S ORENSEN This paper uses detailed weekly data on sales of hardcover 64257ction books to evaluate the impact of the New York Times bestseller list on sales and product variety In order to circumvent the obvious problem of simultaneity of sales and b ID: 38387
Download Pdf The PPT/PDF document "BESTSELLER LISTS AND PRODUCT VARIETY LAN..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
BESTSELLERLISTSANDPRODUCTVARIETYT.SORENSENThispaperusesdetailedweeklydataonsalesofhardcoverctionbookstoevaluatetheimpactoftheNewYorkTimesbestsellerlistonsalesand behaviorinthepresenceofsocialeffects,withbestsellerlistsservingasaformofcoordinatingmechanism.Forexample,teenagerswhowanttolistentomusicthatishotcanlooktotheBillboardchartstondoutwhatispopular,andpeoplemayfavormoviesatthetopofthebox-ofcechartsbecausetheywanttobeconversantinpopularculture.Inthespeciccaseofbooks,bestsellerstatusalsotriggersadditionalpromotionalactivitybyretailers.Forthesamereasonsthatbestsellerlistsmaydirectlyaffectconsumerspurchasedecisions,theymayalsocausesalestobemorehighlyconcentratedonthefewbestsellingproducts.This,inturn,couldinuenceproductvariety:iftheadditionalsalesaccruingtobestsellersasadirectconsequenceofthepublicationofthelistcomeattheexpenseofnon-bestsellingproducts,theoptimalnumberofproductstooffermaydecrease(relativetowhatwouldhavebeenoptimalintheabsenceofabestsellerlist).Forexample,ifpublicizedboxofcerankingscauseticketsalestobemoreconcentratedonblockbusters,theymayalsomakeitunprotabletoincurthexedcostsofproducingalmwhosepopularityisexpectedtobeonlymarginal.Thispaperexaminestheseissuesinthecontextofthebookpublishingindustry,lookingspecicallyattheimpactoftheNewYorkTimesbestsellerlistonsalesofhardcoverctiontitles.Theempiricalanalysisaddressestwoquestions.First,doesbeinglistedasaNewYorkTimesbestsellercauseincreaseinsales?Second,doestheinuenceofthebestsellerlistalsoaffectthenumberofbooksthatarepublished,andifso,inwhichdirection?Obvioussimultaneityproblemsmakeansweringtherstquestionanontrivialempiricalexercise;however,subtletiesintheconstructionandtimingoftheNewYorkTimeslistcanbeexploitedtoidentifyitsimpact.Theresultssuggestamodestincreaseinsalesforthetypicalbestsellerwhenitrstappearsonthelist,andtheincreaseappearstoreectaninformationaleffectratherthanapromotionaleffect:theeffectsareconcentratedintherstweekabookappearsonthelist,andtheyaremostsubstantialfornewauthors.Regardingthesecondquestion,theimpactofbestsellerlistsonproductvarietyistheoreticallyambiguous,sinceitdependsonwhethermarketexpansionorbusiness-stealingeffectsdominate.Althoughthedataarelessthanidealforaddressingthisquestion,Ipresentindirectevidencesuggestingthebusiness-stealingeffectsofbestsellerlistsareunimportant:ifanything,bestsellerlistsappeartoincreasesalesforbothbestsellersnon-bestsellersinsimilargenres.Althoughthispaperisthersttoexploretheimpactofbestsellerlistsonproductvariety,similarquestionshavebeenpreviouslyaddressedinanumberofcontexts.EarlytheoreticalworkemphasizedthetradeoffbetweenSee,e.g.,(BeckerandMurphy[2000],Banerjee[1992],andVettas[1997]).ThislineofreasoningwillbeclariedinsectionII.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. quantityanddiversityinthepresenceofscaleeconomies(DixitandStiglitz[1977])andtheeffectsofmarketstructureonproductvariety(Lancaster[1975]).Empiricalstudiesofproductvarietyhavebeenundertakenfortheradiobroadcastingindustry(BerryandWaldfogel[2001],andSweeting[2005]),themusicindustry(Alexander[1997]),andforretaileyeglasssales(Watson[2003]),tonameafewexamples.Inthefollowingsection,Ioutlineabasictheoreticalframeworkforunderstandinghowbestsellerlistscaninuencethenumberofbooksthatgetpublishedinequilibrium.Section3providesabriefdescriptionofthebookindustryandthedataset.Theempiricalanalysis(Section4)proceedsintwoparts:rst,thedataareusedtoidentifyandquantifythedirectimpactofthebestsellerlistonsales;second,substitutionpatternsinthedataareanalyzedtodeterminethelikelydirectionofthelistsimpactonproductvariety.Somebroadimplicationsoftheresults(aswellasalternativeinterpretations)arediscussedintheconcludingsection.II.ASIMPLETHEORETICALFRAMEWORKConsiderahighlysimpliedmodelinwhichasinglepublisherchooseshowmanymanuscriptstopublish.Therearemanuscriptsunderconsidera-tion.Printingandmarketingamanuscriptrequiresaxedcost,,inadditiontothe(constant)marginalprintingcost.Priortopublication,themanuscriptscanberankedinorderofexpectedpopularity,withbeingtheindexofthe-bestbookamongthealternatives.Themarketpriceofapublishedbook,,istakenasgiven,andthepost-publicationpricedoesnotadjusttoreectabooksrelativepopularity.Theexpecteddemandforthebestbookisgivenby),where)canbeinterpretedasthelevelofaggregatedemandforbooks(whichmaydependonhowmanyareoffered),and)isafunctiondetermininghowaggregatedemandisallocatedamongbooksdependingontheirrelativepopularity(e.g.,wecouldhave1).Onlybookswithpositiveexpectedprotswillbepublished,sothenumberofbookswillbethemaximumsuchthat,asillustratedinFigure1.Nowconsiderthepotentialimpactofpublishingabestsellerlist,sothatconsumersobservethesalesranksofthetopbooks.SupposethattheThemodelcouldalsoapplytomoviestudiosdecisionsabouthowmanylmstoproduce,ortorecordlabelsdecisionsabouthowmanyartiststosign.Thisassumptionwouldbeabsurdinmostcontexts,butinthiscaseitisatleastdescriptiveofacuriouspracticeinmultimediamarkets.Pricesofbooks,movies,andCDsalmostneverreectthepopularityoftheindividualproducts.(Clerides[2002])providesdirectevidenceandathoroughdiscussionofpricingissuesinthebookindustry.)Typically,pricepointsforbooksandCDsaredeterminedbeforetheyaremarketed,andsubsequentadjustmentsareextremelyinfrequent.Movieticketpricesareevenmorerigid:inthesummerof2002,forexample,itcostthesameamounttoseeChicago(whichwontheOscarforbestpictureandwasabox-ofcesuccess)asBoatTrip(whichoppedattheboxofceandwasuniversallyridiculedbycritics).BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial aggregateconsumerresponsetoabestsellerlistleadstoanincreasedconcentrationofsalesonbestsellers:letting)denotetheexpectedmarketshareofthe-rankedbookamongalternativeswhenabestsellerlistoflengthispublished,and)denotethemarketsharewhennolistispublished,weassumethat)if.Ifbestsellerlistshaveanydirectimpactonconsumerbehavior,theeffectwouldalmostcertainlyhavethisfeature.Themostobviousmechanismforthiseffectisinformational:ifconsumersareuncertainaboutbooksqualities,andtheybelievethatatleastsomepastpurchasershadmeaningfulinformationaboutthebookstheypurchased,thenbestsellerstatuswouldbeasignalofquality.Alternatively,socialeffectsmayleadtohigherdemandforbestsellers:consumersmaywanttoreadwhateveryoneelseisreadingintheinterestofkeepingupwithwhatispopulare.g.,theydontwanttobeleftoutoftheconversationwhentheygotothecocktailparty.Inthemarketforhardcoverctionbooks,anadditionalmechanismpushessalestowardbestsellers:retailersroutinelydiscountbestsellersandpositionthemprominentlyintheirstores.Iftheoveralllevelofdemand)isindependentofanybestseller-listeffects,thenthelistunambiguouslyreducesthenumberofbooksthatcanbe Figure1Thepublish/no-publishmarginALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. protablypublished.ThisisillustratedinFigure1:theincreasedsalesforthebookscomeattheexpenseofthenon-listedbooks,shiftingthepublish/no-publishmargintotheleft(from).Morerealistically,however,thepublicationofabestsellerlistcouldincreaseoveralldemandinadditiontochangingtheallocationofdemandacrosstitlesi.e.,theadditionalpromotionandinformationaboutbestsellerscouldattractconsumersthatotherwisewouldnothavepurchasedanybookatall.Inthiscase,theimpactofbestsellerlistsonthepublish/no-publishmarginisambiguous.(Figure2illustratesacaseinwhichmorebookswouldgetpublishedinthepresenceofalist,eventhoughthelistleadstoarelativelyhigherconcentrationofsalesamongbestsellers.)Thisframework,whileobviouslyoversimplied,illustratestheprincipalideasunderlyingtheempiricalanalysestofollow.Ideally,wewanttoexaminesalesdatatoseeifindeed)forthatis,toseeifbestsellerlistsanincreaseindemandforbestsellersrelativetonon-bestsellers.Inordertosayanythingaboutwhetherbestsellerlistsaffectthenumberofbooksthatgetpublished,wemustthenaskamuchmoresubtlequestionofthedata:howaresalesofrelativelyunpopular(non-bestselling)booksaffectedbythepublicationofbestsellerlists?Thatis,howLK* Figure2ListmayincreaseoveralllevelofdemandBESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial does)compareto)forveryclosetoUnfortunately(butnotsurprisingly),theavailabledataareinadequateforansweringthisquestiondirectly.Instead,wewilllookforindirectevidenceofsubstitutionbetweenbestsellingandnon-bestsellingtitles,whichwouldsuggestthepotentialfor(andthelikelydirectionof)productvarietyeffects.III.BACKGROUNDANDDATAIII(i).TheBookIndustryIntheU.S.,thevastmajorityofbooksareproducedbyasmallnumberoflargepublishinghouseslikeRandomHouseandHarperCollins.Theoddsagainstamanuscriptsbeingacceptedbyoneofthesepublishinghousesarelong,especiallyinthecaseofction.Thirtypercentorfewerofavailablemanuscriptsinanygivenyearareinprint,andalthoughninetypercentofpublishedbooksarenonction,seventypercentofthemanuscriptssubmittedtotraditionalpublishersarection(Suzanne[1996]).successfulmanuscriptsarebrokeredtopublishersbyliteraryagents.Theseagentsaretypicallyreluctanttotakeonrst-timeauthors,andtheirfeestendtobesteep(around15percentofauthorsroyalties).However,usinganagentgreatlyincreasestheauthorschancesofsuccess:unsolicitedmanu-scripts(manuscriptsreceivedoverthetransom)areestimatedtohavefteenthousandtooneoddsagainstacceptance(Greco[1997]).Manu-scriptsaresometimessoldtopublishersbyauction,butthismethodistheexceptionratherthantheruleandisusedprimarilybyestablished,brand-nameauthors.Thedecisiontoextendacontracttoanauthorismadeonlyafterreviewofthemanuscriptsqualityandsalabilitybyseveralstagesofeditors.Ifamanuscriptsurvivesthereviewprocess,thepublisherofferstheauthoracontractgrantingroyaltypaymentsinexchangeforexclusivemarketingrightsaslongasthepublisherkeepsthebookinprint.Royaltiesaveragesevenpercentofthewholesalepriceonhardcoverbooksbynewauthors,NotethattheillustrationinFigure1makesitappearthatsalesofthe-rankedbookand-rankedbookareequallyaffected,whichneednotbethecase.Forinstance,itisquiteplausiblethattheimpactisadecliningfunctionofabooksrank.Moreover,onlytheeffectonbooksatthemarginisrelevantfordeterminingthenumberofbooksthatgetpublished.Becausethismodeldoesnotexplicitlyspecifyconsumerspreferences,itsaysnothingaboutthewelfareeffectsofareductioninproductvariety.Thispaperwillbedeliberatelyagnosticonthispoint,sincethewelfareeffectscouldplausiblygoeitherway:ontheonehand,fewerbookscouldmeanforegonesurplusfromtitlesthatwouldhaveappealedtoreaderswithdiversetastes;ontheotherhand,fewerbookscouldmeanthatlesstimeiswastedreadingbadFortherstquarterof2003,thetopsixpublishingconglomeratesaccountedforover80percentofunitsalesinadultction.Thesenumbersimplythat10percentofctionmanuscriptsgetpublished.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. andmayincreaseoncethebookachievesacertainlevelofsales(Suzanne[1996]).Uponthedeliveryofacompletedmanuscriptalargeportionofexpectedroyaltiesaregiventotheauthorintheformofanadvance;manyauthorsneverreceiveadditionalpaymentsbecauseadvancesoftenexceedroyaltiesearnedfromactualsales.Publishersretainalargeshareofroyaltiesinescrowaccountstocompensateforreturnedbooks;booksellersmayreturnunsoldbookstothepublishersforfullprice,andthereforereturnratesexceedingftypercentarecommon(Greco[1997]).Inspiteofthedifcultythatauthorsseemtofaceingettingtheirmanuscriptspublished,thebookindustrygeneratesanastonishingowofnewbookseachyear.Acrossallcategories(ctionandnonction)andallformats(hardcover,tradepaper,andmass-marketpaper),over100,000titleswerepublishedintheyear2000alone.Inadultction,thenumberofnewbookspublished(calledtitleoutputwithintheindustry)hasincreaseddramaticallyoverthepastdecade.Theindustrystradepublication,Annual,reportsthattitleoutputforhardcoverctionmorethandoubledfrom1,962in1990to4,250in2000.Incontrast,therateofincreasewasmuchmoregradualpriorto1990.Infact,Bowkerreportsthatthenumberofctiontitlesin1890wasover1,100,sotitleoutputhadlessthandoubledinthe100yearspriorto1990(Bogart[2001]).Thedramaticincreasesintitleoutputinthe1990swereroughlyconcomitantwithanincreaseintheconcentrationofsalesamongbestsellers.Fromthemid-1980stothemid-1990s,theshareoftotalbooksalesrepresentedbythetop30sellersnearlydoubled(Epstein[2001]).In1994,over70percentoftotalctionsaleswereaccountedforbyamereveauthors:JohnGrisham,TomClancy,DanielleSteel,MichaelCrichton,andStephenKing(Greco[1997]).Publishersandauthorsemployanumberofmarketingstrategiesintheirattemptstoachievethekindofsuccessenjoyedbythesetop-sellingauthors.Publishersmarketingbudgetsmayrangebetweentenandtwenty-vepercentofnetsales(Cole[1999]),andauthorsareexpectedtoappearpubliclyinpromotiontours.Bookreviewsarehighlysoughtafterbutdifculttoobtain:tensofthousandsofbooksarepublishedeachyearintheUnitedStates,buttheNewYorkTimes(forexample)reviewsonlyoneperday.Publishersalsomaypayretailstoresforshelfspaceorinclusioninpromotionalmaterials.Bookmarketersconcentratetheireffortsoncreatingasuccessfullaunch;retailstoresmayremovelow-sellingnewreleasesfromtheirshelvesafteraslittleasoneweek(Greco[1997]).Rosen[1981]providesaclassicexplanationforthepresenceofsuchsuperstars.Acriticalpieceoftheargumentisthattheconvexityofreturnsresultsfromtheimperfectsubstitutabilityoftheproductsofferedi.e.,readingseveralunremarkablebooksmaynotbeagoodsubstituteforreadingasinglegreatone.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial Inspiteofalltheresourcesspentonmarketing,oneofthebestkindsofpublicityappearingonabestsellerlistcannotbebought.Bestsellerlistshavelongplayedanimportantroleinthebookindustry.Regularpublicationofbestsellerlistsbeganin1895,whenaliterarymagazinecalledTheBookmanstartedprintingamonthlylistofthetopsixbest-sellingbooks.NewYorkTimesBookReviewbeganpublishingitsbestsellerlistasaregularfeaturein1942.Althoughmanyotherprominentlistsnowexist,NewYorkTimeslistisgenerallyconsideredthemostinuentialintheindustry(Korda[2001]).III(ii).DataThemaindatasettobeanalyzedconsistsofweeklynationalsalesforover1,200hardcoverctiontitlesthatwerereleasedin2001or2002.ThesalesdatawereprovidedbyNielsenBookScan,amarketresearchrmthattracksbooksalesusingscannerdatafromanalmost-comprehensivepanelofretailbooksellers.Additionalinformationabouttheindividualtitles(suchasthepublicationdate,subject,andauthorinformation)wasobtainedfromavarietyofsources,includingAmazon.comandavolunteerwebsitecalledOverbooked.org.Table1reportssummaryinformationforthebooksinthedata,brokenintothreesubsamples.Theoverallsamplerepresentsarelativelylargefractionoftheuniverseofhardcoverctiontitlesreleasedinthistimeperiod,thoughitislikelytobesomewhatskewedtowardpopularbooks.Booksthatneversoldmorethan50copiesinasingleweek(nationwide)weredroppedfromthesample,sincetheirweeklysalesnumbersappearedtobemostlynoise.Also,somebooksareexcludedfromtheempiricalanalysesiftheirreleasedatesweredifculttodetermine.Insuchcases,thenumberofbooksisreducedto799.AswillOccasionallyanauthorhastestedthispropositionbypurchasingnumerouscopiesofhisownbook,inhopesofpushingitontoabestsellerlist.Noneoftheseattemptshaseverbeentrulysuccessful;thetypicaloutcomehasbeenconsiderableembarrassmentfortheperpetratoroncetheschemewasuncovered.MostmajorU.S.newspaperspublishtheirownlocallist(sometimesinadditiontotheYorkTimeslist),andanumberofnationallistscompetewiththeNewYorkTimeslistforattention(e.g.,PublishersWeeklyWallStreetJournalUSAToday,andBookScancollectsdatathroughcooperativearrangementswithvirtuallyallthemajorbookstorechains,mostmajordiscountstores(likeCostco),andmostofthemajoronlineretailers(like).Theyclaimtotrackatleast80percentoftotalsales.Inconstructingthesetofcandidatebookstotrack,wehadtorstlocateabook(anditsISBNnumber)inordertoconsiderit.Obscure,slow-sellingbooksare(bydenition)hardertoThreesourcesofinformationwereusedindeterminingtheexactweekofrelease.Formostliststheexactdayofrelease.Ifnot,BookScanreportsthemonthofrelease,andeyeballingthedatausuallyrevealsanobviousreleasedate.Ifforanyreasonthereleasedatewasnotobviouse.g.,theandBookScanreleasedatesdidntmatch,and/orthereleasedatewasntobviousfromlookingatthesalesdatathetitlewasexcludedfromthesamplewhenevertheresultsmightbesensitivetotheaccuracyofthereleasedate.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. beshowninthenextsection,fornearlyallbooksthevastmajorityofsalesoccurintherst4 6months,andsubsequentsalesarerelativelyuninformative.Inlightofthis,onlytherst26weeksofsalesareincludedinthesampleforanygivenbook.Dataonbestseller-liststatuscomedirectlyfromtheNewYorkTimes.TheanalysisfocusesontheNewYorkTimeslistbecauseitisanationallypublishedlist(andthesalesdataarenational),andbecauseitisalmostuniversallyregardedasthemostinuentiallistintheindustry.Amajorityofretailbooksellers(includingonlinebookstores)havespecialsectionsdevotedtoNewYorkTimesbestsellersandofferpricediscountsonthesetitles,andauthorsaresometimesofferedbonusesforeveryweektheirbookappearsontheNewYorkTimesToconstructitslist,theNewYorkTimessurveysnearly4,000bookstoreseachweek,inadditiontoanumberofbookwholesalerswhoserveadditionaltypesofbooksellers(likesupermarkets,newsstands,etc.)Thereportedsalesguresfromtheserespondentsarethenextrapolatedtoanationallyrepresentativesetofsalesrankingsusingstatisticalweights.BecausetheYorkTimeslistisconstructedusingsamplingmethods,itoftenmakesmistakesi.e.,booksthatshouldhavemadethelistinagivenweek Table1SummarystatisticsAllbookswith26weeksofdata(nMinMedianMaxMeanStd.Dev.Listprice10.9524.9568.8524.282.63Releaseweek1578349.6424.18Newauthor001.15.36Sales,rst6months873,9601,443,34535,183111,444Max.one-weeksales50486360,1336,90324,453Bookswithreliablereleasedates(nMinMedianMaxMeanStd.Dev.Listprice10.9524.956024.442.38Releaseweek1568348.7024.52Newauthor001.14.35Sales,rst6months878,0631,443,34551,396134,445Max.one-weeksales50972360,13310,22729,626NewYorkTimesbestsellers(nMinMedianMaxMeanStd.Dev.Listprice10.952529.9525.052.25Releaseweek1408341.3323.89Newauthor001.05.22Sales,rst6months20,96393,5071,443,345174,987222,530Max.one-weeksales2,10818,438360,13336,04750,058Weekrstappeared44344.52.4Weeksonlist14305.85.3Forreleaseweek,therstweekof2001isweek1,andtherstweekof2002isweek53.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial (becausetheirsalesexceededthesalesofthebooklistedatrank15)aresometimesomitted,andtheorderingoflistedbookssometimesdoesntreectthetruerankingofsales(asindicatedbytheBookScandata).Also,assemblingthelisttakestime,sotheprintedbestsellerlistreectsrankingsfromthreeweeksprior.BothofthesefeaturesoftheNewYorkTimeslistthemistakesandthetimelagsarecriticalintheempiricalanalysisofitsimpactonsales.IV.EMPIRICALANALYSISANDRESULTSIV(i).SkewnessinBookSalesThemoststrikingpatterninthedataisthatbooksalesareremarkablyskewedintwoimportantways.First,thedistributionofsalesacrossbooksisheavilyskewed.Figure3plotstotalsalesintherstsixmonthsagainstsalesrankforthetop100booksinthesample.Evenwhenlookingonlyatthetopdecileofbooks,theskewnessofthedistributionisstriking.Ofthe1,217booksforwhichatleast26weekswereobserved,thetop12(1percent)accountfor25percentoftotalsix-monthsales,andthetop43(3.5percent)accountfor50percent.The205booksthatmadeittotheNewYorkTimesbestsellerlistaccountfor84percentoftotalsalesinthesample.Themostpopularbookinthesample,SkippingChristmasbyJohnGrisham,soldmorecopiesinitsrstthreeweeksthandidthebottom368booksintheirrstsixmonthscombined.Second,booksalestendtobeskewedwithrespecttotimeforanygiventitle:thatis,salestendtobeheavilyconcentratedintherstfewweeksafterabooksrelease.Ofthe1,217booksinthesampleforwhich26ormoreweeksareobserved,898(73.8percent)hittheirsalespeaksometimeintherstfourweeks.Themedianpeakweekisweek2.Somewhatsurprisingly,thispatternalsoseemstoholdfordebutauthors,forwhichonemightexpectgradualdiffusionofinformationandthereforeanS-shapedsalespath.Fornewauthors,112of182books(61.5percent)peakintherst4weeks,andthemedianpeakweekis4.Thissecondformofskewnessisimportanttokeepinmindwheninterpretingthemodelsandresultsinthefollowingsections.Thesteadydecayofabookssalesovertimeisthedominantpattern:withtheexceptionofseasonaleffects(e.g.,Christmas),anyotherchangesinabookssalestendtobesecond-orderrelativetothisdecaytrend.Essentially,thesalespathsBooksforwhichthereleasedatewasquestionableareexcludedhere,sincegettingthereleasedaterightisntcriticalforthisexercise.Decay-likesalespatternsarenotuniquetothebookindustry;forexample,Einav[2003]reportssimilarpatternsformovieboxofcesalesandincorporatesexponentialdecaydirectlyintohisempiricalmodel.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. typicallyresembleexponentialdecaypatterns.Eyeballingthetimepathofsalesforallthebooksinthesample,onerarelyobservesabookssalestakeoffafterithitsthebestsellerlist;ifanything,makingthelistappearstotemporarilyslowthepaceofdecline.IV(ii).DoBestsellerListsDirectlyAffectSales?Theoryclearlysuggeststhatbestsellerlistsmaydomorethansimplyreectconsumerbehavior:thelistsmaydirectlyconsumerbehavior,sothatabooksappearanceonthebestsellerlisthasanindependenteffectonitssales.However,measuringsuchaneffectisadifcultempiricalproblem:thesetofbooksthatreceivethetreatmentofbeinglistedasbestsellersisclearlynotrandom,andanaiveempiricalapproachwouldlikelyconfusethedirectionofcausality.Thereisobviouslyacorrelationbetweenthelevelofsalesandbestsellerstatus(bytheverydenitionofabestsellerlist),butwecannotinferfromthiscorrelationthatbeinglistedasabestsellercauseshighersales.However,giventheavailabledataandthesubtletiesintheconstructionofNewYorkTimeslist,thereareatleasttwowayswecanattempttoidentifythelistsdirectinuence.Onestrategyistoexploittheso-calledmistakesthataresometimesmadeinthelist.Asmentionedpreviously,the 500 1000 1500Sales (thousands), first 6 months 1 100Overall sales rank Figure3SkewnessofbooksalesBESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial processusedtogeneratetheNewYorkTimeslistisinexact.AlthoughthelistisbyandlargequiteaccuratewhencomparedwiththetruesalesnumbersavailablefromBookScan,itisnotuncommonforabestsellingbooktobemissedi.e.,abookmaynotappearonthelisteventhoughitssalesexceededthesalesoflistedbooks.Inprinciple,thesemistakesprovideameansofidentifyingtheeffectofappearingonthelist,byservingasanappropriatecontrolgroup.(Comparinglistedbookstounlistedbooksisabadexperiment,sincewhetherabookislistedisanearlydeterministicfunctionofthedependentvariable;butcomparinglistedbookstobookshavebeenlistedis,inprinciple,avalidexperimentaslongasthemistakesarerandomoccurrences.)Duringtheyears2001 2002,therewere182instancesinwhichahardcoverctionbookwasnotlistedasaNewYorkTimesbestsellerwheninfactitshouldhavebeen,representing109differentbooks.(Inseveralcases,thereweremultipleweeksinwhichabookshouldhavebeenonthelistbutwasnot.)Themajorityofthese(roughly70%)werenarrowmisses:hadthebooksbeenlisted,theywouldhavebeenranked13 15onthelist.Inordertoconstructafaircomparison,Ifocusontwosetsofbooks:thosethatwerelistedatrank13,14,or15whentheyrstappearedontheNewYorkTimeslist(44),andthosethatshouldhaveappearedat13,14,or15whentheyweremistakenlyomitted(75).Table2summarizessomeobservablecharacteristicsofthebooksinthetwogroups.Iftheomissionswerenotrandom,butratheranattemptbytheNewYorkTimesateditorializingthelist,wemightexpecttoseeadifferentsubjectcompositionamongtheomittedbooks.Thedistributionofsubjectsandlistpricesappearstobemostlysimilarbetweenthetwogroups,lendingsomecondencethatthe Table2Characteristicsoflistedbooksvs.omittedbooksPercentofmistakeslistingthisgenre(PercentofbestsellerslistingthisgenreLiterature/Fiction51641.39(0.17)Mystery/Thriller51361.53(0.13)Romance21111.47(0.14)ScienceFiction1270.96(0.34)Religion101.00(0.32)Horror320.13(0.89)Averagelistprice:24.9824.541.09(0.28)Onlybooksthatwereranked13 15whentheyrstappearedonthebestsellerlistareincludedinthebestsellersgroup.NumbersarebasedonbooksgenresaslistedonAmazon.com.Percentagesaddtomorethan100becausebookstypicallylistmorethanonegenre.Pricesarepublisherslistprices,fromBookScan.Tobeprecise,Iwillsayabookshouldhavebeenlistedif,fortheweekthatwasrelevantforgeneratingthelist,thebookssales(accordingtoBookScan)exceededthesalesofthebookthatwaslistedat#15.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. omissionsareindeedrandommistakes.TheonlynotabledifferencesarethatthegenreLiterature&Fictionismorelikely(andRomanceorMystery&Thrillerslesslikely)amongbooksmakingthelistthanamongomittedbooks,butthesedifferencesarenotstatisticallysignicant.Moreover,whenconsideringtheoverallcompositionofthelist(notjustpositions13 15),theproportionsofRomanceandMysterynovelsmatchalmostperfectlywiththeproportionsamongtheomittedbooks,soitseemscleartheomissionswerenottheresultofanybiasagainstparticulargenres.Table3reportsacomparisonofsalesforthetwogroups.ForbooksthatwerepublishedforthersttimeontheNewYorkTimeslist,salesdeclinedbyanaverageof7.6percentrelativetothepreviousweek.Formistakenlyomittedbooks,salesdeclinedbyanaverageof22.7percent.Takenatfacevalue,thedifferenceimpliesthatintherstweek,beinglistedleadsto19percentmoresalesthanwouldhaveotherwiseoccurred.However,thedifferenceisstatisticallyimprecise:forsignicancelevelslessthan.08,wewouldfailtorejectthenullhypothesisthatthelisthasnoeffectusingaone-tailedtest.ThecomparisonreportedinTable3doesnotcontrolforanycovariates,butdoingsohasverylittleeffectontheestimate.Usingsimplelinearregressionstocontrolforseasonaleffectsandtime-since-releaseeffectsyieldsestimatesofthesameapproximatemagnitude.Giventheavailabilityofpaneldataonbooksales,asecondstrategyforidentifyingtheeffectofbestsellerlistsistousealltheavailabledata(notjustmistakebooks)tomeasuretheweek-by-weekchangesinsalesassociatedwithchangesinbestsellerstatus.WhereastheresultsinTable3arebasedoncomparisonsbooks(withmistakenlyomittedbooksservingasthecontrolgroupforthelistedbooks),thissecondapproachisbasedonwithin-bookvariationovertime(sothecontroliseffectivelythesamebookssalestrajectorytoitsappearanceonthelist).Observingsalesoverseveralweeksforeachbookmakesitpossibletocontrolforbookxedeffects,thusabsorbingtheobviouslyendogenousdifferencesinsaleslevelsforbestsellervs.non-bestsellertitles.Moreover,thetimelaginvolvedintheNewYorklistmeansthatwehaveatleastthreepre-treatmentobservationson Table3Comparisonofsaleschanges:listedbooksvs.omittedbooksMeanStd.Dev.sales,rstweekappearingonbestsellerlist(n.076.702sales,weekinwhichmistakenlyomitted(n.227.157.151.One-tailedt-testfor:differenceiszero:1.404,Recallthatthedominantpatterninsalesovertimeisasteadydecay:evenforbooksappearingonthebestsellerlist,itisrareforsalestoincreasefromoneweektothenext.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial eachbestsellerbeforeithitsthelist.(Duetothetimelag,thesoonestabookcanappearonthelistisatthebeginningofitsfourthweekonbookstoreshelves.)Inprinciple,therefore,thedatacanbeusedtoestimateamodelthatcontrolsforbook-specicdifferencesinthelevelofsalesbook-specicdifferencesinsalestrendsobservedpriortoappearanceonthebestsellerlist.Anempiricalmodelofbooksalesovertimemustaccommodatethedominantdecaytrendinsalesovertime(asdescribedinsectionIV(i))andallowforappearancesonthebestsellerlisttodirectlyaffectsales.Moreover,goodnessoftisevenmoreimportantthanusualhere,sincethecounterfactualwewantfromthemodel(howmanyunitsabookwouldhavesoldifithadntappearedonthebestsellerlistinacertainweek)isessentiallyaforecast.Giventheseconsiderations,onesimpleapproachistomodelbooksalesasanautoregressiveprocessinwhichtheautoregressionparameterisafunctionofcovariates:withasetofcovariatesforbookinweek(includingbestsellerstatus)andademandshockthatisindependent(butnotnecessarilyidenticallydistributed)acrossand.Theautoregressivestructureisappealinginpartbecauseittrackssalesquitewell,sothatthepredictioninanyperiodwillneverbeoffbymuch.Moreover,thisspecicationfocusestheestimationonchangesinsalesforagivenbookratherthandifferencesinthelevelofsalesacrossbooks,andallowsforbook-specicdifferencesintherateofdecayinsales(viabook-specicconstantsin).Inotherwords,themodelcanaccommodateunobserved,book-specicheterogeneityinboththelevelandtrajectoryofsales.Inorderforanyremainingendogeneitytobeimportantitmusttakeapeculiarform.Inparticular,thetimingofabooksappearanceonthebestsellerlistwouldhavetocorrespondtoweeksofidiosyncraticallyhighdemandinordertobiasourestimateofthelistsimpact.Giventhatthecurrentlistreectssalesfromthreeweeksprior,suchendogenoustimingseemsveryunlikely.Althoughpublishersaresurelystrategicaboutwhentheyreleasebooks,itseemssafetopresumetheydontsystematicallyreleasebooksexactlythreeweeksinadvanceofanticipatedpositivedemandshocks.Thoughimplementedsomewhatdifferently,theempiricalstrategiesemployedherearesimilarinspirittotheanalysisofReinsteinandSnyder[2000],whoexploitquirksinthetimingofmoviecriticsreviewstoidentifytheirimpactonboxofcesales.Onepotentialcaveatisthatpublishersmaystrategicallyavoidreleasingbooksatthesametimeasmajorblockbusters,whosereleasedatesareannouncedwellinadvance.(Thisisdiscussedinmoredetailinsection4.3below.)However,eveninthiscaseitisnotclearwhypublisherswouldtimetheirreleasessuchthatthepositiveshockcomesexactlythreeweeksafterALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. Table4reportsestimatesofthismodelforfourseparatespecications,usingweeks2 26foreachbookinthesample.Allfourspecicationsincludeafullsetofweekdummiestocontrolforseasonalvariationinbooksales(thereisalargeincreaseinsalesinmidtolateDecember,forexample),aswellasafullsetofbookdummiesin.ColumnIreportstheestimatedcoefcientonanindicatorthatequalsoneforeveryweekinwhichthebookappearedontheNewYorkTimesbestsellerlist.Thepointestimatesarestatisticallysignicantatthe5percentlevel,andimplythatsalesdeclineabout4percentagepointsmoreslowlywhenabookislistedasabestseller.Giventhemanytheoreticalreasonswhyappearancesonabestsellerlistshouldhaveadirectimpactonsales,itisnotsurprisingtondastatisticallysignicanteffect.Themoreinterestingquestionisperhapsthelisthasani.e.,whatistherelevantmechanismbywhichlistappearancesboostsales?Oneclassofexplanationsisbasedontheideathatbestsellerstatusisinformative:forexample,bestsellerappearancescouldsignalhighquality,orsignalwhatotherpeoplearereading(andtherebyfacilitatesocialcoordination).Anotherpossibilityisthatsalesincreasesresultfrompromotionalactivitiesundertakenforbestsellers(e.g.,prominentplacementonbookstoreshelves).Arathermundaneexplanationwouldbethatthe Table4Regressionestimates:booksalesandliststatusIIIIIIIVNYT-listed.043.030(.015)(.019)NYT,week1.087NYT,week2NYT,1weekoffrank.004.004.042(.016)NYT-listed[6 10].040(.027)NYT-listed[11 15].068Oprah6.3566.2866.3446.346(.111)(.111)(.114)(.112)GMA.108.100.118.109(.085)(.083)(.086)(.085)Weeksout(.022)(.021)(.024)(.023).977.978.977.977#books799799799799#observations19,02419,02419,02419,024Robuststandarderrorsinparentheses.Allexplanatoryvariablesareinteractedwithlaggedsales,asexplainedinthetext,andeachspecicationincludesweekxedeffectsandbookxedeffects.Thenumberofobservationsdoesnotequalthenumberofbookstimes25becausesomeweeksaremissing.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial increasesaremerelypriceeffects,resultingfromtheautomaticdiscountthatmanystoresofferoncurrentbestsellers.ThespecicationsreportedincolumnsII IVofTable4attempttoshedlightonthesourceofthebestsellerlistsimpact,andspecicallytrytodistinguishwhetherinformationorpromotionisdrivingtheeffects.ColumnIIseparatestheeffectofalistappearancebetweentherstweekandsubsequentweeksonthelist.Ifthesalesincreasesaredrivenbyinformation,weshouldexpectthemtobeconcentratedinweekone,whentheinformationisnewtoconsumers.Ontheotherhand,ifsalesincreasesmostlyreectpromotionalactivities,thentheeffectsshouldbeapparentaslongasthebookremainsonthelist.Theestimatesaremoreconsistentwiththeinformation-shockexplanation:theyindicatean8.7percentage-pointchangeintheweekthebookrstappears(statisticallysignicantatthe5percentlevel),withanyeffectsinsubsequentweeks(combined)beingstatisticallyindistinguishablefromzero.Moreover,abooksremovalfromthelist(which,accordingtothebookstoremanagerswithwhomIspoke,abruptlyterminatesanyspecialin-storepromotionalactivityforthebook)haslittleimpactonitssales:thecoefcientonthedummyforrstweekoffisneitherstatisticallynoreconomicallysignicant.Giventhisresult,itisveryunlikelythatthemeasuredimpactofbestsellerstatusrepresentsapurepriceeffect.Justasinformationaleffectsshouldbeconcentratedintheweekofabooksrstappearanceonthelist,theymayalsobeconcentratedamongbooksatthetopofthelist.Thereisobviouslyabigdifferenceinthebuzzgeneratedbyanumber-onebestsellerthanbyabookbarelybreakingthroughtothebottomofthelist,andsocialeffectsinconsumptionarepresumablystrongestforthebooksattheverytop.Bycontrast,theeffectsofin-storepromotionsshouldbeequallyimportantforallbooksonthelist,becauseallofthemareputonaspecialrack,and/orallofthemarediscounted.ColumnsIIIandIVofTable4reportestimatesfromregressionsinwhichtheimpactofalistappearanceisallowedtodependonthebookspositiononthelist.Basedontheseestimates,thereisnoevidencethatthelistsimpactislargerforbooksatthetopofthelist:theinteractionofthelistdummywithlistrankisstatisticallyindistinguishablefromzeroinspecicationIII,andthecoefcientsonlistdummiesforseparategroups(books1 5onthelist,books6 10,andbooks11 15)arestatisticallyindistinguishableinspecicationIV(0.13,-value0.88).Ifanything,thepointestimatessuggestlargereffectsforbooksattheofthebestsellerlist.Ininterpretingtheseresults,itisimportanttorememberthattheinformationcontentofanappearanceonthelistisquitelowforsomeauthorseveryoneknowsthatGrishamsnewestnovelwillmakethelist,sothereisnoinformationshockwhenitrstappearsandtheestimatesinTable4reporttheeffectoverallbooksthatmadethelist.Indeed,ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. itcouldbethatformostbooks,appearingonthelisthasvirtuallynoeffectonsales;onlyappearancesthataresurprisesmakeadifference.TheresultsreportedincolumnsIIIandIVofthetable,suggestingthatappearancesatthetopofthelisthavenomoreimpactonsalesthanappearancesatthebottomofthelist,maysimplyreectthatmosttop-rankedbestsellersarebooksthatwereexpectedtobethere.Inordertoexaminethisissuemorecarefully,themodelwasre-estimatedallowingforbook-specicheterogeneityintheeffectofbeinglisted.Thatis,themodelwasestimatedwithisanindicatorequaltooneifbookappearedonthebestsellerlistforthersttimeinweek.Themoststrikingfeatureofthesishowtheyrelatetotheauthorshistories.Figure4showskerneldensitiesfortheestimatedsseparatelyforwell-establishedauthorsvs.newerauthors.Forestablishedauthors,thedistributioniscenteredat0.02andisapproximatelysymmetric.Forauthorswhowerelesswell-established,thedistributionhasapositivemodeandalongrighttail,whichisconsistentwiththeideathatbestsellerstatushasanappreciableimpactonlyforsomebooks.Perhapsthemostinterestingresultrelatestothesmallsampleofelevendebutauthorswhomadethelist.Fortheseauthors,theestimatedsareconsistentlylargeandpositive:oftheeleven,nineareinthetopquartile,andsevenareinthetopdecile.Theaveragevalueofamongthesenewauthorsis0.35(comparedwith0.05fortheremainderoftheauthors).Althoughthesamplesizeisobviouslysmallhere,thepatternsareclearlyconsistentwiththehypothesisthatappearingonthebestsellerlistonlyhasanimpactwhentheappearanceisinformative.TheheterogeneityevidentinFigure4isimportanttokeepinmindwheninterpretingthequantitativeresultsfromTable4.Takenatfacevalue,theestimatedcoefcientsimplyamodesteffectoflistappearancesonsales,eventhoughthespeciedautoregressivemodelallowstheeffecttopersist.Forexample,comparingabookthatappearedonlyonceonthebestsellerlist(initsfourthweekfromrelease)toabookthatstartedwiththesameinitialsalesbutneverappearedonthelist,andassumingaconstantTwooutlierswereomittedintheestimationofthedensity.One,anextremenegativevalue,wasforabookthatwasannouncedasatelevisionbookclubpickonthesamedayitrstappearedonthebestsellerlist,sothedatahavedifcultydistinguishingthetwoeffects.Theother,anextremepositivevalue,wasforabookwithamysteryauthor.TheDiaryofEllenwasnominallywrittenbyanew(butctitious)authornamedJoyceReardon,anditwasrumoredthatthebookwasactuallywrittenbyStevenKing.Salesspikedconsiderablywhenthebookrsthitthebestsellerlist,soitsestimatedisverylarge.Thegroupsofauthorsweresplitbasedonthenumberofpastadultctiontitleseachauthorhadpublished.Thetopthird(havingpublished35ormoretitles)weredesignatedtheestablishedauthors,andtheremainingtwothirdswerecalledthenewerauthors.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial equalto0.7,the8percentage-pointincreaseinweek4translatestoa13.7percentdifferenceinexpectedsalesovertherst52weeks.However,theaverageeffectsseemtobemaskingtherealstory:forestablishedbestsellingauthors,thelisthasnodiscernibleimpactonsales;butfornewauthors,appearingonthelisthasarelativelydramaticimpact.(Aone-timeincreaseof0.35inweek4wouldleadtoa57percentincreaseinsalesoverthefollowing52weeks.)Notethatevenattheirlargest,theeffectsofbestsellerlistappearancesaresmallwhencomparedtotheeffectsoftheannouncementvariablesincludedintheregressionsreportedinTable4.TheOprahindicatorisequaltooneifthebookwasannouncedasaselectionforOprahWinfreysbookclubinthatweek,andtheGMAindicatorisequaltooneifthebookwasannouncedasthepickfortheGoodMorningAmericashowsbookclub.Theestimatedeffectsoftheseannouncementsareverylargetheydwarfanyeffectofthebestsellerlist.Theinuenceoftheannouncementscouldderivefromvarious 0density .4 0 .8Beta established authorsnewer authors Figure4Heterogeneityintheimpactofbestsellerlistonsalesisinitialsales,thentotalsalesoverweeksis.Ifincreasesbyanamountinweek(butthenrevertstoitspreviousvalueandisotherwiseconstant),thenthepercentageincreaseinsalesisequaltoALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. sources:theannouncementscouldsimplyalertarelativelylargenumberofconsumerstotheexistenceofthebook;theycouldserveasaqualitysignal;ortheycouldactasacoordinationmechanismbywhichalargenumberofconsumersagreetoreadthesamebook.(Thelattermechanismistheapparentobjectiveofthetelevisionbookclubs.)Finally,althoughthespecicationsreportedinTable4shouldadequatelycontrolforunobservedheterogeneityrelatedtobooksvaryinglevelsofsales,itisneverthelessusefultoconstructarealitycheckforthekeyestimatedcoefcients.Forexample,insteadofregressingsalesonaninteractionoflaggedsalesandanindicatorforrstweekonthelist,wecaninteractlaggedsaleswithanindicatorforrstweekalmostonthelist,denedasanyunlistedbookwithsalesgreaterthan90percentofthe15rankedbestseller.IfthecoefcientsreportedinTable4aremerelyanartifactoflatentheterogeneityinsalesdynamicsthatiscorrelatedwithdifferencesinthelevelofsales,thenthecoefcientonthisvariablewouldbepositiveandsignicant.Infact,runningthisregressionyieldsacoefcientestimateof0.028withastandarderrorof.018,lendingsomeadditionalcredibilitytotheresultsreportedinthetable.IV(iii).DoBestsellerListsAffectProductVariety?TheresultsoftheprevioussectionindicatethatsomebooksappearingonNewYorkTimesbestsellerlistenjoyaconsequentincreaseinsales.Buttowhatextentarethoseextrasalesstolenfromnon-bestsellingtitles?AswasexplainedinsectionII,whether(andinwhichdirection)abestsellerlistinuencesproductvarietydependsonwhetherthelisthasanyimpactonbooksnearthepublish/no-publishmargin.Iftheconsumerwhobuysabookbecausehesawitonthebestsellerlistwouldotherwisehaveboughtabooknearthemarginofprotability,thenonecanarguethatthelistreducestheprivatelyoptimalnumberofbooksbyfurtherconcentratingdemandonbestsellers.However,itisalsopossiblethattheconsumerwouldhaveotherwiseboughtnobookatallinthatcase,salesaremoreconcentratedonbestsellers,butthiscomesatnoexpensetonon-bestsellingtitles,andthenumberofbookspublishedisunaffected.Moreover,itisalsopossiblethatbestsellerlistsincreasedemandforbooks,forexamplebysimplybringingmoreconsumersintothebookstore.Bestsellersandnon-bestsellerscould,inprinciple,becomplementarygoodsifconsumersbuymultiplebookswhentheyvisitthebookstore.Thedataavailableareclearlyinsufcienttoprovideaconclusiveanswertothisquestion.Theidealexperimentmightbeoneinwhichalargesetofconsumersmakespurchasesinthepresenceofabestsellerlist,andanothersetofconsumersmakespurchaseswithouthavinganyexposuretothebestsellerlist(eitherthroughthemediaorattheretailoutletitself).BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial TheubiquityoftheNewYorkTimeslistmakesitvirtuallyimpossibletondanygroupofbookpurchasersthatresemblessuchacontrolgroup.Whattheavailabledatapotentiallyrevealisindirectevidenceofsubstitutionbetweenbestsellersandnon-bestsellers.Eventhisisdifcult,however,sincethedatacontainnopricevariationthatwouldenableestimationofcross-priceelasticities.Theonlyuseablevariationistimevariationinthesubjectcompositionofthebestsellerlist.Ifsubstitutionbetweennon-bestsellersandbestsellersisimportant,thenpresumingthatbooksinthesamegenreareclosersubstitutesthanbooksindifferentgenressalesofnon-bestsellingbooksshoulddeclinewhenthebestsellerlistiscomprisedofbooksinsimilargenres.Forexample,salesofanon-bestsellingdetectivenovelwoulddecreaseinaweekwhenthreedetectivenovelssimultaneouslyhitthebestsellerlist.Inordertocapturethesekindsofsubstitutionpatterns,avariablesummarizingeachbookssimilaritytothecurrentsetofbestsellerswasconstructedbycomparingthebooksgenre(s)(aslistedbyAmazon.com)tothegenresofallbooksonthecurrentbestsellerlist.Specically,thepairwisesimilaritybetweenbooksAandBisdenedas #ofgenressharedbybooksAandBnumberofgenreslistedforAnumberofgenreslistedforBThismeasureisequaltooneifthetwobooksgenresareidentical,andzeroifthereisnooverlapatall.BookAsaveragesubjectsimilaritytothecurrentbestsellerlististhencomputedas1,whereindexesthecurrentbestsellersbyrank.Table5showstherelativefrequenciesofthegenresinthesampleforallbooksandforbestsellersonly.Asisclearfromthetable,mysteriesandthrillersarethemostcommonbestsellers,and Table5SummaryoflistedgenresPercentofbookslistingthisgenrePercentofbestsellerslistingthisgenreLiterature/Fiction6844Mystery/Thriller3956Romance1119ScienceFiction69Religion23Horror23Basedonbooksgenresaslistedon.Percentagesaddtomorethan100becausebookstypicallylistmorethanonegenre.Amazon.comoftenlistsmultiplegenresforthesamebook:e.g.,ContemporaryctionandRomance.So)willbelessthanoneunlessbookslistexactlythesameALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. romancenovelsarerepresentedmoreheavilyamongbestsellersthanamongbooksoverall.Importantly,thereissubstantialvariationinthecompositionofthebestsellerlistovertime.Figure5showsaseriesofstargraphstoillustratevariationinlistcompositionforasample36-weekperiod.Totheextentthatthereismeaningfulsubstitutionbetweennon-bestsellersandbestsellers,weshouldobservethatsalesrespondtochangesintheaveragesimilarityvariableinducedbyvariationinthelistcomposition.Table6reportsestimatesofamodelanalogoustotheautoregressivemodeloftheprevioussection,butwithsimilaritymeasuresincludedintheInsteadofabusiness-stealingeffect,theestimatedcoefcientsseemtosuggestacomplementaritybetweenbestsellersandnon-bestsellers.Weeksinwhichthegenre-compositionofthebestsellerlistisclosetoanon-bestselling Figure5VariationinbestsellerlistcompositionovertimeBecausethepurposeistoevaluatethepotentialresponseofnon-bestsellerstochangesintheirsimilaritywithbestsellers,onlybooksthatnevermadethebestsellerlistareusedintheestimation.Thetablereportsspecicationswithandwithoutsubjectxedeffects;withoutthesexedeffects,thecoefcientsonthesimilaritymeasurescouldpartlyreectaveragedifferencesintheprevalanceofcertaingenresonthebestsellerlist,whichwouldobviouslybecorrelatedwithsalesinthatgenre.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial booksgenretendtobeweeksforthenon-bestseller.ConsistentwiththendingsinTable4,theeffectseemstobemostpronouncedforbooksappearingonthelistforthersttime:whenthesimilaritymeasureisdenedonlyforthesubsetofbestsellerswhoserstappearancewasinthegivenweek,thepositiveeffectofthesimilarityisstronger.Thispatterncouldplausiblyreectmultiple-bookpurchases:forexample,weeksinwhichanewromancenovelhitthebestsellerlistdrawmorethantheusualfractionofromanceenthusiastsintobookstores,andtheymaybuyseveralromancenovelswhenvisitingthestore.Althoughtheestimatedcoefcientsonthesimilaritymeasuresarestatisticallysignicant,theyshouldbeinterpretedcarefully.Onecaveattokeepinmindisthatpublisherschoosetheirreleasedatesstrategically.Forexample,thereissomeevidencethatpublishersofnon-bestsellingbookstendtoavoidreleasedatesthatcoincidewiththereleaseofablockbustertitleinthesamegenre.TheregressionsinTable6maybebiasedagainstndingbusiness-stealing,becausepublisherschoosetheirbooksreleasedatestoactivelyavoidbusiness-stealing.Thatis,ifbookAwereexpectedtostealbusinessfrombookB,thenbookBsreleasedatewouldbestrategicallydistantfrombookAsrelease,sothatpotentialinstancesofbusiness-stealingarelesslikelytobeobservedinthedata.Anotherthingtokeepinmindwheninterpretingthecoefcientsisthattheyrepresent(atbest)indirectevidenceofthesubstitutionpatternsbetweenbestsellingandnon-bestsellingtitles,andtheresultsmaynotgeneralizeeasilytoothercategoriesofbooks.Forexample,non-ctiontitlesonsimilartopicsareperhapslesslikelytoexhibitcomplementarities.Aconsumerlookingtobuyabookoninvestments,forinstance,seemsmorelikelytochoosejustone(instead Table6Regressionestimates:booksalesandsimilaritytocurrentbestsellersIIIIIIIVAvg.Similarity:Allbestsellers.044(.033)(.080)Newonlist(.100)(.102)Weeksout.005.005.003.002(.002)(.002)(.002)(.002)Weekdummies?yesyesyesyesSubjectdummies?nonoyesyes.945.945.951.951#books595595595595#observations14,11514,11514,11514,115Robuststandarderrorsinparentheses.Allexplanatoryvariablesareinteractedwithlaggedsales,asexplainedinthetext.Avg.Similaritymeasureshowsimilarabookistothecurrentsetofbestsellers,basedonthebookslistedgenres.Onlybooksthatnevermadethebestsellerlistareincludedintheestimation.Thenumberofobservationsdoesnotequalthenumberofbookstimes25becausesomeweeksaremissing.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics. ofbuyingmany)thanwouldaconsumerlookingforasuspenseorromancenovel.V.DISCUSSION&CONCLUSIONSBasedonthesalesdataanalyzedhere,itseemsclearthatappearingontheNewYorkTimesbestsellerlisthasadirectimpactonabookssales.Estimatesbasedontwoalternativeidenticationstrategiesshowthatsalesincreasewhenabookappearsonthelist.However,themagnitudeoftheeffectismodest,andappearstoreectspikesinthesalesofbooksthatweresurprisesonthelist(e.g.,booksbynewauthors).GiventheamountofattentionpaidtotheNewYorkTimeslist(andthedesperateschemesoccasionallyemployedbyauthorstosecureapositiononthelist),theestimatesofitsimpactmayseemsurprisinglysmall.However,thepresentanalysisignorestwoeffectsthatcouldbequitesignicanttoauthorsandpublishers.First,whileappearingonthebestsellerlistmayhaveonlyamodestimmediateimpactonthebookssales,itmaydramaticallyincreasethepopularityoffuturebooksbythesameauthor.Second,paperbacksalesmaybeinuencedbywhetherthehardcovereditionwasabestseller.(Indeed,paperbackversionsofbooksthatwerehardcoverbestsellerstypicallyannouncethatfactprominentlyonthefrontcover.)Althoughtheseeffectscantbemeasuredwiththeavailabledata,anecdotalevidencesuggeststheymaybeimportant.Althoughthereareavarietyofreasonswhythebestsellerlistmightdirectlyinuencesales,thepatternsinthedataaremostclearlyconsistentwithaninformation-basedexplanation.Retail-levelpromotionscannotrationalizetwoofthestudyscentralndings:whereasretailerspromotionspersistforthedurationofabookstermonthelist,theestimatessuggestthattheimpactofappearingonthelististransitory,withthebulkoftheeffectrealizedintherstweek.Moreover,theimpactonsalesismostpronouncedamongrelativelyunknownauthors(newauthorsinparticular),apatternthatfavorsinformationoverpromotionasanexplanationfortheeffect.Theimpactofthelistonsales(andtheconsequentincreaseintheconcentrationofdemandonbestsellers)raisestheinterestingcounterfactualquestion:Wouldmorebooksbepublishedifitwerentforbestsellerlists?Whether(andinwhichdirection)thebestsellerlistshiftsthepublish/no-publishmargindependsonthenatureofsubstitutionbetweenbestsellingandnon-bestsellingtitlesi.e.,aretheextrasalesofbestsellersonesthatInunreportedanalysesusingsalesdatafornonctiontitles,coefcientsonsimilaritymeasureslikethoseinTable6indeedsuggestedapatternofsubstitutionratherthancomplementarity.Nonctiontitleshavetheadvantageofbeingabletocategorizesubjectsmuchmorenelythaninction,butitwasdifculttoincludenonctiontitlesinthebroaderanalysisbecausethecoverageofnonctionbestsellersinmysamplewassparse.BESTSELLERLISTSANDPRODUCTVARIETY2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrial wouldhaveotherwisegonetolesspopularbooks?Theresultshereofferindirectevidencethatforhardcoverction,bestsellersandnon-bestsellerswithinthesamegenremayinfactbecomplements:weeksinwhichbooksofaparticulargenrerstappearonthebestsellerlisttendtobestrong-sellingweeksfornon-bestsellersofthesamegenre.Althoughtooindirecttobeconclusive,thisresultsuggeststhatmarketexpansioneffectsdominateanybusiness-stealingassociatedwithbestsellerlists,sothatbestsellerlistsmayinfactincreasethenumberofbookspublishedinequilibrium.REFERENCESAlexander,P.,1997,ProductVarietyandMarketStructure:ANewMeasureandaSimpleTest,JournalofEconomicBehaviorandOrganization,32(2),207 214.Banerjee,A.,1992,ASimpleModelofHerdBehavior,QuarterlyJournalofEconomics107(3),797 817.Becker,G.andMurphy,K.,2000,SocialEconomics:MarketBehaviorinaSocial,(HarvardUniversityPress).Berry,S.andWaldfogel,J.,2001,DoMergersIncreaseProductVariety?EvidencefromRadioBroadcasting,QuarterlyJournalofEconomics,116(3),1009 1025.Bogart,D.(ed.),2001,TheBowkerAnnual,46thedition.Clerides,S.,2002,BookValue:IntertemporalPricingandQualityDiscrimination,InternationalJournalofIndustrialOrganization,20(10),1385 1408.Cole,D.,1999,TheCompleteGuidetoBookMarketing,(AllworthPress,Chicago).Dixit,A.andStiglitz,J.,1977,MonopolisticCompetitionandOptimumProductAmericanEconomicReview,67(3),297 308.Einav,L.,2003,GrossSeasonalityandUnderlyingSeasonality:EvidencefromtheU.S.MotionPictureIndustry,workingpaper,StanfordUniversity.Epstein,J.,2001,BookBusiness:PublishingPast,Present,andFuture,(W.W.Norton).Greco,A.,1997,TheBookPublishingIndustry,(Allyn&Bacon).Korda,M.,2001,MakingtheList:ACulturalHistoryoftheAmericanBestseller,,(Barnes&NobleBooks,NewYork).Lancaster,K.,1975,SociallyOptimalProductDifferentiation,AmericanEconomic,65(4),567 585.Milgrom,P.andRoberts,J.,1986,PricesandAdvertisingSignalsofProductQuality,JournalofPoliticalEconomy,94,796 821.Reinstein,D.andSnyder,C.,2000,TheInuenceofExpertReviewsonConsumerDemandforExperienceGoods,workingpaper,GeorgeWashingtonUniversity.Rosen,S.,1981,TheEconomicsofSuperstars,AmericanEconomicReview,71(5),Suzanne,C.,1996,ThisBusinessofBooks,(Wambtac).Sweeting,A.,2005,TooMuchRock&Roll?StationOwnership,ProgrammingandListenershipintheMusicRadioIndustry,workingpaper,NorthwesternUniversity.Vettas,N.,1997,OntheInformationalRoleofQuantities:DurableGoodsandConsumersWordofMouthCommunication,InternationalEconomicReview,38,Watson,R.,2003,ProductVarietyandCompetitionintheRetailMarketforEyeglasses,workingpaper,NorthwesternUniversity.ALANT.SORENSEN2007TheAuthor.Journalcompilation2007BlackwellPublishingLtd.andtheEditorialBoardofTheJournalofIndustrialEconomics.