II FYI 1WNOVEMSER/DECEMSEE 1995

Donald I. Baker is a senior partner at Baker & Miller PLIC in Washington, D.C.

Shared ATMNetworks—The AntitrustDimension

Donald I. Baker

WHERE NAVE ALL THENETWORKS GONE?The automated teller machine (ATM) hasbecome an increasinglymore importantpart of the average modern consumer’sfinancial life and for this reason presentsan ever-more pressing set of difficultantitrust questions for networks, financialinstitutions and government decision-makers. Thirty-two years ago, the SupremeCourt emphasized in its landmarkPhiladelphia National Bank decision thatconvenience and location were key tocompetition in the retail banking sectorand ruled that “the fact that banking is ahighly regulated industry critical to theNation’s welfare makes the play of compe-tition not less important, but more so.”Today ATMs are the key to consumerconvenience in banking, and yet it is notat all clear that the Supreme Court’sPhiladelphia National Bank message aboutthe importance of competition has gottenthrough to the policymakers and courtsmaking decisions about ATM networks.

In the last decade, large and successfulArM networks have been created, andsuch names as MAC, NYCE, STAR andHONOR have become familiar landmarksin different parts of the country Meanwhile,the number of ATM network alternativesavailable in any particular region hascontinued to decrease.

Why has antitrust policy been sucha random—and not particularly construe-

tive— factor as mergers have helpedcreate dominant ATM networks in variousregions of the country? Natural monopolyeconomies? Or just honest uncertaintyand indecision? ‘[he Board of Governorshas told us that “the largest regionalnetworks now account for 80 percent of allregional ATM network transactions in theUnited States” and used this as a reason forapproving another network merger in theElectronic Payments Services, Inc/NationalCity ease.1

Meanwhile, Department ofJustice(DOJ) officials have influenced the courseof events over the years with what theyhave said, done or—more often—notdone. They have thus left the field largelyto private plaintiffs, the private attorneysgeneral who are necessarily more interestedin winning battles than in establishingpolicies and precedents. It is hard to digout of the resultant patchwork of DOJpress releases and Business Review Letters,private complaints and a few consentjtndgments any coherent vision of themodern ATM network as a competitoror of the markets in which networksoperate. Probably the most serious effortat illumination was entirely private:Former U.S. Assistant Attorney GeneralThomas E. Kauper’s long opinion, asa private arbitrator in the 1988 FirstTexas/Pulse arbitration.1

DQJ has not been a factor in thenetwork merger area. Its most instructiveenforcement effort—its 1994 complaint,consent decree and competitive impactstatement in United States v ElectronicPayments Services, Inc—dealt ~vith theconsequences (rather than the creation)of monopoly power. It was DOJ’s firstantitrust enforcement action in theelectronic funds transfer (EFT) sector in17 years. Significantly DOJ’s complaintasserts that branded ATM network access isa relevant market and recognizes that ATMnetworks may enjoy substantial marketpower on a regional basis. (The DCJJ

I United States v. PhiladelphiaNail onal Bank in 374 UnitedStates Reports (U.S.) 321(United States Supreme Court,1963).Banc One Corporation, et of,March 6,1995, p.17 (theEPS/Notional City opinion).See also Banc One Carp.,Federal Reserve Bulletin,vol.81(1995), pp.491-2.I have a client who is veryinterested in this proceeding.The views expressed here,however, are my awn.

1 First Texas Assa./Financiol

Interchange in 55 Antitrust &Trade Regulation Report 340(1988). Professor Kauperwrote that opinion becouse theparties had agreed that hewould do so as part of theirarbitration agreement. (I waslead counsel for Pulse in thatunique proceeding.)

-, UnitedStates v. ElectronicPayments Services, Inc,Na. 94-208 [U.S. District Courtfor the District of DelawareApr. 21,1994159 FederalRegister (Fed. Reg.) 24711(May 12, 1994).

FEDERAL RESERVE BANK OF ST. LOUIS

5

OFYIIWNOVEMBER/DECEMBER 1995

See EPS/Notm’onol City opinion,p.17.

‘See FPS/Nohbnol City opinion,p.14.

consent decree enjoins one MAC networkpractice—concerning ATM driving—butleaves other restraints in place.)

By chance, the Federal ReserveSystem’s most illuminating effort alsoinvolves the same ATM network—EPS orMAC—but the Board’s extensive effortalso turned out to be rather timid onboth structural and conduct issues. OnMarch 1, 1995, the Board of Governorsannounced that it was allowing a majorexpansion of this already dominant Penn-sylvania network by allowing it to includeNational City Bank of Ohio in the owner-ship group. In a 28-page statementreleased five days later, the Board rejectedthe objection of a competing ATM networkto “undue concentration of resources”with a statement that “the Board believesthat, as a result of economic and marketconditions, regions are likely to have onedominant ATM network.” At the sametime, as Vice Chairman Alan Blinder’sdissent made clear, the Board declinedto follow up on the perceptive sets of ques-tions concerning allegedly anticompetitivenetwork rules that the Board staffhad pre-sented during its investigation. The Boardseemed to take comfort from DOTs priorinaction on some subjects in its consentdecree proceeding.~

As observers and students, we can allbe grateful to the MAC network for havingbeen such a patron of antitrust learning inthe ATM field. As critics and consumers,we may c9me away with some doubtsabout how well government agencies havedone in meeting the challenge.

In sum, those sporadic governmentactivities—plus a few private caseschallenging leading ATM networks’ pricingpractices and operating rules—cannot hesaid to have added up to the kind of com-prehensible body of law that PhiladelphiaNational Bank spawned in the bank mergerarea. It simply has proved difficult for theFed, DOJ and the courts to come to gripswith the market realities and the competi-tive effects of these new and importantfinancial enterprises—branded ATMnetworks. Like the federal budget deficit,this accumulated learning deFicit may pose

larger burdens on the next generationof consumers and regulators: the realcompetitive issues are not likely to goaway even if we end up with largely anantitrust regime of rule making forregulated monopoly

NETWORK REALITIES:JOINT PRCOUCTS~O6VE?SEINTE RESTS

Creation of any new network requiresa great deal of cooperation among theproposed members. An ATM network isnecessarily the reciprocal commitment byparticipants to issue cards, deploy machines,honor each others’ cards at their machinesand settle the resulting transactions.

A new network needs access to facilitiesfor switching transactions among members,rules governing the terms and conditionson which members will honor each others’transactions and a recognizable trademarkthat will tell consumers where networkservices are available. The switchingfacilities may be (and often are) providedby a third-party processor under contract,whereas the rules and trademark maybe created with extensive help from expe-rienced professional advisors. Even so,somebody—presumably the investorsand users—must work through a wholeseries of difficult business decisions toturn the network idea into an effectiveoperational network.

The main reason for a depository insti-tution to create or participate in such anetwork is competitive advantage—or atleast the fear of being left behind in thechanging payments marketplace. Differentpotential participants may have quitedifferent visions of what they want anetwork to do for them. An aggressivegrowing bank may want a quite differenti-ated network offering its customerssomething not available from its sleepycompetitors across the street or in thenext town. Meanwhile, the comnpetitormay want the opposite: a teniversal orleast-common denominator network thatguarantees the proverbial level playingfield and ensures that nobody is left

FEDERAL RESERVE BANK OF ST. LOUIS

6

OFYIEWNOVEMBER/DECEMBER 1995

behind in the new world of ATM-basedbanking convenience.1 Thus there is agreat deal of room for business andcompetitive disputes over what thenetwork should have its members do,what it should do, and what it was goingto charge metnbers for doing it.

..AN.TITRUST RE.AI.ITJESA shared ATM network requires

collective action involving competitors.This can be in the form of: (1) a joint ven-ture of participating institutions or (2) aseries of contracts between a single systemoperator and the other participants in itsnetwork. Either way the networkarrangements fall within the “contract,combination, or conspiracy ... inrestraint of trade” rubric of section 1 ofthe Sherman Act.8 Therefore, disputesamong members of a network overbusiness questions that in any way relateto competition can he framed as antitrustcases. Because the Clayton Act automati-cally awards treble damages, disputes arevery likely to end up as antitrust cases,especially when the plaintiff’s hand isstrengthened by being able to show thatthe defendant is a regionally dominantnetwork or controls such a network.’Effective network planning and antitrustcounseling becomes ever more essentialas regional ATM markets becomemore monopolistic.

The treatment of joint ventures hasbeen one of the most confusing areas ofantitrust jurisprudence in the UnitedStates, as well as in many foreign countries.Agreements among direct horizontal com-petitors have been treated severely underantitrust law and, where they involvedpricing restraints or customer allocations,could be labeled cartels. Meanwhile, fullintegration by merger has generally beenanalyzed under market structure standardsin which results depended primarilyon the nature of the market and theparties’ share of it. Joint ventteres havehistorically—hut unevenly—beensubjected to cartel rules.” The moderntrend, however, has been to look at joint

ventures that created something newunder merger rules or under the fact-basedrule of reason theory” Market power(rather than form) is critical to suchan inquiry

This conclusion is not so confusing asit might seem at first blush. Today it ispossible to say both that: (1) under modemantitrust rules, it is likely that creation of ajoint venture and corresponding rules willbe looked at in a context that includesevaluation of market power and relevantmarket, and (2) we have very little ideaprecisely what definition of a relevantmarket will be used to evaluate thecompetitive effect of an ATM networkmerger or of a rule imposed by a majorATM network. A narrow marketdefinition, of course, may produce highmarket shares (and even an inferenceof monopoly power), whereas a broadmarket definition may make all activityappear benign.

DEI-INING NIARKETS EORNETWORK SER.VICES

Commentators, regulators, enforcersand the courts have not developed anyconsistent way of defining mnarkets forATM networks, or indeed for networksgenerally They have not really figured outhow to factor the brand element into theanalysis of the relevant market for ATMnetworks. The presence of a brand wouldseem to make an ATM networkvery differentfrom a joint venture pipeline or electricpower pool (hut not necessarily from along distance telephone network or acredit card network). Both DQJ and theFed have finally seemed to recognizethis difference.

Recognizing the differences betweenbrand and nonhrand markets is not just amatter of intellectual curiosity Determina-tion of the relevant market is often criticalin antitrust litigation. TI’he plaintiff tries todefine the market as narrowly as possible tostrengthen the claim that the defendant hasmarket power or the ability to increaseprices above the competitive level. Thedefendant argues that the market is almost

‘In mare than 20 states, smallerbanks hove successfully ledpopulist political campaigns tomandate universality solutionsby passing compulsory sharinglegislation thot requires eachbank to share its ATM facilitieswith every other bank onequal, or nondiscriminatoryterms. See Baker and Brondel(1995)1(25.03141 [a).

815 United States Code (U.S.C.)§ I.

‘15 U.S.C. § 15.

“See, for example, UnitedStates v. Topro Assocs.,405 U.S. 596 (1972); UnitedStates v Sealy, Inc., 388 U.S.350 (1967) (territariol andprice restraints); United States

Columbia Pictures Indus.,507 Federal Supplement (F.Supp.) 412, 430 [U.S. DistrictCourt far the Southern Districtof New York (S.D.N.Y.) 1980,affirmed in a memorandumdecision (aff’d mem.,) 659Federol Reporter 21 Series(F.2d) 1063 U.S. Court ofAppeals for the Second Circuit(2d Cir.) 19811.

‘‘See generally, Chapter 21‘Special Antitrust Prohlems ofJoint Ventures’ in Baker andBrandel (1988), discussingespecially Broadcast Music, Inc.

CBS, 441 U.S. 79(1979)and NCAA v. Boardof Regents,468 U.S. 85(1984).

See, for example, CB&TBoncshores, Inc. (1984),Atlantic Bancorpomtion(1983), d CenterreBoncorporolion (1983) (thepravision to unaffiliated finan-cial institutions of Iota process~ing services, particularly theoperation of an ATM networkexchange), Interstate FinancialCorp. (1983) (same).

FEDERAL RESERVE BANK OF ST. LOUES

OFYIIWNOVEMRER/ORCEMRSR 1995

“See, for example, Citicorp(1986) (provision of AIM seevices), Sovron Financial Corp.(1986) (same), Bardays BankPLC (1985) (competition inthe provision of ATM or POSservices).

682 F. Supp. 269 (U.S. DistrictCourt for the District of NewJersey (D.NJ.), aff’d men.,853 F.2d 921 (3d Cir. 1988).

Id. at 279 (emphasis inoriginal).

“59 Fed. Rag. 24713 (May 12,1994). Similarly, in the PULSEarbitration, Professor Kauperrejected proposed markets ofoIlpoymentsystems and allmeans ofobtaining cask ondinstead identified a narrowretail market of ATM servicesan the grounds ‘that there is asignificant group of ATM userswho value the characteristics ofATMs and for whom othermeans of obtaining cash arenot reasonable substitutes.’He thus concluded that “AlMsore themselves a relevant, iffragile, market for antitrust pur-poses.’ 55 Trade keg. Rep.(BNA) Na. 1380, at 356.

“59 Fed. keg. 24712 (May 12,1994). In PULSE, ProfessorKouper identified a wholesalemarketof network sv.’itcl ring,and concluded that PULSF hadmarketpower because ‘exist’ing subnetworks, regional net’works and national networksdo not presently provide a rea-sonable substitute far the[switching) service PULSE pro-vides to its members.” Id. at355. Other regional networkswere found no be only potentialalternatives for lexas ATM own-ers and substantial harriers(including tIre national net-works’ nntiduality mernhershiprules, the preference by banksfor local networks and the factthat PULSE was very efficientand well established) were said

boundless. The relevant market typically isdetermined on the basis of a factualinquiry into the practical alternatives avail-able to customers—which in the ATMnetwork business can mean either institu-tions or their depositors.

The decisions on relevant ATM networkmarkets have been all over the board. Inits early orders approving hank holdingcompanies’ acquisitions of voting stock inshared EFT networks, the Federal ReserveBoard typically defined the relevant marketas the provision of data processing services totcnaffiliated financial institutions.” By thexnid-1980s, however, it began to define themarkets with reference to consumerpayment networks (but it never found acase in which a merger would createmarket power).” Similarly in TheTreasurea Inc. v Philadelphia National Bank,the District Court adopted a broad defini-tion of the relevant product market inrejecting a private challenge to the acquisi-tion of the Mellon Cashstream network bythe owner and operator of the then propri-etary MAC network.” The court definedthe relevant product market as “electronicdata processing to all ATMs plus all ofthose institutions which have unaffiliatedATIVI systems and those institutions whichdo not currently have ATMs but have thecapacity to install them and utilize unarkettechnology to its fullest.”

Recently the market definition issuehas become tnore focused—thanks to DQJ,the Fed and the MAC network. In its El’Scomplaint, the Antitrust Division of theDOJ defined two relevant markets. Thefirst was a market of regional brandedATM acc:ess, based on the needs of banksto provide their depositors “ubiquitousaccess to their accounts.” It observedthe following:

While a bank can deploy itsown ATMs, the advantage to ashared ATM network is that a hank)depositors will be able co use ATMsat many more locations than onebank alone could practicably support.The areas a hank seeks to servethei ugh a shared ATM network

include the areas in which its deposi-tors live, work and sleep, and thebroader areas in which they moveregularly A bank~ability to offer itsdepositors access to other banks’ATMs, and thereby to offer its depos-itors convenient access to theiraccounts, is in most bankers’ viewnecessary to attract and retaindeposits. - - . Because no otherservice constitutes a reasonably closesubstitute for regional ATM networkaccess, regional ATM networksconstitutes a product market.

The DOJ’s second market was ATMprocessing, which involved “providing thedata processing services and telecommuni-cations facilities and services used” inproviding regional ATM access.”

The Board revisited the market defini-tion question in its recent FPS/NationalCity decision:

The Board notes that ATMnetworks have been recognized asencompassing separate productmarkets - - . - On the basis of theseconsiderations and all the other factsof record, the Board concludes thatnetwork access, network services,and ATM processing constitute therelevant product markets Jor evaluat-ing the competitive effects ofthis proposal.

On the question of geographic market, theBoard pointed to a Federal Reserve staffstudy that “suggests that the geographicmarket for network access is an area signif-icantly larger than local banking markets.”The Board then added that “the marketsfor network services and ATM processingare at least regional.””

Ultimately how a fact finder analyzesthe relevant product market in casesinvolving bank networks depends in parton how much weight is accorded to thevalue of the network trademark. If onelooks only to the data processing functionof shared ATM networks, it may he plausibleto conclude, as the Board of Governors has

FEDERAL RESERVE RANK O~ST. LOUES

8

OFYIIWNOVIMERE/DECEMSIR 1095

in the past, that the data processing industxyis unconcentraced, that entry barriers arelow, and that there are numerous alternativesavailable to financial institutions that wantto do their own data processing and that anetwork—even a dominant regionalnetwork—does not have market power.On the other hand, if the ATM network isviewed as the purveyor of a unique brandedproduct marketed under the network logo,the fact finder should probably reachvexy different conclusions about theproduct market.

There have been very few new entriesinto the branded ATM network market any-where in the country It requires a criticalmass of cards and ATMs. Participatinginstitutions have a lot of reasons to be con-cerned about having to switch from onenetwork to another—in part because itwould involve reissuing cards, reassigningATMs and, perhaps more important,reeducating customers.

Market analysis has been further com-plicated because networks are clearlysubject to economies of both scope andscale. Indeed, DOJ noted the economiesof ubiquity in analyzing ATM networkarrangements back in the mid-1980s—atime when it was not particularly active asan antitrust enforcer.”

Continuing uncertainty as to whatmarket concepts should be used to evaluateATM network arrangements has no doubtcaused government decisionmakers to becautious about taking structural antitrustaction in the ATM network area, whilecausing ATM networks to be cautiousabout accepting the risks of private litiga-tion. Both effects have contributed to theregional monopolies that we have seenemerge in some key areas. Let us turn tothat subject.

MARKET STRUcTUREISSUES—THE ROAD TOREGIONAL MONOPOLY

Many ATM networks were started bythose who sought competitive advantagefor themselves over local competitors orout-of-state institutions that could become

increasingly competitive from the distancein the plastic and electronics mode.Despite this competitive advantage impetusin the early stages, network consolidationhas turned out to be the trend. Twofactors have led to this trend: (1) lack ofgovernment antitrust enforcement of ATMnetwork mergers; and (2) dejacto networkmergers brought about by the ability ofsome banks to use antitrust boycott claimsto force entry into competing networksand thereby defeat network exclusivity

Mergers. Since the mid-1980s there hasbeen tremendous consolidation amongATM networks. For a generation, DOJ andthe Board of Governors have allowed everyATM network merger they reviewed, evenwhen the result was a regional monopoly”The most striking example was the 1988MAC-Cashstream acquisition (in whichPhiladelphia National Bank acquired Mel-lon’s branded ATM network). These twonetworks together controlled virtually allthe branded ATMs in Pennsylvania, wherethey competed vigorously for membersand transactions. Their merger was chal-lenged—unsuccessfully—by a thirdnetwork, the acquisition of which (byMAC) eliminated most competition insouthern NewJersey” The DOJ haddeclined to act, and the Board was notinvolved in the transaction.

In the 1990s, the Board of Governorshas continued to approve major networkmergers. An important 1994 example,scrutinized carefully by both the Board andDOJ, was the merger of NYCE and Yankee24, which competed in parts of New Eng-land and were together joined by Citibank’sATM network. In approving the merger,the Board noted that “a number of factorsshould mitigate the loss of Yankee 24. - -

as an independent competitor.”4 In partic-ular, the Board relied on the network’soperating rules, which permit (1) third-party processors to participate; (2) membersto participate in other networks; (3) cardissuers to determine routing; and (4) insti-tutions to participate on a nondiscriminatorybasis. Of particular importance may be thecard issuer routing rule, which might

to impede competition from thenational networks, PLUS andCIRRUS. Id. at 353-4.

EPS/Nolionol City opinion, p. 7(footnotes omitted.)

Id. at 8-9.

“ Id. at 9.

“Former U.S. Assistant AttorneyGeneral Charles Rule suggestedin 1985 that, in analyzing ATMnetwork consolidation, theantitrust division focused moreon the economies of ubiquityand the resulting consumerbenefits achievable by wide-spread shaing of ATMs, and heindicated that the divisionwould not challenge shoredATM networks based an sizealone. See (holes F, Rule,‘Antitrust Analysis of JointVentures in the BankingIndustry — Evaluating ShoredAIMs,’ Remarks Before theFederal Bar Association andAmerican Association (May 23,1985), reprinted in Baker ondBrondel (21. ed. 1988),at A’ 139.

Balto, ‘Payments Systemsand Antitrust: Con theOpportunities for NetworkCompetition be Recognized?”speech before the D.C. BarAssociation, January 25, 1995.

“The Treasurer, IncPhiladelphia Naianol Bank,682 F. Supp. 269 (U.S. DistictCourt far the District of NewJersey 1988), off’d mem.,853 F.2d 921 (3d Or. 1988).the court dismissed the suit onthe grounds that the plaintiffhad suffered no antitrust injuryas required under sections 4and 16 of the Clayton Act andhence locked standing to sue.

Decision of Oct. 3, 1994, at 8.

FEDERAL RESERVE SANK OF ST. LDUES

9

IIIYIEWNOVEMEER/DECEMEER 1995

“EPS/Natianal City opinion, at17 (footnotes omitted).

“See Baxter, Caatner and Scott(1977).

permit banks to choose lower cost networksif the merged network attempted to raiseprices.

Then, in the even more recentEPS!National City order, the Board seemsto abandon regional network competitionas an important factor, even where thenetwork rules were open to more seriouscompetitive questions. In this decision,the Board said:

It has been recognized thatMAC has a significant position inATM network access services incertain states in the Mideast region.However the significant position of aregional ATM network is not, standingalone, contraly to the public interest.Network externalities, such as theeconomies ofubiquity, tend to pro-mote consolidation of regionalATMnetworks. As a result, in variousgeographic areas, like the Mideastregion, dominant ATM networkshave been emerging throughout theEFT industry. One recent studyindicates that the ten largest regionalnetworks now account for 80 percentof all regional ATM network transac-tions in the United States. In thislight, the Board believes that, as aresult of economic and market struc-ture conditions, regions are likely tohave one dominant ATM network.”

Inaction by the DOJ and the regulatorsappears to have been driven by two factors:(I) recurring doubts over whether brandednetwork service is. in fact, a relevantmarket; and (2) uncertainty about theeconomics of ubiquity—which looselytranslated means, the more coverage thebetter. Under this concept, a mergerbetween ATM systems seemed procompeti-tive because it increased the number ofcardholders and ATMs in a network,improving accessibility for consumers.They do so at the price of eliminatingcompetition for institutions that issuecards and deploy ATMs and creating a riskof monopoly pricing and market cornering.

In many regions, all this merger activity

leaves only the national ATM networks(PLUS and CIRRUS) as possible challengersto regionally dominant networks. Althoughthe national networks offer a degreeof coverage comparable to a dominantregional network, they depend for coverageon the same regional leaders which controlthe regional network and hence may beunlikely to be seen as vigorous competitivealternatives. The competitive impact state-ment in the EPS/National City case notedthat national ATM networks are “by designnetworks of the last resort.”

Do facto Mergers Resulting FromBoycott Claims

Threats of private litigation based on“boycott” theories have also tended toreduce the differentiation of ATM networks.A member of one network may attempt tojoin a competing network, perhaps to gainsome sort of competitive advantage. If itand others are admitted, the result maybe a defacto merger between the twonetworks and a concomitant loss of inter-system competition. Because the antitruststandards have never been very clear andprivate litigation possess the threat oftreble damages, these cases usually end upwith the admission of the complainingnonmember—followed by similarlysituated competitors.

This reality is well illustrated by whathappened in Texas in 1982—when DOJdeferred action based on uncertainty Atthe time, there were two separate, verycompetitive ATM networks in Texas:PULSE and MPACT. PULSE was anonprofit joint venture and MPACT was ashared propriety network owned byMercantile Texas Corporation. A largethrift institution, First Texas Savings andLoan Association, was an MPACT memberand wanted to join PULSE, which had anexclusive membership rule. PULSE resistedthis, and counsel for the parties eventuallyagreed that the issues would be resolvedthrough the DOJ business reviewprocedure, rather than litigation. PULSEasked whether the division would takeenforcement action if it: (I) admitted as a

FEDERAL RESERVE BANK OF ST. SeIZES

10

DIYI[WNOVRMBER/DECEMRRE 1995

member First Texas; (2) generally admittedmembers of competing networks; or (3)barred its members from participating in acompecing network such as MPACr. U.S.Assistant Attorney General WilliamF Baxter,a leading author in the field, answered onlythe first question.” He said that the incre-mental consumer convenience that wouldresult from admitting the savings and loanassociation appeared to outweigh therestraint on rivalry that unight occurbetween the two competing networks.”The other two questions were notanswered because DOl did not considerthem ripe for review Thereafter, PULSEadmitted virtually all the depository insti-tutions in Texas, and DQJ did not doanything about it. Thus PULSE becamethe universal network in Texas, andMPACT became a substantial competitiveparticipant in it.

Similarly in i986, BayBanks, at thetime operator of one of the largest propri-etary ATM networks in the United States,sued Yankee 24, a new joint ventureATM network, when it was denied accessto Yankee 24.” At the time, competitionbetween the established BayBanks andthe nascent shared network was veryactive. Yankee 24 offered an aggressivepricing structure to attract banks, andboth networks offered low fees toconsumers to attract accounts. Theparties settled; BayBanks was admitted,and Yankee 24 eliminated its incentivepricing structure.”

The recent, high-visibility Dean WitterIVisa decision by the 10th U.S. CircuitCourt of Appeals may provide much neededguidance in this area.” The case involveda boycott challenge to a Visa bylaw thatdenied membership to any institution thatissues Discover cards, American Expresscards “or any other card deemed competi-tive by the Board of Directors.” After longpretrial sparring, the case went to trial infall 1992 and the plaintiff prevailed beforethe jury In September 1994, the appellatecourt reversed because it determined thatVisa lacked market power in a properlydefined market (in which Visa was treatedas separate from MasterCard) and because

the Visa rule promoted intersyscem cocnpe-tition in card issuance. If followed, thisdecision should reduce the likelihood chata competitor can use the threat of a boycottclaim to reduce the ability of competingnetworks to differentiate themselves. Thismessage may he a little late in the brandedATM network markers where only onealternative exists.

~ ~c~”~i(

~“ ~ N

01in, 0’ 0, ~0’i”V 5: ruSeewe-er- ,Seefe e.*,Sa~55,e5e’S,er,

to 0” 0’-,” 55’ it’S it’ S 5N~e-r- ,~ ‘fl’S~_r-N %..._ ,,._1~

Modern antitrust analysis of jointventures since Broadcast Music has tendedto turn very heavily on market power con-siderations under the rule of reason.” Insimplest terms, this means that a smallnetwork without market power should heable to do chings that a big network withmarket pow-er would he prohibited fi’omdoing. The message has great practicalimportance as fesver and fewer ATMnetworks are becoming ever moredoaninant in their regional markets.

The message was reetnphasized lastyear when, after more than a generation ofsilence, the antitrust division reentered theEFT network enforcement market with asuit against the largest ATM network in theUnited States—MAC, operated by EPS.”Originally started by Philadelphia NationalBank (the most famous anticrcast defendantin the financial sector), EPS had become ajoint venture of four leading Pennsylvaniaand Ohio hank holding companies. It hadapproximately a 90 percent market sharein Pennsylvania and a strong position inadjacent mid-Atlantic states (as a restult, inpart, of prior mergers that DOJ had notchallenged). The DQJ complaint allegedthat EPS barred banks thatbelonged to itsnetwork From buying data processingservicefrom third parties and used its control overATM processing to prevent memberbanks from connecting with competingnetworks. It alleged violations underboth sections 1 and 2 of the Sherman

“See also letter from WilliamBaxter, assistant attorney gen-erol, antitrust division, toDonald I. Baker (Aug. 3,1983), discussed in Baker andBrandel (l995)~24.D6 [31.

“See BayBonks, Inc v. NewEngland Network, mc,Na. 86-3532-K (U.S. Districtcourt for the District ofMassachusetts) filed Dec. 9,1986). I was one of thedefendants’ attorneys in thisproceeding.

“Similar access issues hove beenraised by nonbank banks suchas Merrill-Lynch, AmericanExpress and Sears, seekingaccess to AIM networks. thesecases were settled with theadmission of the nanbankbanks to the network. SeeHousehold Bank, FS.B.Cirrus System, Inc., No.87(2353 [filed U.S. DistrictCourt for the Northern Districtof Illinois, Mar. 2, 19871;Household Bank, PSI. v.Money Station, Inc., No. C-2-88-D274 [filed U.S. DistrictCourt for the Southern Districtof Ohio, Mar. 2, 19881.

In 1993, BuyPnss Corp., theowner of the i~(AIM network.sued the NYCE AIM networkseeking access for its processingsubsidiary. The case wassettled with an agreement thathoth MAC nnd NYCE could actas a third-party processor ineach other’s network. SeeBuyPoss Corp. v. New YorkSwitch Corp., No. 93-C V-32D1[U.S. District Court for theEastern District of Pennsylvaniafiled June 15, 19931.

‘SCFCILC Inc v. VISA U.S.A.,

36 F3d 958 (10th Cir. 1994).

“Broadcast Musk, Inc u. CBS,

441 U.S. 111979).“United States x Electronic

Payments Services, mr., No.94-208 ID. Del. Apr. 21,

FEDERAL RESERVE BANK OF ST. SeIZES

11

DpdIII~NOVEMSER/DECEMSER 1995

19941,59 Fed. Reg. 24711

(May 12, 1994).

“504 U.S. 451 (1992).

First Texas SovingsAssn./Finanriol tntecchonge,Inc, 55 Antitrust g trade leg.Rep. 340 (1988) (Thornos E.Kauper, arbitrator).

“Valley Bank ofNevoda mc PlusSystems, mc, 749 F. Supp.223 (U.S. District Courtfor theDistrict of Nevada ID. Nev.)19891 affirmed 914 F2d1186(9th Cir. 1990)(Commerce Clouse lMgotion).

“Southmrust Corporotion v. PLUSSystem, CV-93-P-2291-S (U.S.District Court for the NorthernDistrict of Alahamu 1994).

Act. The tying violation alleged thatregional ATM network access and ATMprocessing were separate products andthat MAC’s rules and practices effectivelyforced its customers to purchase ATMprocessing from EPS. The monopolizationclaim alleged that EPS “willfully hasmaintained its monopoly power in themarket for regional ATM networkaccess in the affected states throughexclusionary practices.”

The consent decree required EPS toopen its network to independent ATMprocessors on a nondiscriminatorybasis. EPS was permitted to providevolume discounts for processing, butthese must he provided on a nondiscrimi-natory basis. FPS was also required tosell its network services “at prices thatwill not varywith the processor selected”and to provide a more open environmentfor third-party processors.

The ATM driving rule was only oneof a series of EPS rules that seemed tobe clear restraints on competition bothby third-party processors and othernetworks, including the national networks.The parallel of the challenged rule to theSupreme Court~sdecision in EastmanKodak v Image Technical Services, Inc. isobvious.” There Kodak tied equipmentservicing to its sales ol new equipment butprovided an exception for those who didself-servicing internally The SupremeCourt sustained the plaintiffs’ tie-in claimagainst the defendants’ summary judgmentmotion. Thus what we see in EPS/NationaiCity is DOJ singling out one particularrestraint and securing its elimination byconsent decree in the context ofsubstantial monopoly power in thebranded ATM network access market thatthe government alleged. -

One can anticipate more of this type oflitigation from the DOJ, private parties andprobably the state attorneys general. Anydominant network rule that discriminatesagainst a particular class of market partici-pants (for example, third-party dataprocessors) is an obvious target, as areprice discriminations and restraints by amonopoly network on participants using

competitive networks. All of these wouldseem to be classic actions of the typespunishable under sections 1 and 2 of theSherman Act when undertaken by adominant firm.

,r-iC,,;it,,-,r,t,tr’wt CoMhcto’ ,‘emc’nru

Nefwc-,-ksThe potential for conflict between

(1) a dominant network; and (2) networkparticipants, network competitors,network users or government enforcersfalls into a variety of categories.They include the following:

1. interchange fees and routing rules;2. direct customer charge on network

transactions, acquirer surchargesand issuer foreign fees;

3. routing rules;4. switch fees in for-profit networks;5. trademark usage rules and fees;6. processing rules (including those

related to third-party processors);7. use of nonbank cards and ATMs on

a network; and8. scope of network services.

Let us look at each of these in turn.

Interchange fees. ATM networks, likeother payment systems, have traditionallyprovided an interchange fee set by thenetwork to encourage activities that thenetwork believes need subsidizing. In anATM network, the interchange fee is paidby the card issuer to the ATM owner. Oth-erwise the ATM owner has no incentive toallow a foreign card in its machine, and thecard issuer has no guaranteed outlet for itscards. Stated more generally the ATM net-work interchange fee is designed toencourage acquirers to commit their ATMsto the network. ATM network interchangefees have been challenged at [east once bya private antitrust plaintiff in the 1988First Texas-PULSE arbitration as horizontalprice fixing or a card issuers’ cartel. Thearbitrator, however, held that the fee wasnot illegal as long as the individual ATM

FEDERAL RESERVE BANK OF ST. LOUIS

12

lltYI[~NOVEMSRR/DECRMEER 1995

owner was free to levy surcharges oroffer rebates.”

Regulating direct customer charges fornetwork transactions—ATM surchargesand issuer fees. Any ATM network has alegitimate (hut not necessarily unlimited)interest when participants levy any trans-action fees that may substantially reducenetwork volume. These fees fall into twocategories: surcharges charged by ATMowners and foreignfees charged by cardissuers. A network prohibition on individualpricing by participants is certainly subject toa price fixing claim and must be defendedby the network on the Broadcast Musicground that uniform pricing is necessaryto make the joint network service work.Interestingly, many ATM networks havesought to restrict surcharges, but apparentlyno network has yet dealt with the parallelsubject of issuer fees. The two seem to falllogically in tandem. Higher issuer feesmay cause issuance of additional cards,whereas surcharges may encourage deploy-ment of additional ATMs. They may ormay not generate additional networkvolume(or may reduce it) depending on the levelof the fees and consumers’ response tothem. To prevail in a challenge to asurcharge or issuer fee prohibition and topass muster under Broadcast Music, thenetwork will have to make a reasonableshowing chat the prohibited (or regulated)fee is likely to reduce network volumesubstantially or reduce the value of thenetwork trademark substantiallyLitigation to date has been indecisive.”We anticipate, however, that more isto follow”

Routing rules. ATM networks have alsodeveloped compulsory routing rules thattell acquirers (and sometimes the issuer,too) how to route every transaction.These two subjects are closely relatedbecause if both the ATM and the card areeligible to participate in two separate net-works, then the acquirer has every incentiveto send the transaction hack home bywhichever network has the higherinterchange fees. This is not necessarily a

form of competition that antitrust lawshould seek to encourage because theultimate effect may be higher charges toconsumers. By contrast, an issuer routingrule gives the card issuer, which pays theswitch and interchange fees, freedom todesignate the routing in such circumstances,presumably based on both feesand efficiency

In any event, serious antitrust concernsare raised if a monopoly network—or evena very strong one—insists that all transac-tions be routed by it wherever possible.This makes the creation of a new networkcompetitor very difficult indeed and shouldprobably be illegal on a tie-in or boycotttheory.” The lower courts have also gener-ally applied the rule of reason in casesinvolving boycott challenges to compulsoryrouting rules. The most illuminating is aD.C. Circuit decision by judges RobertBork and Ruth Bader Ginsberg, whoapplied a rule of reason analysis to routingrules in the household moving industryIn this case the defendant had a clear freeriding problem and lacked market power.”

Although ATM routing rules have gen-erated considerable controversy only oneprivate case, BayBanks, Inc. v. New EnglandNetwork, Inc., has challenged a networkrouting rule on antitrust grounds.” In thatcase, BayBanks, the operator of the XPress24 network in Massachusetts, sued theYankee 24 network and several New Eng-land banks that were instrumental inorganizing the network. Bay Bankschallenged a compulsory routing rule thatrequired the major organizing banks torun all the transactions between themthrough the network switch, rather than asubswitch. In March 1988, the partiessettled the case, thus leaving theseinteresting antitrust issues unresolved.

Switch fees. Obviously a network switchhas to he supported, and therefore a reason-able switch fee is easily within the BroadcastMusic standard. In the context of a nonprofitjoint venture, any switch fee presents littleproblem for antitrust. The situation isquite different where there is a proprietarynetwork with market power that is owned

“See Baker and Brondel 11988)§23.07 [3] [6] (Bypass andRouting).

“Rothery Storage & Von Co. v.Atlas Von Lines, 792 F.2d 710[U.S. Court of Appeals for theDistrict of Columbia Circuit(D.C. Cir. 1986) petition forcertiorari denied 479 U.S.1033 (1987). this caseinvolved a rule of Atlas Vonlines that required that anyAtlas order received by one ofits carrier agents be transportedunder the operating authority ofAtlas. The rule was generatedby deregulation of the movingindustry in 1979 that mode iteasier for the local cacreragents to obtain their owninterstate authority and com-pete against the national vonlines (such os Atlas). Thuscarrier agents could freeside onAtlas’ efforts while cuttingprices to attract business thatotherwise would have gone tothe von line. In response tothis threat, Afas imposed o rulethat required its carrier agents,if they chose to take their ownorders, to do so only through aseparately owned enterpriseusing its own operating outhodty;the new entity could not usethe facilities or services ofAtlas, nor could it use the Atlasname. Applying the rule of rea-son, the D.C. Circuit concludedthat the restraint was lawful—both because the defendantlacked market power andbecouse the rule was justifiedas a legitimate attempt byAtlas to eliminate free-riding.792 F.2dat229.

“No. 86-3532-K ID. Moss. filedDec. 9, 1986). As discussedbelow, the complaint also chal’lenged the trademark licensefee that Yankee 24 proposedto lew~on transactions thatcould have been routed to theYankee 24 switch.

FEDERAL RESERVE BANK OF ST. LOUES

13

DIIIE~NOVEMSER/DRCEMRRR 1995

See Oelowore & Hudson Ry,Co. sc ConsolidotedRoil Corp.,9D2 E2d 174, 179-80 (2dCit. 1 99D), cert. denied 50DU.S. 928 11991) (high mtetreoted as on access issue).

“See Baker and Rmndel (1995)¶24.D7[4](c).

“This practice was abandonedapparently as part of theBoyBonks settlement.

“See Radiant Burners, Inc. v.Peoples’ Gas Light & Coke Co.,364 U.S. 656 (1961).

“See Associated Press tc UnitedStates, 326 U.S. 1(1945)and its progeny. I have a lotofdifficulty with the sweep of theAssoriated Press rule. SeeBaker (1993), hereinafterCompulsory Acress.

United States TerminalRoilroodAssn., 224 U.S. 383(1912).

by a few major banking institutions (oreven a single one) but necessarily used byeveryone. In this context, the switch maylook like a monopoly toll bridge, and theswitch fee may be characterized as oppor-tunistic gouging, or at least an importantrevenue source for the shareholder-owners.In these circumstances, a very high switchfee could be credibly attacked by nonownerusers or DOJ under a horizontal pricefixing, monopolistic price squeeze or essen-tial facility theory under the Sherman Act.”Similarly a fee that discriminated stronglyagainst some group (for example, smallbanks or out-of-state issuers) could also besubject to a monopolistic abuse theoryunder section 2 of the Sherman Act. Asmore regional ATM networks becomemonopolies, their switch fees are likely tocreate a more pressing antitrust issue.

Trademark usage rules and fees. An ATMnetwork has to create a consumer productand establish the organization and infrastruc-ture necessary to make it work. It is logicaltherefore for the network to view its prin-cipal asset—namely, its service mark—asa device for supporting the promotionalcosts of the network.” Several servicemark licensing approaches have beentried, including a royalty on (1) everycard accessing the network, (2) everyATM accessing the network and (3) everynetwork transaction, regardless of whetherit uses the network switch.

The per-transaction royalty fee has hadan uneven history When ConnecticutSwitch expanded in 1986 to become theNew England Network and adopted a ser-vice-mark licensing fee of several cents pertransaction, BayBanks promptly sued, chal-lenging this fee as price-fixing andmonopolization.” By contrast, in the BPSconsent decree, DOJ allowed the defendantto reserve the right to charge a per-transac-tion trademark fee that was as high as itsswitch fee—a right that would obviouslydeter routing of transactions to otherswitches and thereby deter entry of anew network.

Even if a high per-transaction feeconstitutes an unreasonable restraint when

imposed by a monopoly network bent ondeterring new entry a competitive networkwith a valuable trademark should be ableto charge license fees on any of a variety ofbases. Competition from another networkshould ensure that such fees would nothe exorbitant.

Processing rules. A network needs tohave detailed specifications and rules tooperate. As long as these are objectiveor technically defensible, there should beno antitrust problems. However, the pro-cessing rules can raise the kinds of issuesthat arise in standards-making cases underthe antitrust laws. From time to time,certain products have been arbitrarilyexcluded from certification or the certifica-tion process has been skewed to favorcertain enterprises over others.”

Moreover, a processing rule thatrequires all ATM transactions to be drivenby the network switch is plainly bad ifimposed by a monopoly (or even a verypowerful) branded ATM network. This isexactly what DOj ordered eliminated inthe El’S/National City consent decree ontie-in and monopolization theories.

Nonbank ATMs. Increasingly supermar-kets and third-party processors are deployingATMs, and a variety of merchants aredeploying POS arrangements with cashback features that function similarly toATMs. Attempts by traditional bankingorganizations controlling a joint venturenetwork to exclude nonbank ATM deployersfrom the network will raise serious boycottquestions, if the network has a significantdegree of market power in its market.” Ina monopoly network situation the compul-sory access principles of St. Louis TerminalRailroad and its progeny may come intoplay as we shall see shortly”

Scope of network services. In the networkjoint venture context, intense disputes canexist over expansion of a joint venture’snetwork services. Often, the joint ventureoffers small institutions certain things thatlarge institutions can offer alone. Thistype of dispute is most likely to occur

FEDERAL RESEEVE RANK OF ST. LOUIS

14

lily IEt~NOVEMBER/DECEMBER 1995

openly in a nonprofit joint venture inwhich different sizes and types ofinstitutions have board representation.It has occasionally turned into an antitrustdispute. (My favorite example wasCitibank’s mid-1970s case against theMasterCharge joint venture to enjoin itfrom issuing traveler’s checks, a market inwhich Citibank—a leading MasterChargeissuer—was a major player. The casenever went to trial.)

Antitrust conflicts of the types justoutlined are inevitable in an environmentwhere diverse competitive interests mustdepend on a facility for which thereare no clear alternatives—especiallywhen it is controlled by a small group ofcompetitors. The dominant network isnecessarily subject to stricter antitrust ruleswhen aggrieved competitors have fewer net-work alternatives. By allowing dominantregional ATM networks, the DOJ and theFederal Reserve have simply created alarger field of antitrust risks for morenetworks. More business disputes havebeen taken out of the market and willbe switched—potentially—--- to thecourthouse, where treble damages andattorneys’ fees under the Clayton Actwill encourage private plaintiffs to frametheir claims in antitrust terms w-heneverpossible. We have seen this in a goodmany private antitrust cases challengingATM network practices during the pastdecade—including cases brought bylarge and aggressive banks againstnetworks. In such an environment,antitrust planning is a prudent exercisefor any important ATM network andits major participants. Most majorATM networks are largely controlled bytheir leading institutions, which issuethe bulk of the network cards anddeploy the bulk of ATMs. In someinstances (for example, EPS), the networkis a profitable enterprise owned by a few ofits largest members, which, as shareholders,elect all the directors and collect all theprofits. In others (for example, Pulse), thenetwork is a nonprofit entity with electionof directors on a town meeting basis,

but with guaranteed seats for certainfounding members.

Advc’nrs’Piunning byi’-<’eiw-orkst’c-Rec,h,re Antitrust Ri.sks

My suggestions for antitrust planningby a dominant network fall into threegeneral areas:

• More representative corporategovernance arrangements;

• Greater flexibility in use of thenetwork facilities or competitivealternatives by participants; and

‘Use of alternative dispute resolu-tion procedures for handlingantitrust disputes amongnetwork participants.

None of these steps depends on the others,but they work well as a package.

Corporate Governance. Whenever anetwork board of directors (or themanagement that it has elected) is obliged

to make a decision that affects differentnetwork participants in different ways, thedecision becomes easier to defend if allaffected interests were represented in thedecision making process. Thus, in principle,it is advisable to have small institutionsrepresented on the board or even to havepublic directors elected from outside theownership or network participants.”

Operational Flexibility. A dominant net-work should be careful that its technicalstandards are technically justifiable—andnot more restrictive of competition amongusers than technically necessary The net-work needs to he especially careful thattechnical justifications are not just a guisefor what is really a monopoly rentenhancement scheme.” Giving membersoperational flexibility to route transactionsover alternative networks is likely to he aparticularly important operational issue inthe case of regional monopoly ATM

“See Baker and Brandel (1995)124.08. Interestingly, whenthe New York Stack lxchnnge(a very dominant network)come under heavy antitrust andpolitical fire for its rate fixing onbroker cnmmissions in the late1 96Ds, the exchange begnnappointing distinguished publicdirectors, who did not representNYSlfirms,

See United States v. ElectronicPayments Services Inc 11994)and generally Baker andBrondel 11995)1124.07[21 and (31.

FEDERAL RESERVE BANK OF ST. LOUES

15

DF~Il~NOVEMEER/DECEMBER 1995

“Mitsubishi Motors Corp. c SolerChrysler-Plymouth, mc, 473U.S. 614 (1985).

“See, Baker and Stabile(1993).

There is one fomous exampleof an antitrust arbitration ofmajor ATM network dispute.The First Texos/FinonriolInterchange, Inc. Arbitrotiog55 ATRR 420 (1988). Iworked on it for the respondentnetwork.

“See Baker (1993) at 1020-25, discussing UnitedStatesTerminal RoilmodAssn., 224U.S. 383 11912).

“224 U.S. at 409.

“See Baker and Brandel(1995), ~j22.0314) [ci, dis-cussing U.S. Rocky MountoinAutomated flooring HouseAssn., CA. No. 77~39l(U.S.District Court for Colorado1977); and U.S.c CaliforniaAutomated Clearing HouseAssn., CA. No. 771 463-ITt(U.S. District Court for the NorthDistrict of Colifornini 977).

networks. This is an issue of not onlycomputer programs and switches, but alsofee structures for gateways and trademarkusage on bypass transactions.

Alternative Dispute Resolution. In 1985,the Supreme Court opened the door toarbitration of antitrust cases between thosein a contractual relationship.” As a result,an ATM network—even a monopoly—caninstitute a reasonable arbitration programas part of its membership and participationarrangements; it can even appoint a partic-ular arbitrator or panel to hear cases. Aslong as the arbitrator or panel is neutral,industry expertise can be injected into theprocess—and some greater certainty canbe generated in the context of proceedingsthat have a limited time for decision. Theprocess is likely to be successful in reducingthe cost and uncertainty of disputes only ifa lot of care goes into designing the processand selecting arbitration panels.”

Of course, arbitration only providesassistance where the antitrust dispute isamong a network and its participants.”By contrast, if the charge is discriminationagainst an outsider (for example, anonmember financial institution, anothernetwork or a third-party processor) thenthe antitrust case will go to the federalcourthouse. The same is true if thepractice is being challenged by DOJ or astate attorney general.

To conclude, any dominant ATMnetwork faces a real antitrust exposure; it

should do everything possible to avoidbeing arbitrary in its decisions or the wayit goes about making them.

CONCLUSWN: THE PASTAS PROLOGUE?

It is fitting—and indeed almostBiblical—that we should come to St. Louisfor today’s discussion. For it is out of yourhistory here that the next chapter of theATM network antitrust history may yetbe written.”

When the railroads came west, theycame to St. Louis, and this crucial crossing

point on the Great River became a vitalbridgehead and terminal. Some 24 railroadsconnected here. Through a series ofacquisitions, a smaller group of presum-ably richer railroads extinguished thecompeting ferry service and acquired allthe rail links to the two big bridges acrossthe Mississippi. Their joint venture com-pany (owned by 14 railroads) monopolizedthe East St. Louis traffic that had no alterna-tives, while being competitive for trafficthat could use the bridges upriver at Alton,Illinois or downriver at Memphis. TheTerminal Railroad Association of St. Louis(Terminal Company) discriminated againstnonowner railroads. Into this thicketcharged President William Howard Taft’sDepartment of justice and urged that themonopoly Terminal Company be brokenup into competitive pieces. It was asimpler age, and the DOj probablyproceeded without economic counsel,let alone expert witnesses armed withintriguing regression analyses.

The two district judges who heard theevidence here in St. Louis could not decidewhat to do and issued no opinion orfindings. The government appealed.

The U.S. Supreme Court thus becamethe court of first (as well as last) resort inthe case. In its celebrated TerminalRailroad Assn. decision of 1912, the courtdecided that breaking up the monopoly (asDOJ had urged) would he inefficient giventhe local terrain. Instead, the court decidedthat the Terminal Company should herestructured so that all market participantscould become owners—or, if they preferred,could be offered access to the terminal, asthe Court said, “on as nearly an equalplane as may he with respect to expensesand charges” as the owners.”

The Terminal Railroad Assn. principlehas been applied to a variety of regionalmonopoly facilities—including a fishmarket, several tobacco markets, sportsstadiums, and even a couple ofFed-supported regional automatedclearinghouse facilities.” It remains avital principle in today’s world of evermore sophisticated nec\vorks. Yet it isa second best principle: competition

FRDERAL RBSRRVE BANK OS ST. LOUIS

16

DF~IlWNOVEMBER/DECEMBER 1995

among networks is generally preferable Baxter, William F, Poul H. Cootner and Kenneth F. Scott. Retoil Bankingto having a federal court sit as a public in the Electronic Age: The low ond Economics of Electronic Fundsutility commission to restructure and Tronsfer(1977).supervise the terms of access to CB&T Bancshores, Inc., Federal Reserve Bulletin, vol. 70(19841,a network.” p. 589.

Terminal Railroad is still the best . -

- . . Centerre Bancorpnraton, Federal Reserve Bulletrn, vol. 69(1983),prtnciple that we have to deal wxth the sut- 643uation where natural monopoly economicsor government decisions based on natural Citicorp, Federol Reserve Bulletin, vol. 72(1986), p. 583.monopoly assumptions have created a Competitive Impact Statement in United Stotes v. Electronic Paymentsjoint venture monopoly for which there is Servkes Inc., 59 Federal Register (Fed. Reg.) 24711 (May 12,no likely substitute. Because the Board of 1994).Governors now’ seems to assume that . . -

- , - . - Interstate Frnoncral Corp., Federol Reserve Bulletin, vol. 69(1983),regional monopoly isa way of life in AIM 560networks, we can expect to hear a lot more -

about Terminal Railroad—as those who Sovcon Financiol Corp., Federol Resenie Bulletin, vol. 72(1986),have been excluded from ownership in the p. 347.monopolies seek o\vnership rights or a The Bank of New York Company, Inc., Federal Reserve Bulletin, vol. 3Dposition “on as nearly an equal plane” (19941, pp.1107-11.with those whom the Fed and the DOjhave authorized to own these vitalelectronic bridgeheads.

REFERENCES

Atlantic Bancocporation, Federal Reserve Bulletin, vol.69(19831,p. 639.

Baker, Donald I. “Compulsory Access to Network Joint Ventures Underthe Sherman Act: Rules or Roulette?” Utah low Revieri (1993),p. 999.

Baker, Donald I., and Roland F. Brnndel. The low of Electronic FundsTransfer Systems, revised ed. Worren, Gorham & Lnmont (1995).

Baker, Donald I., ond Stabile. “Arbitration ofAntitrust Claims;Opportunity and Hazards for Corporate Caunsel,” Business lawyer,vol.48(19931, p. 395.

Bonc One Corp., Federal Reserve Bulletin, vol. 81(1995), pp. 491-2.

Barclays Bonk PLC, Federal Reserve Bulletin, vol.71 (19851, p. 113.

See Baker (1993), especially1076-8.

FROBRAL RESERVE BANK OF ST. LOUIS

17