Electronic commerce revisited.
While Bill has had a very extensive practice in the area of electronic commerce, he has published only a few academic works--albeit substantial ones.
Electronic commerce in the sense I am using it is rather distinct from its contemporary meaning of all business transactions done over the Internet. Bill's publications have, more strictly, dealt with payments systems. But the U.S. payments system is not a minor or inconsequential business. In 1997 it comprised an estimated 650 billion payments, worth approximately $22 trillion.(1) Commerce over the Internet represented only 1/2 of 1% of all retail purchases, although it is expected to grow rapidly.(2) The revenues derived from the payments system were about $115 billion, shared among banks, third-party service providers, and technology vendors.(3) And revenues aside, the sheer volume of transactions makes the efficiency of the system important to the smooth functioning of the economy.
Bill's first contribution in this area was the outgrowth of a study project that he, the late Paul Cootner, and I undertook in 1975 for Citicorp which resulted in a book on electronic funds transfer.(4) Following that book in 1983, Bill authored a related law review article on the economics and history of the exchange and clearance of checks, credit card charges, and debit card charges.(5)
My focus will be on the book, which was an attempt to look forward from 1975 to the potential future of electronic payments systems in light of the economic forces and legal constraints affecting them as well as to offer some policy recommendations. In collaborating on the book, while of course we criticized each other's drafts, I wrote the portions on bank and consumer regulation, Paul did the models of bank and consumer costs and retail selling, and Bill was concerned with the economic and antitrust issues for ATM and point of sale (POS) networks. For this occasion, the latter will be my primary concern.
The perspective of the book was that electronic funds transfer (EFT) issues had to be analyzed in terms of the existing competition among cash, checks, credit cards, and newly emerging debit cards, for consumer use in purchase payments, and how that competition would be affected by advancing technology that enabled immediate access over telephone lines to vast amounts of data on consumer accounts. In particular, technology was making feasible the creation of networks of POS terminals. Sitting on retail counters next to the cash register, these terminals were directly connected to banks that had previously issued the customer a credit card or held his deposit account. These networks were potentially going to reduce greatly the cost of transacting, for both the merchant and the customer as well as the bank. Networks would play an increasing role in the future payments system. Debit cards would experience significant growth, displacing cash and also checks.
A central issue of the book was whether these nets would develop as natural monopolies, to be regulated by the government like public utilities, or as rivalrous firms, regulated by market forces. An attempt to answer that question required, first, an estimation of the economies of scale in network operation and the demand for terminals. That involved a series of factual inquiries. At what size did the long run average cost curve for a network become nearly flat? At that scale, what would be the annual cost of a terminal? What would be the savings per transaction it would produce? What annual sales volume would be required to generate enough transaction savings to cover the terminal cost? How many retailers have annual sales sufficient to employ one or more terminals? How many terminals, under such conditions, might be installed in each state? How many efficient nets could that number of terminals sustain?
Obviously, all of those inquiries were answered on the basis of then available data and cost figures, and with the aid of assumptions intended to be conservative. To simplify, Bill's conclusion was that, for the nation as a whole, there could be from 2.5 to 5.5 million terminals, and from 40 to 230 nets, creating a competitive environment.(6) But those numbers viewed the nation as a single market, and if an individual bank owned and ran the network, bank regulation would come into play.
Even where there is no natural monopoly, if such exist, one can be created by law. Banking law in the mid-1970s confined a bank's branches to its home state, or even more narrowly, and, in thirteen states, permitted no branches at all.(7) An ATM terminal, or POS debit card, accesses a customer's bank account. Did that make the terminal a "branch" of the bank? If a bank's ability to deploy ATM or POS terminals was constrained by treating them as branches, in most instances the bank could not fully serve the potential local market and the attainable number of efficient nets would decline. Unfortunately, that was the position that the federal courts adopted in 1976, as to national banks' ATMs.(8) So in many states, a single bank would not be able, by itself, to form an efficient net. It would have to be done by a multibank association or a nonbank operator dealing with banks as members or customers.
If the nets involved multiple banks, to what extent would there still be competition? On the state level, legislatures for the most part did not treat terminals as branches but often did require that the network be shared with any other financial institution requesting access.(9) The consequence of mandatory sharing is that it facilitates the development of an early or leading network, and makes it harder to form a viable rival. Thus we viewed mandatory sharing statutes as unwise economic policy that would ultimately lead to situations of unnecessary monopoly.(10)
So much for the two page summary of a two hundred page book. The analysis was actually a bit more complicated and nuanced than I have just made it sound. How was it received, and what impact did it have?
I have gone back and looked at some of the reviews it received, an activity that did not engage me at the time. The longest by far--it was almost half the length of the book--thought it was absolutely terrible.(11) The author thought our methodology was all wrong because we assumed consumers would act to economize on costs and maximize benefits, where costs include both explicit charges and the opportunity cost of time spent on effecting transaction payments.(12) This "sterile" model led us to conclude that EFT systems would afford significant savings and be advantageous to consumers, whereas there was overwhelming evidence that consumers were well satisfied with the existing currency and check system but would be hostile to the EFT system and resist it.(13) The "central failure" of our report, therefore, was its prediction of wide acceptance and growth-of EFT systems.(14) The author's own "richer model" of human behavior, derived from Brehm, Dostoevsky and Sartre, led him in the opposite direction.(15) As may be apparent, the review was in large measure simply an ideological attack on law and economics, especially in the regrettable Posnerian tradition. But it did have an empirical side, generating sharply contrasting predictions about consumer reactions to electronic payments systems and their future growth.
Our analysis fared somewhat better at the hands of the National Commission on Electronic Fund Transfers, created by Congress to study the policy questions posed by EFT and provide recommendations.(16) The Commission's 1977 Report agreed, after examining network cost data, that POS systems did not exhibit the characteristics of a natural monopoly calling for public-utility-type regulation.(17) It agreed that the deployment of EFT terminals should not be regulated as if they were branches, and recommended legislation to authorize national or state banks to establish ATMs on an expanding basis--first state-wide and then across state lines in natural market areas. POS networks serving debit and credit cards should not face restrictions on terminal deployment.(18) The Commission also recommended Congressional legislation to nullify state mandatory sharing laws, leaving questions of access or exclusion to existing antitrust standards.(19) In some sense, the Commission's Report was the best review we received.
But the book, the criticisms, and the Commission's report all appeared two decades ago. The advantage of an occasion like this is that one can now bring hindsight to bear, to see how EFT systems actually turned out. (That is one of the risks of engaging in forecasts early in one's career.)
The evolution of the legal environment was not as favorable to network competition as we would have wished, although it has slowly moved in that direction. As already mentioned, the courts said bank-owned ATM terminals were branches under the National Bank Act, and that position stood until its repeal by Congress in 1996.(20) The consequence was that an ATM in a network would often have to be owned by some other entity. In that case, its access to the bank would not transform it into a branch.(21) This gave impetus to the formation of joint ventures by groups of banks. The joint venture, or an independent firm, would own the off-premises ATMs and fund the system through transaction charges to the banks sharing its use.
Congress did not enact compulsory sharing legislation, but neither did it follow the National Commission's recommendation and preempt state sharing statutes, which had been adopted in a mandatory form in fourteen states.(22) That inaction left the outcome in any particular state to be determined by an interaction between the law of that state and federal antitrust law. The federal antitrust doctrines governing illegal boycotts or access to essential facilities have never been clear(23) and, as applied to EFT networks, the Department of Justice (DOJ) has not been of much help.
The most celebrated illustration of the lack of DOJ guidance is Worthen Bank & Trust Co. v. National BankAmericard, Inc.(24) Visa and MasterCard developed out of local proprietary bank credit card systems to become national competing joint ventures, organized as non-profit corporations owned by their members. Card-issuing banks could belong to one system or the other, but not both. The Worthen Bank, a Visa member which wanted to issue both cards, challenged Visa's exclusivity bylaw as a group boycott illegal under Section 1 of the Sherman Act. Worthen won summary judgment in the district court as a per se violation, but the Eighth Circuit reversed for trial under the rule of reason.(25) Contemplating the possibility of treble damages, Visa settled and then sought DOJ letter clearance for a bylaw prohibiting all dual membership. After pondering for a year, the DOJ declined, because it was uncertain whether the bylaw would be procompetitive or anticompetitive.(26) Visa then dropped all restrictions on duality, banks rushed to join both POS systems, membership became substantially identical and so did the card products. Banks compete vigorously to issue cards or sign up merchants, but competition between Visa and MasterCard is muted.
The pattern was repeated in the context of ATMs, which often began as competing local networks which then combined, either through direct merger or the prospect of antitrust litigation. Admittedly, there are difficult judgments between permitting an efficient scale of operations to be reached and allowing acquisitions beyond that point that lead to creation of monopoly power. One example is the history of PULSE in Texas. One network (MTECH) had been built by a major bank and another (PULSE) was a joint venture formed in response by other Texas banks for non-MTECH participants. First Texas Savings Association, which was in MTECH, embarked on a program of ATM deployment and demanded admission to PULSE, arguing that exclusion would be an illegal boycott. PULSE sought the views of the Antitrust Division of the DOJ, and the Assistant Attorney General (one W.F. Baxter) replied that there would be no objection to admission. PULSE dropped its rule against duality, and in effect the two systems became one.(27)
So the legal evolution has been less than optimal from the standpoint of attaining competition, innovation, minimum costs, and maximum usage. Nonetheless, where have we arrived? In 1976, cash was used in about sixty-six percent of all payment transactions, but they were mostly small. Some twenty-eight billion checks accounted for over ninety percent of dollar payments.(28) Credit cards represented 5 billion transactions with $71 billion in value. Debit cards were essentially nonexistent.
There were at the time many predictions, with which we disagreed, that we were on the verge of an alarming cash-less, check-less society, but we have yet to go over that verge. Today cash and checks are still used in perhaps ninety-five percent of all payment transactions(29) but constitute only seventy-seven percent of total dollar value. The greatest growth (in dollar volume terms) has come in credit cards, which doubled from ten to twenty percent of the total between 1975 and 1995, reducing the share of checks accordingly (Figure I). Cash held its own, and debit cards have made hardly a dent. But debit cards have finally started to take off, particularly at gas stations and grocery stores (Figure II), so the prediction is still one of accelerating growth and a sharp decline in the use of cash.
[Figures I-II ILLUSTRATION OMITTED]
The technology continues to drive the process. For example, cards are moving from magnetic strips to integrated chips, which can store far more information. That makes possible the all-purpose card, which can perform credit, debit, ATM, check authorization, stored value functions, and contain additional information--if, of course, the terminal has a chip reader that can handle all that. And Internet shopping will become the next setting for payments transactions. The cost and value analysis in the book still affords, I believe, a logical and comprehensive structure for thinking about whether and how quickly these potentials are likely to be realized.
When it comes to network competition, the story is complicated, and not easy to depict or measure. There are 2 national and 43 regional POS and ATM networks that link over 15,000 financial institutions and over 500 million cardholder accounts with over 4 million merchants (Figure III). Despite constantly advancing communication technology that has been lowering costs and enlarging geographical markets, then, the number of regional nets has come in at the low end of the range Bill gave. I have not been able to obtain data that measure accurately the extent to which these nets overlap or directly compete with one another, but it is substantial (Figure IV).
[Figure III ILLUSTRATION OMITTED]
The networks are comprised numerous players focusing on different geographical regions of the U.S.
Top Debit Networks Network Network Market Membership(*) Interlink (Visa) 50 States 1,226 Maestro (MasterCard) 50 States NA Mac/EPS 40 States 2,716 Star/Explore 50 Stales 978 NYCE 10 Northeast States 1,316 Honor 18 Southeast States 2,730 Pulse 8 Midwest States 2,021 Network 1996 POS Transaction Volume Interlink (Visa) Maestro (MasterCard) 170,000,000 Mac/EPS NA Star/Explore 96,000,000 NYCE 51,648,858 Honor 60,000,000 Pulse 86,000,000 56,000,000
Source: Debit Card Directory, 1997, Faulkner & Gray, Card Industry Director 1999 Edition
(*) Bank, Savings & Loans, and Credit Unions
On the national level, there are only two major POS networks, Visa and MasterCard, each of which has an ATM subsidiary (Plus and Cirrus). In 1997, Visa accounted for fifty-three percent of general purpose cards in the U.S. and fifty percent of the dollar volume; MasterCard for thirty-three percent in number and twenty-five percent in volume. American Express (five percent of number and eighteen percent of volume) and Discover (8.5 percent of number and six percent of volume) make up substantially all of the balance.(30)
Since the Worthen case, as already mentioned, direct competition between Visa and MasterCard has been modest at best. But even if the Antitrust Division has come to appreciate the error of its prior ways on duality and its anticompetitive effects on innovation and efficiency, it would be hard to restore the long-gone separateness of membership. So the DOJ has just filed a Sherman Act Section 1 suit seeking separateness of governance. To be on the board of directors or committees of Visa or MasterCard, a bank would have to be "dedicated" to that brand.(31) It would be interesting to know how the Division believes the choice process would work, when one brand has twice the size of the other. The suit also seeks to enjoin Visa and MasterCard from preventing member banks from also issuing Discover or American Express or other competing cards.
In conclusion, the intervening almost quarter century has not brought us as much network competition as seemed possible in 1975. But the general mode of analysis, as opposed to the specific numerical outcomes, has stood the test of time reasonably well, at least in my view--a verdict with which I suspect Bill may agree, and for which he deserves much credit.
(1.) See BANK ADMINISTRATION INSTITUTE, PROFITING FROM CHANGE IN THE U.S. PAYMENTS SYSTEM 2 (1998) [hereinafter BAI].
(2.) See id. at 3-4.
(3.) See id. at 6 (listing the stakeholder shares of the payments business).
(4.) See WILLIAM F. BAXTER, PAUL H. COOTNER & KENNETH E. SCOTT, RETAIL BANKING IN THE ELECTRONIC AGE: THE LAW AND ECONOMICS OF ELECTRONIC FUNDS TRANSFER (1977).
(5.) See William F. Baxter, Bank Interchange of Transactional Paper: Legal and Economic Perspectives, 26 J.L. & ECON. 541 (1983).
(6.) See BAXTER, supra note 4, at 114-16 (discussing the geographic scale of network operations).
(7.) See id. at 175 (discussing the geographic limitations that apply to branch banks).
(8.) See Independent Bankers Ass'n of America v. Smith, 534 F.2d 921, 951-52 (D.C. Cir. 1976), cert. denied, 429 U.S. 862 (1976) (affirming that all customer-bank communication terminals were branches).
(9.) See BAXTER, supra note 4, at 127 (describing the provision for mandatory sharing).
(10.) See id. at 145 (listing reasons why these statutes would be anticompetitive as well as slow the innovation and deployment of EFT systems).
(11.) See Edward A. Dauer, The Policy Implications of Neutral Scholarship: A Case Study of Electronic Funds Transfer and the Baxter, Cootner and Scott Report, 2 CARDOZO L. REV. 397 (1981).
(12.) See id. at 408 (proposing that this model of consumer behavior ignores components of human behavior).
(13.) See id. at 420-21 (indicating that there is overwhelming evidence of consumer hostility towards the institution of EFT systems).
(14.) See id. at 442 (pointing out that our sanguine view of an avid consumer response to EFT systems seemed to be false).
(15.) See id. at 424, 442 (proposing that people stubbornly resist what is often in their best interest).
(16.) See Act of Oct. 28, 1974, Pub. L. No. 93-495, tit. II, 88 Stat. 1500, 1508-11 (codified as amended at 12 U.S.C. [subsections] 2401-2408 (1994)).
(17.) See NATIONAL COMMISSION ON ELECTRONIC FUND TRANSFERS, EFT IN THE UNITED STATES: POLICY RECOMMENDATIONS AND THE PUBLIC INTEREST 163 (1977) [hereinafter NC REPORT].
(18.) See id. at 83-87.
(19.) See id. at 97-98.
(20.) See 12 U.S.C. [sections] 36(j) (Supp. 1997).
(21.) See Independent Bankers Ass'n of N.Y. State, Inc. v. Marine Midland Bank, 757 F.2d 453, 459-63 (2d Cir. 1985) (holding that ATMs owned and operated by food stores did not constitute unauthorized branch banking), cert. denied, 476 U.S. 1186 (1986). The prohibitions on interstate branching in general were greatly attenuated by the Riegle-Neal Act of 1994 authorizing interstate bank mergers and permitting interstate branching on a state opt-out basis. See Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994, Pub. L. No. 103-328, tit. I, 108 Stat. 2338, 2339-68 (amending scattered sections of 12 U.S.C.).
(22.) See NC REPORT, supra note 17, at 92.
(23.) See BAXTER, supra note 4, at 138-42 (raising interpretive questions of the sharing statutes).
(24.) Worthen Bank & Trust Co. v. National BankAmericard, Inc., 345 F. Supp. 1309, 1322 (E.D. Ark. 1972) (holding that the bylaw imposed a horizontal restraint on trade and commerce and constituted a per se violation of the antitrust laws), rev'd 485 F.2d 119 (8th Cir. 1973), cert. denied, 415 U.S. 918 (1974).
(25.) See Worthen Bank & Trust, 485 F.2d at 126.
(26.) See Phillip Brooke, Justice Dept. Warns NBI Dual Proposal May Bring Suits, AM. BANKER, Oct. 16, 1975, at 1.
(27.) See Donald I. Baker, Compulsory Access to Network Joint Ventures Under the Sherman Act: Rules or Roulette?, 1993 UTAH L. REV. 999, 1063-64.
(28.) See NC REPORT, supra note 17, at 333-45. These figures do not include interbank networks (Fed Wire and Bank Wire) and automated clearing houses.
(29.) See BAI, supra note 1, at vii.
(30.) See Complaint at 9-10, United States v. Visa U.S.A., Visa Int'l Inc., and MasterCard Int'l Inc. (S.D.N.Y. Oct. 7, 1998) (No. 98-7076).
(31.) See id. at 41-42.
Kenneth E. Scott, Ralph M. Parsons Professor of Law and Business, Emeritus, Stanford Law School. I would like to express my appreciation to Mr. Bennett Katz and Visa International for the figures used in this article.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Baxter Symposium|
|Author:||Scott, Kenneth E.|
|Publication:||Stanford Law Review|
|Date:||May 1, 1999|
|Previous Article:||Bill Baxter in the antitrust arena: an economist's appreciation.|
|Next Article:||The learned intermediary doctrine and patient package inserts: a balanced approach to preventing drug-related injury.|