Financing drug discovery for orphan diseases

Professor Roger Stein

DRUDIS-1300;NoofPages6

DrugDiscoveryToday??Volume00,Number00??December2013PERSPECTIVE

feature

Financingdrugdiscoveryfororphan

diseases§

DavidE.Fagnan1,2,AustinA.Gromatzky1,3,RogerM.Stein1,4,

Jose-MariaFernandez1andAndrewW.Lo1,2,5,*,alo@mit.edu

Recentlyproposed‘megafund’?nancingmethodsforfundingtranslationalmedicineanddrug

developmentrequirebillionsofdollarsincapitalpermegafundtode-riskthedrugdiscoveryprocessenoughtoissuelong-termbonds.Here,wedemonstratethatthesame?nancingmethodscanbeappliedtoorphandrugdevelopmentbut,becauseoftheuniquenatureoforphandiseasesandtherapeutics(lowerdevelopmentcosts,fasterFDAapprovaltimes,lowerfailureratesandlowercorrelationoffailuresamongdiseasetargets)theamountofcapitalneededtode-risksuchportfoliosismuchlowerinthis?eld.NumericalsimulationssuggestthatanorphandiseasemegafundofonlyUS$575millioncanyielddouble-digitexpectedratesofreturnwithonly10–20projectsintheportfolio.

Introductionamoreattractiverisk-adjustedreturnandacancer-drug-developmentprograms.TheirThedrugdevelopmentprocesshasbecomehigherlikelihoodofsuccess.This,inturn,allowssimulationresultssuggestthatRBOstructuresexpensive,lengthyandrisky.Inresponsetothemegafundtoraisetherequiredamountofcan,inprinciple,generatereasonablereturnsforthesecharacteristics,andtothelacklusterper-capitaltoachievesuf?cientdiversi?cationbydebtandequityinvestors,whileatthesametimeformanceofinvestmentsinthebiotechandissuing‘research-backedobligations’(RBOs):providingabridgefortranslationalresearchinpharmasectorsoverthepastdecade,traditionalbondsthatarecollateralizedbytheportfolioofthedrugapprovalprocess.Inafollow-onstudy,sourcesof?nancingforsuchendeavors–pri-potentialdrugsandtheirassociatedintellectualFagnanetal.[2]proposedananalyticframeworkvateandpublicequity–arewaningascapitalproperty.BecauseRBOsarestructuredasbonds,forevaluatingtheimpactofthird-partyguar-shiftstolessriskyinvestments.Fernandezetal.theycanbedesignedtoappealto?xed-incomeanteesonRBOtransactions,andfoundthatsuch

[1]arguedthatthisproblemcanbeaddressedinvestors,whocollectivelyrepresentamuchguaranteescanimprovetheeconomicsofRBObyincreasingthescaleofinvestmentandlargerpoolofcapitalthandoventurecapitalists,transactionsatverylowexpectedcosttothepoolingalargenumberofdrugdevelopmentandwhohavetraditionallynotbeenabletoguarantor.However,theexamplesinFernandezeffortsintoasingle?nancialentityor‘mega-participateininvestmentsinearly-stagedrugetal.[1]andFagnanetal.[2]relyonverylargefund’.Thebene?tsofdiversi?cation—lowerdevelopment.Forexample,thetotalsizeoftheportfoliosofhundredsofcandidatecompounds,aggregateriskwithmore‘shotson’goal—yieldUSventurecapitalindustryin2012wasUS$199whichraisesseveraloperationalchallenges

billion,whereasthecomparable?gurefortheinthepracticalimplementationofsucha

§Thisisanopen-accessarticledistributedunderUSbondmarketwasUS$38trillion.megafund.

thetermsoftheCreativeCommonsAttribution-ToillustratethemechanicsofmegafundInthisFeature,weexploretheapplicabilityofNonCommercial-NoDerivativeWorksLicense,

which?nancingusingRBOsecurities,Fernandezetal.theRBOapproachbyextendingtheframeworkpermitsnon-commercialuse,distribution,

and[1]providedananalyticframework,simulationtoaccommodatedrugdiscoveryfororphanreproductioninanymedium,providedthe

originalsoftwareandempiricalexamplesinvolvingdiseases.Becauseoftheuniquepathologicalauthorandsourcearecredited.

1359-6446/06/$-seefrontmatterß2013TheAuthorsPublishedbyElsevierB.V.Allrightsreserved.http://wendang.chazidian.com/10.1016/http://wendang.chazidian.com1

http://wendang.chazidian.com/10.1016/j.drudis.2013.11.009EVITCEPSREP??serutaeF

Professor Roger Stein

DRUDIS-1300;NoofPages6

PERSPECTIVE

DrugDiscoveryToday??Volume00,Number00??December2013

characteristicsofmanyorphandiseases,aswellastheconsiderablesupportprovidedbytheOrphanDrugActof1983(ODA),orphandrugdevelopmentprojectsfrequentlyhavehighersuccessratesandshortertimestoapprovalbutstillgeneratepotentiallifetimerevenuesthatarecomparabletonon-orphandrugsdespitetheirmuchsmallertargetpatientpopulation.TocapturearealisticrepresentationoftheRBOs,weusenumericalsimulationtechniquestocomputetheinvestmentreturnsofahypotheticalportfoliooforphandrugdevel-opmentprojects.Givenempiricallyplausibleassumptionsforrevenues,costsandprobabil-itiesofsuccessfororphandiseases,we?ndthatmuchsmallerportfoliosthanthoseofFernan-dezetal.[1]–containingonly10–20com-poundsandrequiringlessthanUS$575millionincapital–aresuf?cientlydiversi?edtoyieldreasonableinvestmentreturnsforRBOinves-tors.AlthoughtheinvestmentreturnsofRBOsarepositivelyrelatedtoportfoliosizeowingtotheimpactof?nancialleverage,forcertaintypesofprojectstherequiredthresholdofassetscanbemodestanditmightbeworth-whiletotargettheseprojectsforaninitialproof-of-conceptofthemegafund?nancingstructure.

OrphandiseasesandtheODA

Inthe30yearssincetheODAwaspassed,theorphandiseaselandscapehaschangeddrasti-cally.Orphandiseases,formallyde?nedasthosethataffectfewerthan200,000individualsintheUSA[3],wereonceanathematothepharma-ceuticalindustry.Today,thisonce-ignored

categoryofdiseasescommandsamarketworthnearlyUS$90billionannually[4]andisbelievedtoservemorethantwicethenumberofallUScancerpatients–atleast25millionAmericansareaf?ictedwithoneofalmost7000recognizedrarediseases[5].Clearlyasacollective,rarediseasesarenotrareatall.

Before1983andtheODA,orphandiseasesposedtoomanychallengesforindustrytoconfrontseriously.Approximately80%ofrarediseasesarecausedbyunderlyinggenetic

defects,whichcanbehardtoidentify[6].Othersarecausedbyexposurestorareandunusualtoxins.Someorphandiseasesaresouncommonthataf?ictedindividualsmightnotbecorrectlydiagnosedformanyyears,andthereare

instancesofaf?ictedindividualsneverreceivingacorrectdiagnosis[7].Additionally,therigorousstandardsoftheFDAforclinicaltrialsoftenmeantthatassemblingpatientpopulationsofsuf?cientsizefortestingwasexceedinglydif?-cult.TheODAhasbeenbroadlyacclaimedforits

effectivenessindiminishingthesebarrierstodevelopment.

TheODAanditssubsequentrevisionspro-videdseveralimportanteconomicincentivestosponsorsoforphandrugs.Tojumpstartthera-peuticdevelopmentintherarediseasecategory,theODAcreatedresearchgrantsspeci?callyfororphandrugresearch,implementedtaxcreditsofupto50%forclinicaltestingcosts,authorizedexpeditedregulatoryreviewfororphandrugsand,mostimportantly,establisheda7-yearperiodofmarketingexclusivitythatprecludesFDAapprovalofthesameorgenericdrugsforthesameorphanindication[8].Theexclusivityprovisionisdistinctfromapatentand,inmanycases,providesadditionalprotectionfromcompetitionbygenericsandotherpotentialmarketentrants.

ThecombinationoftheODAincentivepro-gramandseveralsigni?cantscienti?cbreak-throughsinmolecularbiologyandgenomesequencinghasresultedinthreedecadesofinnovativeandfruitfulorphandrugdiscovery.BeforetheODA,theFDAhadapprovedfewerthantendrugsfororphandiseases;today,that?gurestandsatmorethan350unique

drugs(http://www.accessdata.fda.gov/scripts/opdlisting/oopd/).Currently,orphandrugsareattheforefrontofglobalpharmaceuticalR&Dtrends.Althoughthecompoundannualgrowthrate(CAGR)between2001and2010fornewmolecularentitiesasawholewasnegative,theCAGRfororphandesignationsoverthesameperiodwasrobustatapproximately10%[9].Theoveralldrugmarketalsore?ectsthistrend.Orphandrugscurrentlycomprise22%oftotaldrugsaleswithaCAGRof25.8%during2001–2010,comparedwith20.1%forthenon-orphanmarket[8].Someindustrydevelopmentssug-gestthatthesestrong?gurescouldcontinuetoriseastheevolutiontowardmore-targetedtherapiesandstrati?edmedicineprogresses.

ThesuitabilityoforphandrugsforRBO?nancing

Orphandrugsareparticularlywellsuitedtoportfolio?nancing.Aprimaryreasonisthesig-ni?cantlyhigherratesofsuccessthatorphandrugdevelopmentprojectsenjoywhencom-paredwiththoseofotherdiseasegroupssuchasoncologyorneurodegenerativedisorders.

Orphandiseasesarelargelycausedbyamuta-tioninanindividual’sgeneticcode,mostcom-monlymanifestedasamalfunctionorabsenceofoneormorekeyproteins.Iftheunderlyinggeneticdefectcanbeidenti?edandcharacter-ized,itisoftenpossibletocreatehighly

targetedandeffectivetherapiestoaddressthe

malfunctionanditssymptoms[9].Similartar-getingmethodologieshavebeenusedtocom-batrarecancers,notablyfordrugssuchasRituxan1andGleevec1.Consequently,theoddsofaneworphandrugreceivingFDAapprovalaresigni?cantlyhigherthanthoseofanon-orphancounterpart.Fororphandrugsthatenteredclinicaltestingbetween1993and2004weestimatetheoverallregulatorysuccessratetobeapproximately22%,whereasthecomparable?gurefornon-orphandrugswasapproximately11%[10]andtherateforanticancercompoundswasevenloweratonly6–7%[10].

Thesuccessorfailureoforphandrugdevel-opmentprojectsisalsolesslikelytobecorre-latedacrossdiseasesasaresultofthelargeproportionoforphandiseasesthatdisplaymonogenicpathologyoractthroughlargelyunrelatedmechanisms[11].Thisobservationisparticularlysigni?cantgiventhecentralrolethatcorrelationhasindeterminingtheriskofaportfolioofcandidatedrugcompounds.

Althoughwearenotawareofanylongitudinalestimatesofhistoricalcorrelationsamongdrugdevelopmentprojects,thescienti?cbasisoforphandrugssuggeststhatcorrelationsarelikelytobesmall,especiallywhencontrastedwithotherdiseasegroupssuchasoncology.Manytypesofcancershavesimilarpathologies,suchasthederegulationofspeci?csignalingpathwaysandmutationsincrucialoncogenes.Asanexample,considertheJanuskinase/signaltransducersandactivatorsoftranscription(JAK/STAT)andtransforminggrowthfactor(TGF)-bpathways,eachofwhichhasbeenlinkedtodozensofoncologicdiseases[12,13].Ofcourse,thereareseveralorphandrugsinoncologybut,incontrasttolargerclassesofoncologydrugsthatshareacommonmechanismsuchastyr-osinekinaseinhibitionorantiangiogenesis,orphandrugs(asadistinctcategory)actagainstawidervarietyoftargets.

Furthermore,orphandrugshavebeenshowntohavealmostequivalentlifetimerevenuepotentialtonon-orphantherapies.AccordingtoThomsonReuters,anaverageorphandrugcanbeexpectedtoattainsalesofUS$100–500mil-lionperyear[14].Smallpatientpopulationsizesareoftencompensatedforbyhighper-patientrevenues.Forexample,Soliris1,adrugtotreatparoxysymalnocturnalhemoglobinuria(arareblooddiseaseaffectingfewerthan6000indivi-dualsintheUSA),ispricedatmorethanUS$400,000perpatientperyear[9].Conse-quently,theblockbusterdrugmodelthatischaracteristicofmanytop-sellingnon-orphandrugsisequallyapplicabletotheorphanmarket:compoundsinthetop29%oforphandrugsare

Features??PERSPECTIVE

2

http://wendang.chazidian.com

http://wendang.chazidian.com/10.1016/j.drudis.2013.11.009

Professor Roger Stein

DRUDIS-1300;NoofPages6

DrugDiscoveryToday??Volume00,Number00??December2013PERSPECTIVEeachexpectedtoearnmorethanUS$1billionin(NDAs)orfocusonparticularsegmentsofthecompoundvaluations.Fagnanetal.(http://revenueperyearovertheirlifetime[14].Asandrugdevelopmentprocess.However,http://wendang.chazidian.com/abstract=2203203)andthesupple-extremeexampleofthepotentialpro?tabilityofexistingdrugroyaltysecuritizationtransactionsmentarymaterialin[1]providedetailsoftheorphandrugs,weconsideragainRituxan1–an(e.g.,RoyaltyPharma,DRI),ourproposalmodelsandestimation.Thepseudo-codefororphandrugthatisexpectedtoattaindis-accommodatesinvestmentinearly-stagepro-thesesimulationsisgivenin[1],andthesourcecountedlifetimesalesofoverUS$150billion,ajectsthatcouldbefarfromFDAapproval;hencecodeisavailablefromtheauthors.

?guresurpassedonlybyP?zer’snon-orphantheyhavenodiscernibleroyaltystreamattheFollowingFernandezetal.[1]andFagnanetal.Lipitor1[14].timeofinvestment.Asaresult,theriskismuch(http://wendang.chazidian.com/abstract=2203203),weconsiderFinally,theODA’smarketingexclusivityclausehigher,creatingtheneedformoresophisticatedanRBOtransactionwithacapitalstructureprovidesakey?nancialincentivefororphan?nancialmodelingoftheeconomicvalueoftheconsistingofaseniortranche,http://wendang.chazidian.complexitiesofthewaterfallandthedrugextendedtheaveragecombinedpatentandFernandezetal.[1]presentedastylizedapprovalprocess,numericalsimulationsareexclusivityperiodbynearlyayear,resultinginanmathematicalexampleofsecuritizationofusedtoevaluatetheRBOsecurities.Wefocusaveragecompetition-freemarketingperiodofexperimentaldrugcompoundsandFagnanetal.onlyonearly-stageinvestment(preclinicaland11.7years[15].Fortherapiesthatreceive(http://wendang.chazidian.com/abstract=2203203)developedPhaseI),whichrepresentstheriskiestportionofapprovallaterintheirpatentlifespans,thethisexampleinmoredetail.Althoughillustrative,thedrug-developmentprocessandwhereincreaseintheexclusivityperiodcanbesignif-thisexampleoversimpli?estheeconomicsofthefundingisscarcest.Wesimulateacquiringanicantlylonger.biopharmaindustry.TheauthorsalsoprovideequalnumberofpreclinicalandPhaseIcom-AssumingaverageannualsalesofUS$200resultsofamoredetailedsetofsimulationpounds,withthegoalofsellingalldrugsthatmillionanda10%costofcapital[16],weesti-experimentsthatincorporatemore-realisticsuccessfullycompletePhaseIItrials.Oursimu-matethattheaveragepresentvalueofanfeaturesofthedrug-developmentprocesslationreliesonseveralkeyassumptionsandorphandrug’srevenueoveritscompetition-freeincludingcorrelatedassets,stochastictransitionsparametersincludingclinicaltrialcosts,valua-lifespanisUS$1.36billion.Inaddition,webetweenclinicaltrialphases,theneedtoman-tionsanddurationofeachphase(SeeTable1).assumethatthepro?tmargin,includingcostofagecashtopayinterestandprincipal,realisticWederiveourpreclinicalestimatesfrom[17],goodssold(COGS)andmarketingcosts,is60%,valuationsofcompoundsthataresoldduringmakingtheassumptionthatthepreclinicalresultingina?nalaveragevaluationofUS$818intermediatestagesoftheclinicaltrialsprocessphaseissimilarfororphanandnon-orphanmillion.Todemonstratethesensitivityofourandtheneedtomanagecashtofundnewtrialsdrugs.KaitinandDiMasi[18]reportedthatsimulationexperimentstovaluationassump-duringtheapprovalprocess.Fagnanetal.orphandrugtrialsinrecentyearstakeapproxi-tions,wealsoemployalessconservativeesti-(http://wendang.chazidian.com/abstract=2203203)extendedmately5.9yearsfromPhaseItoNDAwithanmationofannualrevenuesofUS$400million[9],thisworkbyanalyzingtheimpactofthird-partyadditional0.8yearsrequiredfortheapprovalwhichresultsina?nalaveragevaluationofdefaultguaranteesforasubsetoftheRBOprocess.ThetimeforeachphaseiscalculatedbyUS$1.63billion.Inthenextsectionweusethesesecurities.Theyfoundthatsuchguaranteescanscalingtherelationshipsusedin[1].

valuesalongwithsomeadditionalparameterstogreatlyincreasetheattractivenessofRBOs,Clinicaltransitionprobabilitieswereestimatedconductaseriesofsimulationexperimentsenhancingtheirfundraisingpotential.from[19]basedonalarge-moleculedataset,demonstratinghowanorphandrugportfolioThesesimulationexperimentsextendthewhichwehaveassumedtobeacloseapprox-mightbeusedascollateralforanRBOstructure.frameworkofthestylizedexampletoaricherimationfororphandrugsowingtotheincreased

multistatemultiperiodsettingthatincludestargetingspeci?citythatcharacterizesbiologic

Orphandiseasemegafundsimulationpath-dependenceandcorrelatedassetvalua-drugdevelopment.Furthermore,webelievetheThe?nancialengineeringtechniqueofsecuriti-tions.Theneedtoextendthesingle-periodresultingsuccessratefrompreclinicalto

zationinvolvescreatingalegalentitythatissuesmodelarisesasaresultofthenatureofthedrugapproval(21.8%)tobereasonablebasedonourdebtandequitytoinvestorsandusesthepro-trialprocess.Ateachstageofthisprocess,largeranalysisofrecentorphandrugsdevelopedbyceedstopurchaseaportfolioofassets.Thedebtandlargercashin?owsarerequiredtofundpharmaceuticalcompanies.Valuationsforeachandequitysecuritiesaresaidtobe‘backed’byadditionaltesting.Importantly,newinvestmentphasewereobtainedbydiscountingtheesti-theassetsintheportfoliointhesensethattheateachstagecanonlyoccurwhenthereismatedescribedabovebasedontheprobabilityholdersofsuchsecuritieshavecertainowner-suf?cientcapitalavailablethatisnotrequiredforofsuccessandusingdiscountratesofupto30%shiprightstothoseassets.Inparticular,thecashotherusessuchasdebtserviceorrepayment.peryear.Uponacquisitionofcompounds,we?owsfromtheassetsareusedtorepaythedebt,Thedominantsourceofcash?owforthefollow[1]andassumethatupfrontandmile-andallresidualvalueispaidtotheequitymegafundcomesfromthesaleofcompoundsstonepaymentsareproportionaltoclinicalcosts.holders.Aprimarymotivationforsecuritizationisfromtheportfolio.Pro?tsorlossesaccruewhenInaddition,http://wendang.chazidian.comstly,weestimateclinicalthroughdiversi?cationandtoallow?xed-phaseandsubsequentlysellsitatanotherphase.trialcostsusingconservativevaluesfortheincomeinvestorstoinvestinassetclassesthatAnalysisoftheportfolioprimarilyinvolvesthenumberofpatientsperclinicaltrial[20]andcostwouldotherwisebetooriskyorfragmentedtospeci?cationoffourquantities:thetransitionperpatient[21].Weassumeahighercostperbeofinteresttothem.InabiomedicalRBO,theprobabilities;thedistributionoftrialcostsinpatientinPhaseItoaccountforexpenses

assetsareadiversecollectionofbiomedicaleachphase;thedistributionofvaluationsforassociatedwithlocatingsuitablecandidatesforprojectsthatcanspantheentirerangefromeachcompoundthatissoldinaspeci?cphase;thetrial,whichisinherentlymoredif?cultforanpreclinicalresearchtonewdrugapplicationsandsomeformofdependenceamongsttheorphandrug.

http://wendang.chazidian.com3

http://wendang.chazidian.com/10.1016/j.drudis.2013.11.009EVITCEPSREP??serutaeF

Professor Roger Stein

DRUDIS-1300;NoofPages6

PERSPECTIVE

DrugDiscoveryToday??Volume00,Number00??December2013

TABLE1

RBOsimulationparameters

Phase

Clinicaltrialcost(US$million)

Clinicaltrialsuccessrate

Clinicaltrialduration(years)

Valuation(US$million)

PreclinicalPhaseI

558

69%84%53%

1.001.662.092.150.80–

7.127.675.6321.5701.9817.6

PhaseII

PhaseIIINDAAPP

43––

74%96%–

Abbreviations:APP,approval;NDA,newdrugapplication;RBO,research-backedobligation;ROE,returnonequity.

Althoughwebelieveourparameterassump-tionstobereasonable,topermitother

researcherstoexperimentwithdifferentvalueswehavemadeallofourassumptionsandthesourcecodeforoursimulationspubliclyavail-ablewithanopen-sourcelicense(forfurtherdiscussionofallparameters,see:SupplementalInformationandParametersprovidedonline).

Simulationresults

Table2comparesthesimulationperformance(usingtwomillionsimulations)ofan(approxi-mately)optimizedRBOstructuretoatraditionalequitymodel,assuminga?xedcorrelationof20%.Forcomparison,weincluderesultsfortheequity-onlystructureusingthesameequityamountusedintheRBO,aswellasasecondsetofresultsforanequitystructure,butusingthesametotalcapitaloftheRBO.

Thesimulationacquirestenor16orphandrugs(dependingonthecapital),withanequal

numberofcompoundsinpreclinicalandPhaseIusingatotalcapitalofUS$373.75orUS$575million,respectively,substantiallylessthaninthecaseofcandidatecompoundsinoncologyasdiscussedin[1].Thesimulationextendsforahorizonof6.5years(in6-monthincrements),withanadditionalyearusedforthesellingofcompoundsuponliquidationoftheremainingdrugsintheportfolio.Thesimulationistargetedtosellcompoundsoncethey(successfully)completePhaseIItrials,butcompoundscanbesoldearlierintheprocessinanticipationofbondcouponorservicepayments.Readersareencouragedtodownloadoursimulationsoft-wareandre-runsimulationswiththeirownparametersandassumptions,e.g.,purchasingdrugsearlierindiscoveryorusinganalternativetarget-sellingphase(http://wendang.chazidian.com/documents/RBOToolbox_112713.zip).(Forsensitivityanalysisofsomeoftheparameters,seeSupplementalInformationandParameters.)

Usinganequity-onlystructurewithcapitalofUS$373.75million,Table2reportsthatthemeanreturnonequityinourexperimentswas10.7%,nearly3%belowtheRBO,whereastheprob-abilityoflossofequitywasalsohigher,resultingin16.1%comparedwith13.1%fortheRBO.Byincreasingthecapitalusedintheequity-onlymodeltoUS$575million,theprobabilityoflosstoequityisreducedto10.1%,withasmallerimprovementtoexpectedreturnonequity,http://wendang.chazidian.comingthehigherlevelofcapital,theRBOandsame-capitalequity-onlystructuressellaround?vePhaseIIIcompoundsonaverage,comparedwithaboutthreeforthesmallerlevelofcapital.Thus,theRBOachievesahigherreturnonequity,amodestincreaseintheprobabilityofloss,butmorethantwicetheprobabilityofreceivingareturnonequitylargerthan25%.Thishigherequityreturniscausedbytheincreaseduseofleverage,whichcomeswithmorerisktoequityholders,ascanbe

Features??PERSPECTIVE

TABLE2

RBOsimulationresultsusingatargetsellingphaseofPhaseIII

Allequity(sameequity)

RBO88

Allequity(samecapital)

NumberofcompoundsPreclinicalPhaseI

55

881.75.0575––57511.8%<0.1bp10.1%59.77%6.27%

Researchimpact

NumbersoldinPhaseIINumbersoldinPhaseIII

1.13.1

2.24.7

Liabilities

Capital(US$millions)

Seniortranche(US$millions)Juniortranche(US$millions)Equitytranche(US$millions)

373.75––

373.75

57586.25115373.75

EquitytrancheperformanceAverageannualizedROEProb.(equitywipedout)Prob.(returnonequity<0)Prob.(returnonequity>10%)Prob.(returnonequity>25%)

10.7%0.2bp16.1%54.7%7.8%

13.4%60bp13.1%66.7%18.4%

DebttranchesperformanceSeniortranche:

defaultprob.,expectedloss(bp)Juniortranche:

defaultprob.,expectedloss(bp)

––

0.8,<0.1

––

56,15

Abbreviations:bp,basispoints;prob,probability;RBO,research-backedobligation;ROE,returnonequity.

4

http://wendang.chazidian.com

http://wendang.chazidian.com/10.1016/j.drudis.2013.11.009

Professor Roger Stein

DRUDIS-1300;NoofPages6

DrugDiscoveryToday??Volume00,Number00??December2013PERSPECTIVETABLE3

RBOsimulationresultsusingatargetsellingphaseofPhaseIII,aPhaseIIvaluationofUS$174millionandaPhaseIIIvaluationofUS$643million,correspondingtosalesofUS$400millionperyear[9]

Allequity(sameequity)RBOAllequity(samecapital)

Numberofcompounds

Preclinical388

PhaseI388

Researchimpact

NumbersoldinPhaseII0.72.31.7

NumbersoldinPhaseIII1.84.65.0

Liabilities

Capital(US$millions)230575575

Seniortranche(US$millions)—115—

Juniortranche(US$millions)—230—

Equitytranche(US$millions)230230575

Equitytrancheperformance

AverageannualizedROE19.6%33.8%23.2%

Prob.(equitywipedout)2bp81bp<0.1bp

Prob.(returnonequity<0)10.4%2.5%14bp

Prob.(returnonequity>10%)79.1%95.4%93.7%

Prob.(returnonequity>25%)40.5%82.9%45.7%

Debttranchesperformance

Seniortranche:

defaultprob.,expectedloss(bp)—1.2,<0.1—

Juniortranche:

defaultprob.,expectedloss(bp)—80,27—

Abbreviations:bp,basispoints;prob,probability;RBO,research-backedobligation;ROE,returnonequity.

seenfromthemuchhigherprobabilityofallriskpro?lefromissuingsigni?cantdebtisofR&Dinvestment–hasdeclinedoverrecentequitybeinglostintheRBOversuseitheroftheminimal.Inparticular,RBOsincreasetheprob-years(http://wendang.chazidian.com/abstract=2203203),andequity-onlycases[60basispoints(bp)versuslessabilityoftotallossoveranall-equitymodel–theirstock-priceperformanceoverthepastthan1bp].withthesameamountofequity–byonly79bp.decade–anannualizedreturnof-1.2%fortheTheleverageisachievedintheRBOstructureTheriskofextremelossesinthisstructureNewYorkStockExchangeArcaPharmaceuticalthroughtheissuanceoftwotranchesofstruc-involvesatrade-offbetweenincreasingdiversi-Indexduringtheperiodfrom2January2002to1turedbonds.Thedefaultrateonthesenior?cationasaresultofalargercapitalpoolontheJuly2013–hasbeenequallydisappointing.trancheoftheRBOisapproximately1bp,whichonehandandincreasingleveragecausedbyDespitetheneardoublingoftheaggregateR&Discomparabletothehistoricaldefaultratesfordebtissuanceontheotherhand.Speci?cally,budgetofthepharmaceuticalindustryfrombondsratedatthehighestlevelsbyratingalthoughtheaveragereturnonequityincreasesUS$68billionin2002toUS$127billionin2010,agencies.Thedefaultrateonthemezzaninefrom19.6%intheall-equity-?nancedcasetotherehasbeenlittleappreciableimpactonthetrancheis56bpwithanexpectedlossof15bp.33.8%intheRBO-?nancedcase,theprobabilitynumberofnewdrugsapproved(http://wendang.chazidian.com/EvaluatePharma_World_beattractiveto?xed-incomeinvestors,giventhebyafactorof40.Evenso,manyinvestorsmightPreview_2016.aspx).

assumedcouponrateonthedebt(5%and8%,stillpreferthe‘leveraged’?nancingstructuretoFernandezetal.[1]introducedtheconceptofrespectively).theall-equityversionbecauseofitsboosttoRBOsasameansofchannelingfundsfromTodevelopasenseoftherangeofinvestmentreturns.However,itshouldbeemphasizedthatglobalcapitalmarketstoearly-stagedrugreturnsthatarepossible,Table3reportsthetheproperuseofRBOsdoesrelymoreheavilyondevelopment.Acomplicationregardingtheoutcomeofthesamesimulationsbutassumingtheaccuracyoftheassumedparameters,givenoriginalRBOmodelwasthatconstructingannualrevenuesofUS$400millioninsteadofthethatthedefaultprobabilitiesofthebondsandportfoliosofthesizedescribedinthemetho-originalUS$200millionforthesamecapitalreturnonequitycanvarymateriallyfordifferentdologywasuntestedandintroducedseveralstructureusedintheexperimentofTable2.Asaparameterassumptions.potentialoperationalchallenges.Inthisarticle,resultofthesehighervaluations,weseewehavetriedtoaddressthisbyreportingtheexpectedreturnsof20–34%withahigherlevelConcludingremarksresultsofaseriesofexperimentsthatsuggestofdebtsupportedbytheRBO.UnderthisAcon?uenceoffactorsisresponsibleforthethatsmallerportfolios–ontheorderofasfewasaggressiverevenueassumption,thereislittlelowernumberofnewdrugsapprovedoverthetencompoundsandUS$373.75millionof

needfordebt?nancing,giventhata19.6%ratepastdecade,causingsomeauthorstosuggestcapital–canstillbeusedascollateralforRBOofreturnisconsiderablyhigherthanthatthatthecurrentbusinessmodelforlifesciencetransactionsanddeliverreasonableinvestmentobtainedbytheaveragebiotechventurecapitalresearchanddevelopmentis?awed[22–24].Thereturns.

?rmoverthepastdecade(butcomparablewithproductivityoflargepharmaceuticalcompaniesThebiggestdriversofthisresultarethethereturnsofthemostsuccessfulones).How-–asmeasuredbythenumberofnewmolecularassumptionsofahigherprobabilityofsuccessever,atthesametime,theimpacttothefund’sentityandbiologiclicenseapplicationsperdollarandamuchlowercostofconductingclinical

http://wendang.chazidian.com5

http://wendang.chazidian.com/10.1016/j.drudis.2013.11.009EVITCEPSREP??serutaeF