianeconomicandpoliticalcontextbeforethepandemic,seeCarvalho(2020).


11. See Roser et al. (2020). Themetrics used to calculate thegovernment stringency index are: school closures, workplaceclosures, cancellation of public events, restrictions on publicgatherings, closuresofpublic transport, stay-at-home require-ments, public information campaigns, restrictions on internalmovements,andinternationaltravelcontrols.12. Australia, Canada, France, Germany, Italy, Japan, Korea,Spain,UnitedKingdom,UnitedStates.13.Argentina,Brazil,China,India,Indonesia,Mexico,Russia,SaudiArabia,SouthAfrica,andTurkey.14.Thisvaluecorrespondstoapproximately60percentof theBrazilianminimumwage.15.ProgramaBolsaFamíliaisthelargestconditionalcashtrans-ferprogramontheplanet,asitreachesoutto13.9millionfami-lies in poverty and extreme poverty. The program’s monthlybenefitreachesupto$38,dependingontheinitialincomeandthenumberofchildrenorpregnantwomeninthefamily.16.InordertocalculatetheeffectofthecrisisontheGiniindex,wehavecomparedwhat respondentsdeclared tobe thevalueoflaborincomeregularlyreceivedandwhattheyclaimedtobetheirlaborincomeinJune2020.17.Pires(2020)reportedinJunethatBrazil’sfiscalcreditmea-suresforbusinessesonlyaddedupto1.9percentofGDP,relativeto29.7percentinGermany,15.1percentintheUnitedKingdom,6.1percentintheUnitedStates,4.1percentinSingapore,and3.8percentinIndia.18.AggregateresultsfromupcomingPNAD-Contínua.

ReferencesAndreoni,M.,E.Londoño,andL.Casado.2020.“BrazilHeal-

thWorkersMayHaveSpreadCoronavirustoIndigenousPeople.”The New York Times,July19.

AMIB(AssociaçãoBrasileiradeMedicinaIntensiva).2020.“AMIBapresentadadosatualizadossobreleitosdeUTInoBrasil.”SãoPaulo:AMIB.Availableat:https://www.amib.org.br/fileadmin/user_upload/amib/2020/abril/28/dados_uti_amib.pdf

Bengtsson,T.,M.Dribe,andB.Eriksson.2018.“SocialClassandExcessMortalityinSwedenDuringthe1918Influen-zaPandemic.”American Journal of Epidemiology187(12):2568–76.

Bucchianeri,G.W.2010.“IsSARSaPoorMan’sDisease?So-cioeconomicStatusandRiskFactorsforSARSTransmis-sion.”Forum for Health Economics Policy13(2):1–31.

Carneiro,S.2000.“Miscigenação”Portal Geledes,August18.Carvalho,L.2018.Valsa brasileira: do boom ao caos econômi-

co.SãoPaulo:EditoraTodaviaSA.___.2020.Curto-circuito: o vírus e a volta do Estado.SãoPau-

lo:EditoraTodaviaSA.CDC(CentersforDiseaseControlandPrevention).2020.

“HealthEquityConsiderationsandRacialandEthnicMi-norityGroups.”CDCwebsite.AccessedAugust5,2020.Availableat:https://www.cdc.gov/coronavirus/2019-ncov/community/health-equity/race-ethnicity.html

Chow,N.,K.Fleming-Dutra,R.Gierke,Ar.Hall,M.Hughes,T.Pilishvili,M.Ritchey,K.Roguski,T.Skoff,andE.Ussery.2020.“Preliminaryestimatesoftheprevalenceofselectedunderlyinghealthconditionsamongpatientswithcoronavirusdisease2019—UnitedStates,Febru-ary12–March28.”Morbidity and Mortality Weekly Re-port69(13):382–86.

CNNBrazilBusiness.2020.“OBrasilprecisaescolher:din-heiroparaasaúdeouparaasaéreas,dizGuedes.”CNNBrazil,March14.https://www.cnnbrasil.com.br/busi-ness/2020/03/14/se-nos-entregarmos-a-psicologia-do-fracasso-havera-desaceleracao-diz-guedes

CordobaE.,andA.E.Aiello.2016.“SocialDeterminantsofInfluenzaIllnessandOutbreaksintheUnitedStates.”North Carolina Medical Journal77(5):341–45.

CoronavírusBrasil.2020.“PainelCoronavírus.”AccessedonAugust15.Availableat:https://covid.saude.gov.br/

COVID-19SurveillanceGroup.2020.“CharacteristicsofCOVID-19patientsdyinginItaly:reportbasedonavail-abledataonMarch20th,2020.”Report.Rome:InstitutoSuperioreDiSanita.Availableat:https://www.epicentro.iss.it/coronavirus/bollettino/Report-COVID-2019_20_marzo_eng.pdf

Duque,D.2020.“Umaavaliaçãodoauxílioemergencialparte1.”InstitutoBrasileirodeEconomia(IBRE)blog,August5.

Furceri,D.,P.Loungani,andJ.Ostry.2020.“Howpandemicsleavethepoorevenfartherbehind.”IMFBlog,May11.

Gao,Q.,Y.Hu,Z.Dai,F.Xiao,J.Wang,andJ.Wu.2020.“Theepidemiologicalcharacteristicsof2019novelcoronavirusdiseases(COVID-19)inJingmen,Hubei,China.”Medi-cine99(23):e20605

Gaskin,D.J.,R.J.Thorpe,andE.E.McGinty.2014.“Dispari-tiesindiabetes:thenexusofrace,poverty,andplace.”American Journal of Public Health104(11):2147–55.


Gomes,M.G.M.,R.Aguas,R.M.Corder,J.G.King,K.E.Langwig,C.Souto-Maior,J.Carneiro,G.Gonçalves,andC.Penha-Gonçalves.2020.“Individualvariationinsus-ceptibilityorexposuretoSARS-CoV-2lowerstheherdimmunitythreshold.”MedRxiv,2020.04.27.20081893.https://doi.org/10.1101/2020.04.27.20081893

Guan,W.,Z.Ni,Y.Hu,W.Liang,C.Ou,J.He,L.Liu,etal.2020“Clinicalcharacteristicsofcoronavirusdisease2019inChina.”New England Journal of Medicine 382(18):1708–20.

Guimarães,C.2020.“Especialistasanalisamadisponibilidadedeleitosnopaísediscutempossibilidades.”Portal Fio-cruz,May8.

Harris,B.,andA.Schipa.2020.“PandemicforcesBrazil’s‘Chicagoboys’torevisereformplans.”Financial Times,April28.

IBGE(InstitutoBrasileirodeGeografiaeEstatística).2013.“PNS(PesquisaNacionaldeSaúde)”Availableat:https://www.ibge.gov.br/estatisticas/sociais/saude/9160-pesqui-sa-nacional-de-saude.html?=&t=o-que-e

___.2019.“PNAD-Contínua(PesquisaNacionalporAmostradeDomicílioContínua)”Availableat:https://www.ibge.gov.br/estatisticas/sociais/populacao/17270-pnad-conti-nua.html?edicao=27258&t=sobre

___.2020a.“PNAD-COVID(PesquisaNacionalporAmostradeDomicílioCOVID-19).”Availableat:https://covid19.ibge.gov.br/pnad-covid/

___.2020b.“PULSOEmpresas.”Availableat:https://covid19.ibge.gov.br/pulso-empresa/

IMFFiscalAffairsDepartment.2020.“FiscalMonitorDa-tabaseofCountryFiscalMeasuresinResponsetotheCOVID-19Pandemic.”Washington,DC:InternationalMonetaryFund.

Ishitani,L.H,G.D.C.Franco,I.H.O.Perpétuo,andE.França.2006.“SocioeconomicinequalitiesandprematuremortalityduetocardiovasculardiseasesinBrazil.”Revista de saúde pública40(4):684–91.

Kim,S.J.,andW.Bostwick.2020.“SocialVulnerabilityandRacialInequalityinCOVID-19DeathsinChicago.”Heal-th Education & Behavior47(4):509–13.

Leite,I.D.C.,J.G.Valente,J.M.D.A.Schramm,R.P.Dau-mas,R.D.N.Rodrigues,M.D.F.Santos,A.F.D.Oli-veira,etal.2015.“BurdenofdiseaseinBrazilanditsregions,2008.”Cadernos de saude publica31(7):1551–64.

Malta,D.,B.Duncan,M.Schmidt,I.Machado,A.Silva,R.Bernal,C.Pereira,etal.2019.“Prevalênciadediabetesmellitusdeterminadapelahemoglobinaglicadanapopu-laçãoadultabrasileira.”Revista Brasileira de Epidemiolo-gia22(Suppl.2):E190006.

MamelundS-E.2017.“Socialinequalityaforgottenfactorinpandemicinfluenzapreparedness.”Tidsskrift for den Norske laegeforening: Tidsskrift for praktisk medicin, ny raekke137(12-13):911–13.

Margolis,P.A.,R.Greenberg,L.L.Keyes,L.M.LaVange,R.S.Chapman,F.W.Denny,K.E.Bauman,andB.W.Boat.1992.“Lowerrespiratoryillnessininfantsandlowsocioeconomicstatus.”American Journal of Public Health82(8):1119–26.

Medeiros,M,A.Cravo,andC.Tatsch.2020.“DadosdoSUSrevelamvítima-padrãodecovid-19noBrasil:Homem,pobreenegro.”Revista Época,July3.

Morgan,M.2017.“FallingInequalitybeneathExtremeandPersistentConcentration:NewEvidenceforBrazilCombiningNationalAccounts,SurveysandFiscalData,2001–2015.”WIDWorkingPaperNo.2017/12.Paris:WorldInequalityLab.

Murray,C.,A.Lopez,B.Chin,D.Feehan,andK.Hill.2007.“EstimationofPotentialGlobalPandemicInfluenzaMor-talityontheBasisofVitalRegistryDatafromthe1918–20Pandemic:AQuantitativeAnalysis.”Lancet368(9554):2211–8.

Nassif-Pires,L.,L.d.L.Xavier,T.Masterson,M.Nikiforos,andF.Rios-Avila.2020.“PandemicofInequality.”PublicPolicyBriefNo.149.Annandale-on-Hudson,NY:LevyEconomicsInstituteofBardCollege.

Orellana,J.D.Y.,G.M.Cunha,L.Marrero,B.Horta,andI.C.Leite.2020.“ExplosioninmortalityintheAmazonianepicenteroftheCOVID-19epidemic19.”Cadernos de Saúde Pública 36(7):e00120020.

Pires,M.2020.“Atualizaçãodaspolíticasdecombateàcrise.”Observatório da Política Fiscal,June16.

Preston,S.H.,andP.Taubman.1994.“Socioeconomicdifferen-cesinadultmortalityandhealthstatus”.InNationalRe-searchCouncil (ed.),Demography of Aging.Washington,DC:TheNationalAcademyPress.


Price-Haywood,E.G.,J.Burton,D.Fort,andL.Seoane.2020.“HospitalizationandMortalityamongBlackPatientsandWhitePatientswithCOVID-19.”New England Journal of Medicine82(26):2534–43.

Rache,B.,R.Rocha,L.Nunes,P.Spinola,A.M.Malik,andA.Massuda.2020.“NecessidadesdeinfraestruturadoSUSempreparoàCOVID-19:leitosdeUTI,respiradoreseocupaçãohospitalar.”InstitutodeEstudosparaPolíticasdeSaúdeTechnicalNoteNo.3.SãoPaulo:InstitutodeEstudosparaPolíticasdeSaúde.

Roser,M.,H.E.Ritchie,E.Ortiz-Ospina,andJ.Hasell.2020.“CoronavirusPandemic(COVID-19)”.OurWorldInData.org(website).AccessedAugust8,2020.https://ourworld-indata.org/coronavirus

Simonsen,L.,P.Spreeuwenberg,R.Lustig,R.J.Taylor,D.M.Fleming,M.Kroneman,M.D.VanKerkhove,etal.2013.“Globalmortalityestimatesforthe2009influenzapandemicfromtheGLaMORproject:amodelingstudy.”PLoS medicine10(11):e1001558.

Souza,P.H.F.,andM.Medeiros.2017.“TheconcentrationofincomeatthetopinBrazil,2006-2014.”InternationalPolicyCenterforInclusiveGrowthWorkingPaperNo.163.Brasília:InternationalPolicyCenterforInclusiveGrowth.

Tess,B.H.,C.F.Granato,M.C.Alves,M.C.Pintao,E.Rizzatti,M.C.Nunes,andF.Reinach.2020.“SARS-CoV-2seroprevalenceinthemunicipalityofSaoPaulo,Brazil,tenweeksafterthefirstreportedcase.”MedRxiv.2020.06.29.20142331.https://doi.org/10.1101/2020.06.29.20142331

Tricco,A.C.,E.Lillie,C.Soobiah,L.Perrier,andS.E.Straus.2012.“ImpactofH1N1onsociallydisadvantagedpopula-tions:systematicreview.” PLoS One7(6):e39437.

Viacava,F.,S.M.Porto,C.D.C.Carvalho,andJ.G.Bellido.2019.“Desigualdadesregionaisesociaisemsaúdesegun-doinquéritosdomiciliares(Brasil,1998–2013).”Ciência & Saúde Coletiva24(7):2745–60.

WorldBank.2018.“GiniIndex.”WorldDevelopmentIndica-torswebsite.Washington,DC:TheWorldBankGroup.AccessedAugust8,2020.Availableat:https://data.world-bank.org/indicator/SI.POV.GINI.

WorldHealthOrganization.2020.“CoronavirusDiseaseSi-tuationReport41.”WHOwebsite.Availableat:https://www.who.int/docs/default-source/coronaviruse/situ-ation-reports/20200301-sitrep-41-covid-19.pdf?sfvrs-n=6768306d_2


About the Authors

LUIZA NASSIF-PIRES is a research fellow working in the Gender Equality and the Economy program. Her research interests include gender and political economy, distributional aspects of gender discrimination, gender and racial aspects of development, and input-output methods. Her recent research relies on statistical equilibrium and game theory to formalize the impacts of gender and racial segregation in the labor movement with an application to the United States. Nassif-Pires has also written on intersectional political economy with a focus on the impacts of social conflict for the labor theory of value and the long-run profit rate. She is also collaborating with Prof. Katherine Moos at University of Massachusetts, Amherst on a feminist input-output project.

Nassif-Pires has taught microeconomics, macroeconomics, and political economy at the New York City College of Technology and at the Eugene Lang College of Liberal Arts at The New School for Social Research. She holds a BS and MS in economics from the Federal University of Rio de Janeiro and a Ph.D. in economics from The New School for Social Research.

LAURA CARVALHO is an associate professor of economics at the University of São Paulo, a lead researcher with the Research Group on Macroeconomics of Inequality (MADE), and a senior fellow at the Schwartz Center of Economic Policy Analysis (SCEPA).

Besides publishing in the Journal of Economic Behavior & Organization, Journal of Evolutionary Economics, Cambridge Journal of Economics, Journal of Post Keynesian Economics, Review of Keynesian Economics, and Metroeconomica, Carvalho was a weekly columnist for the Brazilian newspaper Folha de S. Paulo between 2015 and 2019. Her 2018 book, Valsa Brasileira: Do Boom ao Caos Econômico, which analyzes the growth and subsequent crisis of the Brazilian economy starting in 2014, became one of the country’s best sellers that year.

Carvalho has recently worked on the distributive impacts of fiscal policy, the multiplier effects of social benefits, and the relationship between wage inequality, employment composition, and consumption patterns. She holds a Ph.D. in economics from the New School for Social Research.

EDUARDO LEDERMAN RAWET is a Ph.D. student at American University. His research interests include growth, distribution, and Post-Keynesian economics. His master’s thesis investigates the relationship between wealth accumulation, inequality, and economic growth through a Kaleckian model. Rawet holds a BS in economics from the Federal University of Rio de Janeiro and an MS in economics from the University of São Paulo.

multidimensional inequality and covid-19 in brazil · public policy brief, no. 153 2 contents 3...

Documents