Chapter 3 Imperfect competition, virtual products, and network industries FUNDAMENTAL ISSUES l. 2, 3. 4. 5. What are alternalive forms of competitioo among evenly matched but imperfectly competitive e-commerce f irms? How do the specialproduction-costcharacteristics ofvirtualproducts influence how they are priced? How do network externalities alfectthe demand for a producl? How can network externalities produce market feedback effects? ljnder what circumstances will a lirm choos€ to offer products that ar€ compatible with those ofcomp€ting lirms? E.COMMERCE TODAY FreshDircct E one af sewralftns that spec ialize in fillins online otderc fat srcceties. Each yeaL the canpany qenetates an avetaqe af 9200 mlllian in annual revenues by shipping about 2 millian atde6packed into appru inately I nillionbaxes. To induceconsune6toprrchase lram amons the 10,00A itens it olfe6 far sale, FreshDnect enphasizes its capability ta cElomize morc lhan one-faurlh ol lhose itens, Custohe6 can, for inslance, choose the thicknesses of steaks- In adlition, they can select the narinades for the steaks and even specify haw the steakswillbe packased. Ihe anline grocer also strives to nainrain at teast a99.9 petcent nte af iten accuncy. FrcshDkect\ eles constltute only about 6 percent of tatal online gtacety sales and less than one-tenth of 1 percent af sales thrcuqhout overall bath clicks and bricks-grocery industty. Thus, F reshoirect is a snall fhn in rclation to both the anline grocety natketplace and the overall narket for qroceries. Nevenheless, the special teatures Fresh Dnect allerc its custonizable online the lirn ftan its otie6, its deti r'ery setvice, and its attentian to accuracy -differcntiate canpetitors. 61 ELECTRON]C COVI IM ERCE_ I\,lARKETS AND PRICES perfe.tly .ompctitim lirm, FreshDire.t: output olgro.crJ producis is lery smallr.latiycto thc industrv as a{hole.The lirn:abilityto diffcrcntiatc its products Like a ard services from those olothcr onlinc and ofllinc groccrs implics, holLcvcr, that its products are cledly distinguishable from those of its competitors.Thus, Fr.shDircct and other firms that sell groceries online are not perfe.tly competitire firms. They operak h aD nrdustN that is inperj6.r, competitile.As Iou will l.arn in this chaptcr, FreshDirect and other online groceH engage in aform of lmp.rfcct compctitb. that economists call mo,op"lint.!Dperi,l.,. In this .hapter you will learn hos to eraluate moDopoli$cally competitire indust.ies, such as thc gro(rrv indust.{ in wh(h a numbr:r ollirms sellproducts that rrc similar but suflictuntl) distinguishablc that consumcrs rcga thcm a\ {sscntially dillercnt itcms. In additb., vox $ill l.ain how to cxamnre th. functioning ol an nrdudh containing onlv a handlul ollirms.You *ill discorcr {hv many economists beliere Gat perfect compedln)n rarely arises am,,ng firms that sell virtual produ.ts. In addltion, you Nill learn ahout how band*agon ellects r.ally can arisc r.ithin nctrvork indushies, therelrv .onfronting lirms in su.h indushies vith important choices regarding * hether to oller produ.ts that ce operate jn .onju.tion with items olTered for sale [,y their .ompetitors. E.COMMERCE WITH IMPERFECT COMPETITION_OLIGOPOLY AND MONOPOLISTIC COMPETITION OtI THE WEB ln the previous .hapter, )ou leatned about horv being li*t mover might allow a companl to emerge * a dominanl firh $ ithin an electroni. marketpla.e. Sohetimes only a lcw firms arc able to mount a successtul challc.gc to a li.st mov.r. Ir mar be that the minimur elEcient scale for a t}?ical6rm jn the industry is relatirely large. There may be signidcant startup.osts entailed in entering the industry. and only a few firns nal be able to raise sufliciot fuds to estabiish a market presen.e. Other barriers to entrl such as gorernment licensing requirements or other legal impedimeDts, ma) also discourage narLet entrv bv all but a le* firms. As a consequence, just ahandful of firms mar comprise an industrv rvithin an electronic mdketplace. a E-commerce oligopolies oligopoly:.an L.onomislr lse th. t.rm oligopoly t{, dcsc.ib. a lerv lirms are prorecred lron comp.titnr. an lndustrr stru.turc ln whictr onlv bv barrnirs to.ntn: Srlatcgicpricing and production are a drtlnguishing charactcristn,ol iimrsln an oligoplistic industrr, each of *hich re.oqniTcs that its own frn,ing and outprt declsions rdll affect the.hoi.es rlleir rivals in tf(i markctplac.. F.onomisrs call this o ligapolin). rnt dd.pendd.e. ol Game theory and oligapoly A problcm that economists face in eeluatiDg oligopolisti. indusbiesis thal therc are manl different theories olthe quantity and price setling shategies that a fcw intcr dependent Erms nay adopt. Suppose, for instaDce, that ts o rimls who share nearly all ::ta: 62 1 IVI P E RF ECT COIl/] P ETIT]ON AN D N ETWORI( ]N DU STRI ES industy outplt focus on qudtitv-based profit maxihization. One .ompanx for instm.e, hight plan alternativc proiit mdimizing output rates under diflircnt scenarnx concerning i1s main riul. Simultdeously. the rivai firm might cLoose its orn pro6t maximizing output rates given alternativc possibilities Lr the output rate oI the first .ompany. Their combined rate of output \ould be .onsistent witL both .ompani.s'produdion sftategies. lfconsumers regard the t*o lirms'products as essentiallv perfect substitules, oncc the compmies' output rates arc dctcrmined the p.oduct price will be th. amount consMers &e $illing to par at this total mrket Itisalsopo$ible tlat the two riva.ls could.ohpctcon thc basis oftLeprj.e charged lor their products. On. nlght set a schedule ol dille.ent priccs it wouid charge iD response to pricing dccisions of its main rival. lte other might do tbc same Then the \ould yieldamarket product price.The harket quantitrsold by the tuo lirns wouldbe determined by thc qumtirv pricing decisions that are consistent with both strategies demdded at that pricc The idea 1hatfirms in oligopolistic industlies mighr.hoose guantities nr maximizc thcir pro6ts giren the productlo. strategies oirival6rms wd proposed in the ealy ni.ct.cnth century br AuSustin Cournot md has since been horn as the Coumot approach. About fou decades latci Joseph Bertrand suSgested that 6rms might instead set prices to maxihize proHts gikn the pli.ihg strategics ol competing firms. Both approachcs, in turns out, are part gane rneoN, a broader approa.h to examidng the bchavior of 6rms that de oligopolisticallv intcrdcpc.dent.This encompassing theor] cxami..s alt rnadle strategies and tactics that intcrdcp.ndcnt nrms may utilize in m cllbrt to maximize their prollts. A key impliGtion ofgame theory is that there is that oligopolistic ilrns a wide rnge ofakernativc stratc€ics may adopt. Consequendy, there also are mmy dilTercnt possihlc lvals that lirms in olisopolistic indusfties ma) behale. Hence, economisrs hc. somethmg olan"embanassmc.t olriches'in theoreticai predi.tions alJout behau)r in such an industry xhich you u,ill see in Chapters 9 and 10 .onph,atcs c*orts to determine appropriate .egulatory policics for markets dominated by just a few Erms. An oligopoly example of interdependent pricing To.on1empla1e the rrpes (,fissu(is prcscnt{al by oligopolhric intcrdcpcnd.nc., lcCs .onsider a setling in {hich an industrr consists olonlr t*o lirms, dcnotcd Firm A ana Firm B, cach of rLhich produces virtualh identical products. As shovn h Figure L l, each 6rm must de.ide whether to .harge a "high" o.'1o$" fri.e dlring 1he currenl period. If Firm A charges a 1ow prta, and Flrm B docs not, tLen Firm A $ill oDsiderably outsell Fim B an.l reap $4 million in pro{its, she..as Firm a xill take in on\' $ I million. But if Fiim A charges a low ptt and Firm B also .harges a low price, eachirm sells the same quartity ofoutput od earns $2 Norv consider the posibilities if Firm A opts to charge millnr.. a high pricc. If Firm A .harges high price and rirm 1l does no1. theh liirm B $ill outsell llrm A, lcaving Firm A ednjng only s 1 miliion, whereas Firm B's sal.s sill bc g4 miLbn.llFirmA dd Firnr B both charge the high price, tlen eacL takcs i. Sl milln,n. a 63 ELECTRONIC COI\4I]/I ERCE IV]AR1(ETS AN D PR]CES ; Wl Figure 3., A qane theaty exanple af oliqopaly pricins.In this example, if Firn A charges alaw ptice and Fitm B daes naL then Firn A'ivilleatn ptafits af $4 million, while Fnm B will earn $1 million- If Firn A and Firn B bath charqe alowprice, eachfirnearns$2 million. In contrast, if FirmAcharqes a high price and Fitn B does nat/ then Fitm B earns ptafils of $4 nillian, and Firn A earns only $1 nillion. If Firn A and Firn B both charse the high price, then each earns $3 nillian- Far both firms, charqinq a low ptice is the doninant nrabgyl sa each firn will end up eatninq prcfits af $2 millian. Clearl1, each lirm's earnlngs outcomes ar. interdepend.Dt: HolL .ach lirm perlbrms dcpcnds not only on its pri.,ing {,hoi.c but also on thc choice ofits rival. (}rnsldcr thc situatioD in s,hich the tso ffrms cannot coordlnate their declsions but must rnake independent choices even though final pa),ofs ther re.eile are interdependeDt. ID this a lo$ pii.e, FirmAi best d..is nisal$to.hoose alo!pricc. il tiirh B sets a high pri.e, Iijrin Al oplimal choi.e is ll 10.hoose a low situalion, ifFirm A sek But dominant strateg), meaning that the lo$ pri.e choice always \ ields the highesl earnings irevectire ol the choi.. ol Fi.m B. Also not. that bJ dre same logic, a los' price srategv is also dominat l;r Firm B Thus, if the firms facing tle po$itrle earnings pa)ofs displaycd in Figur. price. For nrmA, therefore. charging l.l a lo{'price is alwqs a cannot coordinate their pricing de.isions, bo1h \Lill chargt firm will earn L,{ priccs, and r:ach rnillion in proiits durjng the.urrent period. Faccd !'ith this po$ibilit! it liould be in rhe interests of the t\o lirhs ro trv to .oordi.atc thclr prlc. choices If thev could a.complish such coordini&,n $ hi.h i@omists call collusion-ach rvould agree ro set a high p.icc and hcnc. agree to produce the same guanlitr and th.rcb! .arn profits ol g3 million during the .urrent period. Thcrc is, of course, a coordination problcm that th. l]rms musr overcome. As llgurc I . I indicates, if both nrms lnitially agrc. to keep thet prices high, there is anjh.entirc ft,r cach 6rn torenege on rhc agrccmcnt and charge a lower price that sould boost its sales and consequentlv its prolits.Thus, a collusive pricing and produ.tnrn agreement can be di$icult lor firms to maintaln. ('lhc difl;.utties entail.d in maintaining a collusi\e pri.ing agrcement do notne.esrarilr prcr.nt nnns lrom attempling to collud., hosrveri see the box Clobdltntian Online:(:olluitng to BaDi th. 64 $2 P.rces aJ LcD s uns.) ]IV]PERFECT COIVlPETITION AND NETWORI( IN DUSTRI ES GL0BALIZATIoN 0tlLItlE: coLLUDIttc T0 800ST THE PRICES OF LCD SCREENS Recently, three Asian firrns-Chunshwa Picture Tub€s, LG Disptay, and Sharp- admitt€d that th€y col uded to r€strain producuon oi LCD screers used in laptop cornputers, compuier monitors, cell phones, and other etectronic devjces used to oldjng down production of L C D screens rcduced rhe quantity availableto purcharers oi thes€ products, such as Apple, Dell, ed [,lotorola, conduct electronic commerce. H This led these and other purchasersof LCD sffeensto bid up pricesto hisher teve s. The resultwas incr€ased profits for the three colluding firms. For ctitical analysis: Haw did rhe collLsive actians of the three Asian prcducerc of LC D scrcens likely affect the ptices of watlucts af canpanies such as Apple, Del|, and Mototala? Consequenlr, wha else had to pay hisher prices as a consequence a{ callusion by these fnn$? Product differentiation and monopolislic competitioI ln tuany indust ies, ther. irc rclatilel! low barrlers to cnlrr.nd. h.n.e, a.umhcr ol .ompeting sellers. ta.h n lli r, howcvcr, dcvotcs ronsnkr.blc cft;rr ro distinguishinS rts products lrom thor.ol ils (ompctitors. In the cl..rronic markepla.e, I.terner {.ll(s tvpicallr strn! to ijnd unique and readilr recognizable\\'eb addresses, such as \ah.ro, GooSle, o.-,\ma,otr, an(it(, dcmlop ar.puratnrn lorh.\ nrg the nost hs.inarin8 \th page, the mosr currnt a.d inktusring iinLs k) orhf sircs. hncst shoppingfeatur.s, in(l thc b.st ostomer $rnr A monopotistic competitive industrv is one h {hich a nunber oftirms sotl ,litferenliat.d produds grrxls and services ci. disti.glish a.u)ss ro .ntrr orr)cxittast.t rnetsellerswith that ct).sumds puluc.rs. It is also an ind usrrr thar is easr .rho$isc catch brand Dah.s su.h as Blaze.com. s,,ldoul com, (liscoro.d bel;rc running p.rsntent economi. lossrs, shutting do$ in(lsclling thcir\\kb names lo. oiher uses. and Dccplclp n (oh thcir opclatt)ns. Short-run pticing at a monopolistically competitive lnternet seller Under nonopolistic comp.titbn, cach firm vithin an indushv produces outpur at a rate that is relatively small compar.d with toral indusbv outpur. In addition, lirhs in thc industrvproducc similar, butnot identi.al, products. Eachis able ro set rhe pricc ol its o{n produ.t, but dcmand depends on the availabilit) of .lce substirufus produced b! other lirhs in rhe industh: Btcausc a monopolistitallr competitire firh sells. produ.t d'at is ar lest slightl! dill'crcnt from thc produ.ts ofallother 6rms,there is a unique curomer demed l;r its produd. Thus, .ach indilidual6rm potentiallv.m set a price rhat is diferent than th. avcr{. price charged by other lirms in the in<1ustry. Nerertheless, each lirm 65 ELECTR0NIC C0lvl[/ERCE I\TIARKETS AND PRICES recognizcs that scttin8 a relatirel) high pri.c for its pDduct.an induce some ofits .ustomers to switch to consming the substitutc produ.ts offered l,) rival lirns. Pancl (a) of Figurc 1.2 shows a po$ibl. short run situation that a monopolisticlh .ompetitive Internet selle. might lacc whcn it takes into account prorluction cosB as well as the deman(l lbr its product md tlE impltul marginal r.v.nue curve. To simplifl, ssme th.t it can als?ys vm al fadors ol production, so that its long ruaterage .ost cule applies.Thc othc. cost arve sho$'n in the figurc is thc tirm\ narginal cost $.e mdimize its economicpro6t, thc lirm produces to th. point at which marginal revenue eguals marginal cost.This is point S where the naiginal rca.nue and marginal cost curyes Goss. Consequently, in the short-run situation shown in Pdet (a), the 6rm's profit-maximizing outputhte is q{ at s'hich m.rgiml revenu. equals m ginal To (r,st.Thc Erm charges the prie P! that the d.mand crrc indicates its .Ntohers are $ illing to pay for this amout of output. Thc aroage total cost of produ.ing qs mits is .lTCs. Thc 6rm's total profit, tlerefore, cquals (P. lTC.), which is the height ofthe shaded re.tangle, times q,, which is thc base of the rectangle. Thus, the shaded rectangle depicts a positive econonic prolit for the lirm. l'1.] aa. a Fiswe 3.2 Denand, ptoductian, and pricing at a fim in amonapotistically conpetitive narketplace. fhe uaducl ol a nonopolistically conpelitive firh is dis tinguishable from those af nuneraus tival ftms in the induslr/, so it laces downwatd-slopinq demand ahd naqiDl revenue curves. Pahet (a) depicts a shoiiun situation in which the fin prctluces to the point wherc maryinal rcvenue equals narqinat cost at a shonlun equilibriun point S. At this point, it ptaduces Q s units, which it sells at the price P e thereby eanins a positive tines Qs- This induces othet fnms b econamic profit equalto (Ps- AfC (b), entry by new tivats has twa effects entet the industry. As shawn in panel on lhe denand and narginal reEnue curues faced by this firm- Fnst, the denand and maryinat reve.ue cunes tend to shift teflwad as sone af the Itn5 cusloners begin consuning ootpul fton new fins in lhe industry. Second, the availability af narc substitute products ca ses the denand and marginal revenue curues to became mare elastic- At a point of |onglun equilibriun, the demand cutve is tansent to the fnn's aeerase total cost curve at point L1 and the ftn eans zero economic ptofit, ! 66 IfulPERFECT COl\lPETITION AND NETWORK INDUSTRIES Long-run equilibrium in a monapolistically campetitive electronic marketplace Inthc short run, a 6rm in a monopolisucalll .ompctitivc i.dustrv can.arn a posilive cconomn, pro6t, slich mcans that its total rocnuc.an .accc(lthc opportunity cost of hcnrg pnrt of that indusht instead of mother industry. ln a monopolisticallJ conrpetitire indusfty,itis easy for new firms to enter the industrl: Positire economic prolits encourage other Erms to enter the jndustrl', thelebl inceasing the rang. ol product choice aeilalle ro consumers. The fact dut tLe lirm depided in panel (a) ofligure 3.2 earDs apositive econonrlc prolit is a signal that ..!rnucs in this indusirl art more than sufli.ienl 10 .o!er the opporlunil) .on ofbeing in thn i.dustrv. In the l()ng run, additional fi.ms will enter dre indust). P el (t ) shows $ hat happens at the p.eviously exisling lirm in pmel (a) folloring the entrl ofne* firms. rirst, the denrdrd lbr 1his nrm's produ.t de.lines, b.cause some ofits .ustohers will L,uy sinilar goods from other firms. Thus, the demand.urre shifts lcftrvard Sc.ond, thc cntrr oi'nce ilrm\ mcans that morc substitutc produ.h arc avallable, so the demud for this 6nn\ productbecomes nore elastic.Thatis, agivcn proportionate price inoease willinduce a larger proportionate de.rease in Ge guanrity ol thc linn's po,luct thar @stomers rvlsh to purchase. Thc ne* denand.urve is thus represcntd as D/. Bccausc thc 6mr's marginal r!.nuc orrc stems from irs dehand cu.rc. it also shilts lcli!'ard and bccomcs nor. elastic, as Asuming that thc 6rm's costs are uaffe.ted b) 1he enlr) oI nerv lirms, the result is a decline in the iirm's c.ononic prolits to zero in the longrun.As panel (b) indictes, this occurs *hen thc iirm's d.mud.urre shfts to a poinl ofrangen.y with the l{)ngru average cost currc at pointl.At dns ponrt, the e.obonic p..fit olth{: Ii.m l:1ls to zero.lt produces a lou=r rate ofoutpul, qL, at a I(i{erprnt, PL, and thc totalrcvcnue it carDs just covers 1he oppo.tunitv cost ol b.ing part of this industry.This removes the in.entire lor an) morc tirms to.ntcr tle iD,lustrv. As nov firns enter the mo.opolisti.allv .r)mpcdtnr industq and capture some of the existing lirm's .ustomers, the demand tor its prodxct (l{rlincs and b..omes nore dastic, and i$ econonri. prolit d(,linc\ to*ard Tcro As lor a firm h a perfe.tl,v compellti\'e indust,I, th. total revenue eaned bI a firn in a monopolist.alli .odp{:titivc lndustryjust covers the opporlunilv cost of rmaini.g in th(i industry: In contrast to a perfectll compelilive industr! iirms in a monopolisricall\' compctitivc lirm arc ahle to chrge consumers a price above marginal prodult cost in long rm equilibrium . In additi on, iI a tl pical lirm's long rm av erage cost .ur!e is u shaped, as in pan.l (b), tlc prolit maximiang price ofits produd also ex.eeds 1he lirm's lo*e( Icasibl. long ru derage cosl a1 the minimum point olth. long- run ave,lgc .Dst curvc. Manopalistic competition and the electronic marketplace: twa perspectives A numbcr ofobservers contend that manl e comme..e ihdusties tehd 10 derelop monopolistjcaliJ .ompetitive sbu.tues. A.cordinS to this view, the costs ofestablishing 67 EL E CT RO N IC CO IVI IVl E RC E _ IM A R I( ETS A N D PRICES m r 'commerce firm are relatirel! l(,x in manr indNtries.l'or me! pot.ntial .ntrants wtl, srr\cr,dorlopingaweb site,sd mark.tinga n.$ or cxisting product could pro\. t() b. rclati(lrinexpensile.At the $mr tim..lnl$nct s.lling olTers a nMbd of \!a!s k) dilTcrentiate produds.Th$. is th. tic(l'.n(l rrue approach of sellinS a "nc\ impror.d" product $ irh diflerent features Inim thosc ofered by other competit{)rs. In a,lditn,n, an Internet seller o strirc to dcsign rn inriguing home page and a distinctir. shopping Da\igation proL.ss. It also.an adr.rtisc dd ser uprmgements fordoss-links withWeb pa$s .l Intcrnct scllcrs in other induslri.s. As you have lcmn,:n, c<nmm$ce firms silhin mo.opolistically coftpctiri!e to such indusrics, establishing a indush-ies are likelJ to fail to prorlucc at the lou,est feasible av.ragc .ost. ThcJ arc also unliL.ly to equalize the priccs ol lhcir produ(1s sith tLe rarginal cosl ofpro(luciog a.hn\! minimum a!.rage cost dd marSinal cost pri.ing in monopolisticall) competitive indu(rfus a( nxial ine{nciencies. Compar((l $ith a pcrlc.th .ompetitile lirm, a monopolistnalll competitire firm is all(,cati!.lv in.lh.icnt. because it does not enSage in m.rginal(ost pricin& It is also techni.all! in.lli.n.nt, bccaus. in the longrm it does not produc at theminimum teasible areragc cost.lt is uclear \rhat miSht bc donc to diminatc su.h inemcien.jes of monopolisti. .ompctition, short of go\ernment int(r\.oln).s r.quiring that all web p.ges look the same and that all I.tcrnct shopping pro.csses thosegoodsorscrriccs. From on. pospe.rile. There is d Lhe liilure to alternatjve perspcclive. howerel A.cording t(i this rn\, monopolistic compctition is not rcally inclli.i.nl. Long run equilibrium under monopolisti. .ompetition, after all,,rickls zcro cconomic profits to all firms in th. industr). Furtherhore, the pri.e ofea(h pro{luct reilects customers' rillingncss to pay lbr th. unit that thry pu.hased. Afte. all, .onsumers will not buI the produ.t il thcy arc un$illing to pay the price charged b' the produte5 ad they may be $illing to pay a pricc that er(eeds marginal cost and arerage cost to .onsume a brand with cha.. a.teristi.s ther desirc. som. ..onomtts irguc that the pric. cost diff.r.ntiil thal arjses under monopolistic compttition amounts to a premium that.onsumcrs pa! in cxch gc for thc product rarn t! th.t cxists in monopoLsticallr compctitn c indusui.s. As disosscd shortlr', $illingncss to pa! is a rerl importet concept in the.codohi.s ol'clc.tronic commcrc.. 7. What are altetnative loms of comp€tition amoug evenly natched hut imperfectly competitiw e-comnerce ltmsTore form of impedect competltior is oligopoly, which can involve a f€w fkms that are relative yevenly malched. This isar rdunrynrlcture inwhichaveryfewrivaksellthe bolk of indlstryoutprt.ln ano isopolisi c indusiry, firmy priclns and produ.tio. decisions are iiterdeperdeft, mearins that the price or production decislons lhat on€ firm makes can affectthe d€cisions at anoth€r f rm. Another form of imperlect compelitlon is monopolistic comp€tition, in which enlry ba(ie6 are low and there are a number ol firms in the I'rdurtry. Firnrs produce similar yet distlnguishable tems, however, so each film determinet the pr ce of its producttaking intoaccountthe lact thatlhe demandfor its own produ.l depends on 68 IMPERFECT COIT,lPET]TION AND N ETWORK ]NDUSTRIES the prices charqed by other firms and on lhe number of Iirms in th€ industry. In the long run, the pricecharged bya monopolistically competitive firmerceeds marsinal cost. Althoush each firm earns zero economlc profit ln the lonq run, each firm fails to produce atthe mirimum feasible averase totalcost. IHE ECOI{OMICS OF VIRTUAL PRODUCTS olc commetu. products ar. virtual products, whicL arc it.hs ofl.r.{l (ligital lor sale in form. Many virtual product are availablt via computtr dorvnloads. Exmplcs include compuk roperatinSs)stcms, of6(e p.oductirit\ sft$are, ele.tonic bools fr lnoks) md cnc).lopedias, (li8it!l music and rirleos, and e<lucational t,l A numb.r Cost charact€dstics oI rirtual products What makes vlrtual prorlucts different from a numb.r (n other good\ an{l scr\i..s ,s that pro(ludng the nrst c(,py often entiils in.urringa si,atle up front.ost. Once th. first copv is produced, ho{.v.., making idditioral.opn s;s ncarlr costlcss. After all, rheputlish.rcansinrphmJccopi.sof.pproprirtclr{ormancdcoproltheauthor\ original digital file arailahlc lbr consum$s to dos nk)ad, at a pricc. \ ii rht lnrerner. Costs of producinq viftual products tht sdl.r ola rirtual product, considcr theproductunmd sal. olan e book.An c.book\ author mustdelote many hoursof labor to $r11ing, organiling, md editing its co.tcnt. Each hour ot labor dtvoted to performing this task cntails m opportunitr cost. Thus. to indu.e someode to r.it. an e book. an c'book publisher tlpicallv must par an author an uP front Parh€nt, called an arl,oce becausc it is norhill) dcducted from ro\alt! Parm.nts that luturc 'Io think about the cost ronditions faced br willgcnerate.To induce an cstablishcd author or a Particularh Promising D.rL althor to write an e bool(, thc publish.r may have to pay an additn,nal lumP-slmP mout, olicn in the form oi a granr. Thr' sum otall uP-l;ont PaYmcnts .onsttutes a sales relarircl! sizable frcl.oi that thc publisher must incur t{) generatc th. first copr of On.. th(: author Las nritten the c book and thc Publisher has transferred its .ontents t() an appropriatelv formatted digital iil. that is r..dable by sPccidl .omPut.r softrvare or e-book rca<line deLi.es, rh. .ost of m.king and distriblting additional copies is \er) lo\. Rc.all from Chaptcr 2 that the cost that a Produ.o incurs in producingdd selling an additional uir oloutputis thc tirn's nd4ri,dl .o!. For aq?ical virtual product, marginal .ost is often cxtremdy k,w. ln th..ase oian e book, it i3 simply . matter of incurring a miniscul. cost to post a do*nloadablc l'ile on thc publishrr'sweb site.nd paug the author a rovalt! a Pavment lor cach book thal consumos purchae. 69 ELECTRONIC COMMERCE_IllARKETS AND PRICES Cost cutves fot a virtual product Suppos that a publishcr .onrinces a srll knotrn author of scic.cc li{:rion norrls to s ritc a ne$ dovcl rhat it vill .clease in e book form. The publisher agrccs to make a $45,OOO upJi.ont pqment to thc author in exchangc for a "digital manuscjpt" ofa ne{ i,icncc fictn). norel.Then thc publisher in.u.s a ffacd d)st ol s2,000 to pa) copy cditors md .ompanr emplovecs to clean up the *ritinS a littlc and ro format tle manusLript br Inrernet do$nbading The publisher hd also d.termined thar it hcrs Iixed opportunitr.ost of S 1,000 for aranging a deal wnh rhis sciencc fictn'n auilor instead ofthe n.xrbcsr author it might co.sido inducing ro write an e book. Hcnce, th. total fixed .ost, inclusiv. of opporrmiry .osts, for the e book produ.l is g 50,000. a Akhoughrovalil palments to authors normallr are a percentage of a mcasurc ottoral dollar sales, letls suppo* tlnr io this cse, th. author md publisher agree to a 6xed paymenr of$0.50 per e book that the publishcr sells on tle Intcrn.r. Conscquentt); th. marginil cost that the publishcr i.curs n)r ca.h copl olthe e-hook tLat .onsumcrs do$nload is a(on(ot mounr equal to s0.50 pcrc.book Figue 3. I displNs tlr publisher's .ost .u!es for this \irtual produ.t. By dc6nition, avcrage fixed cost is total fixed cost di\ided br th. qumtiry produ.ed md sold. Hence, the a\erage lixed cost of thc lirst c book is S50,000. But iI the publisher sells 5,000 copies, the alerage 6xed cost drops to $ 10.00 pcr c,book. Il the total qudtitv sold is 50,00O, average lixed cost de.lines to $ l.00pd c.book. Hcncc, thc arerage iixed cost cuve slopcs dos n\ard ov.r the entir. rrngc ofpossibh quetities otc-books. (/fq Avcrage variable cost equals rotal vdiable.ost divided bv thc hrmb€r otunirs ofa product that i lirm sells.lfthis fublishcr sclls onll one.opy, rhen the cost it incurs is the rolall\ parnent $e author so rhc ayerage variablc (ost of thc 6rst copr sld is s0.50. Il.onsumers dounknd s,ooo copics oftht e-book from the publisher's Web site ea.h year, rhen total rolaltv pryTcnts to thc author.qual t0.50 mu[iplied s 0. 50 1(, by 5,000 copies, or g2,500.The publisher's average variable cost ofsclting this .luantity ofc-books thcn cqoals its total lariable cost ol $2,500 divided by the 5,000 copies it produces md s.lls, or s0.50 po c-book. Thus, .s sholr in Figlrc t.t, rhc akrage rariable .ost ol producing ed selling this e.book is alwars equal to the constant ma.ginal cost oi so. s0 pcr . hook that thr publish.r incurs. The average va.iahte cost curve is thr: sane the marsinal cost (.,l.IC) Luvc, B hn,h lbr this publisher is the horizontal tinr dcpicted in Figure 3.3. (,.1vC) as Bv .lefinition. arerage btal (xt cquak th( sum of av$age lixcd cost ad average variable cost.The alerage total cost (t7(t) @rve lor rhis e book publisher thcr.for., slopcs downward ovcr its enti.e rang.. Redll Fom Chapter 2 thar alonS the downtr ard sloping raoge of a firm's tolg rh av.ragc cost ork, thc lirn cxp$ien..s eL.nomies of $le. I-or &e producer ofa lirtual produ.r such 6 a.. book, tle short run averagc total cost orvc slopes downsard. Cons.qu.ntl), sellers of irtual produds typiolly experience sho.r ru economies ofs.ale.The average toral cosr of producing and stlling a rirtualprodud declines as horc units ofth€ product &e sold. Short rm economies of scalc are a distinguishing chara.terbtic of virtual prr u..ts &at *ts rhem aparr from host other 70 I[4PERFECT COIMPETITION AND NETWORK INDUSTRIES Un t (5) 10.50 I' 1 5,OOO IO,OOO 2O,OOO 3O,OOO 4O,OOO 5O,OOO Quanttv If the total fixed cost af publishins an e-boak is $5a,0a0, then the cost of producinq the firct e boak copy is $5a,a00, but if the publisher selh 5,aa0 copiet the avetase fixed cost falh to $la.AA per e baok, and if sales rise to 50,aaa, the publisher'saverasefixed cost deyeases to $1.aa pere-baok. Thus, the pub ,,t::: Fisute 3.3 Con curves fat a producer of a virtual product- lisher's average fixed cast Gr-C) curve slapes downward. If the perunil cast of producins each e-book is $0.5a, then bath the narqinal cost (MC) and averase variable cost (AVC) cu(ves are hatizontal at $a-50 pere-baakAd.linq the AFC anrl ATC curves yields the publisher's shoh-run averaqe total cast (ATC) curve- Because the ATC curve slapes clown||ar.l, the publishet experiences short-ron econanies af scale. Impertect competition and yirtual products ln the examplc dcpicnd in Figlrc ] the rtual produd is a scic.cc Ection novcl distributed in e L,ook form ria Intc.nct downl{rads. Thdc arc numcrous scicnc€ li.tion audorsddmany sci.nc liction nolch among rhich mnsumcrs.an choo!:. Hcn..t th.re arc man)- products that are close substitutes in the market lbr scicncc iiction norels.Yet, no science fictioDnovel jsexa.tly the same.This meas that the parnolar e book produ.l nnd by the publisher in our example is ,listinguishable from other 3, .ompeting produ.ts. For the sake of argunent, therelore, let's suppose that this publisher fac.s a monopolistically .ompetitive market for this e book. Panels (a) and (b) of ligurc 3.4 display a posible demad cur ve tor lhe e book * rittd bl this partiolar science fiction 7t ELECTRONIC CO|llIVIERCE I\IARI(ETS AND PRICES :10 40 50 Quaiity Ithousand.) =a (a) 10 20 ll] =! 4a 50 Q!.nt ry (lhousands) (lr) .a::, Figurc3.4 Marginal versus arerase castpricinq af vi ual praducts. ln panel (a), if the publisher with the aveftge lotal cost antl narginal cost cutres shown in Fiqurc 3,3 sels the price af the science fictian e book equalto its canstant naryinat cast of$a-50 per e-boak, then consumerc willdawnload 2O,AAO copies. This yields $1a,000 in rcvenues ta the publisher. The publishe/s ayeraqe total cast of 20/0aa e books is $2_50 per e-boak, so its rord/cosr al sellinq that nunber af capies is $3.0a tines 2a,aao = $6a,aAO. Ihe fnst best sacial optinum af narginal casl pticinq yields a $5O,AAO lass ta the publisher, which is the tatal fi\et1 con of producins the e boak. A secondben socialoptinum exisls at the pointwhere the demand cDtee is tanqent Ia the average total cast cutvel sa that the firm practices aretage cost pricing. At a price of$5.50 per e boak that is equalta avetase totalcosL the publishe* revenues of $55,00a just caver its $55,a00 btal cost of praducinq and dlstributing the boak anthe Internet- llarginal cost pricinq and vittual praducts in Chapter 2, (rrinomists often regard narrrndl.oll pri.rD, as a dcsirable out.ome. Marginalcost prtingimplies allocatire em.ien.JrThc prnrc tlat the con sumer pays is equal to thc .ost in.urred to p.oduce the last unit of output sold, so that the seller: explicit &d impli.it opportunit\- costs ofprodu.ing the product arc just corcred by the consumer's payhrnt. Bccause marginal cost pricing is all()cativch ellicient, e.onomi(s $mctincs call marginal .ost pri.ing th. ffrst-best optimum As discussed iere to behave as St it rvercapolcctlv .omP.titircErmbye.gaginginmdginal.ostprn,ing?Panel(a)olFigureJ.4prx.j the eswer ro thn question . If the pul,lisher sets thc price ofthe science li.rion . book equal to marginal cost, tlen it vill charge onlr s0.50 pcr e hoolr it sells. Naturalll,, a nunber ol p{-plc desjre to purchase e books at rhis price, md given the demand curvc What if the publisher ofthis parli.ula c-bool in thc ligure, the pul,lisher could scll 20,000 cop]es of t]ris e book. The publisher would fa.e a problcn, hos!!€r At a pri.e of $0. s0 pc. c book, it *1ruld earn g I 0,000 in relenues on salcs ol 20,000 copies.'l he averag. 6xcd cost of 20,000 copics.quals $50,000 72 / 20,000, or s2 50 pcr e book. Adding this t(, rhc IIMPERFECT COI\,lPETITION AND NETWORK INDUSTRIES constdt s0.s0 alcragc lariabl. cost implics an avcragc total .ost ol sdling 20,000 copn:s of $ 3.00. Undcr marginal.ost pricing, th.rclorc, thc publisho woul(lca.n a. alerage lo$ of s2.50 (pri.e average lotil cosr = $0.50 51.00 = $2.50) p.r c book lor all 20,000.opi.s sold.Tbc publishcr's total ..{)n{,ni( loss liom scllinS 20,000 c-books ataplice cqual to marginal.ost would amountto s50.000. Hcncc, t}re publisher wodd {ul ro reoup the full anomt ofiixed cosr ol proJucins thc c -book. It wouklnot hare cvi:r uanted t" pr"ducc Ge c'b""! in Ge lirst place ifit plmncd to set its pricc.qual to th. e book'smarSinal produ(:thn.ostl Thc failurc of nargiDal cost prn:ing t,) all,w lirmssclliog virtualp(xlucrs to(orer the lixed costs ol producing rhose products isinrrinsicto rhtnarureofsuch products. ln the presen.e of shorl rD economics ol sLnle in p.oducing rirtual prdlucts, marginal .ost pricing is no1 a frasible outcome in the marketpla(. Recall that marginal cost pricingis ass@iatc(l {ith Frlictcompctifi'n.An imp(rt&t impliG )nisthatmarkets for rirtual products cannot fun.tion as p.rL.tl) .omp.titi!c mark.ts- Imp.rli.t competition isthe rule, not the ex.eption, iD th. markctplacc lirr a rirtualproduct. Average cost pricing and virtual products ol Figure l.a illustates n alternati\e pri.ing.honc hr the publish$ ol thc science fiction e book.The publisher sets tle price oith...t ook cqurl to arcrag. 'ts cost of production, thereb! eDgaging in ave.agc cost pricing. Ciren th. dchind Panel (b) orFdepictcdinFiguel.4,arapriceots;.S0pere,book,consumcrsarcrrillingto puchae md dornload 10,000 copies lhe publishrr's arcragc t<,tal cost ofolTering 10,000 copies ii). salc is also cqual to s 5.50 p$. book. H.ncc, this is rh. pri.c rith arc..ge cost pricing At a pri.c ot s 5 50 pcr c book, thc publishcr'r rorcnu.s f'rom s.lliDg I 0,000 (opi.s equal t 55.000.This anomt ofrcunucs is just sulliLi.nt to cor er chc publisher's total Iixed cost ofs50,000 dd rhe s;,00o bral yariable.osr ir incu.s io producing 10,00O copies at d ave.agc variablc (ost ol i 0. 50 pcr ( booL. Thus, th. publish.r .arns rcro ccononic profit br engaginS in alrrage cost pricin& Economists often regard aEage .ost pri.ing as a second-best optimum fo. society. The lirst best optimum oldarginal cost pricing ol yirrual p.oducts is.ot attainable, because sellers ol such products loow that marginal cost pricing rrould entail negatire e.onohic profts. udder arerage .ost pri.ing, s.ll.rs ol virtual produ.ts .harg. tht mininum p.i.c rcquirc{t r{i covrr rhrir prcducrn}n cosrs, i.(lu(ling rhc rclativclvhigh lixod costs thcJ must in.ur k,{lcv.l(,p thcir products in th.lirst pla(e. Consumcrs th.rcbv pav thc l,nvcst pric. nrccssarl to ioducc s.llcrs 1() prcvi(le the consistcnt l. )ou lcarncd .arltur i. this chaptcr, thc situatD. illust.ated in panel (b) conesponds to a iong run monopolistically.ompctitivc cquilibrium in the market lbr s.ience liction e bools. If this and other publishers la.c a situatbn su.h as th. diagrm depi.ts, then th.re is no incentive for additional publish$s to entcr or lcavc la.t, as the marketplace. Consequendy, in the long run amonopoli$n:all) ()mp.titiv. industr) composed ol sellers ol virtual producrr fa<ing short run cconomics ol sctle, thc second- bcst optimum will naturally tcnd to cmcrgc. lo other impcrlcctly competitivc 73 ELECTRON IC COI\1 IV] E RCE _ I\IARI(ETS indus[n]s, s!.h as AN D PRICES olilopoln.s {hcft.harrnrs 1.) (ntrr ir. {il$ih.int, this is l.ss likelv 2, How do lhe spechn produclion<osl charactetistics ol'tirtual producl influence how yiced? Becaull's.llers provlde !irlualproducts disitalforms, the nrarqinal cost of producins aid distributiig such items is very row. Flrms typica ly incu/ they ate in relativelysls.illcantfixedcostsinprovidingvirt!alproducis,however.Corsequently, producers of virtual product9 face a downwa/d4lopin9 averaqe tola cost curle, so thatthey experience short-run economies o{ sca e, In this s luation, attaiiing a firclbestoptimum of marglnalcost pricing lails, becausesettifg price equarlo marqinal cost falls lo cov€r the fix€d cost of producinq a vi/rual product. The averaq€ cost pricinq equilibr um thal aris€s in the lonq rur u.d€r monopolistic compelilion might be reqard€d as a second-besl optimum, because s€ttln9 pric€ equa to average total cost siles lirms the incentive to provide virtual prodlcts but ylelds zero economic NETWORK EXTERNALITIES diso$ed in Chapter 2,6rms in.(tain markets comprise,.r*ort indunr,.r ln such d indBtry, thc laluc dlat consuhers assign to the goo<t or scrvicc pulucd bv firns depends to somc cxtent on the .umber of ot-her peoplc sho consumc the prcduct. Economists sy that su.h pftxh.ts ar. subject ro network €xternalities, meoirg As that the benefit that a consumer anticipatcs r...iring froh a good or on the number ofconsumers usingthc itcm.An cxample of such a s.r!ic! dcpeDdr good is a $ft*arc program for sharing a parti@lar twe ofdigital 6le. If 1-ou wcrc thc only onc who Nncd the software, then the prcgram would halc little value to vou, except perhaps ifit also performs some other usetul tunttion. Alto all, vou would be umble, on thc one hand, to transmit the specific type ofdigital Iilcs to shar€ with other individuls or, on thc other hand, to receive these digital iilcs from an),one ehe. As hor. pcoplc with thom vou wish to share the particular typc ofdigital liles acquir. thc soltware p,(,girm Ge benefir \ou derirc fr^m util,T,ngrr incr.a..s Product demand in the ahsence 0I network ellernalities ho\ r\tltrrlnn.s ( ,nllL,.n.r rhr (l($in(| 1;r i fro(tu.t, it isl,ehlulio cxpl(n.a siht)lr cramplc libl{. l.I(li\plr\sih.nr.rtn. nouDlsrharli\r consun.rs $r wllling to l,av lnr r slngh ,,1,r ,)l (l,Nnl,,i,lihle (,lli.c pro(lu.ri\ir\ solt$ir. pickagc. lritiallr .r lun(l n.nlrl(hIi(ronrn,,lrhis{,li\a,c,srh.rtcopn. nho pur.h.s. .url rr. ,t (in ()nl! \1,rk $irh (,)D,tut.r lll,\ l,(.ir.(l ,{r rhci, o{n pcrsonal .oDr|ur.rs lh.! rinn,n \h.l r thr lil.s $ irh,nl,.rs Suppose d,at thc lirc r,rurDtr! s itr l.rl)l(. 1.1.,1(rnn.,l.1, r, (1,,r), rn{lt, ar. th. onlr (,nsunr.rsjnth.nr.trk.t.ltrserlr th.irrIrn,.rrirrrrrl;l)hl.l,irispi)s\ibl.rotil)u1atc Ti) 74 und.rstand n.1s1irL ] 8 I\,1 P ERFECT CONl PETITION AN D N ETWORK ]N DUSTRI ES tarre 3.1 Willinsness ta pay fat a softwa re product without netwotk externalities I Consumer A wiliingness ro pay ConsLrmer A willlnqness to pay Consumer Cw llinsness to pay Co.slmer Owi linsness to pay Consumer E wil inqness ro pay $50 50 60 7A a0 quallilies olthe soliwarc pa.kage demmded at varn)us priccs.Thesc pria quantitl combinations are displa\c(l inTable 3.2. IIthc product pri.! is sso per uit, rhc. ajl li( (,nshss &e iilling r) purchasr the product, l)ccaus. ea.h is willing to pa) at lcasr lhis ahout for rhe soli{arc Il the price is s5o pcr prrxtuct. r-hen onsumcs ,1an(lB are unwilling to bu! n, so rh. qua.tnl demndedinthc markctplace declines units.ll'the pricc risesto $70 per unir,thcn neith$consumers,t,B, nor coDsuher the producr, and the rotal quantitrdenunded ta s ro 2 uhiB Su(crssire pricc inoeases bring about dcclincs in the number .,t.onsumcrs who arc willing ro bur the so|l$are. At or abolc a pri.e ot s90 per un, no ont r,, p".cha.e"the "ishcs produ.t, so qudtitv demanded equals aro 'Ihe pri.e-quantnl combinarions for the solis,arc market gileD h hble 3.2 appear to 3 cNillpucha'c in Figure l.s th! med s(hedute D, Be.ause sofrlqre p& k.gs e" ",lda.s;;gle, indivisihleunits, this dcmand s.hedul(, isnotasmooth cure.Thc1a.. "i,1.,-D,lh,}-l(l,. noncthclcss: As the price ol 1he sofhrare pa.kage in.rcucs, rhc quantitr dtmande<l {tcclincs. Hucc, thc (lemdd relationshjp DL slopes do*nward. as d( Product demand in the preience of a network externalit, \o\ sufpo{r that pr,)plc.an usc th. sann. ,nH.c pr uclnir\ soli$Je pact.ge ro sha(. lilcs rirh orh.rs, su h s spolscs an,t otho l:hih mrmlre.s, liien([:dd co $,rkos As.r.!h, mm\,,ith. apptidtnr:s.onraincdinrhj\soli$ire pa.lL.qe til. {ha.ing progran,s, Nord |.ocesihg softr\ ir., an,trhc rir. c nrtrr liluil;fu k, och comunrm rvhur oher consuners purLhase thc sottwift.pa.kagr Ea(I (,n rn,tr's rrll;rgnes r. parl dk refore, nr.rea*s as mo.. indi, i,r.;k ptr; has( rn,r ,*. the !;trErc. ($re ssme, br rhe r..!, rhat th. sottrvaR.N suth.x.ntlr (,)pv pnncctdl that il cannor bc pirar.d or eir .n arla\ to oth. r users, \ hi.h m,N ttur ati .;n\rmos nust pr! lh. pur.ha{.prt. ri, use n ) ; l.bte 3.2 Purchases af saftwarc product at variaus possible prices withaut netwotk $50 B,C,D,E Quantity a $60 C,D,E l $7a 980 $90 190 l 0 D,E 2 75 ELECTRONIC COfu1IV]ERCE IV]ARI(ETS AND PRICES 80 1o 50 30 20 10 5 I Figurc 3-5 ,lenand for softwate packases without network exte.nalities- ]f an office prcductivity software packase has no facilitv far e,changing files amons different usery then each potentiat cansumer antv prrchases an amount that he ot she desnes tu use in isolatian kon athet users' ln the exanpte illustrated in Tables3.t and 3-2, the demand relationship D t isthe The Productdemand with a network externality Yari.s s ith thc numb'r of otbcr .onsumers $ ho alv) choosc to Purchde and use thc ofiic€ Produ'-tilir.-' soft$ &e packag.. The li.st column of-l-ablc l 3 shosr the sanrc rvillingness to pav lirr ecch co,sumer that appears i" Tablt l . l From cach consumer's Point of rics, il h' or she tur.s out t, be th. only uscr of thc sottware, th.o the lact that lilc sha ng is L,bk L I {tisplals ho{ .a.h .onrumer's willingness to Pav norv possible makcs no diffcrcn.e. Each consumer will halc th' samc willingn'ss to shc rvould hare had if sharing files across computos remained p"y th"t h" ". Table 1.4 indicatcs Table 3.l, fable t{, purchasc 3.3 *hich.onsunds are \rillin8, aunitofthe ofiice Produ.tiritv sof't{arc at th' inlormation in larious Pri'es' lfthe willinsness to pay lor a software prcduct with a network extetnalitv I Consumer A Willingness to Pay Consumer AWil ingnessto Pay Coisumer C Wi insness to Pay Consumer DWillinsness to Pay Consuner E wil inqnessto Pav 76 based on 3 $50 50 70 80 $60 70 a0 90 100 $70 $80 110 $90 100 110 130 t2a 140 110 160 r50 130 170 ]IVPERFECT COMPETITION AND N ETWORK INDUSTRIES price olthe vnhrare packi{f is 950, thcn narurrllr all 1l ve .onsumers are willing t(, pu.hase thc produ.t. Ilerrl., th. tahle sho*s rhat rhc t{nal.luaDlil! demanded in th. marlerplr(. is ; uits l,rx\.ols;0 per uDrr ( in.reases to s60 per unt. \..ording (, co.sum.rs ,), ind I a(. N illinS to pav al La\r s60 per ufi.y.D itno \o\ ar a supp,'sc rhal the pri( to Trb[. ].l. oDe elsc buvs aDduses thc soliware. Thes.' rhree consumos, rhcr.lbre, pur(hisc thcproduct. thusi l.ast thrrc pcoplc can shatu tilcs among thc!..omputos. Il a li,urth starts using thr soft\Lar.,1l!.h thc $illingnrs r,) par for thar in,ln idual md ft)r !h.litth \ill in...ar. .{s sho$ n inTahle l.I, th. s illingnes to pi\ ()l .onsunro i iNr.rs.s to S8O on( ( rt least lour pc"ple purchast rh soft\! are. Ont c .onsumer . l burs thc product, s(, rhrt tburpeopl{ (an slDre liles, (onrumera's willn,gness ro p:\ rs r tiith user nro.iscs ro $ I10. so rhis (onsum.r al*) i\ willingto pu..hin. th{: soft},arc t)a.kag{: at a s60 t)rrr.. Thc total quintitr deman.Lll it the higher p.(\, oi s60 per unit, therefore, .emaih equal t() ; unirs, rs indnlnrl inTablc 1..t. I pri.. inc.ase to S70 prr unit $ould i.(lu(c odlv.onsum.rs D rnd E k) bu\ rh( .olrsd, tu.ld8"ruu.. ,Lri.r.rJll) {'n,,.r\,,., r'.,.f,..h,...,r..,,.,",*..tr..,,,. tregnr .xchanSing iil(.s, an,l rhr nurnber ofuscrs rises to two. r'h. {iltingnes ri) p.! f;r connrhrr (, as a third unr, thrn in.ft.ans 1() Sl00_ \r this .onsumer s.ill iln) pur.hasr tlr soli\rar. packrlc Jr aprice ols70 CoDsumer, s.utd then L,tcom. rhe Iourth us.r il she bu\ s rbe l)ro( lud, ed il th.r( .tu l;ur us.r\ TJbt. ]. I in(li.at.\ rhat this consuh.r's $illing.es\ r{, piv is s I 10, nJ shc is rvilling t., l,uv the producr ar rhc t70pric..li1rD consum$.1's pcrspedile, th.r.ti)rc, ,oini.g fi. {,lhrr tour and making the total nuDrho ofus.rs (rlLrrlt() filc rais.s his \illingness r,, pa\ to s90, y, h. al!, issillnrgt(,l)urtheproducrrr i pricc ols70pd unrt.Thu5, thr rinatguantir\ ol uDn5 dem&dcd ar r prtuc ol s70 is itso equal ro ;. Lr prolJleh 6 al rh( tnd oith. .hrpr(r Lu g.t th.,)pporrunitr to r.ain oul thar allli\r consumcrs !t",,,vish to pur.has( rh. product it r prke oI $80 per uDil No\r thnrk itrour {h.1 hJppcns at a pfur.ol s90 per unir Assumirg ,at rlL consumrr, hr.r no $ar to ronrnrunicate ro oD( inolhcr thc t;cr rh.r.i.h consunr.r.s silhrgdesr 1,, par incrca\r\ rs nrore indi\iduals use rhe solr*rrr packagc, thc tirst ,,.lrn,n ,l l,'hl. {., i,,J,.J,,.,r,,,.r.un,,ri.s,.t,ner. l ,tr., .*p,..*., r.., result, tdal (luintitv dcman(l(\l d.clines to 7cn), !s sho\n n lirhlc J.4. lndo(l, thr z.n)it orabovr a $e0 pricc (inproLtrh 7 atth..nd olrh. chapter, rou gcr 1he opporlurn\ r(r .ontcmpl (. \ hafihe deDrr.(I r.tationship mishr l(xik 1iI( il .ll (onsumer\ c\(hiDge i.lir.nurnh ibout $iltinlnrss to pi! an,l t,, quantitv (1. nr purchast th. S Tahh Quaniitr andcd equils vniln prhgr.l 3.4 Purchases of software product at various po$ibte ptices with a network extemality $50 $60 $70 $80 A,B,C,D,E A,B,C,D,E A,B,C,D,E A,B,C,D,E 5 5 $9A-19A 0 77 ELECTRONIC COI\,1[,1ERCE N]ARKETS AND PRICES Implications of network externalities for product demand Figrrc 3.6 displars both the dt'mand rclationsh;p D, tha.rc obtained in tbe casc rthcrc thcrc brre no ncr$brk rxternalitics md the demud rclationship implied b) l ablc 1 .4 wh.. a nctsork externality is pres.nt, dcnoted Dl Wh.n co.sumers (an use the software package to transfer and shte 6lcs, thcn all consum.rs de not silling to purcLse r-he pr.lu, r ar mt fnc. belos sco p(r un,r An immcdiatc implication is that adding a ffle exchangiog capabilit\ ro this partiolar softu are package induccs a rights ard shift in tht dcmand rclatlonship. from D, to Dr.tevery pri.c below s90p.r unit.Thar is, the prcscn( oia n.tworl cxl..nality lcadstoe i,.rcas in the demand lor the product. This is a tlpical chra.-teristi( ofproducts \yith net\\ork cit(rnalincs- B.causc each c'onsumer\ rvillingness to pa1 rises as othcr consumers buv ard use such a produa, the orcnll dcmand ior tLc product tends to dillcr from the ltv.l it $ould havc cxhibited ilno netFork cxternality existed. ln this Partkular cxamplc, there is a "bandwagon cllect":The prcsencc ofthe ncrutrk cxtcmalitv inducesmore people toconsumethc product at nost pri.es \rtere Positile consum Ptbn occurs. Another impli.atio. ol Figur. 1.6 is thar rhe Price elasticity of demand $,hi.h economists define as the absolute valu.,,lthe ProPortn)Dak changc inquantity ola product dcmandedresulting tom aproPortionatc chang. in the Product's Pri.e is afi.cted hr th. presence ofa n.t\rorkexternalir!.Alongthe down$dd sloPing.dge -l 80 50 l0 20 10 723456 al 78 Figure 3,6 fhe denand lot software Wkages without and with network extemalities. ll the capability to share af{ice Uodudivity software files is added as a feature of the packase, the^ consumers' willinsness to pav fo. the softwate packase increases when othes also use it. Based an the exanple sunnarized in labtes 3.3 and 3.4, this qires the demand relationthip D,. I n the prcsence of this netwark ertenality, the denand fot the uoduct increases and beromes relalively morc inelaslic, as conparcd with the o.iqinal denand cuNe D ) in which there was no network externalitv. ]I\IPERFECT COMPETITION AND NETWORK 1N DUSTR IES ofrh. demand relatnrnship D, that applics in thc absen.e otlilc sharing.:pabilities, okr th. rdgc ol pr((s liom s i0 to s90, d iooease in the product prnr (luses the quantitl (lcmandcd t,) de.linr.This . neriork externalnv is not true $hen 6le ex.hangc is possibl{,, so that is present. As thc producr pri.e rises orer this rangt, ofpric.s Dl rhe quantitr (l.mn dcd docs not chage, and the prnr clasticit\ olilcmrnd is 7tro. In this situarion, therefore, thc presence ot a nrr$ork ext.rnal,tr mak.s (lchind perJedl) ti.lostE. along the demand rclationship Hen.., this cxampl. illustrat.s another common charad.risri. ofp(iducts with nctwork cxtcrnalirn.s. The pri.e clasticitv of(l.mand tor su.h products usually is altcrctl uhen a nctuork r:xternalirr is present. In this parti.ulir cxmplc, ihc pricc eldticir! ol d.mmd i\ r.larirch l.^v.r, a compar.(t $ith thc pri(.tasricit\ olik.mmd Ibr thc prcduct ilthen.*rre no ncr{ork externaliries. As you will learn in th. next chapfur, tLe pri.e elasticitv of dthand can havc an importar bcaringon a produ.er,s de.ision about the prolit maximi,iDg pri.. to chrrSc Ior its pi ucr. 3, How fu rctwotk extemalities aftect the ctefiand for a product ? I n th€ pres€nce of externalily, thebenefit rhar an indivtduatreceives fromconsuminq a aood orsedice increaseswhen additionat indi!iduats ahoconsume the irem_ Each penon/s willinsn€ss to pay for the item, therefore, rises as the number of orher consume^ of th€ item increases. As a cons€quence, rhe posit on of rhe demand for the item h a nelwork affected bylhe presence ofanetwort exlernatity, as istheprice elasticily ofdemand TIETWORK IXDUSTRIES A]{D ELECTRO TC COlliiIERCE we halc alreadv discuss(l ,.r,".t ,nJ6,rier. Thes., arr induftn.s in $.hich thc !atu,that consumosplacc on thc producrs sold bvfirms in thosc indusbjcr lariesdirecth- s.ith nuhh.r ot , oh.um, r .. l'hdrN.n,r$.rt,\r,.,nJli,ie.drFI,rcscnr.\u,.un\umer! \rillioSncss to pa\ in(reases $ith rhc cxrcnr tt, rrhich oth$ consumers usc tirms, pro{tu.ts. As {! shrll dis.uss shorrly, maq cconomists aho suggcst thar ncts.ork induslrns tcnd ro shar. other .hara.rristics thar ilso tcnd to r(tatc to th. n ure of r,rr, l{etwork externalities and market leedback Thc bandsagon.li.r in the exa.rtne dis.ussd abore is an (xample ol market feedback that t\?i(allr.xists in a ncxrork indusr.\. This is a l<,grolling et)ect, s trtch .m bc (ithcr positiv. o. negatite, as a 6rm's or i.du(rv's produd eider.!nt!hes on" or b..ohcs a"has becu" wirh co.sumers rvhos. wiUingncss t() pav mrics with thc nunbf oiother usrrs ofthe produ.l ofthe lirn orthc i.dustN 79 ELECTRONIC COI/][T]ERCE-MARKETS AND P RIC ES Positive feedback in network industries ln a situation su.h s thc band\?gon .lTe.t i.rolring groter purcha{ s of a solt$are product rLith ffle sharinS .:apabiliti.s $at is illushrt.d in Figur( 1.5, a ncbrork cxtcrnalitv indues po'rrd markct fccdback. ln thc s} cifi. examp l( s e consi(lcrcd, introducing the abilitr l;r uscrs to exchmge lites induced mort ptople t,, use thc produa, thcrebl cxp:nding demand. Positi!. markctfc.dba.k c allect thc prxpccts ol an entire industrrThe mark.t for Int€rn.t (Jricc pro dcr (lSP) ser viccs is a. cxinl Plc. The growlh of this industrl hr rcughlr paralleled tht growth ol Inltrnd hosts $orldlrid. that is disPla!((l in Figur I .2 on p.ge 20. UDdoubtedh, po{itir. m!*.t lec'dback resulting from netNorh cxternalilics hade possiblc bv lnhrnct .ommmi.aln,s and inttrailn)ns indu..d a signilicant l,)grolling dlcct as morc pcople soughl ,)nlinc a($s, thereb) girinS additional p()plc th. inccntirc to obtain ac.ess 10 th( lnt.rn.t In additn)n, ho{c'cr. !s i.dnidual lirm cu reap the beneits ol mtkct fculback arisinS from $e prestncr,'l a nct\(rrk .xternalitr. A good examPl. ol aiirm th.r has benefited In)m positivc markct fccdback is the C2C auclioh lirn.lla\'. An indivnlual is morc likc'ly to use the servi.es ofan auctbn sit( il thcrc is a signilicant likelihood that mD\ other potential burers or scllors also tradc items at that sne. Hen.e there is a nenrorL erlernalitt in tht C2C audn,n industrr. and .B.iv hd .xPerienced Positne mdket fecdback as a resull. (To learn how a.umbcr olAsian comPanies arePro6rin8 &om offering to let people play multiplaycr rolc plaving onlinc gam.s at no charge. Genedt\ A.renuesJot Onln see the box Glob,/izdron o nhle: Frce PlnJ { nhv htdl lok ^ GLOBALIZATION ONLINE: FREE PLAY WITH VIRTUAL TOKEI{S GETIERATES REVENUES TOR Ot{LIt{E GAI{ING FIRMS Ga.nins firms in Chlna and severa other Asian nations hav€ sp€cialized ln developing role-playing games that people can register to plav at noexplicit charge. The sames are desisned to be more fu. to play as more and more people join in, which makes lhe games more valuab e as addltlonal people become lnvolved n v iltual as virtual ol tokens, tuch interactions is also erhanced, however, via the use online role olayinq. T he satlsfactlon that qame partic ipants experience from weapons, that players can employ in rhe games'vi.tual environmenls. To obtain tokens, however, p ayers must provide money paymerts via online transfersto lhe gamlng firms- Thus,lh€ companiesseek to qenerate revenuesby inducins more and they can play ior free, which iioeasesihe vaLoe that p1aye6 assign to virtual tokens they caa ulilize in the qames. This boosls the more peop e toloin into gameethat prices that players are willing to pay for the tokers, which in turn lrcreases the profiisthatthe{kmsreapfromprovidirstheorinegameenvlronmenls"forfree.'/ I 80 IIV]PERFECT COIMPETITION AND NETWORK ]NDUSTRIES citical analysis: Whr do you suBosethat online ganins conpanies seek to devetop playing rc|e sames that are norc enjoyable to ptay as nore and more Wople join into For Negative feedback in netwotk industries gescnce ofnctut,rk cxternaliries is not al*ays ds@iated {ith inocases in demod for the product ola firm or ao industry Th.re .an also be ncgarire marktt fcedback: People can reduc. th.i. .o.sumptn)n ol a product with a negatile cxternatiq, thercby iDducing oth.rs to redu.e thcir consumprion as ErU Ttrc An examplc ol an industrr thar has rxperiehc.d negatire mdkct teedbacks of late is the tclecommLrnt,arions industrr Traditional t.lecomnuni.atnrns lirms such as AT&T, MCI, Sprint, and the like expericnccdpositilc market fecdback during rhc lare 1980s mdearly 1990s u wirelcss phones and Iax machincs proliferated andindividuats and lirms began haking long distoce phonc olis Fom cclt phones or yia hx machincs. Sinr tLe mnl 1990s, as morc peoplc halr acquired I.ternet .ccess ,ia lo.al ISps, e mail .omnuni(ations and ( mail dooment attachments havr sL:pplantcd large volumesofphon. md lax communications. For the tclecommui.ations indusrr]., the greater use olc-mail and c mail artachm.Dts bJ soh. indiriduals induccd orhers to Iblkx logrollinS c$e(1 rlEt has reduc<l thc orall tlemand Ior traditionallong distoce phonc serlices. An individual l)rm withina nctwork industrr can also experience negative market fcedback. Manv carly C2C auctioo rjvals of€Bay fcll by ttrc waysi<te ai the posnive ha.ket fecdback thar cBal.expcrien.rd rcduced the sharc ofthe oreral dcmand suit. This has resultcd in a lor lnternct au.tion serri.es arailablc ro th.h. As more consum.rs utilized eBay's servt.cs, thc redu<rn,n in users of oth.r audn,n sit.s impljed a fall in the number ofpotential buy$s and srll€rs of any giveD ttcm oflcrcd tfi sate on those sitcs.-l his fcd a..garire logrolling eff.ct at mady ri,al sites singled out by users as "hu beens."Auction rolumes at those sites gradualh ds iDdl€d a$.a! Thm, during the economt. ilo{ nturn that lollowed thc linan.ial turmo;t ot 2008. cBar iounJ ,r'ell "rrug{lh8 r,, main,din pu!riI mrit,1 te.,tba. } r,r ir. aud,,,n...r,i,e, Lre.tuallx it embarked on a plan to boost the demdd ior its auction services bv t un in8le"s plus aughcnrinS ih ire $1th mok trx.,t pr*. pun h^e op,ron,. r On.. a coDp.ny selling an item subject ni netrork cxternahies has bcnetitcd trom positive narket f'oedba.k. a kcv task is pr.vrnling reversion to a negatire fcedback situationi scc the box ,/d, ogen.nt Online: In th. tqtst of Bnlioh, alT||eet, rwntet Se.ns to Keep thc Boodwogan Rolhnl.) 81 E LECTRON IC COI\T1 N'I E RC E _ IVIARKETS AN D P RICE S irAl{AGEMEI{T 0NLINE: IN TflE MIDST 0F BILUoI{S 0F IWEETS, TWITTER SEEKS TO KEEP THE BANDWAGOT{ ROLLING Twitter is a social neiwork ing and mkro-b ossins s€rv ce. An indlviduaican utilize the Twitter seruiceto send and read text updates of 140 characte6 or fewer, called "tweets," which are immediately displayed on the prolile pase of the author and simultaneously delivered lo other subscribers to the author's profile paqe. Bv its productexhibliing network ex&rnaliiies, naiure, thereforc, tie Twitierservice in which increased utilization ofthe service rais€s€ach user's valuation of the *rvice The resultfor Twitter was a is a positlve market feedback eltect. wilhir two vears after its creation in 2006, Twitter developed intoa potenlially po\^,erfu I tool for p€rsonal commun lcalion and oroduct mark€linq. By 2008, therewere 1.6 million registered Twitler userg. A year later, there were l2.1ml lior' In responleto the apparen! positlve market feedback effeci enjoved bv Twitter, investors offered a total of mo/e than 9150 million li finarclns to Twitter between 2oO8 and early 2009. Soon, financiers had valued lhe companv in the hundreds ol millions ol dollars, and Apple and oiher larqer firms were contemplaling purchai lnq the firm for more than $1 bil ion. By late 2009, however, some investorc were becominq concerned about th€ cornpanv's fai ure to generale revenues-a consequence of offerins the seryice free to users and free of advedisements ln an eflort to attract more people to register for the service. oiher investorsworied over siudies finding that 1O percent oi T\titler users senerate ahost 90 percent of all tweets and that only 40 perceni of reqistered users ev€r qet around to uti i?lnq the seflice-many no more than aslngle lime before 1osin9 interest. Fot ctilicat a6lliis: why is ary prcduct that exhibits netMrk extenaliti$ potentiallv subiect to an eventual revesal ol fortnet after an initial bust af positive na*et 4. How can network externalities woduce market feedback effects? tl a praduct externallty, lhen market feedback eflects can occur. A posilile market feedback eflect arises if a prodtct becomes used bv a sLrfficlen! number of possesses a network individuals to induce othe6 to purchase it. A negative marketfeedback effect can take place if a lalloff ir usage of a product by som! consume/s causes oihe6 to slop purchaslng the item. [,larket fedback effects can influence lhe performance of ar entir€ network industry or ind vidual I rms within that lnduslry. I 82 IMPERFECT CO|llPETITION AND N ETWORK INDUSTRIES Multiproduct oligopolies and conpatibility decision. in the Iace ol mtwork externalities N.twork cxtcrn.lnics expla,n rhe existen.( oi nctworl indusfties. lir]nomists argue rhat selrral oth.r factors n)ust bc tak.n into a..ount *hcn clalualing (hoices that llrms md.ohumers halr in their int(.ra.tions !ri1h lirms $nhin n.tltork i.(lu(rics. In pa.tn.uhr, netlrork .ffects influ(n.. dccisn,n! rhat firms nnk. r.gard ing compatibility (ilsol<nos. as Drercr*oarry) oipn u(rs that is, r.hether to ollcr p odu.ts tha1 hmction whcn used tog.th.r {ith complementary products ot lssues of compatibility and complementarity Ea.h tine thatAdol{ lau.hes a nol lersion ol its popularA.robal program, itsWeb td,m. -Cumpt, m.nrs Mi, r;,"t' om.ej" -k. mrr [.r rng flr, r,. .nd hot .,,\(,, tr,,. Clean!, Adobe regards Mi.rosoli Otfice as rhc dominmt product within thc ofncc productivity softwarc indusfty, and it choos.s to makc its o{.n oiEc. ptuductilitr .oli$ar. hiehl! cohprr,hlc sirh Mh ro$h uth.,. E.onomists.all products that arc oft en consumcd simuttrn.ouslr compteDents. Normall): *e c.risx)h consumcrs haking th..hoic. about wh.r products to vie* as compL'nrentary: fo. nNtan.e, bfuad ard buuer, tca and $eer.,ncr, or coffe. and cream . Adobc has made thc .hoicc l?,r consumcrs ot its product, hower€r Anlonc who purchasd and used Adobe Aoobat sill hare ao euier timr using it $ i$ Micosoft Worl and othcr Microsoft Oflicc progans, as compared *ith comp.ring ofii.e produ.tirity soliwar.. (h. posiblc indication of this decision is ErAdobe r.gards Microsoft Office .omp.rrl,lF 5o'rsdr, a, rh. srandard producr,n $" ",n.i. p,"a,n^,r. .am,industrv.This is a good or serrirc that producers otcomplentntarr prortuers regarrl as thc most {idclv unal a.d, h.ncc, most ihportant for purposcs ofasuring c;n patibilit\' with their compl.mcntary products. Failingto.njurc conpatibilit! ;ith rLc standard product,lirms such asAdobe oticn (,a!,., s1)uld guarantcc dcprcssins thc dcmaniilor$eirosn,ompkhrnra p,odu.t,. Issu$ rclating to.ompatibility. complemenr*ity, md shndard indusr.) producrs ariseoaturalll i. nerwork industr nr. In su.h in(lustries, ont or two lirms mav benest lrom pnnltrehdrtur t,r.dbr, k ansing hnm n(rs. r (\lrrnJnn. this,n.,"Jrr, fi, likelihood that lirns produ.ing compl.mcntar! goods or nr\i.cs $ill r.sard thcir produ(rsdsrsdJr.lprodLrr\$rrhindk,,,nJunr\ Ot .,,uh(. b,, om,"g produ( would r.inbrce rhe positive market l.cdback that these one or ".ruoa,,t tu.o 6rms already eDjov.At lhc same tihe, hiling roemc.g( s r stodud product cm tta soure ofncgatnc market lcedback for their rnals. Haw ptaduct compatibility might emerye as an industry outcone Intcnkpcndcnt oligopolists cannot make choices abour produ(1 comparibilit\ in isolatio. lrom rhe d.(isiohs ol lh.ir cohperir(,rs. Thq musl also rak i.1,,.c.our rcsponscs ol other iirms. 83 EL ECT RO N IC CO I\,1 I\,I E RC E I\,1 A R K ET S A N D PRICES To.onsnlcr ciromstances mdcr *hichlLms opt tbr.ompatiblc iilrmats, begin bv taking a look at Figure l.7. rhis figur. displa\s s.'ts ol prolit parollis ior an indusr* composed oltuo multiprodud lirms, FirmA and l:i.n B, \rLich are contcmplating *hether to olltr thclr produ.ts in litrmat I or Format 2, cithd of whi.h is in(om patible with the other lormat. Pancl (a) displals th. simpl.st posible outconr., in $ hi.h Firm A ed Firm B cach anticipate earninS th. tnost prolits il both choosc the same tormat. ln this cxample, both firns ca.n thc highest protits. s; million, during thc relelant period, b) adopting Format 2 d d.!.lopingil as thc sta.dard produ.l ot siiuation of lormat igrccmcnt. lt is still the casc that thc llrms earn morc prolits ;l horh choose the sanc, tompatible Panel (b) of Figur 1.7 depicts a less clear l 2 (a) I E e (b) S Ful/tre fims to seek to cootdinate prcduct fonat' Iwo payoffs in panel (a) earn qreatest ptofits bv profit fnns conlronting the 3,7 Ptofit payof{s inducins cootcli nati^s choi ce of F o mat ?. l n panel (, e ach of the two f irns e arns hiqhest ptofits if both fitns'products have the sane fornats Fitn A, haweltet, earns greatest Ptofits if bath fins apt fot Farmat 2, while Fim B earns the highest ptolits, if both agrce lo use Fanal 1- Hence, both panies will desire to caordinate the fomat choice bd inniallv nav not aqrc on how to da so, which is the re$on why this type of sane is naned he Aattle ofthe seres. 84 I|llPERFECT COIV] PETITION AND N ETWOR]< INDUSTRIES l;rmat. Firm A, ho$cr.r, earns thc highrst proffts if the firms huru,lly adopt th. Format 2, rvhereas Firm B carns greatest prohts if the 6rms ag.ee to choosc Formar l.Thus, each psty desircs to coordinate with th. other, vet agr(ment regarding thr sptcilic fortuat h Dot ihmcdiatelr forthcoming, \rhich is $h! in game thcorr this is olien relerred ro $ th Edrda ol.hd J.r6. Hos might th.lirms facing thc pro{it payoils in pan.l (b) of Iiigure 3.7 s.(le on a parnala.lbrmat? On. possibilit! hightbc lb. firn A ro agre. tl) makeasidr parmenr o{ S0.5 milln)n to Firm Il to agre. to adopt krmar 2, therebv yirlding each firm th. sa'ne net pro6t paJoff.l $4.5 million. Anothrr possibilil! might bc {i)r Fir.r B to oiler s0. 5 million to Firm { r{) igrce n, adopt Forhrt I , reruking in thc samc nct paloff for each lirm. Altcrnanv.l!, the t*o li.msmighr agrce todevelop on. hrmartogetherdd split profits tlcnly. A dilhcuity wirh this approa.h, howcver, is that as long ar one tirm standsto Sain a great.r proiit paloll from.hoosing oD( l;rhar orc. anoth.r, it has n i.entile to.heat on thc rgr..m.nr. (Cohpanies hav. ind.cd b.en known to chcat oo agreemcnts to co tlerclop prorlucts; see the box dndg.nent ODttne: rt ,ltictochjp paloIlconfiguration Daclopn. ,Udth Nor.Uddei, Htorcn.) IlAt{AGEMENT 0t{LIt{E: A MICR0CHIP DEvELOPME T iTATCH NOT MAOE IN HEAVE}I In 2001, Sony parhered with Ioshiba and Inrernatiorat Bustness Machines (IArM) lo develop the Cell processor, inrended to be a mtcrochip so powerfut that ir would redef ie the computins power ofcomputing dev ces produced bya[ threecompanies. commited themservesro spending more than g4oo mititon to design the Cell, which Sony intended to integrare into its new ptaystation 3 videogame The companles the midsl of Cel developmeir work, ]\1 .rosoft asked I a M if it woutd produce a newmicrochipforthe xbox 360, its own newvideogame consote that wou td compete with Sory/s Playstation r. I B[45 agrcement w th Sony did notru]e outeventuatsates I n of the Cell miffochlp ro oth€r {rms. Neverthetess, as Cell devetopment work progressed, IAM lvr 1^ IBM hid from Sony the lact lhat its parrn€lship with I8M was assisti.q innediately eryasins i a side vefture wiih Soiy,s videoqame comoetitor lcrosoft. I B [,] employe€s began hiding fronr Sony ensiieerss€cretefioristo boruw t€chnolosy from the Cell to creare a similar microchip for use in rhe Xbox 360. By the ume work on th€ Cell microchip project was comp eted, IBt/l had enabtished producllon lacillti€t ln p ace to produce chtps for bolh Sony,s p aysrarton 3 and rvllcrosoft's Xbox 360. Sony had retiedsotety on rBM to produce irs mi$ochip, but N,lifiosoft had aiso esrab ished a conrract with aroth€r chip producer. Thus, when a producrion problem hetd up IB[/t product]oi of rhe microchips, Sony had to detay its i.troduction of the P ayna!on I for ayear. tn conirast, Sony,s rivat rM ioosoft was abletoobiain sufficient xbox.hipsto beab ero successfu|y taunchrhe Xbox 160 usinq a mkrochipihat Sory had hetped devetop. 85 E LECTRON IC COIV] I\,1 E RCE - IVIAR(ETS AN D PRIC ES For ctitical analysis: Ultinately/ on net all thrce fins ,ony, T6hiba, and IBMprafited frcn devetopins the technolosy fot Cell micrachip Uaduct farnat. IBM, hawveL enhancetl its own profit payofl and nay have cut inta Sony's payofl by wotkirg out its side dealwith Mictosoft. What type of gane do you conctude that Sony and IBM faced in nutuatty developins the Cell nicrochip product fornat? How product incampatibility can be an industry outcome whv might tirms opt 10 produ.c products that ar. incompatible? Take a look at the profit palorfs displaved in Figurc 1.8. In additb., assumc for no$ that the firms act on the asrumpti{). that the products tbel, produ.. arc ,or subject nr nctsorl theit produds in Fo.mat I, eachlirm cxP.cts that il till ear. s2 million in proEts during th. relevant pcrxd.lfboth 6rmsollr thcir Produ.ts in Format 2. .ach frrm {ticipatrs carning $ 1 millt d in Profits. If Firm A diilcrcntiat$ its Product s.t bv offering ir in Format 2 s hile Fird B s.lccts Form.t ], thcn cach lirm eticiPat6 Ilboth 6rms o{Tcr nillion in prolits. If Firm A seleds klrmat I $ hile Firh B oPls for Formar 2, then each iirm expects ro earn $5 millioD in prolils. Firm A, tLerefor.. sill oPt fi)r k,rmat 1, and Firm B $ill opt for Format 2. Thus, the lltms favor mutual in.ompatibilit\ that rli{Ierentiates thtir pftxluct sets .nd rields thc hiShest anticiPated Prolits In an) event, rach 6rm $ill choose a format thar is incomPariblc Nith itsrilalk format earning 54 choice as a consequen e ol this configuratlon of pro6t pavoll-s, which cconomists call rhe T*eedle Dee.T*eedle Dua pne- The complicating impacts of network effects lf FirmA d Firtu B d. .orect in their dsmcd anticiPanon that netwo.k cflects are negligible, Figu.e 1.8 soggests that both lirms can achiele higher Profft le\els by 2 '6 ^, E Ea a 86 Figut€ 3.8 Ptolit payoffs rcsultins in prcduct inconpatibilitv- Il two firns face the depicted profit payoff natti,t known as a Tweedle Dee'Tveedle Dum gane situation, they always anticipate earnins the highest ptofits if they adapt i nco n patib le p r od uc t fo m als. I[,lPERFECT COfulPETITION AND NETWORK INDUSTRIES adopting inconpatiLlc produd li)rmats.Whar happens. howcver ifD.tworks elTects arc present and importa.t? In that case, a "I;rmit $ar" ma! takc pla.. in which Firo A fights for Format 2 to.m.rge as the in(lustrv stmdard and Firm B battles for Format I toassumr that role.That is. dl€r opting for an incomparibleforhat, cach lirm aftulll! is likelr to expend mor. rcsources than preiouslr anricipaled ropromot irs format and a.hie\ e positne m.rk.r lcedba.L. For idstmce, suppose that Firhs A dd B seu digital husic nnd varn,us .omple mentarl acccssorics. After the 6rms havc optcd lor incomparibl. formars, businesses su.h as restauants. tarcrns, md collee houes b.gin installiDg wireless devi.cs alloxing groupsofotomers to ph thcir pr r$dal msic collections at thrirboorh. CoNmers mexpectedll der ek,p a stro.g prcliren e lbr tu..ring in groups in restau&ts, talerns. aDd cofcc houscs t(, rngage in musjc listening scssions.Thus, utilization otthe nes. wi.clcssdclices is likel! to be subjectto network ctlects, .rcatinS the lilclihood that one forhat lLill.xpoicn.e positne naker fecdback $hil. thc otler $itl languish as a.onsequenceofnegatiye feedba.k.Thc rcchnologl asociared *ith these d.vicc onh, perhits one of the tso digital husic formars to be utilized, howevel Consequently, one lormat is more likelrto em$g. as o indu(rv stmdard. while the othcris nore Iike\ uhimatelv to s ither au ar. Each 6rm mal r.'spo.d bvoff.ring fimcial indu.emems to osrers ot restaurants, taverns, and co{Iee housestouse ncwwirelcss d.vices using irs own lormar. Doing so xould, ofcouse, pushupthe lirms' operatiDg .osts. In additnrh,the lirns mav rn to indu.c consmers r{) choose then product b! r.ducing pri.rs, rhe.ebr rcducins thcir rF,Fnuc.Thu,.tu,hhrm\a,rudlh\rilledrnlo$erpronrrh,.rt..anrrrp"r.J"t., thty sclccted thcir lirrmats. Ilon. Ilrm ukiharely loses th. banle to gain positi!. markct ftedback and btcoht thc industrv srandard forhar, its profits tikdy wilt Switching casts and the potential for ,'consumer lock-in,, On( rcstaurants, tarcrns, a.d..ller housr\ ha\e installed \rireless music-plNing deric.s utilizing one 6rmat, repla.ing thc d.!i(cs sirh rhos. using the akernatire. in.oopatible forhal would entailcosrs. One such busine$.osr *!uld bc th. explicit .osts asyiciatcd wilh rcmoli.g rh. preriousll insralled deri.es md repla.inS thcm lLith dcric.s utiliTing thc othr:r inr mat.,{aorher souid be an implicitopportunrtrcost resuking Irom a potcntiil loss ol r.(.ipts of rer.nues tiom customers v ho licqucnted the busincss mainlv bccausc thL.y .ni()yed li.rehing to one aorh.r\ digital music tibrarv whilc risiting durnr8 hn,als or ortr rlrinks or cups of colTee. the sum ofthes. cxplicit ud nhplicit h opportunitv costs coukl .onstitun, 3 signilicor su.itching cost ior the arguable thar $vrt.hing costs are likcl! to h( parricularlr high for the .onsumers ot g,,"h or +r,r,,. t,rodu,,,1 l{ r,rm. in n,rs,.,l ,ndu.rne. ssrr.h,ng Lo Jn altcrnati\e produ.t is likel) to rnrail parti.ularlv high suit(hing costs, such as tle reinnall:ho. .osts that .estauraoh, ra\erns, ed colTee houscs in thc cxamplc sould coDlront. Forconsumds ofproducts ofnetlvork industriesj th$diirc, it is more like\ that consuDer lock-in may rak. place, mennihg tLat thc rwit.hitrg cosr i$ociated is 874 ELECTRONIC COMi/lERCE_T/lARKETS AND PRICES with @nsuming my other dE thc product ofthc nctwork irdlstry will almost always cxceed the perccived benelis Fo- such a *itch. As a resuh, thc 6rm olTering an initially adopted product format beDe6ting nost Fom positilc narket feedb..k may establish a position ofdomiDoce in the marketplace. 5. UNet what circunEances will a lirm choose to ollet products lhal ate conpalible ith those ol conpeting fii4s? Firmy products are comparible when each filmt products funclion when us€d alonqside a riva! firm's product. Firms will be more willinq to contemplate producirg compatibl€ items if they anticipate that aspeciric oroduct format will qenerate positive marketfeedback and create asrandard product withln the neiwork irdustry. Nevertheless/ in some irstancesa "Battle_oithe_Sexev' qame mishlapply, resulting in initlal dilaqreement regarding a common lormat a.d thereby requirinq firmsto coordinal€ side paymenls or to develop products iointly and share profils. In some instances, a "Iweedle Dee Tweedle Dom" game mav applv, resullinq in faillre off rmsto aqree to a common productformat. The corsequ€ice may be a "formai war" inwhlch firnrs oflerexp€.se{aising inducem€.ts and revenue- reducing price cuts in an eflort to lnduce consumers to adopt their cho*n formats. Firmsenqage insuchactl!ities io hopes oI brinqins abo!tconsumer lock ln resulting fromconsumers'switchnqcostsexceedingbenefitsava ablefromsubstituleproducts oflered by compeling firms. CHAPTER SUMMARY Alternative forms of comp€tition amonq evenly matched but imperfectly competitive e-commerce firms: In an oligopollstic marketplace, no more than a fewfirms sel the bulk ol industry output. 0ligopolistic interdep€ndence resu ts:Th€ price or production makes can affect the decisions at competing firms. Under monopolistic competition, it lseasyfor firms to enter or leavethe industry, and many firms produce similar yel dislinsuishable soods or seruices. Coisequently each firm decisions that one liim d€termines th€ price o, its producttaking into accountthefact that lhe demand Jor its own product&p€ndson the prirescharsed byotheriirmsand on the number of firms Firmsearn zeroeconomic profits ln the long run/ but the price charged monopolistical y competitlve flrm erceeds marglnalcost, and firms produce at an in the industry. by a averag€ cost that exceeds the minimum feasible average cost. ow the sp€cial production-{ost characteristics ol ltirtual product influences how they are priced: virtualproducts exlst in d iqital forms that may be produced and dlstributed H at relatively low margina cost. Cr€ating the product normal y requires rcurrins a relativ€ly sizable fix€d cost. as a resull, the average tota costcorvesoffirmss€llinq virtualoroductssloo€downward,andproducersofthe*productsexperiencesho.trun economies of scale. lf a producer were to set the price of a virtual product equal to marginalcost, itwould fai to recoup theflxed cost ofproducingthe item, sothefirct- 88 1I\,lPERFECT COMPETITION AND NETWORK INDUSTRIES nrarqinalco+ pricing is unaitainable. To lnduce the firmto incurthe fixed cost of fiealinq th€ product/ pric€ must equal average cost. Economists often best optimum of reqard average cost prlclnq as asecond'best optimum forsociety, because sttlns prlce equaLto averagetotalcost induc€safirm to provide avidualproduct but at an economlc 3. How network externalities affect the demand for a product: ll a good or service is subject to a network externa ity, this means that the be.efit thar a given indiliduat receivesfrom consumirgthe liem rises when others atso consume rhe good. As a resutr, each consumer s will rq io pay more for the ltem as the number ol other consumeB ofthe liem rises, th€rebyaffeclinq the overal demand forthe itern aswellasthe price elastic ty oi demand for the item. 4. How network externalities produce marketfeedback effects: tn a network indu+ry, loqrollinqellects can take place inrhe form of market feedback etfecrs. Ifasufiicienr number of people purchase a product subjectto network exrer.atitiesto induce many othersto boy the item, ther a posirive marker feedback elfect occurs. In contra(, ila sufficient number of people slop purchasing a productto sive a number ol indiv duak an rc€rtive to cut back or their 5. pLrrchases, then ther€ is a negative market feedback effect. Market feedback €ffecis irflue.ce fortunes of both entire nerwork industrtesand of indiv dualfirms wilhin those indusiries. Circumslances under which a firn choos€s to offer products thai are compatible with those of competing firms: Product compat bility r€fers to the capabttity of a fi/m,s product to lunction when usediogether w th aconrpetinq ftrmb product. Ftrnrs are nrore likely to olfer conrpatible products when they anttcipaie rhat a particutar prodlct formai is llkely to experience posit ve nrarket feedback, cauring ir to become the standard productwithinthe network industry. Sometimes, however, a\\Battie of the, Sexey' sltuation might arise, with firms disasreeins about which common f0rmat ro ulilize and hence must coordinate a side-paymenl arranqement or a ptan for ioint productdelelopmenl and profit sharing. In some cases, firms, an|ctpated profit payoffs may sive ris€ to a\\Tweedle Dee Tweedle Oum,,situarion inwhich firmscannot agree to a common lormat. This can resu t in a .'forfiat war,/ €niai|og cost boosting iq price cuts thaterod€ previousty aiticipated protirs. Firms may eigage n soch batt es in effo(s to qain tniriat consum€r adopiion and inducements and revenue{ed!c subsequenlly achieve corsumer lock-in thar resutts when consumers/ swilchins cosrs exceed benefits derived from rival firms' substirute producrs. QUESTIONS AND PROBLEMS I The consulting tkm Foriester Research conducts reqular ,'power,, su rveys of dot.com industries, in which t attempts to identily compan ies with the stronqest opportuniries lor siqnificant profitability. Keyfactors inftuencing firms,rankinss are characterisrics that dlstingu sh the quality oftheir products aird s€rvtces and set rhem apadfrom r va firms. appropriate diaqrams to exptair why successfut product difterenriation is likely to €rhance an Interretselier/s protitabiltry. Use 89 I ELECTRONIC COIVlIVERCE I!]ARKETS AND PRICES 2- There is a large markel ln oiline dating and matchmaking seruices. Each reardozens of fkms enter aid exitthls mark€|. The irdustry is composed of hundreds ol sites that, in exchanse for monthly subscription fees and permessase charges, provlde paqes rtere sing le people can provide p€lsonal information, andensase in online chats. Each site attempts to distinguish itself Jrom others by offerias safe forums for onaltached people to meet, often by deveioping sp€cia features {or preseruinq anonymity and screening out unscrupulouscustomers. Take a posilion on howlhe prices of these o.llne services are iikely to be det€rm ned/ and draw on on€ of th€ theories discussed in Chapters 2 and 3 to suppo( your position. 3. Consid€r the e book publisher whose costcurues are depicted in Figure r.4. Suppoe thatb€forelhepublishermakesadeawiththeauthorofthise-book,itdeterminesthat ihis isthe lons run situation itwillface if it paysthe aulhor $45,000 in advance. The author of the e book, however, indicates that she is willlnq to write the book for an tryto work out numerical values, butexplain the baslc sholt{un impllcations for the publisher's profit-maximizlns price, sales, revenues, costt and profits. advance ofonly $25,000. Do not 4, Suppos€that,asinquedionl,theebookpublisherlindsthattheauthorwillsettlefor a rmaller up{ront payment to write an e book. ln addiiion, suppose that other publishersiilhee-bookmarkelplacefindthatauthorsaccepi ower up frontpaymentsWhat are the basic impllcations for lons{un equilibrium in this monopo istically compeiitlve industry? 5- When a lirmS lonq run averaqe total con cune s opes downward over a! ranqes of outpot/ economists sometimes call lhe Itm a naturat nonopoi, meanins lhat it is most efficient for a sinqle firm to s€ll the product in queslion. Explain why firms thal !€ll 6. 7. virtualproducts in monopolistical y compeiitive industriesare not .atural monopo ies. Iake a look back at Tables 3.3 ard 3.4 on pases 76 and 77, and explain whv all five conrumers willchoose to buy the software packase wher ils pric€ ls $80. Take another look back at Tables 1.3 and 1.4. Table 3.4 is conslrlcted under ihe assumption that the five consumeB do not shar€ with one another information about willinqnessto pay. Suppose that allfive consumers getloqetherfor luncheach dav, ard during a convebation discover how much each one is wlllinq to pav depending or how many olhers purchase and use the product al every possible price. Suppos€ that thev act or this lnformaiion by coordinatinq thelr purchas€s ofthe software packaq€, and redo Table 1.4 under this assumpiion. What k th€ matimum price at which at least one unit oI the software packase ispurchased in this is the 8. five{onsumer marketplace? what totalnumber of urits purchased atthis price? Explain why ietwork exterrallties can causelhe demard fora product eilhelloexpand o/to contract relative lo what it would 9. be lf there were no network ext€rnalitiesproducls you that think are subjeci to.etwork exlernalities. [lake a list of five In your view, are these five producls sold in neiwork industrles? Explain. a. b- For indu.tries in part (a) lhat you hav€ classitied as network industries, do you believe that all firms within each industry experience market feedback effectt or do jusl one or two firmsexperience markel feedback effects? In yourview, are the nrarketfeedback effects currently posltive or regative? lO. 90 Economles ofscale normally re ate to afirm's cost of production, butsome economists INIPERFECT COI\,IPETITION AND NETWORK INDUSTRIES that netwo rk externalilies can be a source of "demand{ide economles of scale/' for firms within n€twork indusries. Speculate abolt what these economists misht ars ue 11. Considerthe sets ofprofitpayoffs anticipated by Firms X and Y in parts (a), (b), and (c) in the figure, which in each case depend on both firms'choices betwe€n Format I and Formal 2. In each part/ exp ain wh€lher the firms will desire to produce and s€tt compatible producrs. (a) 2 Ea ii (b) PN E Ed (c) 91 ELECTRONIC COI\,II!1ERCE MARKETS AND PRICES ONLINE APPLICATIOT{ lnlernet URti www.nolo.com Ii1l€r Nolo.com Law for All Navigailonj Go direclly to the above Web sit€. Applicatian: Perfotm lhe lo lowinq operat ons, and arswerthe fo lowinq questions. 1 lvlost economistswould arque that Nolo.com operates i. either a perfecl y competitive or monopoListlcally competitive market, because there are now many online provideE ol Iegalseruices that can easily enter or leavethis eleclronic marketplace. What facto19 are likely to determine which ofthese two market structures is relevant? 2 Forthesakeofarqument,slppo*thalth€marketforo,rlinelesaseruicesisperfectiy competliive. Tak€ a Iook al the larqe numbers of lesalseruices offered atlhis web site. What factors are likely to deternine Nolo-cornb shutdowr polrt (see page 40 in Chapter 2) for oftering any one ofthese servlces? 3 Suppose that a lesal-services firm offers downloadable soflware that consumert can draft wills and oiher legal documents. If the I rm incurs hish up-fronl costs ir designing the software but very low costs of distributing it on the web, what can you us€ to say abDutthe e likelihoodthatlhe firm wil sellthis software ln a perfectly competitive ectronic mark€tp ace? For group study and analysis: Assisn sroups io search the Web and complle a lisi of at least threeaddltionalorlinelegallirmsardtocomparetheseruic€stheseflrmsoffer.Reconvene theentire classaiddiscusswhetherthe marketforonline lesa servicesshou d be classified d5 perte.rly.o4pFr'rive o' noropol n,cal y coTp.ril vP. SELECTED REFEREI{CES Bel €flamme, AID FURTHER REAOIT{G Pao, "Versioninq Informatlon Goat1s," n lndustrial Orqanizatian and the Digital Economy, by Gethard ll inq and lvartin Peitz, Cambridse, [4A: [4lT Press, 2006,pp. ),75-245. Choi, Jay Pi, Elrik Gaard Kristiansen, a.d Jae Nahm, "Preannouncing Informatlon Goadt," in Indurtial 1ryanization and the Diqitat Econany, by Gerhard llling and [4art n Peilz, Canbridge, tvlA: VllT Press,2006, pp. 206 228. Economid€s, Nicholas, "The lnternet and Network Econom cs," in Intetnet and Disital Economics: Principles, Methods, and Appt ications, b\ E(ic Erousseau and Nkolas Curien, Cambridge, UK: Cambridse tlniversity Press, 2007, pp.239 267- coldmaris, [4aris, Ali Hortaesu, Chad Syverson, and dnsel Emre, "E Commetce and the I\4arketStructur€ of Retaillndustries," NationalBureau of Econonric Research Worklns Paper No. 14166, Ju y 2008. Liebowitz, Stan, "Economists Examine Flle Sharlrg a.d i\,lusic Saes," in Indrstrial 92 I IV1 P ERFECT COI!1 PETITION AN D N ETWORK IN DUSTRI ES Aryanization and the Dlsital Ecanony,by Gethatl ltlins and [4artin Peitz? Cambr]dgg [4A: N]lT Press,2006, pp.145-17t. Roh fs, Jeffi€y, Bandwason Effects in High lechnology Industries, Canfi dse, MA: Mll Pres,2001. Shy, Oz, fhe Econonics of Netwotk Industries, Canbtidge, UK: Cambridqe University Press,2000. Peitz, [,lart]n, and Patrick Waelbroeck, "Disita Music," in Industrial 1ryanization andthe Disital Econony, by Gethard I lins and lvlatin Peitz, Cambridqe, [,lA: lV]lT Press, 2006, pp.7l 744. Smlih, [,lichae, Jo*ph Ba ey, and Erik BrynjoLfsson, "Understandirs Diqital [larkets: Revlew a.d Assessment," in Undetstanding the Digital Econony: Dala, Toals, and Researct edited by Erik Bryijolfsson aid Bria. Kahln, Cambridse, rMA: lvllT Press, 2000, pp.99-r36. Varian, Hal, "Competition aid rMarket Porer," in Ih€ Econanics of Infornation Technology, by HalVarlan, Joseph Farrell, and CarlShapiro, Cambridge, UK:Cambridge University Press,2004, Dp. 1 47. 93