The Regulation of Interchange Fees: Australian Fine-Tuning Gone Awry


THE REGULATION OF INTERCHANGE
FEES: AUSTRALIAN FINE-TUNING GONE

AWRY

Richard A. Epstein*

I. Introduction ................................................................ 551
II. Methodology and Conclusion ..................................... 556
III. Interchange Regulation ............................................. 562

A. M erchant Choices ................................................ 562
B. The Not So Subtle Credit Card Benefits ............ 568
C. The Impact of Interchange Fees ......................... 571

1. Debit Instrum ents .......................................... 571
2. Closed-System Payments ............................... 575
3. Distributional Consequences ......................... 579

IV. The No Surcharge Rule .............................................. 582
V. The Uses and Limits of Credit Card Regulation ..... 586
VI. Antitrust Im plications ............................................... 591
V II. C onclusion .................................................................. 595

I. INTRODUCTION

This symposium provides a notable addition to the
extensive literature that has developed on the operation of
two-sided markets. Two-sided markets are generally defined
as markets in which the value attributed to the goods and
services received by parties on one side of an exchange
depends not only on the intrinsic properties of those items,
but also on the number of parties located on the other side.
In the credit card context, which I shall address here, the
story plays out as follows: merchants accept credit cards only

* James Parker Hall Distinguished Professor of Law, The University

of Chicago; The Peter and Kirsten Senior Fellow, The Hoover Institution;
Affiliate, LECG. The research for this paper was conducted with the
financial support of Visa. My thanks to Tom Brown of Visa for his
constant advice and to Howard Chang and David Evans of LECG for their
continued involvement in my work. The views expressed herein are my
own.



COLUMBIA BUSINESS LA W REVIEW

if customers are willing to use them to make purchases.
Customers use credit cards only if merchants will accept
them. Today, there also exists extensive literature
describing the theoretical features of these markets, which,
as this Article will explain, do not operate with the
matchless efficiency of ordinary competitive markets.' Given
this starting point, the question of industrial organization
faced in two-sided markets is that of selecting the second
best option. Because the ideal solution cannot be achieved,
we must locate the closest approximation to that ideal that
can be attained under practical circumstances. Stated more
generally, this inquiry can be reduced to the question of
whether state regulation can improve the operation of two-
sided markets.

Fortunately, we do not have to examine two-sided
markets solely from a theoretical perspective. We have
available at least one real world example, namely the
interchange fee restrictions that the Reserve Bank of
Australia ("RBA") imposed on the credit card industry in
July of 2003.2 In a report entitled Reform of Credit Card
Schemes in Australia, the RBA implemented a series of
recommendations that had been announced in its
comprehensive 2001 regulatory initiative.'

1 William F. Baxter, Bank Interchange of Transactional Paper: Legal
and Economic Perspectives, 26 J.L. & ECON. 541 (1983).

2 For a more detailed discussion of the RBA's regulatory initiative, see

Howard H. Chang et al., The Effect of Regulatory Intervention in Two-
Sided Markets: An Assessment of Interchange-Fee Capping in Australia,
REV. OF NETWORK ECON. (forthcoming 2005), available at
http://papers.ssrn.com/sol3/papers.cfm?abstractid=820044 [hereinafter
Chang, Australian Interchange].

3 RESERVE BANK OF AUSTRALIA, REFORM OF CREDIT CARD SCHEMES IN

AUSTRALIA I (2001), available at http://www.rba.gov.au/PaymentsSystem
Reforms/CCSchemes/IAConsultDoc/index.html [hereinafter RBA,
AuSTRALIAN REFORM]. The RBA relied heavily on Michael L. Katz,
Reserve Bank of Australia, Network Effects, Interchange Fees, and No-
Surcharge Rules in the Australian Credit and Charge Card Industry, in
REFORM OF CREDIT CARD SCHEMES IN AUSTRALIA II (2001), available at
http://www.rba.gov.au/PaymentsSystemi/Reforms/CCSchemes/IICommissio
nedReport/index.html.

[Vol. 2005



To assess this report and the regulations it has spawned,
it is useful to review the basic features of the credit card
system. The credit card system is one component of a larger
payment system, which includes cash, checks, debit and
credit cards, and charge cards-the last of which provide
credit only for purchases made from the business that issued
the card.4 Bankcard, an Australian operation, as well as
Visa and MasterCard, employ what are known as open credit
card systems, which stand in contrast to closed credit card
systems, both of which are explained below. The RBA's key
initiatives focused on two essential features of the open
credit card systems used by Bankcard, Visa, and
MasterCard. The first of these was the regulation of
interchange fees. The second was a contract provision that
prevents merchants in these systems from offering lower
prices to their cash customers.

In order to assess these two initiatives, it is helpful first
to understand how the closed, or integrated, systems of
American Express and Diners Club operate. Both of these
companies operate simpler three-party, or standalone,
systems involving cardholders, card companies, and
merchants. In closed systems, the credit card company acts
as the sole intermediary between its cardholders and the
merchant base, and it profits by retaining some portion of
the payment from the cardholder to the merchant in
exchange for its services. On the other hand, the more
complex four-party systems employed by Visa and
MasterCard operate a comprehensive network that consists
of a large number of member banks. Therefore, these
systems need to employ the services of two banks instead of
a single company intermediary. The four roles in these
systems line up as follows: the cardholder receives the card
from the issuing bank, which transacts with the acquiring
bank, which in turn has connections with the merchant,

4 For accounts of these systems, see DAVID S. EVANS & RICHARD

SCHMALENSEE, PAYING WITH PLASTIC: THE DIGITAL REVOLUTION IN BUYING

AND BORROWING 1-21 (2d ed. 2005).

No. 3:551] THE REGULATION OF INTERCHANGE FEES



whose customers pay for their purchases with either credit
or debit cards.

Visa and MasterCard serve as cooperative or coopetitive
arrangements of the various member banks that perform
recordkeeping and orchestrate the interactions between the
issuing and the acquiring banks. Using the term
"coopetitive" highlights that each system's member banks
must cooperate in setting standards for credit and debit card
transactions, even as they simultaneously compete with each
other to sign up consumers and merchants on either end of
the market.5 Their "interchange fees" refer to the amounts
that issuing banks collect from acquiring banks for
processing a credit transaction, which the acquiring banks
then pass along to merchants as additional service charges.
Functionally, the interchange fee is usually deducted from
the total amount owed from the issuing bank (for its
customer) to the acquiring bank (for its merchant) for the
relevant transaction. Before Australia enacted its regulatory
scheme, the interchange fee for open credit card systems was
set at 0.95% of the value of a transaction. The RBA's first
initiative reduced that figure to 0.55%, a decrease of
approximately 43% relative to the original charge.

The RBA's second initiative, dealing with price cuts for
non-credit card users, is of lesser import than the first.
Typically, contract terms in open credit card systems
prohibit the merchant from offering any discounts to those
customers who prefer to pay by check, cash, or debit card. In
effect, these terms prevent merchants from inducing
customers to favor one of these three modes of payment, each
of which the RBA found to be cheaper in total cost than
credit transactions.

Ultimately, these regulatory interventions are intended
to address what the RBA claims are efficiency and
distributional concerns, thereby reducing the implicit
subsidies that cash, check, and debit card users pay for
credit card users. The RBA's program is explicitly designed
to improve the position of merchants (and their broad

5 Id. at 161-84.

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



THE REGULATION OF INTERCHANGE FEES

customer base) and acquiring banks relative to that of
issuing banks and their cardholders (which are only a subset
of all customers). This Article shall examine whether
implementing these reforms has allowed the RBA to make
good on its objectives. In evaluating the outcome, I do not
wish to reexamine the detailed empirical issues raised by the
RBA's initiatives. For these purposes, I am quite content to
rely on three main sources. The first is the empirical
evidence assembled by Howard H. Chang, Daniel S. Evans,
and Daniel D. Garcia-Swartz.6 The second is testimony
offered by Ian Macfarlane, Governor of the RBA, in defense
of the RBA's 2001 initiatives.7 The final source consists of
the various submissions made to the RBA on these matters.8

That evidence provides a useful framework with which to
reexamine the conceptual arguments that led the RBA to
adopt these proposals in 2001, in light of their impact over
the two years.

Accordingly, the remainder of this paper is divided into
seven parts. Part II sets out the basic intellectual
framework for examining the RBA's arguments in defense of
its reforms. Part III applies this framework to the
arguments that the RBA made in support of capping
interchange fees. Part IV examines the RBA's prohibition on
credit card surcharges. Part V examines the use and limits
of credit card regulation. Part VI offers preliminary remarks
on the relative merits of regulation and antitrust litigation
for controlling the use of credit cards, should such controls be
necessary. Finally, Part VII offers a brief conclusion.

6 See generally Chang, Australian Interchange, supra note 2.

Review of the Reserve Bank of Australia's Annual Report 1997-98:
Hearing Before the House of Representatives Standing Committee on
Economics, Finance and Public Administration (2005) (statement of Ian
Macfarlane), available at http://www.aph.gov.au/HANSARD/reps/
commttee/r2365.pdf [hereinafter Macfarlane Testimony].

' See generally Reserve Bank of Australia, http://www.rba.gov.au/
PaymentsSystem/Reforms/CCSchemes/index.html (last visited November
1, 2005).

No. 3:55 11



II. METHODOLOGY AND CONCLUSION

My basic thesis runs as follows: the RBA reforms have not
delivered their promised benefits because they are incapable
of doing so. This conclusion depends as much on the
underlying intellectual framework as on the particulars of
the reform effort. Quite simply, payment systems can never
achieve the results attainable in perfectly competitive
markets, so the best option under the circumstances is the
one that best minimizes social losses. In light of this caveat,
any evaluation of the RBA's reforms must be nuanced
enough to ask this question: are the benefits that result from
the regulations large enough to overcome the obstacles to,
and costs of, achieving the stated ends? The key feature of
the payment card industry is that the demands of network
formation require a high industry concentration within
limited affiliated groups. Even when the issuing banks do
not collude with each other, prices on either side of the
market could easily exceed marginal cost. But owing to the
interdependence between both sides of the market, there
exists no effective remedy for this imperfect state of affairs.
Direct systems of price regulation create as many distortions
as they remove. Moreover, antitrust or competition policy
cannot fashion a useful remedy for an industry that cannot
operate in pure competitive fashion. In the world of second
best options, it is unwise to invest limited resources in
proposing and resisting regulation, where they yield a low
rate of return.

This conclusion is not based, either largely or exclusively,
on any appeal to public choice theory, which suggests that
political factions can capture public administrative bodies.
Even though key interest groups lobbied the RBA hard for
their preferred positions, there is no real evidence that the
RBA did anything other than seek to maximize its vision of
social welfare. Although merchants as a class benefited from
this proposal, the reforms were not promoted as an effort to
achieve that partisan end. Rather, the RBA seems to have
followed the highest standards of public responsibility in
promulgating these regulations. Nevertheless, the RBA's

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



THE REGULATION OF INTERCHANGE FEES

proposal still lacks the intellectual and institutional
soundness to make it work.

The key intellectual objection to the RBA's reforms
stresses the limited potential for social gains under this
regime. Because the RBA chose to regulate those credit
arrangements that are both "consensual" in origin and
"durable" in effect, it effectively placed a sharp upper
boundary on the potential for social gain under its
intervention.9 Consensual transactions routinely describe
those situations in which all parties to the arrangement are
better off than they were before entering into the
arrangement. 10 These payment transactions therefore differ
sharply from situations where the "unilateral" actions of one
party have negative external effects upon a second party,
such as in instances of aggression or pollution. Moreover,
"durable" transactions eliminate certain risks that often
infect voluntary transactions, including mistake,
nondisclosure, and misrepresentation. While these risks
could be present in one-shot transactions (e.g., the private
sale of a used car), they are not likely to exist among repeat
players, especially the sophisticated parties involved here.

In the payment industry, for example, the feedback
mechanisms on information are exceedingly strong because
the entire system can operate only if the parties preserve
complete and accurate records of all transactions and share
this information with both cardholders and merchants on a
routine basis. While the element of surprise may infect some
individual transactions, it is not the norm in any situation
involving either checks or payment cards." To be sure, most
individuals do not have perfect knowledge of the full range of

9 This conclusion holds whether such changes result from direct
regulation or from private litigation, but particularly under the antitrust
laws in the United States or the competition law, as it is described,
elsewhere in the world.

10 Baxter, supra note 1, at 552-56.
" A truth in lending remedy may prove appropriate in these cases.

For a useful discussion concerning the topics of misrepresentation and
disclosure, see Rossman v. Fleet Bank Nat'l Ass'n, 280 F.3d 384, 389-91
(3d Cir. 2002).

No. 3:551]



available card alternatives, even though competitors will
constantly try to place better terms in front of them.
Nonetheless, these individuals do have enough information
to be sure that they are not made worse off by their
particular pattern of doing business than they would have
been if they had not used any credit device at all. The
purpose of the RBA's regulation, therefore, was not to
prevent negative sum games from emerging among the four
parties to the standard open-system credit transaction.
Rather, its goal was to improve the outcome of positive sum
games in which all individual players shared at least some
fraction of the gain under the previous institutional
arrangements.

In principle, it is always possible to conceive of some
regulation that can improve upon the status quo, but, as
chess players concede, it is much more difficult to find that
winning move on the board. In particular, there are two
systemic obstacles that make this endeavor more difficult
than would otherwise be the case: administrative costs and
error costs. The first obstacle is that the proposed system or
regulation always adds additional costs. At a minimum,
these include the costs of deciding what should be done.
Moreover, within the modern administrative state, new rules
cannot be imposed by fiat, but instead require a period of
reflection, research, comment, and debate before
implementation, and in many cases, cry out for constant
revision thereafter. Agencies justify this elaborate ritual as
a means to acquire the information necessary to prevent
their regulations from being, at best, hit-or-miss affairs.
Even if the RBA's relatively simple price and term
restrictions come only on the heels of this lengthy process,
they are surely less costly than those setting out, for
example, the requirements for workplace or environmental
safety. In light of the extensive commentary and debate, the
total costs have likely run into the millions of dollars on a
cumulative basis, in what has proven to be an ongoing

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



THE REGULATION OF INTERCHANGE FEES

project.12 These costs are borne in part privately, by the
firms that participate in the administrative process, and in
part publicly, by all of the citizens whose tax dollars support
these initiatives. Given this first obstacle, regulators should
not pursue a program unless they have some reason to
believe that the total gains realized therefrom are greater
than the full administrative costs required to implement it.

On the other hand, evaluating the potential error costs of
a proposal leads one to favor the status quo even more than
evaluating the potential administrative costs. The success of
a proposal depends critically on whether its ultimate design
takes into account all of the relevant variables. That task is
not easy to discharge because any proposed regulation,
regardless of its superficial simplicity, will have ripple effects
on multiple margins. More concretely, any regulatory
initiative could be said to promise some form of social gain if
the only consequences that mattered were those that fell
within the terms of reference in the original study. But it is
one thing for regulatory review to consider a limited set of
consequences; it is quite another to be confident that the
regulator has included within the frame of reference all of
the economic consequences that matter. As I will explain,
one key margin on which the RBA's proposal should be
evaluated is its impact on the competitive balance between
open and closed systems, which the RBA has systematically
and incorrectly refused to undertake.

In conventional academic terms, the basic challenge of
evaluating regulatory reform requires a close look at the
persistent tradeoff between validity and reliability. The
study with the greatest validity seeks to examine the impact
of a regulation along as many margins as possible. For
example, a study of the minimum wage law would be
insufficient if it only considered possible short-term
decreases in employment levels. It would also be necessary
for the study to contemplate the impact on other terms of the
employment relationship, including shift lengths and the

12 See Reserve Bank of Australia, supra note 8. I do not know of any

effort to sum the costs of these various efforts.

No. 3:551)



amount of training. Even a study that considered these
additional facets of the problem would be incomplete if it did
not also consider the impact on the level of capital
improvements undertaken, the educational efforts of low
income workers, the relative wages of workers in the middle
of the wage pyramid, and so forth. But the further the range
expands, the more difficult it is to control for the relevant
variables, and the reliability of the conclusions is
compromised as the set of relevant considerations expands,
seemingly without limits. Finding the golden mean between
the two extremes raises a persistent methodological problem
that is easier to state than to resolve.

While an assessment of the minimum wage law is more
complex than an assessment of many regulations governing
the credit markets (e.g., what counts as an "hour" is only one
of many issues to be faced), this example illustrates the
necessity of considering all aspects of that problem before
making any empirical judgments. Defining the proper scope
for assessing the likely consequences of regulation is also
difficult. The wider the net is cast, the more complete the
analysis, but the less reliable the data. Nothing guarantees
that we can find some combination of breadth and
concreteness that makes the enterprise worth undertaking.
On the one hand, the task of policymaking becomes ever
more complicated. On the other, reliance on simple solutions
with low administrative cost rightly becomes the implicit
default position, especially when efforts are made to override
consensual arrangements.

One minimum condition for this regulatory effort is to set
a legislative frame of reference that considers the position of
all competitors within the same overall market niche. That
was not done in this instance, given the RBA's decision to
refrain from regulating merchant discounts within the closed
systems of American Express and Diners Club. In effect, the
RBA relied on a partial equilibrium analysis with a big piece
missing, namely the operation of the direct competitors of
the regulated parties. If the RBA had performed a general
equilibrium analysis, it would have recognized that the
different forms of payment mechanisms operate at some

COL UMBIA B USINESS LA W RE VIE W (Vol. 2005



THE REGULATION OF INTERCHANGE FEES

level of (imperfect) competition with each other. Such
omissions need not be fatal, but in most cases they are likely
to seriously skew the study of a regulation whose upside
potential is limited by gains achievable under the present
consensual and durable relationship. In this environment,
poorly considered proposals are especially improvident
because the benefits of a program must be large enough to
cover both the anticipated administrative costs of the system
and the extensive error costs driven by the artificial
limitations in the study's initial frame of reference.

These general remarks are not confined to intervention in
the credit system alone, but reflect upon the broader
purposes of regulation within the financial markets
generally. Yet before reaching the difficult issues that the
RBA faced here, it is necessary briefly to consider two other
types of regulation. The first type seeks to regulate
overreaching and fraudulent or misleading behavior in
transactions. These risks are not present here. The second
type, which concentrates on the regulation of cartels and
monopolies, is intended to prevent anticompetitive behavior.
This risk is present here only in an attenuated fashion and is
largely inapplicable to ratemaking questions.

In light of the strong background features that exist in
the payment card system, any alterations to the incentive
structure necessarily will affect the behaviors of the key
players in the system. However, the behavioral responses
are unlikely to correlate with any overall sense of social
improvement because of the inability to identify a large
target of opportunity: the case is not about stopping cartel
conduct, for example. In addition, clearly there can be no
strong Pareto improvement because the entire purpose of the
reforms was to make the position of the credit card holders
and issuing banks more precarious than it had been under
the status quo. In itself, this criticism is hardly decisive,
because the exacting Pareto standard is commonly thought
to be too rigorous with respect to any social reform that does
not begin with a blank slate. It is simply too much to
demand that no one be made worse off when at least one
person (or group of persons) is made better off. But,

No. 3:551]



conscientious social reforms should satisfy, at a minimum,
the somewhat weaker Kaldor-Hicks standard. Under this
standard, the RBA regulations would have to produce more
gains for the merchants and their non-credit card
customers-an ambitious and ambiguous objective 3-than
losses for the issuing banks and their customers, especially
since there is substantial overlap among the populations of
the various groups. Yet that too is an uphill battle. A closer
look at the allocative and distributional consequences belies
any reason to think that this standard has been achieved.

III. INTERCHANGE REGULATION

A. Merchant Choices

The more important of the RBA's two reforms is the sharp
restriction on the level of interchange fees. The RBA
established this restriction to remove the set of subsidies
that led to the systematic overuse of credit cards relative to
other forms of payment. The RBA argued that the high
interchange fees forced all consumers of covered merchants
to subsidize a payment system that only benefited credit
card users. This conclusion contains both an allocative
component (that the fees in question led to a distortion of
relative prices) and a distributional component (that
subsidies were extracted from non-credit card users without
sufficient social justification).

The allocative claim rests on what can only be termed the
odd view-especially odd for a national bank system-that
credit cards provide few if any benefits over alternative
payment systems. That conclusion seems to conflict with the
obvious historical fact that credit cards have expanded their
market share of payment transactions both before and after

13 For discussion of these problems, see infra the discussion of
passthroughs in Section V.

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



No. 3:551] THE REGULATION OF INTERCHANGE FEES

the 2003 capping of interchange fees. 4 To illustrate, Figure
1 shows a continuous upward trend that is so smooth that no
one could figure out from visual inspection when the RBA
imposed its cap on interchange fees.

Figure 1. Real Interchange Revenue per Card"

30

1 25

a

5 ...... But-For

0 1 . . . . . . . . .... . . . . . . . . . . . .. . . . . . . . . .* -

Notwithstanding this obvious trend, the RBA's basic
claim is that "[hiigher sales on credit cards for individual
merchants do not, of themselves, give rise to overall
merchant benefits if: (i) those sales would have taken place
anyway using other payment instruments; or (ii) the sales
have merely diverted business from one merchant to
another." 6

The background assumption underlying these two
conclusions is that credit cards do not add much net benefit
to the operation of the economy and only divert transactions
into more expensive modes of payment. In support of this
conclusion, the RBA notes that credit card fees comprise a
much heftier percentage of sales than fees for other forms of
payment, as shown in Figure 2 below.

"4 For the overall progression, see EvANS & SCHMALENSEE, supra note
4, at 84 (noting the steady rise in use of payment cards from about 2% to
25%).

"' Chang, Australian Interchange, supra note 2, at 21.
16 RBA, AusTRALAN REFORM, supra note 3, at 23.

THE REGULATION OF INTERCHANGE FEESNo. 3:551]



COLUMBIA BUSINESS LA W REVIEW

Figure 2. Payment Costs to Australian Retailers as a
Percentage of Average Transaction Value17

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0
Debit Card Cash Check Open-System Charge Card Closed-

Credit Card System Credit
Card

The last line, which refers to closed-system credit cards
(American Express and Diners Club), was not in the RBA's
original report. That bar can be added in, however. More
generally, what is most instructive here is how this rate
compares with that of the open-system credit cards (Visa and
MasterCard), both before and after the imposition of rate
caps. It seems as though the figures are derived from a
different metric because the absolute values differ. But the
relative values are the same, and they speak to a real price
advantage for both MasterCard and Visa. Before the rate
caps were imposed, MasterCard and Visa (the open-system
credit cards) had total interchange fees of approximately
1.45%, relative to 2.58% for American Express and 2.35% for

17 RBA, AuSTRAiAN REFORM, Supra note 3, at 22. In this figure, the
first five columns were taken from the RBA report, but the sixth column,
representing closed card systems like American Express and Diners Club,
has been added in because the fees for these systems are higher than those
of the four-party systems. See EVANS & SCHMALENSEE, supra note 4, at
152 fig.6.1 (reporting the figure for American Express at around 3%). The
implications of this omission will be discussed presently.

[Vol. 2005



THE REGULATION OF INTERCHANGE FEES

Diners Club (the closed-sytem credit cards)."8  The
interchange fees for MasterCard and Visa were thus
approximately 1% lower than for American Express and
Diners Club. Once the rate caps were imposed, the
interchange fees for MasterCard and Visa dropped to
approximately 0.96%, while those for American Express and
Diners Club remained at 2.38% and 2.31%, respectively. 9

The unregulated sector barely budged in response to the rate
caps in the regulated sector, so that the gap in interchange
fees between the two types of cards only increased. It seems
odd indeed that the regulatory initiative is directed
exclusively at that form of business which does better under
the RBA's own test for relevance. The attack on the four-
party, open-system, credit cards seems at best misplaced.

To bolster its attack on these four-party credit card
systems, however, the RBA takes the position that the
transactions involved in these cases are not "really"
consensual at all. Rather, merchants, both large and small,
succumb to the blandishment of credit cards because "credit
card use has become so widespread that most merchants
believe they have no choice but to accept credit cards."2" I
use the word "really" in quotations to stress the RBA's
questionable insistence that merchants have "no choice" in
this critical matter. A far simpler, albeit more cynical,
interpretation of the phrase "no choice" is that merchants
like to use credit cards, but, like the rest of humanity, would
rather pay less for something from which they already derive
net benefits.

Moreover, a closer analysis reveals the conceptual
confusion that surrounds the easy invocation of this "no
choice" model. Start with this question at the institutional
level: Why would consumers and merchants choose to
support a system of credit that is both clumsier and less

18 See RBA, Bulletin table C3, http://www.rba.gov.au/Payments

SystemlPaymentsStatistics/paymentsdata.html (last visited Nov. 21,
2005).

19 Id.
20 RBA, AuSTRALIAN REFORM, supra note 3, at i.

No. 3:551)



efficient than the systems already in place, i.e., cash and
checks? The slippery phrase "no choice" only obscures the
RBA's central claim. For example, one colloquial meaning of
the phrase "I had no choice" refers to the common situation
where a robber comes up to you and says: "Your money or
your life." In one sense, you remain, to misuse a phrase from
Milton and Rose Friedman, "free to choose" either
alternative. Yet, anyone who decides to hand over the
money can rightly say this "choice" did not count as
vindication of free will, but rather provided conclusive
evidence of illegal duress. The standard libertarian theory,
which treats the use of force as the primary evil, has always
adopted this anticircumvention principle as its corollary: to
limit the effectiveness of force, the threat of force must be
subject to the same prohibitions and coupled with the same
exceptions. In other words, a threat of force in self defense is
as legitimate as the use of force.

Any threat that takes the form "your money or your life"
is especially dangerous because it leaves the putative
wrongdoer better off than he would have been had he just
used the force itself. You value your life more than your
money. The robber values the money more than your life.
Individual differences are not important in this stark
calculation. So the deal is struck, leaving both sides better
off than under the alternative under the circumstances. But
this happy account misses the obvious rejoinder. The robber
has no right to put you to that choice; he is not like the
taxman who says: "Give us your cash or we will seize your
specific assets to pay your lawful debt," which is another
offer that works to the benefit of both sides. Rather, when
facing the thief, you are entitled to both your money and
your life, so, therefore, his "choice" gives you only a means of
minimizing loss, not securing gain from the prethreat state
of the world. Contrast that position with the ordinary
business offer in which you have to choose between your
money and his goods. The mere presence or absence of
choice is neither here nor there. The choice given to you by
the robber does not generate systematic win-win outcomes.
Market exchanges do.

[Vol. 2005COL UMBIA B USINESS LA W RE VIE W



THE REGULA TION OF INTERCHANGE FEES

Against that backdrop, the phrase "no choice" in the RBA
report gains no traction from the common law cases of
duress. Rather, all credit transactions fall within the
domain of win-win transactions, where the only quarrel is
over the size of the gain for each party. Nor could it be
otherwise. If the credit card system is as inefficient as the
RBA supposes, then any merchant could effortlessly decide
to accept only cash, checks, or debit cards and eliminate the
other options. This unilateral threat (or option) to withdraw
is perfectly credible no matter what decision other
merchants make regarding which payment methods to
accept. To avoid this conclusion, the RBA takes the position
of the marginal merchant, who must decide whether to stay
in the game or forego the enormous benefits that credit cards
have "of attracting sales from, or not losing sales to, rival
merchants."21 But if the basic assumption is that credit
cards are inefficient, then this dodge fails because it does not
acknowledge the last crucial step: Why is it necessary for any
merchant to join an inefficient system to attract customers
from rival merchants or to keep them from sliding off to
competitors?

To see why, think of this as a prisoner's dilemma game in
which each merchant, acting out of narrow self-interest, has
the option to participate in or opt out of the credit card
system. Opting out is the dominant solution. If other
merchants cling to the expensive system, then the single
merchant that lowers prices by one or two percent has a real
advantage for himself. Thus, he should defect from credit
cards. Alternatively, if the other merchants opt out, then the
no-credit card strategy is needed to keep pace with others
who have previously adopted the no-credit card strategy. He
should defect again. If the credit card system offers a more
expensive payment system without offsetting benefits, then
the migration from it should be fast and furious, as each
merchant follows the dominant strategy of ditching credit
cards in favor of cheaper cash, check, or debit card payments.
There is no need for government action to eliminate the

21 Id. at i.

No. 3:5511



COLUMBIA BUSINESS LAW REVIEW

dreaded cross-subsidy. Merchants can do this unilaterally.
When they do, the people who invest in credit cards will cut
them in half precisely because they have no reason to clutter
up their wallets.

Yet, that is just not in the (credit) cards. Instead, we see
the reverse flow taking place as lower transaction costs, in
the form of faster computers, more reliable scoring methods,
and other improvements, have led the credit card market to
expand into niches that it had not previously penetrated.
Five or ten years ago, credit cards were not accepted at fast
food outlets like McDonald's and Subway. The sums were
too small and the per transaction costs were too high. If the
RBA's assumptions were correct, this equilibrium should
have proven stable because every firm could stand pat
without having to play follow-the-leader. However, the
opposite result ensued. Special price deals, one suspects,
brought this segment of the market into the credit card orbit.
It is unlikely that the decision was just a bad mistake.
Clearly something is deeply amiss in any analysis whose
major premise predicts the demise of credit cards in the
midst of an explosion in their use. To see what this problem
is, we have to examine more closely the proposition that
credit cards have no advantages over other systems of
immediate payment. The key is finding benefits that
differentiate all types of credit payments from all other
payment systems.

B. The Not So Subtle Credit Card Benefits

To highlight the benefits of credit cards, divide the
landscape of payment methods into two categories. Charge
and credit cards, including those issued by both open and
closed systems, are on one side of the line. Checks, cash, and
debit cards, which I collectively term "debit instruments,"
are on the other side. The credit and charge instruments,
which are more expensive than the debit instruments, also
perform additional functions. Because the debit instruments
have no built-in credit component, they only allow the
purchaser in a given transaction to spend the money that he
or she already has in his pocket or salted away in some bank

[Vol. 2005



THE REGULATION OF INTERCHANGE FEES

account. These stores of wealth do not allow purchasers to
buy goods or services that are desired today with a promise
to pay for them tomorrow out of future earnings. Nor do
they give their users the prospect of free float for a month,
which is attractive to high-income borrowers who have
sufficient funds to pay off their bills when they become due.
With debit cards, it is as though one wanted to use savings to
do the work of life insurance. Yet, it takes little imagination
to realize that most people would like to make their
consumption patterns smoother than their earning
patterns.22  No profound demonstration of this point is
needed at this time. Just recall that, even as adults, people
borrow while in school yet simultaneously save for
retirement. Thus, the earnings in an individual's peak
period are moved both forward and backward in time.

Debit instruments cannot perform this function of
tapping future income. In contrast, credit and charge cards
facilitate this goal on a shorter time frame by allowing
individuals to pay over time for goods and services, whose
value to the consumer is captured over time. Few people buy
homes with cash; instead, they take out home mortgages,
where the monthly payments closely track the benefits
derived from the initial purchase during each pay period.

A question then arises over which type of card is superior,
the store charge card that is valid at only one merchant, or
the credit card from either a closed or an open system that
may be used at any merchant who has agreed in advance to
accept the card. Clearly, if one method was dominant over
all others, then we would see either credit or charge cards
driven from the marketplace. But the simple fact that
Neiman Marcus does not accept general credit cards provides
evidence that each card has its own distinct advantages
enabling it to survive. Charge cards, for example, may
entitle their users to certain privileges that non-charge card
holders cannot receive, including the receipt of information

22 For a basic discussion of the permanent-income hypothesis, see

Robert Schenk, Permanent-Income Hypothesis (2002), available at http:ll
ingrimayne.saintjoe.edu/econ/FiscalDead/PermIncome.html.

No. 3:551]



about various sales and special offerings; free services such
as gift wrapping and holding or delivering parcels; the ability
to purchase, or purchase more quickly, by phone; and the
ability to go to the head of the line at the (quaintly termed)
cash register. The obvious point here is that the charge card
signals a form of loyalty that generates loyalty discounts in
return. In a competitive marketplace, there is no reason
whatsoever to regulate the pricing structure of these charge
cards any more than there is to regulate the price of pajamas
or alarm clocks. After all, reductions in the price of pajamas
are said to create an implicit cross-subsidy from purchasers
of alarm clocks to those of pajamas because of an implicit
shift of some joint costs from the former to the latter.

Nonetheless, there is a simple but compelling explanation
that explains why credit cards have gained ground in recent
years relative to charge cards: the credit card in question
may be used at any one of a number of stores, which allows
its holder to compare prices and goods across different
merchants in an effort to obtain the best available
combination of price and quality." It is a mistake to assume,
as the RBA did, that the net result of any product or service
substitution should be scored as a wash. Each substitution
should generate some additional level of consumer surplus.
Merchants, therefore, find themselves in an unhappy
position. If they stand aloof from the network, they lose the
ability to compete for credit purchases that may benefit
them. So, they join the greater number of merchants who
accept credit cards and thus make the market more
competitive for everyone else, because they sensibly expect
that more shoppers will come their way. A similar reason
explains why jewelers or diamond dealers all like to work in
the same district. Their close proximity generates higher
overall traffic that more than compensates for the stiffer
competition. The consumer benefits in both cases are
unambiguous.

Credit cards also have other advantages for which debit
instruments are an ineffective substitute. Try to rent a car

23 See EvANs & SCHMALENSEE, supra note 4, at 122-24.

[Vol. 2005COL UMBIA B USINESS LA W RE VIE W



THE REGULATION OF INTERCHANGE FEES

at the airport or check into a hotel without a credit card.
Usually, it is necessary to provide a credit card to cover the
expenses of a trip. The protection here is imperfect because
the card limit may be lower than the customer's potential
tab. This recognition explains why most consumers seek to
raise their credit limits, as opposed to keeping them low (in
an effort to guard against binge buying, which people can
indulge in on store credit). But any card is likely to prove
superior to leaving wads of cash on the table or signing over
a blank check that may or not be backed by sufficient funds.
Credit cards, therefore, also have their distinct market niche.
The question remains, though, how the regulation of
interchange fees in open systems will manifest itself.

C. The Impact of Interchange Fees

1. Debit Instruments

The obvious implication of the RBA report is that the
removal of the cross-subsidy will increase, to some extent,
the frequency of use for all debit instruments. The extent of
this effect depends in large part on the relative price
differences observed by consumers. In this regard, it is
critical to look back to the relative costs of the two
transactions as outlined in Figure 2 above.24

Here, the costs portrayed are not the total costs to the
user, but only the transfer payments that users make to
merchants in order to participate in these transactions. But
the situation from the customer's point of view is decidedly
different than this figure indicates because cash and checks,
in particular, have costs that are not reflected in these
transfer payments.

The private costs associated with the use of cash and
checks are plentiful and should not be ignored. Most people
withdraw their cash from an ATM. When cash is drawn
from a machine owned by another bank, consumers
frequently pay a flat fee that, for most withdrawals, will be

4 See supra note 17 and accompanying text and graphic.

No. 3:55 1]



equal to about one percent of the money received.25 In
addition, to maintain ready access to their funds, consumers
must keep their money in checking accounts that pay little
or no interest. Often there are also fees for maintaining
checking accounts or for cashing individual checks. Another
private cost is risk: both cash and checks can be lost or
stolen.26 With regard to cash, what is gone is gone; there is
simply no way to stop payment on this ultimate negotiable
instrument. All of these aforementioned costs must be added
back into the cost of transacting by cash or check.

One response to this concern argues that the amount of
money lost or stolen is trivial in relation to the total amount
of cash in circulation. While this may be true, it is not a
decisive criticism because the amount of cash lost is typically
small because most people choose to carry only small
amounts of cash. They are able to do this because they have
credit cards at their disposal for most transactions,
especially for big ticket items. Few people will opt to carry
around the thousands of dollars that it would take to
purchase a plasma television when they can simply take a
credit card instead.

The costs associated with the risk of theft also applies to
the use of checks in stores. Anyone who has ever had his or
her checkbook stolen knows how difficult it is to stop
payment on forged checks. A theft may go undetected for

25 Greg McBride, Checking Study: ATM fees keep climbing,

BANKRATE.COM, May 11, 2005, http://www.bankrate.com/brm/news/
chk/20050511cl.asp. The study reports that the average fee for
withdrawal from a "wrong" bank ATM is now about $1.35 and that
estimated total annual fees exacted from these transactions approximate
almost four billion dollars in 2005. A second recent survey placed the
average ATM service charge at $1.55. Mike Licker, NYPIRG Plans to
Warn Students About Expensive ATM Service Fees, THE DAILY ORANGE,

Apr. 26, 2004, http://www.dailyorange.com/media/paper522/news/2004/
04/26/News/Nypirg.Plans.To.Warn.Students.About.Expensive.Atm.Service
.Fees-671154.shtml. In some cases, a percentage fee is charged depending
on the size of the withdrawal, but these surveys did not report as such.

26 See EvANs & SCHMALENSEE, supra note 4, at 30-31. Evans and

Schmalensee report that $850 million in cash was stolen from 84,000
robberies and pickpockets in 1995 alone.

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



THE REGULA TION OF INTERCHANGE FEES

hours or perhaps even days. In the meantime, large
amounts of money can be charged against a consumer's
account, particularly if merchants fail to seek verification of
the check writer's signature. Thus, the obvious strategy is to
leave the checkbook at home because it is safer to carry a
credit or debit card. This analysis suggests that debit cards,
due to their inherent safety value, have an even greater
transactional advantage than the overall numbers suggest
and may explain why their use has increased so rapidly.

Moreover, future developments will continue to influence
the mix of payment instruments for debit transactions. One
issue worth examining is how the use of online bill payment
fits into the overall scheme of payments.27 While these
online transactions do not count as credit payments, they do
transfer some portion of the payment volume from checks
into electronic exchanges, since the same banks that run the
credit systems support the applicable networks. The benefits
of online payments are manifest: time savings,
instantaneous financial position reporting, and other such
personal and financial benefits. One important advantage of
this ongoing development is that it allows individuals to
better manage their credit card balances and, as a result, to
control their interest rates. Even if banks had preferred the
older systems of payment, competition would nonetheless
have brought about a shift towards online payment because
of these kinds of advantages for the consumer. And, in fact,
most banks also prefer online banking transactions because
of their own cost savings. Furthermore, the whole system
contains within it no element of government subsidy because
it more closely resembles a private form of money, albeit one
parasitic on the general currency. It is thus clear that
checks, once considered universal payment options, are
becoming obsolete. Why perpetuate such obsolete systems?

27 Terry Savage, Why online bill payment is booming, MSN.COM (2005),

http://moneycentral.msn.com/contentlBanking/BetterbankingfP38218.asp
(anticipating that this year forty million Americans will pay bills online to
avoid the inconvenience and cost of writing checks).

No. 3:55 1]



Aside from efficiency and theft concerns, another
important issue relates to the number of subsidies that are
involved in the respective payment systems. The entire
clearance system for credit cards is privately financed; it is
as though the dream of a private money system has been
realized by indirection. The same cannot be said for either
the cash or the check system, both of which receive extensive
government subsidies." The national treasury bears the cost
of printing, distributing, and protecting currency. The entire
system of check exchanges relies upon the operation of the
various federal banks. Hence, one way to look at the RBA
initiative on its own terms is to see that it substitutes one
implicit cross-subsidy (for cash and check users) for one
explicit cross-subsidy (for credit card users). The net
resource implications of such subsidies are not clear. How is
the state subsidy, which falls on all persons alike, better or
worse than a cross-subsidy that applies within an elaborate
industry network of consenting participants? It is far from
clear whether the RBA's dismal view on credit cards is an
accurate one.

The final issue relates to the utility of cash and checks.
Cash represents a national currency, and persons who travel
overseas incur significant conversion costs that can easily
consume one to two percent of the money's overall value.
Checks, for their part, have an even narrower scope of use.
They are of little use in foreign countries and, even within
the United States, have limited value outside of the
community served by the issuing bank. It is harder to run a
credit check on a checking account than on a credit card, and
people do not want to suffer the costs associated with bad
checks. Credit and debit cards are both immune from these
relative disadvantages. Given all of the concerns associated
with the use of cash and checks, we should expect that, as

2 I have not been able to find exact estimates of the cost of these

subsidies, but the number is commonly stated to be in the billions. For
one estimate, see Oc6, Check Clearing for the 21" Century Act: Printing
Implications for Financial Institutions and Service Bureaus 4 (working
paper, available at http://www.ondemandjournal.com/whitepapers/
Check_21.pdf).

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



technology continues to improve over time, the use of both
credit and debit payment options will further increase
relative to traditional modes of payment.

2. Closed-System Payments

The substitution from the open systems, Visa and
MasterCard, to the closed systems, Diners Club and
American Express, is more critical to this systematic
analysis. The economic dislocations here stem from the
truncation and inadequate frame of reference of the RBA's
interchange inquiry. Interchange fees arise only in a credit
card network that utilizes both acquiring and issuing banks.
In contrast to the open systems, the closed systems utilize a
single company that serves as the sole intermediary between
its customers and its merchants. The more compact nature
of a closed system's business obviates its need to levy any
interchange fees. However, such a system is not free from
the underlying economic issue that prompted the RBA's
initial effort: Were merchants as a group-and, indirectly,
their non-credit card customers-being asked to foot too
much of the cost of running the entire credit card system,
thereby creating an implicit subsidy for credit card users?
That question arises solely because the merchant, through
the payment system, receives only a portion of the face
amount of the credit card charge. The issue is actually more
acute for merchants (and their customers) transacting with
the closed-system players, who charge a larger percentage of
the face amount of the transaction than the open-system
players, while avoiding the use of interchange fees.

Why then limit the inquiry to interchange fees when the
merchants' complaint applies equally to both forms of credit
transactions? Surely merchants cannot object to the high
costs of their own store cards. At this point, moreover, the
rate of substitution between the two rival card types (open-
system and closed-system) should be greater than that
observed between credit cards and any debit system. If these
credit advantages prove tangible, then we would expect
closed systems to garner a larger share of the overall market
after the cap on interchange fees than they had prior to the

THE REGULATION OF INTERCHANGE FEESNo. 3:5511



cap's imposition. If this swing does occur, then the
ostensible subsidy that the closed systems' cardholders
receive from non-credit card users should increase
accordingly.

This prediction seems to have been borne out by the
recent data collected by the RBA, which shows an increase in
the use of closed-system cards relative to open-system cards,
even though the gap in the merchant discount has grown
steeper since the advent of the RBA's selective
administrative rule. It is a fact that three-party (closed-
system) cards now account for 17% of transactions by cash
volume as of August 2005, up from 13.7% as of September
2003.29

In the face of this criticism, the RBA retreats to a set of
unpersuasive bromides. The RBA claims that "at this stage,
such regulation [of both types of card systems] would not
improve the overall efficiency of the payments system," that
it would have "relatively little effect on merchant charges,"
and that any effort to reach this problem could "only be
addressed through considerably more extensive regulation
than that currently existing in the credit card schemes."3

The RBA advances these defensive claims without support
from any empirical or theoretical evidence. In particular, the
RBA does not explain why no regulation of either closed or
open systems is better than the partial regulation of one type
of competitor with its attendant distortions. Nor does the
RBA offer any explanation as to why the admitted shift in
transaction volume has taken place, if not for the competitive
boost.

The simple explanation for the problem lies in the
peculiar economics of two-sided markets, which include

29 Letter from Leigh Clapham, Senior Vice President, MasterCard
Australia, to Dr. John Veale, Head, Payments Policy of the Reserve Bank
of Australia (Aug. 25, 2005), available at http://www.rba.gov.au/
PaymentsSystem/Reforms/CCSchemes/SubmissionsDCCCIStd/mc_250820
05_l.pdf [hereinafter Clapham Letter].

" See id. at 3; see also Press Release, Reserve Bank of Australia,
Payments System Reform (Feb. 24, 2005), available at http:/!
www.rba.gov.au/MediaReleases/2005/Mr 05_02.html.

[Vol. 2005COL UMBIA B USINESS LA W RE VIE W



credit card platforms. Both closed-system merchant charges
without regulation and open-system merchant charges with
regulation must satisfy this minimum condition: all relevant
parties must desire to remain in the network. Two-sided
markets are distinctive in that the charges levied on all
players in the system do not equate with these players'
marginal costs, as they generally do in competitive markets
where marginal costs are constant. In competitive market
situations, there are plenty of bargains, but no bargaining,
because of the unique competitive equilibrium that sets price
equal to marginal cost.

That unique equilibrium does not exist in two-sided
markets. Hence, a pricing problem arises that is not
dissimilar to the problem that arises in connection with the
general marginal cost controversy: How does one price an
item when the cost of the first unit is exceedingly high (e.g.,
a new bridge or a new drug), but the cost of each subsequent
unit is very low?31 The decision to charge only marginal cost
for all units means that the developer cannot recoup its
upfront costs over the useful life of its asset. Some form of
subsidy, therefore, must be administered without any clear
sense that the ultimate social value of the product will
justify the subsidy's imposition. The only alternative pricing
method, which in most cases is thought to yield fewer
distortions, is to allow prices on subsequent units to be set
above marginal cost, allowing for some loss of the user base,
so that the company may recover its upfront fixed costs. The
point here is that even in this relatively simple scenario,
there is no optimal solution that reaches the happy outcome
of an ordinary competitive market where-in the limiting
case-marginal costs are constant for all levels of production.

At this point, it should be clear that the two-sided market
problem shares a common property with the high fixed, but

31 For the original proposal, see Harold Hotelling, The General Welfare

in Relation to Problems of Taxation and of Railway and Utility Rates, 6
ECONOMETRICA 242 (1938), as criticized in R. H. Coase, The Marginal Cost
Controversy, 13 ECONOMICA 169 (1946), and, in relation to intellectual
property, in John F. Duffy, The Marginal Cost Controversy in Intellectual
Property, 71 U. CHI. L. REV. 37 (2004).

THE REGULATION OF INTERCHANGE FEESNo. 3:55 1]



COLUMBIA BUSINESS LAW REVIEW

low marginal, cost situations. Given the excess wealth
within the system, it is not possible to set all prices equal to
marginal costs, which explains why everyone (merchants
included) jockeys to shift a larger fraction of the costs of the
system to other market players. To see why this is so, think
of the familiar problem facing the pub that knows that its
success depends upon its ability to attract male and female
customers in roughly even proportions. It is a safe
assumption that the cost of service for both classes of
customers is the same, yet charging the same price to both
sexes will not get either men or women their desired mix of
goods and services. Men do not want beer and pretzels as
such; they want beer and pretzels with women present in
sufficient numbers. Uniform charges based only on the cost
of the provision of service thus results in too few women
relative to men; the men withdraw because there are not
enough women, and the few remaining women withdraw
because there are too many men. The market shuts down on
both sides. The only way to overcome this problem is to
charge different prices to each sex based on the intensity of
their demand. Men will remain in the market at the higher
price, and more women will enter it at the lower price, thus
reviving the underlying market.

The credit card market shows the same imbalance, with
the merchants as the eager participants and the customers
as the reluctant ones. We should therefore, in most markets,
see an outcome in which a greater fraction of the common
costs are borne by the merchants, which, for open-system
cards, are mediated through the interchange fee and, for
closed-system cards, are expressed through a lower
reimbursement rate for merchants. The RBA's imposition of
regulation makes a difference because it leaves the closed-
system card companies with more money "in the till" to
recruit customers to their payment systems. Since these
closed-system cards specialize predominantly in higher-end
customers, the observed outcome is predictable-a larger
proportion of transactions are funneled through the closed-
system cards, representing a still higher proportion of the
total dollar amount of credit transactions.

[Vol. 2005



THE REGULATION OF INTERCHANGE FEES

3. Distributional Consequences

The RBA's concern with unregulated interchange fees
also has an important distributional component. Let us
suppose (contrary to fact, I believe) that credit cards are a
less efficient way of doing business. There is an inherent
unfairness in asking users of cash or checks to subsidize the
use of these instruments. Thus, the critical question is to
ask to what extent, even accepting the RBA's basic
assumptions, do these cross-subsidies actually exist?

The answer is that these subsidies, while they do exist,
tend to be small. The distributional point is usually framed
to say that credit card holders as a class receive a subsidy
from individuals who use cash or checks. But one mistake in
framing the point this way is that it assumes that the
various users fall into discrete categories without overlap in
class membership. This assumption is most decidedly not
valid. Most people use all three systems of payment in
different proportions and, hence, appear on both sides of this
imaginary line. Assume, in the limit, that all persons,
regardless of economic status or consumption patterns, used
the six common payment systems-cash, check, debit,
charge, and the two kinds of credit-in identical proportions.
No one person would receive a subsidy from any other
person. Each would benefit in his or her role as a subsidized
person, while each would also be harmed in the role of a
subsidizing person, but always in the same proportion. In
this situation, the only concerns hearken back to the
efficiency question. If the systems are mispriced relative to
each other, then some systems will be overconsumed and
everyone would benefit from a shift in prices that captured
the relevant cost variables.

The actual situation on the ground is, of course, more
muddied. There are some people who do not use credit cards
and still others who use them in greater or lesser
proportions. It seems that people with both credit and debit
cards tend to favor one over the other. Accordingly, it would
be necessary to figure out the transfer payments that some
consumers on balance receive from others. But the net size
of those transfers among consumers has to be decidedly

No. 3:551]



COLUMBIA BUSINESS LA W REVIEW

smaller than any reduction in the collection of fees, and,
certainly as between credit and debit, many consumers have
the easy option to shift, at least in part, to their favored
system. Thus, the most recent testimony from Governor
Macfarlane calculates the annual savings to non-cardholders
from capping interchange fees at $580 million. 2 But of
course much of the wealth is simply a transfer from one
pocket to another among persons who use all of the different
types of payment systems simultaneously. We know that in
Australia, consumers use the various instruments-cash,
check, credit card, and debit card-in roughly equal
proportions. Allowing for differential spending patterns, it
seems unlikely that even half of the purported savings
represent net transfers to persons who predominantly utilize
debit instruments.3

There is one final irony in this argument concerning the
extent of transfers: there are two ways in which the RBA
could eliminate the ostensible transfer payments. One
option, which is the RBA's dominant strategy, would be to
crack down on the use of credit cards. A more sensible way
to achieve this result, however, would be to allow market
forces to expand the use of credit cards, so that fewer
transactions would run through cash and checks. If that
route were taken, the extent of the transfer would
necessarily shrink as the percentage of credit card payments
rose. Ironically, if all transactions took place via credit
cards, then interchange fees could not possibly contain an
element of cross-subsidy. In light of these considerations, it
may well be that the public systems of cash and checks are
inferior technologies propped up by the imposition of the
RBA's interchange rules.

32 Macfarlane Testimony, supra note 7, at 29-32.

" Although the calculations would be difficult, it seems likely that
most people who record the greatest dollar volume on credit cards also
make extensive use of other payment systems. For these people,
redistribution is merely from one pocket to another. In contrast, the gain
to those consumers who do not use credit cards, or use them sparingly, is
likely to be small.

[Vol. 2005



These criticisms do not, of course, apply with respect to
debit card systems, which share many of the efficiency
features of credit card systems, but are cheaper to operate
precisely because they do not have to quantify or respond to
credit risk. Indeed, debit cards often have uses that cash
and checks do not (e.g., as payment in e-commerce) and, as a
result, we may well expect their use to grow. Indeed, one
reason for Visa's stronger market position relative- to
MasterCard was its early and forceful moves into the debit
card business-yet another example of how successful firms
can survive only by cannibalizing their own high-margin
businesses.

What then of the possible cross-subsidies between debit
and credit cards, which are often run by the same banks
from the same platform? The answer in many ways parallels
the cross-subsidy answer given above in relation to cash and
checks. Both systems are part of a consensual network in
which there is no unique fee structure that prices all
transactions at marginal cost. If it turns out that debit card
users as a class are unhappy with the association of these
cards with credit cards, even when they carry both, then
some form of market segmentation will occur. New
companies will offer debit-only cards at lower rates than
those at which existing debit card users receive the blended
product, and the incumbents will have to follow suit-unless
the synergy of the two cards has advantages that debit-only
plus credit-only cards could not equal.

Nor would it be necessary that the pressure for market
segmentation come from customers. It might also come from
merchants who wish to negotiate debit card-only deals with
their banks at proportionately lower rates. In California, for
example, payment at the gasoline pump is by debit card only
(a true nuisance!) for precisely this reason. Given this range
of responses, it is at best an open question as to how many
people using both systems simultaneously will migrate from
the current platforms. But as long as that possibility
remains-and any third person could start a debit-only
network-then regulation is not needed as a form of

No. 3:551] THE REGULATION OF INTERCHANGE FEES



industrial policy to force market institutions to march down

an intelligent path.

IV. THE NO SURCHARGE RULE

The second major RBA initiative, disallowing surchages,
is surely less important than capping interchange fees, but it
is nonetheless worthy of comment. The basic objection in
this case is that standard contractual prohibitions on
surcharges prevent merchants as a whole from adjusting
prices for their various classes of users in order to reflect the
relative costs that each class imposes on the overall system.
As with interchange fees, the potential gain from any such
regulation is strictly limited because we know from the
outset that the differential is one that does not meet with
stern market resistance-which would surely be the case, for
example, if the differential cost of service were closer to five
percent instead of the one or two percent that we see. In
part, there is no resistance because customers often prefer to
pay with credit cards than with cash, and these key
customers would be angry with the small surcharges that
occupy their time and do little else. The customers who pay
by check and cash do on occasion ask for and receive certain
discounts, particularly on larger purchases, and thus have
little bloc interest-assuming that they are a bloc-to
override the clause. Thus, the porous nature of this rule
accounts for part of its durability.

Another objection is that the argument for a surcharge
rests on the assumption that non-credit card users gain no
particular advantage because others use credit cards at the
same place of business. There is some reason to think that
this assumption is overstated. Consider the following
thought experiment that could, in principle, become the basis
of systematic empirical research. Suppose that you are
intent on paying for a product by cash or check and could
enter into either of two competitive establishments, one that
announces that it will accept credit cards at no additional
cost and another that announces that it will not take any
cards at all. If one takes the RBA assertions seriously, then
the cash purchaser should, ceteris paribus, make a beeline

[Vol. 2005COLUMBIA BUSINESS LA W REVIEW



THE REGULATION OF INTERCHANGE FEES

for a place that, on average, will charge about one percent
less for each individual purchase.

There are three reasons why the behavior observed will
not prove to be all that simple. First, from the outside, the
two stores are in some sense black boxes, especially to
tourists, who have no knowledge of what types of goods they
carry or the clientele that they serve. In these situations,
the credit card logo over the door operates as a useful
signaling device that tells the potential buyer that this
business has sufficient stability and organization to
participate in a credit card operation. This type of
information is valuable not only to credit card holders, but
also to cash purchasers who see it as a general indication of
the scope and reliability of the business' full range of
operations. There is no direct way for credit card companies
to collect for this bonding function, in essence a form of co-
branding, from non-credit card users. The imperfect
maintenance of the surcharge helps counteract such a free
riding effect.

The second reason is more restrictive, and it applies only
to those people who are happy to use cash, but carry credit
cards as well. When they enter a store, they may well plan
on making some modest purchase. However, if they then
decide to buy an unexpected and expensive item, they may
prefer to switch to a credit card for payment. Thus, they will
gravitate to the place that provides them with the added
flexibility of paying by credit card. Once again, allowing the
credit card company to impose a surcharge is the only way to
overcome free riding by cash-only customers.

Finally, there is the standard concern with the ease of
transacting, which looms especially large in any setting with
a high volume of low-cost transactions. Customers hate to
wait in lines, and anything that complicates the checkout
process will often result in a reduction in sales, with
attendant detriments to both sides. Multiple price schedules
are sure to slow down business, even if the delays only come
from customers who have to check first to see whether they
have sufficient cash in their pockets or funds in their
checking or debit accounts. In any individual case, the time

No. 3:551 ]



costs look trivial, but their aggregate sum over the course of
days and weeks, like that of interchange fees, is not. If
multiple price schedules serve to slow down lines by even one
or two percent-a conservative estimate-this would
translate into higher labor costs and longer waiting times,
which might equal or exceed the surcharge differential.

These costs are, of course, apparent to the merchants
themselves, which is why in many instances they avoid the
adoption of separate price tiers even when given the
opportunity to do so. In this regard, it is instructive that the
surcharge arrangements found in the Australian context do
not support the view that similar arrangements are essential
to deal with the implicit cross-subsidy of cash customers.
The best data suggests that the percentage of merchants
that use surcharges is between 2.3% and 7%, and that, of
these merchants, none impose surcharges in all cases.34 In
addition, in some instances merchants impose surcharges on
credit transactions that exceed the interchange fee, a
strategy that suggests that these surcharges are imposed
with a modest eye toward price discrimination. Credit card
customers are, on average, likely to have higher demand for
goods than are non-credit card customers, and the extra
surcharge is a way to single them out. Again, therefore, the
no surcharge rule does not promote competition, but instead
allows merchants in the Australian markets (which are more
concentrated than U.S. markets) to exercise some modest
degree of monopoly power.

The general unease with surcharges at the merchant
level also provides the best explanation of why issuing banks
have not resorted to tacking on variable charges to
individual credit transactions, even when interchange fees
are capped. The main argument in favor of having issuing
banks impose surcharges on their cardholders is that it
eliminates the ostensible cross-subsidy that credit card users

" See Chang, Australian Interchange, supra note 2, at 19. Note that
in the Netherlands only about 10% of transactions involve merchant
surcharges. The figure is just 5% in Sweden, and is also small in Great
Britain.

(Vol. 2005COL UMBIA B USINESS LA W RE VIE W



receive when the interchange fee is obscured by the uniform
prices charged to all purchasers. In such a case, bank-
imposed surcharges should create uniform prices for
products and services precisely because the merchants need
not build interchange fees into their cost structure. Yet that
alternative system has never been used in the credit card
business, even during its infancy. Credit card systems that
were devoid of any semblance of market power used
interchange fees in order to help issuing banks to woo credit
card customers. This history is consistent with the view that
two-sided markets require these types of internal
adjustments, wholly without regard to grand considerations
of monopoly and market structure. Similar to the way small
bars may charge different prices to men and women, small
(and large) credit card companies have used the same
strategy for predominantly the same efficiency-driven
reasons.

Any surcharge imposed by issuing banks on cardholders
has two distinct disadvantages. First, it forecloses the
possibility of merchant subsidies for cardholders: once the
lines of communication are cut, it is impossible to run cost
subsidies. Second, this new fee structure creates yet another
level of uncertainty and confusion. In principle, a potential
purchaser in a store would like to know the total transaction
price at the moment that goods or services are purchased.
That determination is not possible if there is a subsequent
service charge of unknown size on the credit transaction of
which the merchant is ignorant. In such a case, the
customer has to wait until he receives his bank statement to
be sure of the exact charges. If there are multiple
transactions, there is a greater likelihood of squabbles over
individual bills, especially if the surcharge in question is not
denominated as a simple fraction of the total bill. Yet
imposing a flat fee reduces the ability to differentiate
between different types of payment arrangements adopted
by merchants-such as electronic and non-electronic
processing. Furthermore, even if the issuing bank
instantaneously disclosed the fee to the merchant, it would
merely replicate the unpopular system of surcharges at the

THE REGULATION OF INTERCHANGE FEESNo. 3:5511



merchant level. The argument in favor of the current
contract provision is not that the no surcharge rule is
perfect. Rather, it is that the set of advantages and
disadvantages are so difficult to untangle that these fees are
not an ideal target of regulation.

V. THE USES AND LIMITS OF CREDIT CARD
REGULATION

In one broad sense, both the interchange and the
surcharge regulations are efforts to regulate the payment
systems industry. However, the two regulations differ in
dramatic ways in their institutional implications. There are
no obstacles preventing the implementation of surcharge
regulation on a permanent basis. The only change required
is to strike surcharge clauses from contracts. There is no
need for any continuing oversight or enforcement of the
provision, as nature will take its own course in the
adjustment of other terms and practices.

Setting the interchange fees, however, is an entirely
different proposition that raises all of the nasty complexities
associated with ratemaking procedures that have
traditionally been invoked to deal with natural monopolies.
The first step of the analysis is to determine those categories
of expenses that should be included in the rate base. The
RBA, in its initial findings, took the hard line that the only
expenses covered were those that related to either the
processing of transactions or the prevention of fraud. The
cost of an "interest-free period," which it makes sense to
include under a sound analysis of two-sided markets, was
excluded from the base on the grounds that it "is a matter
exclusively between individual card issuers and their
customers."" The same arguments apply to "loyalty
programs" that issuing banks use to attract and keep
customers 3 and, more dubiously, to various fees for starting

3" RBA, AUSTRALIA REFORM, supra note 3, at 49.
36 Id. at 49-50.

COL UMBIA B USINESS LA W RE VIE W [Vol. 2005



THE REGULA TION OF INTERCHANGE FEES

and maintaining a network, even though these relationships
have no exclusivity features.37

In effect, the RBA's solution fixed the interchange fee and
thus eliminated it as a source of competition between the
various payment systems. At that point, the operating costs
had to come out of bank profits or cardholder fees, whether
on a per transaction or annual fee basis. The interchange fee
had to account for the differential demands found in two-
sided markets in order for the merchants to capture the full
range of the surplus. The RBA thought that further
adjustments would take place on both sides of the markets.
Merchants would have some incentive to pass their lower
costs through to consumers in the form of lower prices.
Likewise, issuing banks would have an incentive to raise
annual fees and transaction fees to cover the shortfall in
revenue. However, dollar-for-dollar passthroughs are
unlikely on both sides of the market, in part because of the
high concentration in both the retail and banking sectors of
the Australian system. In fact, the Chang, Evans, and
Swartz study estimated that about one half of the merchant
savings could be expected to be passed through to
consumers, if-and it is a large if-the rigidities of the price
system do not overwhelm the small sums involved. " There
is no systematic guarantee that when all of these changes
are netted out, one could point to any net allocative
improvement in the overall operation of the payment system.

Choosing this limited rate base is only the first step in the
overall process. The RBA is also required to set the actual
rates, and that determination gives rise, in turn, to two
difficult questions. The first involves the duration of the rate
decision. Presumably, the 0.55% fee chosen represented a
good initial estimate of the average cost of running the
interchange system. Such a presumption is accorded
notwithstanding the fact that the RBA has offered no clear

37 Id. at 51-52.

31 See Chang, Australian Interchange, supra note 2, at 18 (noting that

a 50% reduction in a 0.21% merchant saving works out to 0.105%, which
on a $40 transaction is about $0.04, an amount easily lost).

No. 3:551 ]



articulation of why it chose this number, much less why,
given the differences in customer mix and service systems,
this number should be uniform across all payment systems.
But with the rapid movements in technology, or changes in
the mix of borrowers, nothing guarantees that the system's
two key cost drivers will remain constant over time. For
example, there is a real risk that an increase in credit card
fraud could require additional infrastructure security
measures that would exceed the allowable costs, especially
expenditures that would be incurred today but would have to
be amortized over future years. By the same token, greater
efficiencies in processing transactions could, depending on
scale effects, reduce the costs of running this system. The
balance is not possible to determine in a priori fashion.

These uncertainties suggest that some rate adjustments,
either up or down, might be appropriate over time. In
figuring out what these adjustments might need to be,
however, the regulator has to guard against the danger that
regulated firms (or some more than others) will pad their
expenses in order to increase their net profit under
regulation. The major tradeoff in regulation is between the
size of the rate base and the rate of return. When the base
includes all investment expenditures, regardless of their
prudence, the rate of return should be lower because the firm
bears less of the economic risk than when such expenditures
are excluded. However, that relationship does not normally
supply sufficient information regarding how either of these
tasks should be undertaken. It is useful to exclude certain
combinations, but no ironclad test points to a single unique
solution.

These problems only arise in regulated markets. No firm
in an unregulated market has the same type of perverse
incentive to control the costs of its operations. Instead, in an
open market, a firm will work to reduce its costs in order to
increase its profits. Yet, by the same token, the risk of
confiscation inherent in any ratemaking scheme has
generated in the American setting a substantial body of

(Vol. 2005COL UMB1A B USINESS LA W RE VIE W



"takings law."39 The danger here is that the regulation will
be so restrictive that the regulated firm will not be able to
recover its initial investment or make a reasonable return on
it. There is, therefore, no "safe" direction in which the
regulator can move. Too little regulation yields monopoly
profits; too much results in confiscation. An unregulated
market does not pose these difficulties.

A second problem is parasitic on the first. At a first
approximation, the operation of one credit card payment
system looks more or less similar to another, but the
technology chosen or the book of customer business could
easily vary across firms. The cost structures of two or more
competitive firms could, therefore, differ in small ways that
add up in a high-volume business. In the absence of
regulation, each firm is saddled with its own cost structure.
Accordingly, there is little doubt that if system A offered
lower interchange fees than system B, we should expect
merchants, ceteris paribus, to gravitate toward the system
with the lower cost-either by taking only one card or by
steering their customers to the preferred card when both are
accepted. If such migration does not take place, then the
best explanation for that stable equilibrium is that each
company is serving the clientele in its own niche in the best
possible way. In the end, all card companies have an
incentive to reduce interchange costs to maximize profits.
There is no obvious need or way for a credit card company to
act opportunistically in an unregulated system.

The incentives under a system of regulation are much less
clear. The first question is whether the rate structure should
allow different interchange fees for payment systems that
bear different costs. (Of course, in unregulated markets, the
inefficient firm just fails to recover its full costs.) This issue

" The courts have long been divided over the best approach for
preventing monopoly profits without allowing confiscation. For the most
recent Supreme Court decision on the matter, see Duquesne Light Co. v.
Barasch, 488 U.S. 299 (1989). See also Jersey Cent. Power & Light Co. v.
FERC, 810 F.2d 1168 (D.C. Cir. 1987). For earlier Supreme Court
landmarks in this area, see Smyth v. Ames, 171 U.S. 361 (1898) and Fed.
Power Comm'n v. Hope Natural Gas, 320 U.S. 591 (1944).

THE REGULATION OF INTERCHANGE FEESNo. 3:55 11



arose in the last round of the Australian discussion, when
the representatives of both Visa and MasterCard noted that
there is no good solution to regulating rival schemes with
different cost structures.4' Any decision to use a "blended
rate" in setting the interchange fee would effectively serve to
distort the incentives of both firms. The firm with higher
costs will receive a rate that will not quite cover its expenses,
even if the costs it incurs are justified. The firm with the
lower rate will receive an extra bonus that bears no
relationship to its own costs. Yet, adopting a system that
allows the interchange fee to rise in response to higher costs
does nothing to control the operational efficiency of the
system or to rein in any component of monopoly pricing. In
the end, therefore, no form of interchange regulation is likely
to generate efficiency gains. Cost-based regulation forces the
firm that cuts costs to lose revenue if rates are calculated
with reference to some fixed rate of return. The regulated
firm would do better by raising costs, not by cutting them.
In unregulated markets, regardless of structure, the opposite
is true. The firm that lowers costs makes more profits
because it can internalize some portion of the relevant gain.
Those forces remain in the interchange fee market.

On balance, therefore, the overall situation is as follows.
The RBA's basic perception is that the high concentration in
the credit card industry suggests that the interchange fees
observed are higher than those that could be sustained in a
competitive market. If this is so, then the rate regulation
does not look like some special intervention tailored to
interchange relationships, but rather like typical rate-of-
return regulation on a natural monopoly thought to be a
product of the new technological age. Regulation is used
because a competitive equilibrium is not attainable in this
network industry. The regulatory initiative, therefore, is
best explained as an effort to offset some perceived
monopolistic imbalance stemming from the small number of
independent networks, but that is not the explicit rationale

40 See Clapham Letter, supra note 29, at 1-21.

[Vol. 2005COL UMBIA B USINESS LA W RE VIE W



THE REGULATION OF INTERCHANGE FEES

of the RBA.4 1 At this point, however, the RBA initiative runs
into the same objection that was raised many years ago by
Harold Demsetz in his classic article, Why Regulate
Utilities?: the public expenses incurred will not achieve the
static competitive solution.42 At the same time, the rate
structure operates as a systematic drag on long-term
innovation by robbing the innovative firm of the additional
revenues it otherwise deserves.43 The criticisms of the RBA
interchange regulations, therefore, must be tempered
because the credit card industry does not operate in a
perfectly competitive market. As a result, some regulation in
principle might create some overall improvement.
Nevertheless, one might wonder whether major regulatory
initiatives ought to rest solely on that kind of existential
theorem in light of the formidable obstacles that stand in
their path. The reforms that are necessary to correct one
admitted market imperfection introduce other costs, errors,
and distortions in their stead, which, in all likelihood, are
greater than those costs that regulation has eliminated. The
proper institutional response should take its cue from the
Hippocratic Oath: primum non nocere. First, do not harm.

VI. ANTITRUST IMPLICATIONS

In sum, the RBA's well-intentioned efforts to counteract
imperfections in the interchange market through regulation
may be worse than the disease itself. It does not follow,
however, that administrative actions are the wrong way to
regulate if some state intervention is required to address
some perceived monopolistic (or at least oligopolistic) power
inherent in these credit card networks. If some public
response is indeed required, then the administrative
approach, for all its failings, might be preferable to its most
common rival-the enforcement of antitrust laws through
private rights of action, as allowed in the United States

41 Id.
42 Harold Demsetz, Why Regulate Utilities?, 11 J. LAw & ECON. 55

(1968).
43 Id.

No. 3:551]



under the Clayton Act, for damages resulting from another
firm's improper business practices."

In the antitrust setting, as in the regulatory arena, there
is good reason to be cautious. Even under the best of
circumstances, it is difficult to design an ideal antitrust law.
The strongest case for intervention on that score is presented
by "hardcore" cartels in ordinary businesses that raise
prices, reduce quantity, and precipitate overall social loss.4"
Yet, even a straightforward question, such as whether to
allow a merger of two firms, is fraught with difficulty, for it
is not easy to trade off the deadweight losses that result from
the extension of monopoly power (if any) against the
efficiency gains that result from the change in operations.
Once one moves beyond these cases, the efficiency/restraint
tradeoffs are as hard, if not harder, to measure in the
judicial setting as they are in the administrative one.

In dealing with interchange fees, the administrative law
approach has major advantages. At the outset, combining
efforts of the various banks to work out one unified payment
system has huge social advantages that could not be
achieved by requiring each separate bank to set up a
standalone credit card system for its own customers, like the
one Bank of America ("BOA") originally tried.46 BOA quickly
refocused its effort when it realized that some collaborative
effort was necessary. Moreover, imposing remedies through
professional bodies improves the chances of finding a
decisionmaker who is cognizant of the costs and benefits of
alternative institutional arrangements. Most critically, an
administrative body is less likely to adopt an extreme
position than a runaway jury that operates free of all long-
term institutional restraints.

Clayton Act § 4, 15 U.S.C. § 15 (1992) ("any person who shall be
injured in his business or property by reason of anything forbidden in the
antitrust laws may sue therefore in any district court of the United
States").

" For a recent discussion, see HERBERT HOVENKAMP, THE ANTITRUST

ENTERPRISE: PRINCIPLE AND ExECUTION 26-28 (Harvard Univ. Press 2005).
46 EvANS & SCHMALENSEE, supra note 4, at 61-67.

[Vol. 2005_COL UMBIA B USINESS LA W RE VIE W



THE REGULATION OF INTERCHANGE FEES

This last point is of exceptional importance because the
entire edifice of antitrust law works best when independent
firms, facing constant marginal costs, conspire together to
raise prices above the competitive level, often by secret
agreements. In those cases, the harsh penalties offset the
risk of nondetection and respond to the imposition of a clear
social loss. But why bring the antitrust laws to bear against
any network industry whose internal structure is
transparent to all and whose configuration is driven heavily
by efficiency considerations? It is well understood that the
member banks in any credit card network compete
vigorously to sign up cardholders or merchants. Must they
also compete over the interchange fees that are passed
through to merchants? The basic premise of the four-party,
open-system arrangement is that any merchant can take a
branded card with comfort, even without knowledge of which
bank has issued it. The uniform fee means that the
merchant knows its costs, without having to learn which of
hundreds of possible banks issued the card in question. The
standard fee thus allows for seamless completion of high-
volume, low-margin transactions, which would stop dead in
their tracks if subjected to any requirement of individuated
negotiation. The competition comes as the issuing banks
seek to sign up customers, and as the credit card
intermediates attempt to adjust their interchange fees to
cover costs and provide the ideal balance between merchant
and user fees.

In this institutional setting, antitrust law is of little or no
use. In and of itself, it cannot be used to set rates, especially
those that are subject to variation over time, like
interchange fees. Its role, therefore, is much the same as in
all industries that have some rate regulation component. It
serves as a backstop to prevent illicit collusion above and
beyond that needed to keep any network intact. To use a
familiar example, the 1996 Telecommunications Act imposes
extensive regulations on the setting of interconnection fees

No. 3:551]



throughout the entire telecommunications network.47 The
1996 Act also contains an explicit provision preserving the
applicability of antitrust laws. Recently, however, the
Supreme Court in Verizon v. Trinko" took the correct
position that the extensive FCC regulation of the
telecommunications industry precluded the use of antitrust
laws to second guess decisions made in the regulatory arena,
especially at the instance of a plaintiff (Trinko) who did not
purchase its services directly from the regulated party
(Verizon).

The interchange fee regulation differs from the
telecommunications cases, at least in the United States,
insofar as there is no system that caps these fees like the
Australian model. Even so, if some effort to deal with any
concentration issue (of less concern in the United States than
in the more concentrated Australian market) is desired,
administrative regulation offers the proper forum in which to
raise those issues, as it does in all ratemaking contexts.
Given the strong efficiency properties of the current
arrangements, any breakup of the network would be wholly
misguided because of the massive dislocations that it would
create. Furthermore, a huge treble damages award (over a
nonexistent baseline) could easily gobble up a company's
entire revenue. For example, trebling the 0.40% reduction
ordered by the RBA would equal 1.20%, which is higher than
the original interchange fee. In short, so long as some
interchange fees are necessary to keep the network
functioning, antitrust laws should not apply in the absence of
price collusion between the different networks.

The critique of the use of antitrust law in this arena must
also consider the fundamental difference between the four-
party and three-party systems. MasterCard and Visa, out of
necessity, operate through combined action which is justified
in light of the network coordination issues that payment

" Telecommunications Act of 1996, Pub. L. No. 104-104, 110 Stat. 56
(1996) (codified as amended in scattered sections of 47 U.S.C.).

4 Verizon Commc'ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540
U.S. 398, 412-14 (2004).

COL UMIA B USINESS LA W RE VIE W [Vol. 2005



systems are designed to resolve. Nevertheless, they are
always faced with the charge that they operate as a
combination in restraint of trade notwithstanding the
manifest need for a uniform interchange fee. Firms like
Amex and Diners Club, which operate standalone three-
party systems, can set their own fees unilaterally, thereby
escaping the charge of collusion, regardless of their (limited)
market power in a concentrated payment card market. In
general, it creates major distortions to subject two competing
modes of business organizations to radically different
antitrust regimes. Right now, MasterCard is busy turning
itself into a proprietary organization, in part to minimize the
potential application of the antitrust laws.49 Other reasons
may also account for the shift, such as an improved ability to
raise capital. But even so, potential exposure to the
antitrust laws should not place a thumb on the scale, which
in some situations at least will motivate parties to adopt
inefficient business forms. If some form of intervention is
required, then a uniform regime of interchange fees, which
may allow for some interfirm variation in actual rates, is far
superior to subjecting the market to antitrust laws that are
unable to do the fine tuning necessary to undertake so
delicate an enterprise.

VII.CONCLUSION

This examination of the RBA's proposal demonstrates
how difficult it is to make and implement government
policies that improve upon the outcome of those
arrangements brought into operation by the ingenious
application of the principles of voluntary association. The
development of the payment card system owes nothing to
government support or government regulation. The system,
as it emerged, developed from private networks which, at

49 Robin Sidel, MasterCard Aims to Shift Ownership, WALL ST. J.,
Sept. 1, 2005, at C3 ("Several people also said that the partial sale of the

card association could help shield MasterCard from a series of recent
antitrust lawsuits over fees that merchants pay to accept and process
cards.").

THE REGULATION OF INTERCHANGE FEESNo. 3:55 11



COLUMBIA BUSINESS LAW REVIEW

their inception, appeared to be of limited scope and utility.
But the growth of networks that link individuals together
can take place in a gradual way. Individuals who have no
information about how networks are organized can sign on
with a bank. The bank then will enter into transactions with
other financial institutions in ways that promote their
mutual gain. Within the past fifty years, the transformation
in payment systems has taken place along multiple
dimensions, so much so that it is easy to lose track of the
gains that these innovations have wrought. But in thinking
back on my own life, I recall that in 1973 my wife and I used
a mix of American Express travelers' checks-We didn't
leave home without them!-and a letter of credit to pay for a
month's holiday in Europe. But that was in an age before
large credit card networks, ATMs, online banking,
immediate credit verification, and a thousand other
conveniences, large and small, became universally available.

The processes that generate these major gains always
look chaotic, and their results often leave short-term losers
that call out for further government regulation. Right now,
in the wake of Hurricane Katrina, gasoline dealers have
mounted a concerted protest against the rise in (fixed)
interchange fees that have cut into their profit margins with
the spike in gasoline prices.5" That being the case, there is
no need for piecemeal intervention, especially when these
same firms will see their interchange fees sink as prices
work their way back to lower levels. The self-help remedy of
taking only cash or debit cards-checks won't cut it in this
market-is far preferable to another round of litigation or
regulation that could be replicated in any other submarket
that experiences substantial price variation.

As this timely example illustrates, the case for state
intervention is not made out by a showing that someone can
devise a hypothetical system of fee restraints that is said to
respond to some higher test of rationality. All working
systems contain imperfections that are difficult to defend in

'0 See Margaret Webb Pressler, Card Companies Are Filling Up At the

Station, WASH. POST, Sept. 25, 2005, at FO.

[Vol. 2005



No. 3:55 1] THE REGULA TION OF INTER CHANGE FEES 597

some nonexistent first-best world. The basic social principle,
however, overlooks these blemishes and starts from a very
different premise: don't mess with success. The rapid
expansion of payment systems has exceeded the wildest
expectations possessed by anyone even a decade ago. This
expansion will not continue into the next decade if state
regulation in Australia or anywhere else expands its hold
over a complex system that has succeeded thus far without
the guidance of an all too visible hand.