ReporttoCongress EffectivenessofHIPAAandStateLawsinEnsuringAccesstoHealthInsuranceintheSmall GroupandIndividualMarkets PreparedbytheHealthCareFinancingAdministration, OfficeofStrategicPlanning, ResearchandEvaluationGroup basedonareportpreparedby RANDCorp.andtheInstituteforHealthPolicySolutions undercontractwiththeOfficeofStrategicPlanning,HCFA ReporttoCongress: EffectivenessofHIPAAandStateLawsinEnsuringAccess toHealthInsuranceintheSmall-groupandIndividualMarkets IntroductionandSummary In1996,CongresspassedtheHealthInsurancePortabilityandAccountabilityAct(HIPAA), legislationintendedtoincreaseaccesstohealthcoverageforsomeemployergroupsand individualswhopreviouslywereunabletopurchaseadequatecoverage. Thelawestablished federalstandardsacrossthelargeandsmall-grouphealthinsurancemarkets. Tomonitorthe effectivenessofthesereforms,TitleI,section191(a)ofHIPAAstatesthat: TheSecretaryofHealthandHumanServicesshallprovideforastudyonthe effectivenessoftheprovisionsofthistitleandthevariousStatelaws,inensuringthe availabilityofreasonablypricedhealthcoveragetoemployerspurchasinggroupcoverage andindividualspurchasingcoverageonanon-groupbasis. Pursuanttothisrequirement,theHealthCareFinancingAdministration(HCFA)contractedwith RANDandtheInstituteforHealthPolicySolutions(IHPS)tosurveytherelevantliteratureand todeveloparesearchdesignforacomprehensiveevaluationofHIPAA. TheRANDprojecthad thefollowingfourcomponents: (1) Developmentofadatabasetodescribetheregulatoryenvironmentineachstatebefore HIPAA'senactmentandtotrackregulatoryactionsstateshavetakentocomplywith HIPAA'srequirements; (2) AreviewoftheliteratureonissuesassociatedwithHIPAAimplementationandwithits anticipatedeffects; (3) AreviewofstrengthsandweaknessesofexistingdatathatcouldbeusedinaHIPAA evaluation;and (4) DevelopmentofaresearchdesignforevaluatingtheeffectsofHIPAAontheaccessibility andaffordabilityofinsuranceinthegroupandindividualinsurancemarkets. ThisreporttoCongressisbasedonfindingsfromtheRANDstudy. PartIofthisreportto Congresssummarizestheliteraturereview,emphasizingHIPAA'spotentialeffectson accesstohealthcoverage,affordabilityofcoverage,insurancebenefitdesign,andemployment decisions. PartIIpresentsapossibledesignforamorecomprehensiveevaluationofaccess issuesassociatedwithHIPAA. Theappendixreviewsexistingdatabasesthatcouldprovide informationnecessarytoconductanevaluationofthemagnitudeoftheseeffects. Eachpartof thereportissummarizedbelow. 2 TheliteraturereviewsynthesizesfindingsfromresearchonHIPAA'simpactonsmall-groupand individualmarkets. Theevidenceavailablesuggeststhatinbothmarkets.HIPAA'seffectsare likelytobesmall. PriortoHIPAA'senactment,moststateshadadoptedsmall-groupmarket reformsthatmetorexceededHIPAArequirements. Thus,itislikelythatnationwide,thelaw's effectsonaccess,coverage,andpremiumsinthegroupmarketwillbelimited. Afewstateshad notpreviouslyadoptedinsurancemarketreformssimilartoHIPA\'s,buteveninthesestates,the literaturesuggeststhattheeffectsofguaranteedcoveragearelikelytobesmallwithout additionalpriceregulation. TheevidencesuggeststhatHIPAA'sindividualmarketreformprovisions~whichgenerallyare morelimitedthanthelaw'sgroupmarketreforms~mostlikelywillhavealimitedeffectonthe numberofuninsuredandonpremiumsintheindividualmarket. However,asignificant innovationinHIPAAisguaranteedaccesstotheindividualmarketforcertainpeoplewhohave lostgroupcoverage. Manystatesareusingtheirhigh-riskpoolstomeetHIPAA's group-to-individualmarketportabilityrequirements. PriortoHIPAA'senactment,anumberof otherstateshadestablishedindividualmarketreformsprovidingportability. DuetoHIPAA'sbroadscopeandtheflexibilityitprovidestostates,RANDwasaskedtoassess thefeasibilityofconductingaquantitativeanalysisofHIPAA'simpactonaccess,coverageand portability. RANDhasidentifiedasmallnumberofkeyimplementationandoutcomequestions relatedtoaccess,premiums,benefits,oremployment—thekeyoutcomesaddressedinthe literaturereview~andhassuggestedoneormoreapproachestoaddresseachquestion. RAND'sproposeddesignemphasizesseverallimitafionsinvolvedinanalyzingHIPAA'seffects withinthetimeframesallowedforthisreporttoCongress. First,animportantlimitationisthat thedatacurrentlyavailableforquantitativeinvestigationarequitelimited. Second,becausemost statespreviouslyhadimplementedreformssimilartothoseinHIPAA,themagnitudeof HIPAA'seffectsareexpectedtobesmallandconcentratedamongsmallsegmentsofthe population,suchasthoseinat-riskindustries,orpersonsinpoorhealth. Thismakesitdifficult tomeasureHIPAA'simpactandtodisentangleitfromtheeffectsofotherongoingdevelopments intheprivatehealthinsurancemarket. Third,thelongtermeffectsofHIPAAmaybedelayed becauseitmaytaketimeforindividualsandbusinessestobecomefamiliarwiththelaw's provisionsandtotakefulladvantageofthem. Foracompleteevaluationoftheimpactof HIPAA,itwouldbepreferabletoallowmoreUmeforthelaw'seffectstoemerge. Animportantlimitationisthatthedatacurrentlyavailableforquantitativeinvestigationarequite limited. Mostexistingsurveysaregeneralpopulationsurveys,whichdonothaveasufficient sampleofstatedataortargetgroups,suchasworkersinhigh-riskindustries,topermitanalysisof changesinaspecificstateorofanarrowerpolicyissue. TheMedicalExpenditurePanelSurvey (MEPS),conductedbytheDepartment'sAgencyforHealthcareResearchandQuality(AHRQ), willprovidenewdatathatwillfacilitateanalysisofsomeHIPAA-relatedissues,suchasthe effectofportabilityprovisionsonthecontinuationofcoverage. Otherimportantissues,suchas theeffectsofdifferentapproachesamongstatestothegrouptoindividualportabilityprovisions, carmotbesystematicallyanalyzedbecauseadequatedataarenotavailable. 3 AnumberofagencieswithinHHS,includingHCFA,areexploringwaystoincreaseour understandingofrecentchangesintheprivatehealthinsurancemarket,includingHIPAA's effects. Activitiesincludethefundingofadditionalresearchandencouragingothersinthe researchcommunitytoexploreissuesrelatedtoinsurancemarketsandreform. Someofthese effortswilladdressissuesraisedinRAND'sstudy. Werecommendwaitinguntiltheresultsof theseresearcheffortsareavailabletodeterminewhetherfurtherevaluationactivitiesspecifically relatedtoHIPAAarewarranted. PARTI.POTENTIALEFFECTSOFHIPAA: AREVIEWOFTHELITERATURE INTRODUCTION TheHealthInsurancePortabilityandAccountabilityActof1996(HIPAA)establishesafederal roleforregulatingtheemployergroupandindividualhealthinsurancemarkets(Atchinsonand Fox,1997). ThegoalsofTitleIofthelegislationaretoprovidecoveragesecurityforthose currentlyinsured. TitleIguaranteestheavailabilityofinsurancetoallsmallemployers(with2 ormoreemployees)andassuresthatindividualswholeaveemploymentareabletomaintain healthinsurancecoverage,providedtheyactwithinthestatutorytimeframes. Thus,HIPAA ensurescontinuedaccesstohealthinsuranceforsomeemployergroupsandindividualswho previouslywereunabletopurchasehealthinsuranceorunabletopurchaseadequatecoverage. Whateffectthiswillhaveonthenumberofuninsuredorthepricepeoplepayforinsuranceis, however,amatterofsomeuncertainty. Moreover,variabilityamongstatesinexistinginsurance legislation,andtheflexibilitythatstatesaregiventoimplementtheindividualmarketreforms suggestthattheanswertothesequestionswillvaryfromplacetoplace. PartIofthisreportexaminestheextantliteratureonissuespertainingtoHIPAAandthegeneral insuranceaccessproblemsHIPAAwasdesignedtoaddress. Thegoalofthereviewisto generatehypothesesaboutthelikelyeffectsofHIPAA,themagnitudeofthoseeffects,howthe effectsarelikelytovaryamongpopulationgroups,andhowbackground,policy,or implementationcharacteristicsmightinfluencethemagnitudeoftheeffects. InthenextsectionofPartIofthisreport,abriefoverviewoftheprovisionsofHIPAAis provided. Thisisfollowedbyadiscussionofthepotentialeffectsoftheseprovisionsinthe groupmarketandtheindividualmarket. Toestimatetheseeffects,RANDsearchedtheliterature inresearchjournals,theemployee-benefitstradepress,andreportsofprivateandpublic organizationsforinformationaboutHIPAAandtheprovisionsitcontains,including:guaranteed issue,guaranteedrenewal,portability,pre-existingconditionexclusionlimits,and non-discriminationprovisions. Theyalsoreviewedliteratureontheperformanceofstate high-riskpools,oneofthemostpopularalternativemechanismchoicesthatstatesareallowedto adopttomeettheHIPAAprovisionsfortheindividualmarketreforms. PartIconcludeswith somesummaryobservationsabouttheimplicationsforanevaluationdesign. 4 BACKGROUND: ABriefReviewoftheLegislation ThespecificprovisionsofHIPAArelatedtothereformofthegroupmarketandtheindividual marketaresummarizedbelow. A. GroupMarketReforms Thegroupmarketreformprovisionsthatapplytomostgrouphealthplans,includingself-insured plans,andtogrouphealthinsuranceissuers,limitpre-existingconditionexclusionsandprohibit exclusionofindividualsfromagrouphealthplanbasedonhealthstatus. Inaddition,health insuranceissuersarerequiredtoguaranteerenewabilityofcoverageforallgroupsandguarantee issueallproductsforsmall-groups. Pre-existingconditionexclusionsforallgrouphealthplans(includinginsuredandself-insured plans)arelimitedto12months(18monthsforlateentrants)forconditionstreatedordiagnosed inthe6monthspriortothefirstdayofcoverageunderthegrouphealthplan,orthefirstdayofa waitingperiodforcoverage. Moreover,grouphealthplansmustcreditpriorpublicorprivate healthcoveragetowardreducingoreliminatingpre-existingconditionperiods,providedthe coveragehasnotlapsedformorethan63days(group-to-groupportability). HIPAAprohibitsallgrouphealthplans(insuredorself-insured)fromdenyingcoverageor charginghigherpricestoindividualsbasedonhealthstatus. Itrequiresthathealthinsurance companiesguaranteerenewalofallcoverageissued,tobothlargeandsmall-groups. Italso requiresinsurancecompaniesthatofferhealthcoverageinthesmall-groupmarket(2-50 employees)tomakeallproductsavailabletoallapplicants(guaranteedissue). Allsmall-groups mustbeacceptedandalleligiblemembersofagroupmustbeaccepted. B. IndividualMarketReforms Unlessthestateimplementsan"acceptablealternativemechanism,"insurersintheindividual marketarerequiredtoguaranteeissueandapplynopre-existingconditionexclusionsto individualswhohadanaggregateof18monthsofhealthcoveragewithoutanylapseslongerthan 63days,themostrecentunderagroupplan,andwhoapplywithin63daysoflosingcoverage (group-to-individualportability). Theindividualalsomusthaveelectedandexhaustedany continuationcoverageavailableunderCOBRAorStatelaw,andnotbeeligibleforMedicare, Medicaid,oranyhealthinsurance. Individualhealthinsurersmustguaranteeissueatleasttwo differentindividualinsuranceproducts. Theseproductsmaybeeithera)thetwomostpopular productsbasedonpremiumdollars,orb)alowandhighpolicyoptionthatisrepresentativeof individualpoliciesofferedbytheinsurerinthestatethataresubsidized,risk-adjusted,orcovered byarisk-adjustmentmechanism. Thelawalsoappliesaguaranteedrenewalprovisioninthe individualmarket. 5 C. StateImplementation HIPAAprovidesforsubstantialstateflexibilityinimplementingtheintentoftheindividual marketreformstoensurethatthosewholeaveagrouphealthplanareabletomaintaincoverage andarenotdeniedindividualinsuranceifthatistheonlycoverageoptionavailable. Statesmay adoptan"acceptablealternativemechanism"tothefederalprovisionsoutlinedabove. Tomeet therequirementsofan"acceptablealternativemechanism",astateprogrammustprovideall eligibleindividualswithachoiceofcoverage,includingoneprovidingcomprehensivebenefits, andnotimposepre-existingconditionexclusionsonthem. Forexample,20stateswillusehigh riskpoolstomeettherequirements,inmostcasesexpandingonanexistingriskpool(IHPS, 1998). Otherstatealternativeprogramsincludemandatorygroupconversionpolicies,or guaranteedissueofdesignatedindividualpolicies. TwoStates(NewMexicoandUtah)are adoptingacombinationofahigh-riskpoolandanothermechanism. D. FederalImplementation AlthoughHIPAArecognizesthatstateshaveprimaryresponsibilityforregulationofinsurance, thestatuteprovidesenforcementauthoritytoDHHSintheeventthatastatehasfailedto substantiallyenforcefederalrequirementsunderHIPAA. FederalenforcementofHIPAA requirementsinthegroupandindividualhealthinsurancemarketsisdescribedinregulations issuedAugust20,1999. Currently,HCFAisdirectlyenforcingHIPAAregulationsinallorpart ofthehealthinsurancemarketsinCaliforniaandMissouri. HCFAisphasingoutitsenforcement ofHIPAAregulationsinRhodeIsland,whichrecentlyenactedintolawlegislationthatprovides fortheprotectionssetforthinHIPAA. Inaddition,HCFAhasdeterminedthatfourStatesare notenforcingHIPAAamendments. Therefore,HCFAwillbeenforcingregulationsrelatedtothe Women'sHealthandCancerRightsActinNorthDakota,Colorado,andMassachusetts;andit willenforceregulationsrelatedtotheNewborns'andMothers'HealthProtectionActin Wisconsin. AnalysesarecurrentlybeingconductedwithinHCFAtodetermineifotherstatesare inconformancewithHIPAAandrelatedfederalrequirements. EFFECTSOFHIPAA'SGROUPMARKETPROVISIONS HIPAA'sgroupmarketprovisionsweredesignedtoeliminateinsurerpracticesthatdiscriminate againstemployergroupsandtheirmembersonthebasisofhealthstatus,industry,orother characteristics. Redlining(thatis,excludingspecifictypesofbusinessesfromcoverage), denyingcoveragetoemployeeswithpoorhealthortotheirentiregroup,andexcludingcoverage forpre-existingconditionsaredocumentedpracticesthatcanposebarrierstoemployers (especiallysmallemployers)thatwishtoofferhealthinsuranceasabenefit(McLaughlinand Zellers,1994;Zellers,McLaughlin,andFrick,1992). HIPAAeliminatesthefirsttwopractices inthesmall-groupmarket,andlimitstheexclusiononpre-existingconditionsforallgroupplans. Proponentsofthesereformsbelievethattheywillexpandcoverage;criticsarguethattheywill leadtoincreasesinpremiumsthatmightinturnloweraccesstocoverage(AtchinsonandFox, 1997). 6 Theresearchevidenceontheeffectofguaranteedissue,guaranteedrenewal,guaranteed coverageforindividuals,andpre-existingconditionlimitsingroupplansisscarce. Butit suggeststhattheseprovisionshaveonlysmalleffectsonpremiumsandcoverage. Moreover, manystateshadalreadytakenstepstoeliminatediscriminatorycarrierpracticesandhave standardsthatmeetorsurpassHIPAAstandards. Thus,HIPAAimposesfewnewrequirements inthegroupmarket. Thisevidenceisreviewedbelow. A. EffectsonAccess 1. GuaranteedIssue Manystates(38)hadsomeformofguaranteedissuelegislationforsmall-groupsatthetime HIPAAwasenacted(LitowandMcClelland,1997). Ofthese,16requiredguaranteedissueofall insuranceproductsand22hadstatutoryplansthatwereguaranteedissue. Inabout12ofthese states,however,thestatelawdidnotencompassallsmall-groupscoveredbyHIPAA. Inthe24statesthatdonotguaranteeissueofinsuranceproductstosmall-groupsorhave guaranteedissuelegislationthatdoesnotencompassallgroupsfrom2to50,HIPAAmayresult inchangesinthenumberoffirmsthatofferhealthinsurance. Theresearchevidenceonthe effectofthistypeofreformisverylimitedbecausemostofthestates'small-groupreform legislationisrelativelyrecent. Oneempiricalstudyusingstatedatafrom1989-1993suggests thatguaranteedissuelegislationmayresultinanincreaseinthenumberoffirmsoffering insurance(Jensenetal.,1995). Thisstudy'sestimateimpliesabouta10percentagepoint increaseinthenumberofsmallfirmsofferinginsuranceinstatesinwhichavailabilityis guaranteed,holdingconstantotherregulationsandmarketcharacteristics. Ananalysisof1994 NationalEmployerHealthInsuranceSurveydatafoundthatsmall-groupinsurancereforms similartothosemandatedinHIPAAresultedinasmallincreaseintheproportionofemployers offeringplans,butnoincreaseintheproportionofemployeesenrolled(HingandJensen,1999). Effectsonemployerofferingswereonlyobserved3yearsafterimplementationofreforms. Anotherrecentempiricalstudyoftheeffectofinsurancemarketreformsonstates'uninsured ratesfoundthatguaranteedissue,whenpartofapackageofsmall-groupreformsincluding portabilityandlimitsonpre-existingconditions,reducedtheuninsuredrate(Marstelleretal., 1998). However,thisconclusionislikelytovaryfromstatetostatedependingonotherstate regulationsconcerningthehealthinsurancemarket.HIPAAforexample,guaranteesaccessto insuranceforallsmall-groups,butitdoesnotguaranteeaffordability. Whetherguaranteed accessleadstoadditionalgroupinsurancepurchasesislikelytodependonwhetherandwhat typeofratingrestrictionsthestateinsurancelegislationplacesonpremiumsthatinsurerscan chargesmall-groups. Forexample,Marstellerandhercolleagues(1998)suggestthatthereareno changesintheuninsuredratewhenratingrestrictionsarecombinedwiththereformsincluding guaranteedissue. 7 4 2. Pre-ExistingConditionsandPortability Moststates(30)havelegislationforsmall-groupplansthatmeetsorsurpassestheHIPAAlimits onpre-existingconditionsperiodsthatcanbeimposed(Ladenheim,1996). Manyofthesestate regulationsalsocreditpriorcoverageindeterminingtheseperiods(GAO,1995). However,the HIPAAlimitsonpre-existingconditionsandgroup-to-groupportabilitywouldalsoapplytolarge businesses,whicharetypicallynotsubjecttostateregulations,andtoself-fiandedplans,which areexemptfromstateregulation. TheGAOestimatedthatabout12millionemployeesand almost7milliondependentsofemployeesmaybenefit-intermsofreducedoreliminated waitingperiod-fromHIPAA-typegroup-to-groupportabilityprovisions(GAO,1995). Whilepotentiallybenefitingmanycurrentplanenrollees,theHIPAAprovisionsonpre-existing conditionsareunlikelytoleadtoanincreaseinthenumberoffirmsofferinginsurance. Jensen andhercolleagues(1995)foundnoevidencethatrestrictingpre-existingconditionwaiting periodsincreasedofferratesamongsmallemployers. Similarly,theresultsfromtheUrban Institutestudy(Marstelleretal.,1998)foundasmall,andonlymarginallysignificant,effectof limitsonpre-existingconditionsandportabilityifguaranteedissueisnotpartofthepackage. In largerfirms,theHIPAAprovisionswouldbenefitonlyafractionofemployeesandsohave limitedinfluenceonthegroupdecision,assumingthepreferencesofthemedianworker dominate(GoldsteinandPauly,1976). Moreover,offerratesinlargefirms(thosewith50or moreemployees)currentlyexceed90percent(Cantor,Long,andMarquis,1995). Whileitappearsunlikelythatthepre-existingconditionlimitsandportabilitywillincreasethe numberoffirmsofferinginsurance,theseprovisionsmayinducesomeindividualstoenrollin groupplansthattheyotherwisewouldturndown. Researchindicatesthatindividualsaremore likelytopurchaseinsurancewhentheexpectedbenefitoftheplanisgreater(seeforexample. MarquisandHolmer,1996). Thisincrease,however,isalsoexpectedtobesmall,becauseabout 90percentofworkersofferedgroupinsuranceenrollintheplan(LongandMarquis,1993; CooperandSchone,1997). Portabilitymayalsobenefitindividualswhowishtochangejobsbutwhocurrentlystayintheir jobbecauseoffearoflosinginsurance. ThisbenefitoftheHIPAAreformsisexaminedinmore detailbelow. 3. Other Thenon-discriminationprovisionsalsoareunlikelytosubstantiallyaffectfirmofferratesor employeeenrollments. Healthinsuranceofferratesbysmallemployerswerefoundnottobe significantlydifferentinstatesthathadlimitsontheexclusionsallowedunderpre-existing conditionclausesnorweretheysignificantlydifferentinthosethathadlawsprohibiting occupationalexclusionsfromcoverage(Jensen,Morrisey,andMorlock,1995). Medical underwriting,thepracticeofexcludingworkersfromcoverageiftheyhavespecificpre-existing conditions,occurredinbothlargeandsmallemployergroupplans,thoughitwasmorecommon 8 inthelatter. However,itwasrarelypracticed;fewerthan5percentoffirmsreportedthat specificindividualswereexcludedfromcoverage(Cantor,Long,andMarquis,1995). Finally,guaranteedrenewalisnotaneffectiveguaranteeofaccesstocoveragewithoutlimitson premiumincreases(BlumbergandNichols,1996). HIPAAdoesnotprovideforaffordable coverage;itonlyensuresthecontinuedrighttopurchaseaplan. Severalstate-specificexperiencesafteradoptingapackageofregulationssimilartotheHIPAA provisionsalsosuggestthattherewillbeatmostmodestaccesseffects,andthatothermarket factorsandregulatoryconditionsmayinfluenceoutcomes. Californiaenactedapackageof reforms,effectivein1993,fortlrmswith3-50employeesthatincludedguaranteedissue, guaranteedrenewal,limitsonpre-existingconditions,andrestrictionsonpremiumvariability. In thepost-reformperiod,about10percentmorefirmswith3-24employeesofferedinsurancethan inthepre-reformperiod(BuchmuellerandJensen,1997). Nochangeinofferrateswasfound amongfirmswith25-50employees,however. Minnesota'sexperiencewassimilarto California's. Minnesotaalsoenactedapackageofreformsforthesmall-groupmarketthat includedguaranteedissueandrenewal,andlimitsonpre-existingconditionexclusions. After reform,thenumberofenrolleesinthesmall-groupmarketincreasedby8-12percent(Nicholset al.,1997). However,BuchmuellerandJensen(1997)didnotfindincreasesinofferratesamongsmallfirms inagroupofstatesthatadoptedreformssimilartoCalifornia'sduringthesameperiod. They concludedthatintensecompetitionintheCaliforniahealthmarketthatexerteddownward pressureonpricesmayhaveprovidedanenvironmentinwhichtheaccessreformswere successftilinprovidingasmallexpansionincoverage. ObserversofMinnesota'smarketdrew similarconclusions(Nicholsetal.,1997). Moreover,Minnesota'sreformsincludedaprovision prohibitingindividualsfrompurchasingindividualpoliciesontheirowniftheywereeligiblefor groupcoverage. Itisbelievedthatsomeoftheincreaseinthegroupmarketisamovefromthe individualmarket(IHPS,1995). B. EffectsonPremiums TheHIPAAgroupprovisionsensureaccesstocoverageforgroupsandindividualsinthose groups,butalsoplacesomelimitsoninsurers'abilitytosegmentrisks. Insurerscannotcharge higherpricestohighriskindividualsinagroup;theycannotexcludeentiregroupsfromtherisk pool. Risksegmentationresultsinlowerpremiumsforthehealthythanthesick;greaterpooling ofrisksmayresultinadverseselectionandincreasingpremiumsthatwillcausethehealthyto dropcoverage. However,HIPAAdoesnotplaceanyrestrictionsonthemannerinwhichinsurers cansetpremiums,andsoinsurersretainsubstantialabilitytosegmentthemarket(Blumbergand Nichols,1996). Thus,theHIPAAprovisionsingeneralareunlikelytoresultinsubstantial premiumincreasesforthoseindividualscurrentlypurchasinginsurance. TheHealthInsurance AssociationofAmericaestimatedthatguaranteedissueprovisionshaveonlyasmallimpacton premiums-2to4percent(Thompson,1992). Theempiricalevidencesupportsthis. Jensen, Morrisey,andMorlock(1995)foundnoevidencethatguaranteedissue,pre-existingcondition 9 limits,orlawslimitingexclusionsonthebasisofconditionoroccupationresultedinpremium increases. Moreover,ananalysisofclaimsexperiencefromalargeinsurerspecializinginthe small-groupmarketfoundnodifferenceinaverageclaimsforgroupsthatwereguaranteedissue andthosethatweremedicallyscreened(Glazneretal.,1995). However,allstatesthathaveenactedguaranteedissueprovisionshavealsoadoptedsomeform ofratingrestrictionsforsmall-groups,andobserversbelievethatotherstatesarelikelytoadopt ratingreformsforsmall-groupsastheymodifytheirlawstomeettheguaranteedissueprovision ofHIPAA(LitowandMcClelland,1997). Ratingreformsareintendedtoenlargetheriskpool onwhichpremiumsforsmallemployersarebased,thusmakinginsurancemoreaffordablefor highriskgroupsandencouragingthemtopurchaseinsurance. Ontheotherhand,small employersthathaveagoodriskprofile(forexample,predominantlyyoungworkers)maypay higherratesafterreformsareenactedbecauseinsurerswillbelimitedintheextenttowhichthey cantakeintoaccountthisfavorableprofile. Theeffectofratingreformsonpremiumswill thereforevaryfromcompanytocompanydependingonthecharacteristicsofthegroup. Onaverage,premiumsmayriseaswell,ashigherriskgroupsareattractedintothemarketand thelowerriskgroupsdiscouragedfrompurchase. Thus,ratingreformsincombinationwith othermarketreformsmayincreasethepremiumsforsmallbusinessesthatpurchaseinsurance, andleadsomeofthemtodropcoverage. Thereareotherforces,however,whichmightlead premiumstofallwithachangeinmarketregulationsandraterestrictions. Reformsmay encouragegreaterpricecompetitionbecauseinsurersarelimitedintheextenttowhichtheycan competeonthebasisofriskselection(BuchmuellerandJensen,1997). Reformsmayalso improveinformationandleadtogreatershopping. Theoveralleffectonpremiumsandaccessislikelytodependonhowstringenttheratingreforms are. Underfullcommunityrating,theinsurercanvarypremiumsonlybyfamilytype,geographic location,andbenefitsdesign. OnlyNewYorkcurrentlyrequiresfullcommunityrating. Modifiedcommunityratingreforms(orcommunityratingbyclass),whichallowinsurerstovary premiumsbyalimitednumberofcharacteristicsoftheenrollees(suchasage),aremorecommon butdonotpermithealthstatusorpriorclaimsexperiencetobefactoredintothepricing. Some stateslimitthevariationinpremiumsattributabletothesefactors. Otherstateshaveenacted ratingreformsthatpermituseofhealthstatusinsettingpremiums,butlimitthedifferentialsin premiumsallowedduetohealthstatusdifferences. TheAmericanAcademyofActuaries(1993)conductedsimulationanalysesthatsuggestedthat averagepremiumsforsmallbusinesseswouldrisebyonlyabout2percentwithmodified communityratingandbyabout5percentunderfiallcommunityrating. Thereislittleempirical evidenceontheeffectsofratingreformsfromstatesthathaveenactedthem. Whatlittle evidenceexistsisconsistentwiththeanalysisoftheAcademysuggestingthatratingreforms resultinonlymodestincreasesinaveragepremiums. Forexample,premiumsinNewYorkin thesmall-groupmarketroseabout5percentduringthefirstyearthatcommunityratingwasin effect(CholletandPaul,1994). Minnesota,whichadoptedrestrictionsonpremiumrate variations,alsoexperiencedpremiumrateincreasesoflessthan5percentintheyearafterit 10