Page 1
HOW WILL THE AFFORDABLE CARE ACT CHANGE EMPLOYERS’ INCENTIVES TO OFFER INSURANCE?
by
Jean M. Abraham, Ph.D. University of Minnesota
Roger Feldman, Ph.D. University of Minnesota
Peter Graven, M.S. Oregon Health and Science University
CES-WP-14-02 January, 2014
The research program of the Center for Economic Studies (CES) produces a wide range of economic analyses to improve the statistical programs of the U.S. Census Bureau. Many of these analyses take the form of CES research papers. The papers have not undergone the review accorded Census Bureau publications and no endorsement should be inferred. Any opinions and conclusions expressed herein are those of the author(s) and do not necessarily represent the views of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed. Republication in whole or part must be cleared with the authors. To obtain information about the series, see www.census.gov/ces or contact Fariha Kamal, Editor, Discussion Papers, U.S. Census Bureau, Center for Economic Studies 2K132B, 4600 Silver Hill Road, Washington, DC 20233, [email protected] .
Page 2
Abstract
This study investigates how changes in the economic incentives created by the Affordable Care Act (ACA) will affect the probability that private‐sector U.S. employers will offer health insurance. Using the Medical Expenditure Panel Survey Insurance Component for 2008‐2010, we predict employers’ responses to key ACA provisions. Our simulations predict that overall demand for insurance will rise, driven by workers’ desire to avoid the individual mandate penalty and the availability of premium tax credits in exchanges. Our analyses also suggest that the average probability of an establishment offering insurance will decline from .83 to .66 with ACA implementation, although there is considerable variation by firm size, industry and union status. Keyword: health insurance, Affordable Care Act; premium tax credits; employer behavior JEL Classification: I1; I13; J3 i
i Corresponding author: Division of Health Policy and Management, School of Public Health, 420 Delaware Street SE, MMC 729; Minneapolis, MN 55455. Phone: 612‐625‐4375. Email: [email protected] . We thank the National Institute for Health Care Reform for financial support. Any opinions and conclusions expressed herein are those of the authors and do not necessarily represent the views of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.
Page 3
3
I. Introduction
In2011,58%ofthenon‐elderlyU.S.populationobtainedhealthinsurancethrough
anemployer(DeNavas‐Walt,Proctor,andSmith,2012).Thepreferentialtaxtreatmentof
premiumsforemployer‐sponsoredinsurance(ESI)andeconomiesofscalecreatestrong
incentivesthatencourageprovisionofhealthinsurancethroughtheworkplacerelativeto
theindividualmarket.
SeveralprovisionsinthePatientProtectionandAffordableCareAct(P.L.111‐148,
hence“ACA”)willaffectemployers’incentivestoofferESI.In2014,employersthatdonot
offercoveragewillpayanannualpenaltyof$2,000perfull‐timeemployee(exemptingthe
first30employees)ifanyfull‐timeemployeebuyssubsidizedinsuranceinanewhealth
insuranceexchange.
AtthesametimethattheACAwillpenalizelargeremployersfornotofferinghealth
insurance,itwillprovideanewoptionforworkerstobuyhealthinsuranceoutsidetheir
placeofemployment.Withtheopeningofhealthinsuranceexchangesin2014,theACA
willprovidetaxcreditstoindividualswithfamilyincomesbetween133%and400%ofthe
federalpovertylevel(FPL)whodonothaveaccesstoanofferofESItopurchaseexchange‐
basedcoverage.Thesepremiumassistancecreditswillreducethecostofhealthinsurance
to3percentofincomeforthoseat133%FPL2,phasingoutto9.5percentofincomeat300‐
400%FPL.
OneofthemostcontroversialACAprovisionsalsogoesintoeffectin2014,when
mostindividualsintheUnitedStateswillberequiredtohaveinsuranceorpayapenalty.
2The2013federalpovertylevel(FPL)is$11,490foroneperson,increasingto$39,630forafamilyofeight.ModifiedAdjustedGrossIncome(MAGI)willbeusedtodeterminepremiumsubsidies,resultinginaneffectiverateof138%FPLaftera5%offset.
Page 4
4
Atfullimplementationin2016,thepenaltyforasinglepersonwillbethegreaterof$695
(uptothreetimesthatamountforafamily)or2.5%ofhouseholdincome.Althoughthe
penaltyisnotlargerelativetothefullpriceofaninsurancepolicy,thisrequirementis
expectedtoincreasedemandforcoveragebythosecurrentlylackinginsurance.Taken
together,theemployersharedresponsibilityrequirement,theintroductionofinsurance
exchangeswithpremiumassistancecreditsforlower‐incomeindividualswithoutaccessto
ESI,andtheindividualmandatehavethepotentialtoinfluenceemployers’incentivesto
offerhealthinsurance.
Inadditiontoprovisionsthatdirectlyaffectemployers,otherACAprovisions–
regulationofinsurers’medicallossratiosandpremiums,andarequirementtooffer
“essential”benefits–areexpectedtoaffectthefunctioningofhealthinsurancemarkets.
Thesepolicieshavethepotentialtochangethetypesofinsuranceproductssold,the
premiumscharged,andthedegreeofmarketcompetition.
WeinvestigatehowthesechangesineconomicincentivescreatedbytheACAwill
affecttheprobabilitythatprivate‐sectorU.S.employerswillofferESI.UsingtheMedical
ExpenditurePanelSurvey(MEPS)for2008,2009,and2010,wepredictemployers’
responsestokeyACAprovisionsinfoursteps.First,weestimateabinarylogitmodelto
identifythefactorsthatinfluenceanemployer’sdecisiontoofferESI.Second,wepredict
howkeyACAprovisionsthatwillbeimplementedin2014altertheeconomicincentivesof
workerstochooseESIversusan“outsidegood”andanindividualexchangepolicyversus
the“outsidegood.”Third,usingthemodelestimates,wepredicttherelativeprobabilities
ofeachchoice.Finally,wesolvefortheunconditionalprobabilitiesofanESIoffer,the
individualexchangeoption,andtheoutsidegood.
Page 5
5
Oursimulationssuggestthattheaverageprobabilityofanestablishmentoffering
ESIwilldeclinefrom.83to.66withACAimplementation,althoughthereisconsiderable
variationbyfirmsize,industry,andunionstatus.Moreover,ourresultsshowthatthe
introductionofanewoption–individualexchangecoveragewithincome‐basedpremium
assistancecredits–willbeattractivetoahighpercentageofworkerswhodonothavean
ESIofferfromtheiremployer.
Section2summarizesthedata.Section3describesourmethodsforconductingthe
simulation.Section4presentsresultsandstudylimitations.Section5discussesthe
findingsandimplicationsforpolicy.
II. Data
TheprimarydatasourceistheMedicalExpenditurePanelSurvey‐Insurance
Component(MEPS‐IC)ListSamplefor2008,2009,and2010.3TheMEPS‐ICisanationally‐
representative,establishment‐levelsurveyofU.S.employersthatcollectsdetailed
informationontheprovisionofhealthinsurance.Thesurveyincludesquestionsaboutan
employer´sworkforceandcharacteristicsoftheestablishment.Employersthatoffer
insuranceprovideinformationabouttheplansoffered,includingthetotalpremium,
employerandemployeecontributions,enrollmentbycoveragetype(single,employeeplus
one,andfamily),andSection125status.
Anemployer’sdecisiontoofferhealthinsurancedependsonhowmuchitsworkers
valuethatcoveragerelativetowages.Thisvaluelikelyvariesbasedonthecompositionof
anestablishment´sworkforce,includingworkers’wages,familyincomes,familysize,
3BecausetheMEPS‐ICisnotavailabletothepublic,weobtainedapprovaltousethedatafromtheU.S.BureauoftheCensus.AllanalyseswereconductedintheMinnesotaCensusResearchDataCenter.
Page 6
6
demographics,andfederalandstatetaxrates,giventhetax‐exemptstatusofpremiums.4
WhiletheMEPS‐ICcontainssomeinformationaboutworkers’attributes,itdoesnothave
thelevelofdetailnecessarytomeasurethepriceofESI.Thus,weaugmentedtheMEPS‐IC
withinformationonworkers’familiesfromtheMEPS‐HouseholdComponent(HC),a
nationallyrepresentativesampleofthenon‐institutionalizedU.S.population,usinga
statisticalmatchingmethod(describedbelow)toimputethetaxpriceandadditional
characteristicstoworkersinestablishments.
Becausetheemployer’sdecisiontoofferESImayalsodependonlabormarket
conditions,wemergedinformationontheestablishmentsizedistributionand
unemploymentrateforthecountyinwhichtheestablishmentislocatedfromtheU.S.
BureauoftheCensusCountyBusinessPatterns(2008‐2009)andtheU.S.BureauofLabor
Statistics(2008‐2010).Werestrictedourattentiontoprivate‐sectorestablishments.Table
1providesdescriptivestatisticsofestablishmentsthatofferanddonotofferinsurance.
III. Methods
A. EstimatingtheEmployerOfferModel
Inourconceptualframework,anemployeroffersacombinationofwagesand
healthinsurancethatminimizesitslaborcosts,subjecttomaintainingitsemployees’utility
atalevelthatkeepstheestablishmentcompetitiveinthelabormarket(Feldmanand
Dowd,1987;DowdandFeldman,1987).Employeeshavepreferencesregardingthe
combinationofwagesandhealthinsurancethatcomprisetheirtotalcompensation.Given
thetax‐exemptstatusofemployer‐paidpremiums(ortotalpremiumsforSection125
4Alltaxrates(federal,state,OASDI,andmarginalpayrolltaxratesforMedicareHI)wereestimatedusingTAXSIM(FeenbergandCoutts,1993).
Page 7
7
plans),workerswhopayhigherfederalandstatemarginaltaxratesfacealower“price”for
healthinsurancerelativetowagescomparedwithworkerswithlowerincomesand
marginaltaxrates.Thus,employerswithworkerswhose“taxprice”ofinsuranceislower
shouldbemorelikelytoofferhealthinsurance,allelseconstant.
Otherworkforceattributesmayinfluenceemployees’preferencesforhealth
insurance,includingworkers’familyincomesandhealthstatus.Weassumethatan
employercanobserveitsworkers’preferences(orvariablesthatareproxiesforthem)and
theemployercanaggregatethosepreferenceswhenitdecidestoallocatetotal
compensationintowagesandinsurance.Themostcommonlyusedaggregatemeasureis
thepreferencesoftheaverageworker(GruberandLettau,2004).
Employersmustalsoconsiderthetransactioncostsofofferinghealthinsurance
whenselectingtheoptimalcombinationofwagesandinsurance.Sincetheadministrative
costsofESI(e.g.,stafftoobtainpremiumquotesandcoordinateopenenrollment)arelikely
tobefixedorquasi‐fixed,largeremployerscanspreadthosecostsovermoreworkers
relativetosmallfirms.Thusweexpectlargeremployerstobemorelikelytooffer
insurance.Otherfactorsthatvarygeographically,includingstateregulationsofhealth
insurancemarkets,mayaffecttheadministrativecostsofESI.
Finally,anemployermustsettotalcompensationatalevelthatkeepsthe
establishmentcompetitiveinthelabormarket.Weexpectseveralfactorstobecorrelated
withlocallabormarketconditionsandcompensationlevels,suchasfirmsize,industry,the
employersizedistributioninthelocalmarket,andmacroeconomicconditions.
Theoffermodeliswrittenasfollows:
Page 8
8
(1) ititit
itititititit
TIMESTATE
LABORESTABWORKTPOfferNotUESIU
65
43210)(ln)(ln
Thedependentvariableisthedifferenceinthelog‐utilitiesofofferingESIandan“outside
good”ofnotofferingESI.IftheemployerdoesnotofferESI,thisdoesnotnecessarilymean
thattheworkerswillbeuninsured.TheymayhaveaccesstoESIthroughaspouseorthey
canbuyinsuranceintheindividualmarket.5However,asnotedabove,theindividual
marketlackseconomiesofscaleanddoesnotofferapremiumtaxsubsidy6,sofewworkers
takeupthatoption(MarquisandLong,1995).Byconvention,thelog‐utilityoftheoutside
goodisnormalizedtozero.
Neitheroftheutilitiesinequation(1)isobserved.Rather,weobservevariablesthat
influencetheutilityofofferingESI,andwepostulatethattheemployerwillofferESIif:
(2) 0)(ln)(ln itit OfferNotUESIU
Ifweassumetheerrortermin(1)hasalogisticdistribution,wecanreplacethe
unobserveddependentvariablewithanindicatorthatequals1ifthefirmoffersESIand
zeroifitdoesnot.Thenweestimatethecoefficients(β)in(1)andexpresstheprobability
thatthefirmoffersESIas:
(3) Xb
Xb
itESIexp1
exp)Pr(
wherebstandsfortheparameterestimatesandXrepresentsthevariableslistedin
equation(1).
5 A small proportion of workers may also qualify for Medicaid or state-based public insurance programs. 6 Self-employed individuals who have a net profit for the year may be able to deduct the premiums paid for health insurance for themselves and qualified dependents from their taxable income.
Page 9
9
Ourkeyexplanatoryvariableisthetaxprice(TP)–thepriceofadollarofhealth
insurancerelativetoadollarofwageincome.TPislessthan1.0giventhetax‐exempt
statusofemployer‐paidESIpremiums(totalpremiumsforemployerswithSection125
plans).Itvarieswithinestablishmentsgivenvariationinworkers’familyincomesand
acrossestablishmentsgivendifferentincomedistributions.Ouridentificationstrategy
reliesonobserveddifferencesintheprogressivityofmarginaltaxratesacrossstatesfrom
2008to2010.7
WhiletheMEPS‐IChasbasicinformationaboutthewagedistributionofworkersin
eachestablishment,itdoesnothavedetailedinformationaboutworkers’familyincomes,
marginaltaxrates,ortaxpricesofinsurance.Belowwedescribeourapproachfor
imputingthisinformation.
UsingtheMEPS‐HCfor2007and2008,weselectedworkersinprivate‐sector
establishments.Wecomputedtotalfamilyincome,definingthefamilyusingtheHealth
InsuranceEligibilityUnit(HIEU8)identifierontheMEPS‐HC.Next,weusedNBER’s
TAXSIMsoftwareandallavailabletax‐relatedinputvaluesforeachworkertoestimate
workers’marginalfederalandstateincometaxrates,OASDI,andMedicareHospital
7Sevenstatesdidnothavestateincometaxesin2008.Amongstateswithincometaxes,thereiswidevariationinthenumberofbracketsandrates.Forexample,themaximumrateinArizonawas4.54percentin2008,incontrasttoVermont’srateof9.5percent. 8TheHIEUincludesallmembersofthefamilywhowouldtypicallybecoveredunderaprivateinsurancefamilyplan:adults,spouses,andunmarriednatural/adoptivechildrenage18andunder(orage24andunderwhoarefull‐timestudents).
Page 10
10
Insurancetaxes.9WethencomputedataxpriceforeachworkerintheMEPS‐HCusingthe
followingformula:
)1(
)1()4(
mcss
mcsssfTP
wherefistheworker’sfederalincomemarginaltaxrate,sisthemarginalstateincometax
rate,ssisthemarginalpayrolltaxratefortheOASDIprogram,andmcisthemarginal
payrolltaxratefortheMedicareHospitalInsuranceprogram.TheOASDIandMedicareHI
taxeswereimputedas7.65%ofpayroll.
WeestimatedaregressionofMEPS‐HCworkers’taxpricesasafunctionofwage‐
coveragetypeinteractions,establishmentsizeoftheworker,employmentatamulti‐
locationfirm,industry,state,andthefamilysizeoftheworker(alsopredictedusinga
regression).AlloftheexplanatoryvariablesinthatregressionalsoexistintheMEPS‐IC.
WethenpredictedtaxpricesfromtheMEPS‐HCtoeachestablishmentintheMEPS‐IC.We
repeatedthismethodtopredictworkers’familyincomeintheMEPS‐ICestablishments.10
Inadditiontothetaxprice,weconsideredseveralotherworkforceattributes:the
percentageofpart‐timeworkers,thewage,ageandsexdistributionsofworkersinthe
establishment,andthepresenceofunionizedworkersintheestablishment.Ourmodel
alsocontrolsforestablishmentsize,singleormultiplelocations,andindicatorsforone‐
digitindustrycategories,businesstenure,andnon‐profitownership.11
9SincetheMEPS‐HCdoesnotincludestateidentifiersandinternetconnectionsarenotpermittedinCensusRDCs,werepeatedthissimulationforeachworkerforeachstate.Wethenselectedthestate‐specificobservationafterbringingthefileintotheRDC. 10SeeAbraham,Graven,andFeldman(2013)foradditionalinformationontheconstructionandpredictionofthetaxprice.
11WeusedmultipleimputationandSTATA’s“nearest‐neighbor”approachtoaddressitemnon‐responseissuesforsomeoftheworkforceandestablishmentmeasures.Thismethoduseslinearregressionstopredictvaluesfortheoutcomesofinterestbasedonasetofexplanatoryvariables.Weusedinformationonfirmsize,
Page 11
11
Locallabormarketconditionsarecapturedbytheunemploymentrateforthe
countyinwhichtheestablishmentislocatedandthepercentageofestablishmentsinthe
countywithmorethan50workers.Finally,weincludestatefixedeffectstocontrolfor
state‐specific,time‐invariantfactorsthatmayinfluenceemployeroffering.
TheemployerofferingmodelwasestimatedwithbinarylogitandSTATA12.0/SE
statisticalsoftware.Allestimateswereweightedtoreflectthenumberofemployeesin
private‐sectorU.S.establishments.TheresultsarereportedinTable2.
B. MeasuringChangesinEconomicIncentivesin2014
In2014,employerswillhavethesamechoicesthatwerepresentbeforetheACA:to
offerESIornottoofferESI.IfanemployerdoesnotofferESI,workersmayseekan
alternativesourceofcoverageorremainuninsured.Butin2014,workerswillhaveanew
option–topurchaseindividualcoverageinnewly‐createdexchangeswithincome‐based
premiumsubsidies.Inotherwords,therewillbethreeoptions:ESI,individualexchange
coverage,andtheoutsidegood.12
BecausetheACAwillchangetheeconomicattractivenessofparticularsourcesof
healthinsuranceaswellasindividuals’overallincentivetohaveinsurance,itisimportant
toconsidertheincentivesaffectingallthreechoiceswhenpredictingemployerbehavior.
Althoughweconsiderallthreeoptions,weareconstrainedbyhavingonlyoneoffer
equation.Thus,weassumethatthesamecoefficientsandunmeasuredfactorsinthe
industryandstateasexplanatoryvariables.Valuesareimputedforobservationsmissingworkforceandestablishmentinformationfromestablishmentsthataresimilaronthesedimensionstomaintainacompletesamplefortheanalysis.
12 In 2014, the “outside good” will include coverage through a spouse, public insurance for those who are eligible, and the purchase of individual insurance outside the exchange. There is still much uncertainty regarding the long-term viability of an “outside” individual insurance market, since many ACA regulations apply both inside and outside exchanges. Subsidies, however, may be obtained only in the exchanges.
Page 12
12
model’sconstanttermthataffectthechoiceofofferingESIversustheoutsidegoodin
equation(1)canbeappliedtothechoiceoftheindividualexchangecoverageversusthe
outsidegood.Thisassumptionisreasonableiftheargumentsintheutilityfunction(e.g.,
taxprice,income,anddemographics)andtheireffectsonworkers’utilityarethesamefor
bothsourcesofinsurance.Wediscussthisassumptionfurtherinthelimitationssection.
Whileweassumethecoefficientsarethesame,theESIandindividualexchange
optionswillhavedifferenttaxprices.BelowwedescribehoweachoftheACAprovisions
weconsidermayaffectthetaxpriceofinsurance.Wemodelthesumoftheseeffectsasa
singlechangeinthetaxprice–oneforESIandonefortheindividualexchangepolicy.We
alsoconsiderotheranticipatedchangestoinsurancemarkets(e.g.,differencesin
administrativeloadingfeesbetweenESIandexchanges).
IndividualMandate:TheACArequiresmostindividualstoobtainqualifiedcoverage
orpayapenalty.Whenfullyimplemented,theannualpenaltyforasinglepersonwillbe
thegreaterof$695(oruptothreetimesthatamountforafamily($2,085))or2.5%of
familyincome.TheemployeescanescapethispenaltybychoosingeitherESIorthe
individualexchangeoption.Wecalculatedtheaveragepenaltyforeachestablishment.
ValueoftheESITaxSubsidy:Undertheexistingtaxcodetheemployer‐paidportion
oftheESIpremium(orthetotalpremiumforfirmswithSection125status)isexemptfrom
incomeandpayrolltaxes.ThisprovisionisnotaffectedbytheACA.Weestimatedthe
valueofthetaxsubsidybyestimatingpredictivemodelsofESIpremiumsandthe
employer‐paidportionofpremiums.Next,weinflatedallamountsto2014dollarsand
thenestimatedthevalueofthetaxsubsidybymultiplyingthepredictedtax‐exempt
premiumamountby(1‐taxprice)asdefinedpreviously.
Page 13
13
EmployerSharedResponsibilityRequirement:Ifanemployerwithatleast50full‐
timeequivalentworkersdoesnotoffercoveragein2014andanyfull‐timeemployee
receivesapremiumtaxcreditforpurchasinginsuranceinanexchange,theemployerwill
payanannualpenaltyof$2,000timesthenumberoffull‐timeemployeesless30.Weused
thefollowingformulatoestimatetheestablishment’spenalty:
(5) Penalty=(numberofthefull‐timeworkersatfirm–30)*$2,000*(numberofemployeesinestablishment/numberofemployeesinthefirm).
BecausetheMEPS‐ICdoesnotidentifythenumberoffull‐timeworkersatthefirmlevel,we
usedinformationonthenumberofpart‐timeandfull‐timeworkersateachestablishment
andassumedthatthedistributionsatthefirmandestablishmentlevelsaresimilar.
IfanemployerthatcurrentlyoffersESIdropsitin2014,itsworkerswilllosethe
valueoftheESItaxsubsidy.Weassumetheemployermustadjustwagesupwardto
maintainworkers’totalcompensationatitspriorlevel.13Weraisedtheincomesof
workersinestablishmentsthatofferedinsurancebyanamountequaltothetax‐exempt
premiumlesstheemployersharedresponsibilityrequirement(“employerpenalty”)and
dividedby1.0765(sincetheemployerwillhavetopaypayrolltaxesontheincreasein
wages).Weadjustedincomesofworkersinlargerestablishmentsthatdidnothavean
offerdownwardbytheamountoftheemployerpenaltydividedby1.0765.
ValueofExchangeSubsidies:In2014,individualswholackanESIoffercanbuysubsidized
insuranceinexchanges.Thevalueoftheexchangesubsidyequalsthepremiumofa
benchmarkhealthinsuranceplaninagivenstate,lessapercentagebasedonthefamily’s
incomerelativetothefederalpovertylevel(FPL).Thebenchmarkplanisthesecond‐
13Weassumetheadjustmentwillbethroughwagesonly,butemployerscouldadjustothertypesofcompensation,suchasmakingadditionalcontributionstoretirementplans.
Page 14
14
lowestcost“silverplan”whichhasa70percentactuarialvalue(AV).TheCongressional
BudgetOfficehasestimatedthatthepremiumofasilverplanwillbe$5,300forsingle
coverageand$15,000forfamilycoveragein2016($4,762and$13,476in2014dollars,
assuming5.5%annualrealgrowth).Whilethereisstillsomeuncertaintyovertheamount
ofthesubsidy,anestimatemaybegeneratedfromoneofthepubliclyavailablesubsidy
calculators.Forexample,usingthesubsidycalculatorontheKaiserFamilyFoundation
website,afamilyoffour(2adultsages35and38withtwochildren)earning$31,750in
2014(133%FPL)wouldpay$635out‐of‐pocketforapolicythatcosts$11,282
(http://kff.org/interactive/subsidy‐calculator/,2013).WeusedtheFPLguidelinesandthe
subsidyschedule,alongwiththeestimatedsilverplanpremiums,tocalculatetheout‐of‐
pocketmaximumandthevalueofthesubsidythataworkerwouldreceivebasedon
his/herfamilyincome.
Weconsidertwootherchangestoinsurancemarketsthatarelikelytobeaffectedby
theACA:changesinadministrativeloadingfeesandproductofferings.
AdministrativeLoadingFees:ThevaluesofESIandtheindividualexchangeoption
maydifferbecauseofdifferencesinloadingfees,definedasthepercentageofexpected
benefitspaidoutforadministrativecostsandprofits.Karaca‐Mandic,Abraham,andPhelps
(2011)foundaloadingfeegradientbyfirmsizefrom37%infirmsoflessthan100
workersto4%infirmswithmorethan10,000workers.Historically,estimatedloading
feesfortheindividualmarkethavebeenquitehigh.Proponentsofinsuranceexchanges
arguethatloadingfeeswillbelowerbecauseoftheeliminationofmedicalunderwriting
andstrongercompetitionamonginsurersresultingfrombetterinformationonpriceand
quality.Additionally,theACAimplementedmedicallossratio(MLR)regulationinJanuary
Page 15
15
2011fortheindividualandfully‐insuredsmallandlargegroupmarkets.TheMLRisthe
percentofpremiumrevenuespaidoutforclinicalbenefits.Theregulationspecifiesa
minimumMLRof80%intheindividualandsmallgroupmarkets(approximatelya20%
loadingfee).Toaccountforthisregulation,weassumeESIloadingfeesof20%forfirms
withfewerthan100workers;15%forthosewith100‐9,999workers;and4%forfirm
sizesof10,000ormore.Basedonindividualmarketinsurers’earlycompliancewiththe
MLRregulation,weassumealoadingfeeof23.5%forindividualexchangecoverage.
SpendingonCoveredServices:IndividualsswitchingfromESItoanindividual
exchangepolicymaychooseasilverplan(with70%actuarialvalue)becausethatplan’s
premiumdeterminestheirsubsidy.Alternatively,theymaypurchasecoveragesimilarto
whattheiremployeroffered.EmpiricalevidencesuggeststhatESIplans,onaverage,have
ahigheractuarialvaluethanplansintheindividualmarket(Gabeletal.2012).We
assumedthatworkerswillpurchasecoveragesimilartoESI.Toaccountforthis,weused
informationonESIpremiumsandourloadingfeeassumptionsforESIandexchange
coveragetoestimatetheexchangepremiumfor“ESI‐like”coverage.
Finally,weaggregatedtheeffectspredictedaboveandexpressedthemaschangesin
thetaxprice:
(6) BaselineESItaxprice=1‐(ESItaxsubsidy/predictedESItotalpremium)
(7) 2014ESItaxprice=1‐((ESItaxsubsidy+employerpenalty+individualmandate)/predictedESItotalpremium)
(8) 2014individualexchangetaxprice=1‐((exchangesubsidy+individualmandate)/exchangepremiumfor“ESI‐like”plan)
Weillustratethesecalculationswithahypotheticalexample.Supposewewantto
quantifytheeconomicincentivesforamedium‐size,self‐insuredemployertoofferESI
Page 16
16
versusthe“outsidegood”in2014.Let’ssupposethepremiumofferedbytheemployeris
$6,000andthatthemarginaltaxratesofworkersresultinanaveragetaxpriceof.697.If
theemployeroffersESI,allemployeeswhoareeligibleforcoveragecanhavea$1,818tax
subsidy($6,000*(1‐.697)),giventhetaxtreatmentofESIpremiums.Also,iftheemployer
offerscoverage,anemployeeavoidsthepenaltythatwouldbepassedontohiminlower
wages($1,394)andtheindividualpenaltyifhegoeswithoutinsurance(assumedtobe
$1,000).14Weneedtoexpresstheseincentivesrelativetothepremium.Inthisstylized
example,thenetcostofanESIpolicy(versustheoutsidegood)is$1,788andthetaxprice
is.298(1‐(($1,818+$1,394+$1,000)/$6,000)).
Wecanalsoconsiderthechoiceofanindividualexchangepolicyversustheoutside
goodin2014.UnlikeESI,thereisnotaxsubsidyforthepremiumpaidbytheworker.Let’s
assumethattheexchangepremiumforanESI‐likeplanis$6,175.Wearrivedatthisvalue
bymultiplyingthepredictedESIpremiumfortheworkerbyafactorequalto((1+exchange
loadingfee)/(1+ESIloadingfee))toestimatehowmuchtheexchangepremiumwouldbe
for“ESI‐like”coverage.Iftheworkerhaslowfamilyincomeanddoesnothaveaccessto
anotherESIofferthroughaspouse,hewillqualifyforapremiumtaxcredit.Forexample,a
singleworkerearningapproximately$23,000peryearorabout200%ofFPLwouldbe
requiredtopayupto6.3%ofhisincometowardthepremium.Ifoneassumesthispolicyis
similartoasilverplan,thenthevalueofthecreditwouldbeabout$3,300.Bypurchasing
coverage,theworkeralsoavoidstheindividualmandatepenalty.Inthiscase,thecostofan
individualexchangepolicyis$1,875($6,175‐($3300+$1,000))andthetaxpriceis.303.
14Thewageadjustmentresultingfromtheemployerpenaltyassumesthateachworkerbearsanequalshareofthepenaltyimposedonthefirm.Thesechangesinwagesmaynotoccurinstantlyin2014.Wemodelthelongruneffectsasifthechangeshavefullyoccurred.
Page 17
17
C. PredictingtheOptionsofESIOffer,IndividualExchange,andOutsideGood
Afteraggregatingtheeffectsdescribedaboveandexpressingthemastaxprices,we
usedthemodelestimatesandthenewESItaxpricetopredicttheprobabilitiesofanESI
offerversustheoutsidegoodin2014.Wethenrepeatedthisstepusingthe2014
individualexchangetaxpricetopredictitsprobabilityversustheoutsidegood.
Inthefinalstepinthisexercise,werecoveredtheunconditionalprobabilityofthe
outsidegoodthroughalgebraicmanipulation:
(9)
)1)exp()/(exp(1
1
)exp(/
)exp(Pr/Pr
bXbXPR
PRPRPR
bXPPR
bX
EXCESINO
NOEXCESI
EXCNOEXC
ESINOESI
WeusedthesamemethodtosolvefortheprobabilitiesofofferingESIandtheexchange
option.
IV. ResultsandStudyLimitations
A. Results
Table3providesdescriptivestatisticsbyfirmsizefortheaveragetaxsubsidy,
individualmandate,employerpenalty,andexchangesubsidyperworkerinestablishments
thatofferinsuranceandthosethatdonot.Allestimatesareweightedbythenumberof
workersintheestablishment.
Asnotedabove,thevalueoftheESItaxsubsidyisafunctionofworkers’incomes
andplanpremiums.Acrossallestablishments,themeansubsidyis$2,728perworker.
Page 18
18
Amongemployersthatofferinsurance,theaveragetaxsubsidyperworkerrangesfrom
$2,487forsmallfirmsto$3,065forlargeones.Theestimatedaveragetaxsubsidiesfor
non‐offeringemployersareconsiderablylower.
Acrossworkersinallestablishments,theindividualmandatepenaltyperworkeris
$1,615.Workersinnon‐offeringsmallfirmshavethelowestindividualmandatepenalties.
Smallfirms(andtheirworkers)arealsoexemptfromthesharedresponsibility
requirementiftheydonotoffercoveragein2014.Incontrast,theaverageemployer
penaltyamonglargerfirmsthatdonotofferinsuranceis$1,566perworker.
Theaverageexchangesubsidyperworkerreflectsdifferencesinworkers’incomes
acrossfirmsizes.Notably,theaverageexchangesubsidyforworkersinnon‐offeringsmall
andmediumemployersisabouttwiceaslargeasthosethatsponsorinsurance($3,197vs.
$1,473formediumsizeemployers),suggestingthatworkersinnon‐offeringemployers
haveconsiderablylowerincomes.
Table4summarizesthetaxpricesforthebaselineESIofferaswellasthe2014ESI
optionandthe2014individualexchangeoption.Overall,theaveragetaxpriceforESIat
baselineis.67,withthelargestvariationbythewagecategoryofanestablishment.The
secondandthirdcolumnsofTable4summarizetheaveragetaxpricesforESIandthe
individualexchangeoptionin2014.RecallthattheESItaxpriceincorporatesthevalueof
theESItaxsubsidyandtheindividualmandate,whereasthe2014individualexchangetax
priceincludesthevalueofexchangesubsidiesandtheindividualmandate.
Theaverage2014ESItaxpriceis.30,whichisconsiderablylowerthanthebaseline
ESItaxprice.Allelseequal,thisimpliesthatanESIofferwillbemuchmoreattractivein
2014whentheACAprovisionsgointoeffect.Workerswhochoosetheindividualexchange
Page 19
19
optionforgotheESItaxsubsidy,buttheymaybeeligibleforexchangesubsidiesiftheyare
lower‐incomeanddonothaveaccesstoanotherESIofferthroughaspouse.Theestimated
averagetaxpricefortheindividualexchangeoptionis.52,suggestingthatittoois
relativelymoreattractivecomparedwithESIatbaseline.
Table4alsoillustrateshowtheeconomicincentivesinthe2014taxpricesvaryby
workforceandestablishmentcharacteristics.Weobserveamodestchangeintheaverage
ESItaxpriceforsmallfirms(0‐49workers)betweenbaselineand2014(.68to.45)
comparedwithlargefirms(.66to.22),dueinparttotheexemptionofsmallfirmsfromthe
sharedresponsibilityrequirement.
Table5reportssimulationsfortheprobabilitiesofanESIoffer,theindividual
exchangeoption,andtheoutsidegoodfollowingtheintroductionofkeyACAprovisions.
Column(1)istheaveragepredictedprobabilityofthebaseline(pre‐2014)ESIoffer.
Columns(2),(3)and(4)areunconditionalaveragepredictedprobabilitiesofanESIoffer,
theindividualexchangeoption,andtheoutsidegoodin2014.
ThetoprowinTable5providestheresultsfortheoverallpopulation,weightedby
thenumberofemployeesinprivate‐sectorU.S.establishments.Theaveragepredicted
probabilityofanESIofferfallsfrom.83atbaselineto.66in2014(seecolumn2).However,
muchofthisdeclineisoffsetbytheindividualexchangeoption,whichhasanaverage
predictedprobabilityof.26.Theoutsidegoodprobabilityfallsfrom.17atbaselineto.08in
2014.
Table5alsoreportstheaveragepredictedprobabilitiesofeachoptionbyfirmsize.
Atbaseline,theaveragepredictedprobabilityofanESIofferforverysmallemployers(<9
workers)is0.35comparedwith0.95forlargeemployers(500ormoreworkers).We
Page 20
20
observeanon‐linearpatternofpredicteddeclinesinESIoffers:decreasesinESIoffersare
modestforverysmallandverylargeemployers(11%fortheformerand15%forthe
latter),butthechangesareconsiderablylargerforemployerswithbetween25and499
workers.
ThebiggestdeclineinESIofferprobabilityisforworkersintheaccommodation,
foodservice,entertainmentandrecreationservicesindustries,wheretheprobabilityofan
ESIofferfallsfrom.69to.42(39percentdecline).However,themodelalsopredictsthat
manyworkersintheseindustrieswillshifttotheindividualexchangeoption,basedonthe
offeringprobabilityestimateof.43.Otherindustriespredictedtohavelargedecreasesin
ESIoffersareconstruction(.70to.49)andretailandwholesaletrade(.86to.65).In
contrast,wepredictthesmallestchangesinESIoffersfortheprofessionalservicesand
finance,insurance,realestateandcompanymanagementindustries.
ThereisaclearpatternofESIoffersbasedonthepresenceofunionizedworkers.At
baseline,establishmentswithnounionizedworkershaveanaverageprobabilityofoffering
ESIof.81comparedwith.93forestablishmentswithaunionpresence.Themodelpredicts
thattheaverageprobabilityofofferingESIinestablishmentswithoutaunionpresencewill
dropto.63(22%drop),butemployerswithaunionpresencewillseeasmallerdeclineto
.82(11%drop).
Thedemographiccompositionofanestablishment’sworkersisimportantforESI
offers.ThesimulationsshowdifferentialchangesinESIoffersbyage,wherebyemployers
withayoungworkforce(lessthan20%oftheworkersage50orolder)havealarger
predicteddecreaseintheprobabilityofofferingESIcomparedwithestablishmentswith
higherconcentrationsofolderworkers.
Page 21
21
B. StudyLimitations
Severallimitationsareworthnoting.First,wedidnotconsiderotherACA
provisionsthatwillaffectemployers.Oneistheavailabilityofpremiumtaxcreditsfor
smalllow‐wageemployerstoofferESI.Whilethissubsidymayincreasetheprobabilityof
ESIoffers,earlyevaluationssuggestthatithasbeenineffective(GeneralAccountingOffice,
2012).
Wedidnotmodelthevalueofcost‐sharingcreditsthatwillbeavailableforworkers
withincomesbelow250%FPLwhoobtainexchange‐basedcoverage.Thesecreditswill
raisetheactuarialvalueofthesilverplantoamoregenerousplanwith94%actuarialvalue
forindividualswithincomeof100‐150%FPL,87%forthosebetween150‐200%FPL,and
73%forthosebetween200‐250%(www.kff.org,2010).Thiswillincreasethevalueofthe
exchangesubsidyandwillmakethisoptionmoreattractiveforemployerswithworkers
whowouldqualify.
AnotherESIprovisionwedidnotmodelwillimposepenaltiesonemployersthatdo
notoffer‘affordable’coverageforcertainlower‐incomeworkers.Weassumeemployers
willofferaffordablecoverage.
Wealsousedanaverageexchangepremiumthatvariesonlybyfamilycoverage
tiers(single,employee‐plus‐one,andfamily).In2014,exchangepremiumswillbe
determinedbymodifiedcommunityrating.Adjustmentswillbebasedonfamilycoverage
tiers,age(3:1rateband),geography,tobaccousestatus,andtheactuarialvalue.Whilewe
accountforfamilycoveragetiersinthemodel,wedidnotconsidereithertheage
distributionofanestablishment’sworkersorgeographywhenestimatingtheexchange
premiumfor“ESI‐like”coverage.Olderworkerswillfacehigherpremiumsforindividual
Page 22
22
exchangepoliciesthanyoungerworkers.Thiswillleadtoalargerindividualexchangetax
priceandreducetheattractivenessofthisoptionforfirmswitholderworkers.
Additionally,inestimatingthecompensationreturnedtoemployeesifanemployer
shiftsfromofferinginsurancetonotofferingit,weassumedthateachworkerwould
receivetheaverageemployerportionofthepremium.Itisdifficulttoknowatpresent
whetheremployerswillbehaveinthiswayoruseamorerefinedapproachwhenmaking
compensationadjustments.Moreover,ouranalysesdonotassumeanystickinesswith
respecttoadjustingcompensationpackages.Ifthisisdifficult,theremaybealagindrops
ofcoverageuntilcompensationpackagescanbeadjusted.Asecondimportantlimitationis
thatweusedtheestimatedparametersfromtheESIoffermodeltopredicttheprobability
oftheindividualexchangeoption.Implicitly,weassumedthatworkersperceivethe
qualityofESIandexchangeinsuranceasidentical.Althoughwedidaccountfordifferences
intaxprices,loadingfees,andcoveredspending(weassumedthatworkerswhopurchase
individualexchangecoveragewill“buyup”theircoveredspendingrelativetothesilver
plan),wewerenotabletoaccountforotherpotentiallyimportantfactors–suchashigher
shoppingcostsintheexchanges.Becausetheindividualexchangeoptiondoesnotyet
exist,onecannotknowhowimportantthesefactorswillbe.
Also,wesimulatedtheprobabilitiesofthethreeoptionssolelythroughmeasurable
changesinthetaxpricewithoutmakingexplicitadjustmentsforthedegreeof
“substitutability”betweenESIandindividualcoverageorthepersistenceofESI,asother
simulationmodelshavedone.IfworkersperceivetheircurrentESIoffersasbetterthan
exchangecoverage,ourmodelmaypredicttoomuchflightfromESItoexchanges.
Page 23
23
Finally,oneoftheACA’sgoalsistoimprovethefunctioningoftheunregulated
individualinsurancemarket(the“outsidegood”inourmodel).TotheextenttheACAis
effectivethismayreducetheprobabilitiesofbothESIandindividualexchangeoptions.
Thebestwaytoestimatethe“full”choicemodelthatwillbeavailablein2014maybeto
usenestedlogit,withESIandindividualexchangecoverageinthe“insurancenest”versus
theoutsideoption.IfemployeesperceiveESIandindividualexchangesassimilar,thismay
alsoreducetheprobabilityofchoosingtheinsurancenest.
V. DiscussionandPolicyImplications
TheeconomicincentivescreatedbytheACAarepredictedtohavealargeeffecton
theoveralldemandforinsurancebyU.S.workersandonthesourcesthroughwhichthey
obtaincoverage.Oursimulationspredictthatoveralldemandforinsurancewillrisein
responsetotheACA,drivenlargelybyworkers’desiretoavoidtheindividualmandate
penaltyandtheavailabilityofsubsidizedprivateinsuranceinexchanges.Notably,the
“outsidegood”optionwhichincludesbeinguninsuredispredictedtobe8%forworkersin
private‐sectorU.S.establishments,aratefarlowerthanthe17%baseline(pre‐ACA)rate.
Employerswhoseworkersstandtobenefitthemostfromtheavailabilityofanew
option–purchasingsubsidizedprivateinsuranceinexchanges–aretheleastlikelytooffer
ESIinresponsetotheACA.Whilesmallandmedium‐sizeemployershavelowerbaseline
ratesofESIoffers,oursimulationshowsadditionalmovementawayfromESI.TheACA
providesfewerincentivesforsmallfirmstoofferESIcomparedwithlargerfirms,giventhe
absenceofasharedresponsibilityrequirement.Medium‐sizeemployersmayshiftaway
fromESIandtowardtheindividualexchangeoptionbecauseworkersinthese
establishmentshavelowerincomesandcanreceivelargerexchangesubsidiesthan
Page 24
24
workersinlargefirms.Establishmentsinaccommodation,foodservice,entertainment,
recreationservices,andconstructionindustriesalsoarelikelytohavehigherpercentages
ofworkersqualifyingforpremiumtaxcredits.Thismayshifttheirmostpreferredchoice
fromofferingESItoofferingtheexchangeoption.
TherearecleartradeoffsforworkersbetweenpurchasingESIandindividual
coverageinexchanges.First,giveneconomiesofscale,ESIgenerallyhaslower
administrativeloadingfeesthanindividualcoverage.Asaresult,individualsgetmore
“bangfortheirbuck”throughgroupcoverage.However,employersmaynotofferworkers
achoiceofplansandtheplan(s)theyoffermaynotbewhattheindividualswouldhave
chosenontheirown.Withtheintroductionofexchangesandtheavailabilityofdifferent
typesofplansbasedonactuarialvalue,workersmayselectplansthatalignmoreclosely
withtheirpreferences.
Second,whileanemployer’sdecisiontoofferinsuranceisbinary,thesubsequent
effectsofthatdecisiononworkersmaybequitevaried.Clearly,low‐wageworkerswho
lackaccesstoESIthroughaspousestandtogainthemostfrompremiumtaxcredits
offeredinexchanges.Incontrast,high‐wageworkersmaynotqualifyforanysubsidyand
paythefullpremium.Further,becauseexchangepremiumswillbeage‐rated,older
workerswilllikelyfacehigherpremiumsthantheywouldinanemployergroupsetting.
Finally,thereisstillmuchuncertaintyregardinghowwellinsuranceexchangeswill
functionintermsofproductofferings,premiums,andtheconsumerexperience.
Employersmaytakea“waitandsee”attitudebeforeswitchingfromESItotheindividual
exchangeoption.Lookingahead,itwillbeimportanttomonitoremployerresponsesand
toassesstheimpactofACAprovisionsonworkers’choices.
Page 25
25
References
Abraham,J.,P.Graven,andR.Feldman.“NewEvidenceonEmployerPrice‐SensitivityofOfferingHealthInsurance.”2013,Unpublishedmanuscript.
DeNavas‐Walt,C.,B.Proctor,andJ.Smith.“Income,Poverty,andHealthInsuranceintheUnitedStates:2011,”U.S.CensusBureau,September2012,http://www.census.gov/prod/2012pubs/p60‐243.pdf
Dowd,B.andR.Feldman.“VoluntaryReductioninHealthInsuranceCoverage:ATheoreticalAnalysis.”EasternEconomicJournal.1987,33(3):215‐231.
Feenberg,D.andE.Coutts.“AnIntroductiontotheTAXSIMModel.”JournalofPolicyAnalysisandManagement,1993,12(1):189‐194.Gabel,J.,R.Lore,R.McDevitt,J.Pickreign,H.Whitmore,M.Slover,E.Levy‐Forsythe.“MorethanHalfofIndividualHealthPlansOfferCoveragethatFallsShortofWhatCanBeSoldThroughExchangesasof2014.”HealthAffairs.2012,31(6):1‐10.GeneralAccountingOffice.“SMALL EMPLOYER HEALTH TAX CREDIT Factors Contributing to Low Use and Complexity.” May 2012. Accessed at: http://www.gao.gov/assets/600/590832.pdf.Gruber,J.andM.Lettau,“HowElasticIstheFirm’sDemandforHealthInsurance?”JournalofPublicEconomics,88:7(2004),1273‐1293.
Karaca‐Mandic,P.,J.Abraham,andC.Phelps.“HowdoLoadingFeesVarybyGroupSize?:ImplicationsforHealthReform.”InternationalJournalofHealthCareFinanceandEconomics.2011,11(3):181‐207.
Marquis,M.S.andS.Long,“Workerdemandforhealthinsuranceinthenon‐groupmarket.”JournalofHealthEconomics.1995,14:47‐63.
Page 26
26
Table 1: Descriptive Statistics
Variable SmallEmployers MediumEmployers LargeEmployers
NoOffer(n=32395)
Offer(24587)
NoOffer(N=1351)
Offer(N=12431)
NoOffer(n=859)
Offer(N=22069)
Averagetaxprice 0.7072 0.6665 0.7226 0.6683 0.6903 0.6624
Averagefamilyincome($10,000s) 4.0537 6.1835 3.8016 6.1088 4.6561 6.5925
Multi‐locationestablishment(%) 2.31 5.43 28.28 36.78 76.29 92.6
Businesstenure<1year 2.05 0.39 0.99 0.29 0.37 0.11
Businesstenure1‐4years 18.11 8.01 11.22 4.37 3.46 1.1
Businesstenure5‐9years 19.89 13.41 15.24 7.45 16.65 2.09
Businesstenure10ormoreyears 59.96 78.19 72.54 87.89 79.53 96.71
Non‐profitownership 6.49 10.66 4.96 19.76 3.11 14.27
Retailorwholesaletrade 14.91 17.35 9.9 13.83 14.11 20.23Accommodation,foodservice,entertainment/recreationservices 23.91 7.49 34 11.66 20.33 9.75Personaloradministrative,buildingsupportservices 12.63 8.88 19.89 7.18 35.7 10.32
Professionalservices 18.83 28.16 15.03 30.16 14.78 26.55
Religious,civilorothernon‐profit 4.13 5.64 1.1 3.01 0.08 0.31Finance,insurance,realestate,companymanagement 5.46 7.51 1.86 7.47 7.87 13.36
Manufacturingormining 3.8 9.25 5.89 15.63 3.78 11.71
TransportationorUtilities 2.58 2.53 2.72 2.92 2.54 6.14Construction,agriculture,forestry,fishing,hunting,orunknown 13.76 13.19 9.61 8.15 0.8 1.61
Someunionemployees 2.41 4.46 6.57 10.33 10.36 28.65
Lessthan20%ofworkersage50orolder 59.54 52.03 75.74 46.6 64.49 40
20‐50%ofworkersage50orolder 26.27 36.25 19.26 46.81 32.79 54.76
Morethan50%ofworkersage50orolder 14.19 11.72 5.01 6.6 2.72 5.24Lessthanorequalto33%ofworkersarefemale 39.7 46.34 30.94 36.77 24.33 22.15
Greaterthan33‐66%ofworkersarefemale 24.58 25.1 36.59 33.96 50.94 49.71
Morethan66%ofworkersarefemale 35.72 28.56 32.46 29.27 24.72 28.14Percentofestablishment'sworkersthatarepart‐time 40.93 17.81 42.08 17.31 26.69 20.8
Countyunemploymentrate 8.21 7.94 8.24 8.03 8.26 8.01Proportionofestablishmentsinacountythathaveover50workers 0.0504 0.0533 0.054 0.0548 0.0556 0.0568Notes:Allvaluesweightedbyemployeesinanestablishment. *DenotesvariablesimputedfromtheMEPS‐HouseholdComponent
Page 27
27
Table 2: Binary Logit Model of Establishment’s Probability of Offering ESI
Variable SmallEmployers MediumEmployers LargeEmployers
MarginalEffect
SE ZMarginalEffect
SE ZMarginalEffect
SE Z
Averagetaxprice ‐.2812 .0227 ‐12.3731 ‐.2319 .0441 ‐5.2556 ‐.1383 .0524 ‐2.6395
Averagefamilyincome($10,000s) .0972 .0011 85.9065 .0665 .0029 22.8243 .0450 .0027 16.6342
Firmsize0‐9 Ref Ref Ref . . . . . .
Firmsize10‐24 .2519 .0055 46.1494 . . . . . .
Firmsize25‐49 .3616 .0079 46.0304 . . . . . .
Firmsize50‐99 . . . Ref Ref Ref . . .
Firmsize100‐499 . . . .0398 .0074 5.3815 . . .
Firmsize500andmore . . . . . . . . .
Multi‐locationestablishment(%) ‐.0246 .0106 ‐2.3305 ‐.0185 .0075 ‐2.4649 ‐.0466 .0247 ‐1.8885
Businesstenure<1year Ref Ref Ref Ref Ref Ref Ref Ref Ref
Businesstenure1‐4years .0583 .0212 2.7449 ‐.0162 .0497 ‐.3261 .0403 .0455 .8863
Businesstenure5‐9years .0981 .0211 4.6430 ‐.0183 .0488 ‐.3760 .0727 .0431 1.6865
Businesstenure10ormoreyears .1653 .0207 7.9665 .0084 .0477 .1763 .0841 .0403 2.0858
Non‐profitownership .0990 .0112 8.8098 .0537 .0101 5.3236 .0845 .0149 5.6632
Retailorwholesaletrade .1108 .0137 8.1130 .0576 .0349 1.6481 .2031 .0414 4.9005
Accommodation,foodservice,entertainment/recreationservices
.0115 .0145 .7941 .0605 .0345 1.7534 .1998 .0422 4.7365
Personaloradministrative,buildingsupportservices
.0502 .0142 3.5414 ‐.0015 .0359 ‐.0406 .0833 .0412 2.0196
Professionalservices .0045 .0126 .3615 .0244 .0332 .7364 .0482 .0399 1.2087
Religious,civilorothernon‐profit
Ref Ref Ref Ref Ref Ref Ref Ref Ref
Finance,insurance,realestate,companymanagement
.0069 .0141 .4929 .0074 .0371 .2007 .0840 .0408 2.0576
Manufacturingormining .1524 .0156 9.7606 .0687 .0357 1.9228 .2139 .0475 4.5058
TransportationorUtilities .0675 .0177 3.8119 .0317 .0394 .8050 .1703 .0445 3.8224
Construction,agriculture, .0792 .0140 5.6608 .0276 .0408 .6762 .2110 .0502 4.2024
Page 28
28
forestry,fishing,hunting,orunknown
Someunionemployees ‐.0322 .0131 ‐2.4588 ‐.0175 .0201 ‐.8679 ‐.0486 .0082 ‐5.9307
Lessthan20%ofworkersage50orolder Ref Ref Ref Ref Ref Ref Ref Ref Ref20‐50%ofworkersage50orolder
.0286 .0050 5.7603 .0224 .0078 2.8813 ‐.0149 .0067 ‐2.2336
Morethan50%ofworkersage50orolder
‐.0482 .0058 ‐8.3323 ‐.0232 .0134 ‐1.7304 ‐.0073 .0123 ‐.5890
Lessthanorequalto33%ofworkersarefemale
Ref Ref Ref Ref Ref Ref Ref Ref Ref
Greaterthan33‐66%ofworkersarefemale
.0782 .0058 13.4395 .0127 .0096 1.3185 ‐.0099 .0089 ‐1.1104
Morethan66%ofworkersarefemale
.0498 .0056 8.8586 ‐.0081 .0102 ‐.7917 .0322 .0101 3.2002
Percentofestablishment'sworkersthatarepart‐time
‐.0023 .0001 ‐36.0305 ‐.0012 .0001 ‐11.0864 ‐.0002 .0001 ‐2.0251
Countyunemploymentrate ‐.0036 .0013 ‐2.7084 ‐.0018 .0018 ‐1.0009 .0003 .0018 .1429
Proportionofestablishmentsinacountythathaveover50workers
.5335 .1616 3.3006 ‐.2657 .2604 ‐1.0201 .5674 .3036 1.8692
Year2008 Ref Ref Ref Ref Ref Ref Ref Ref Ref
Year2009 ‐.0034 .0067 ‐.5077 .0057 .0109 .5262 .0045 .0093 .4811
Year2010 ‐.0108 .0070 ‐1.5351 ‐.0002 .0110 ‐.0137 .0117 .0102 1.1455
Page 29
29
Table 3: Economic Incentives of Establishments to Offer Insurance in 2014
AveragePerWorker Overall SmallEmployers
MediumEmployers
LargeEmployers
NoOffer Offer NoOffer Offer NoOffer OfferESITaxSubsidy $2728
$1789 $2487 $1734 $2709 $2021 $3065
IndividualMandate $1615
$1106 $1667 $1004 $1621 $1192 $1734
EmployerPenalty $1033
$0 $0 $601 $1156 $1439 $1566
ExchangeSubsidy $1573
$2982 $1289 $3197 $1473 $2368 $1330
Page 30
30
Table 4: Estimated Tax Prices Overall and by Selected Establishment and Workforce Characteristics
CharacteristicBaselineESITaxPrice
2014ESITaxPrice
2014IndividualExchangeTaxPrice
Overall 0.67 0.3 0.52
FirmSize
0‐49 0.68 0.45 0.37
50‐99 0.68 0.33 0.41
100‐249 0.67 0.28 0.56
250‐499 0.67 0.26 0.56
500ormore 0.66 0.22 0.61
Industry
Retailorwholesaletrade 0.69 0.31 0.49Accommodation,foodservice,orentertainment/recreationservices 0.74 0.4 0.21
Personaloradmin/buildingsupportsvc 0.67 0.29 0.44
Professionalservices 0.65 0.28 0.66
Religious,civilorothernon‐proforgs 0.65 0.35 0.66
FIREorcompanymanagement 0.65 0.25 0.69
Manufacturingortiming 0.65 0.26 0.58
Transportationorutilities 0.67 0.28 0.6
Construction 0.66 0.36 0.37Agriculture,forestry,fishingorhunting,unknown 0.69 0.39 0.17
UnionStatus
Nounionemployees 0.67 0.31 0.49
Someunionemployees 0.65 0.24 0.67
AgeCategory
Less20percentoverage50 0.68 0.31 0.43
20‐50%overage50 0.66 0.28 0.6
Morethan50%overage50 0.66 0.35 0.58
SexCategory
Lessthan33%female 0.67 0.31 0.5
34‐66%female 0.68 0.28 0.52
Morethan67%female 0.67 0.32 0.55
WageCategoryofEstablishment
Morethan50%under$11/hour 0.72 0.42 0.19Lessthan50%under$11/hour&less50%over$26/hour 0.66 0.29 0.57
Morethan50%over$26/hour 0.63 0.2 0.72
Page 31
31
Table 5: Simulation Results Overall and by Selected Establishment and Workforce Characteristics
Characteristic
BaselineProbabilityofESIOffer
2014ProbabilityofESIOffer
2014ProbabilityofIndividual
ExchangeOption
2014ProbabilityofOutsideGood
Overall .83 .66 .26 .08 FirmSize 0‐9employees .35 .31 .24 .4510‐24employees .66 .44 .39 .1725‐49employees .78 .48 .43 .0950to99employees .87 .59 .37 .03100‐499employees .93 .68 .30 .02500ormoreemployees .95 .81 .19 .01Industry Retailorwholesaletrade .86 .65 .28 .07Accommodation,foodservice,entertainment,orrecreationservices
.69 .42 .43 .15
Personaloradministrativeandbuildingsupportservices
.74 .62 .27 .11
Professionalservices .87 .74 .20 .06Religious,civil,orothernon‐profitorganizations
.75 .59 .25 .16
Finance,insurance,realestate,orcompanymanagement
.90 .79 .16 .05
Manufacturingormining .94 .75 .23 .02Transportationorutilities .90 .75 .20 .05Construction .70 .49 .35 .16UnionPresence Nounionpresence .81 .63 .28 .1Someunionpresence .93 .82 .16 .02Age Lessthan20%overage50 .79 .60 .3 .0520%to50%overage50 .89 .73 .22 .18Morethan50%overage50 .73 .60 .22 .18Sex Lessthan33%female .81 .63 .27 .134‐66%female .87 .69 .26 .05Morethan66%female .81 .65 .24 .10