The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Working Paper Series
Demand Externalities and Price Cap
Regulation: Learning from the U.S. Debit
Card Market
WP 13-06R
This paper can be downloaded without charge from:
http://www.richmondfed.org/publications/
Zhu Wang
Federal Reserve Bank of Richmond
Demand Externalities and Price Cap Regulation:
Learning from the U.S. Debit Card Market
Zhu Wang∗
July 2014
Working Paper No. 13-06R
Abstract
This paper studies unintended consequences of price cap regulation in the presence of demand externalities in the context of payment cards. The recent U.S. debit
card regulation was intended to lower merchant card acceptance costs by capping
the maximum interchange fee. However, small-ticket merchants found their fees
instead higher after the regulation. To address this puzzle, I construct a two-sided
market model and show that card demand externalities across merchant sectors
rationalize card networks’ pricing response. Based on the model, I study socially
optimal card fees and an alternative cap regulation that may avoid the unintended
consequence on small-ticket merchants.
Keywords: Price cap regulation; Demand externalities; Two-sided market
JEL Classification: D4; L5; G2
∗
Research Department, Federal Reserve Bank of Richmond. Email: zhu.wang@rich.frb.org. I thank
Wilko Bolt, Huberto Ennis, Darren Filson, Boyan Jovanovic, Grace Bin Li, and participants at the
Economics of Payments VI Conference hosted by the Bank of Canada, 2013 International Industrial
Organization Conference and various seminars for helpful comments. The views expressed herein are
solely those of the author and do not necessarily reflect the views of the Federal Reserve Bank of Richmond
or the Federal Reserve System.
1
1
Introduction
Credit and debit cards have become an important part of our payments system and they
affect a large number of consumers and merchants. Recent Federal Reserve studies show
that 80 percent of U.S. consumers have debit cards and 78 percent have credit cards. In
a typical month, 31 percent of consumer payments are made with debit cards, and 21
percent with credit cards.1
However, the pricing in the payment card markets has been controversial. As Rochet and Tirole (2006) pointed out, payment cards are so-called “two-sided markets,” in
which card networks serve two distinct end-user groups, namely, cardholders and merchants.2 In practice, card networks and their issuers typically charge high interchange
fees to merchants for card acceptance, but provide rewards to consumers for card usage.
Many industry observers and policymakers have become concerned that this highly skewed
pricing structure may distort payments efficiency by inflating merchants’ costs of accepting cards. Meanwhile, more than 20 countries have regulated or started investigating
interchange fees.
In the U.S., the Durbin Amendment to the Dodd-Frank Act has recently required
the Federal Reserve to regulate debit card interchange fees. Under the regulation, the
maximum permissible debit interchange fee for covered issuers is capped at a half of its
pre-regulation industry average level. As a direct impact, card issuers lost multibilliondollar annual interchange revenues to merchants. However, the regulation has also generated unintended consequences on certain merchant groups. Particularly, prior to the
regulation, merchants were charged differentiated interchange fees based on their sectors.
Post regulation, however, card networks set a uniform interchange fee at the maximum
cap amount. As a result, small-ticket merchants who used to pay lower interchange fees
found their rates instead increased. In essence, the price cap has become a price floor.
The unintended consequence on small-ticket merchants made headlines and resulted
in a lawsuit filed by merchant groups against the Federal Reserve’s debit interchange
1
Kevin Foster et al. (2010).
The research on two-sided markets recently has gained wide attention (Rysman 2009). Other examples include HMOs (patients and doctors), operating systems (computer users and software developers),
video game consoles (gamers and game developers), and newspapers (advertisers and readers).
2
2
regulation.3 This presents a puzzle: Why would card networks raise fees on small-ticket
merchants in response to a fee cap? If each merchant sector is independent in terms of
card acceptance and usage and networks find that they maximize profits by charging lower
fees to small-ticket merchants, it is not obvious why they would abandon this strategy in
the face of a cap that is higher than the fees they were charging.
This puzzle is not readily explained by the existing two-sided payment card market
models (e.g. Rochet and Tirole 2002, 2011, Wright 2003, 2011). Those studies find that
privately determined interchange fees tend to exceed the socially efficient level because
of the wrong incentives at the point of sale, i.e. consumers pay the same retail price
regardless of the payment instrument they use. However, those models typically treat
merchant sectors independent from one another in terms of card acceptance and usage,
so they do not predict or explain why some merchants would be adversely affected by an
interchange cap that is not binding for them.
In this paper, I address this puzzle by introducing card demand externalities into a
two-sided market framework. In the model, merchant sectors are charged differentiated
interchange fees due to their (observable) heterogenous benefits of card acceptance and
usage. In addition, consumers’ benefits of using cards in a merchant sector are positively
affected by their card usage in other sectors, which I call “ubiquity externalities.”4 This
type of demand externalities is shown to drive card networks’ response to the cap regulation: Before the regulation, card networks were willing to offer subsidized interchange fees
to small-ticket merchants because their card acceptance boosts consumers’ card usage for
large-ticket purchases from which card issuers can collect higher interchange fees. Once
a cap on interchange fees was imposed, however, card issuers profit less from this kind of
externalities so they discontinued the subsidy.
3
E.g. see “Debit-Fee Cap Has Nasty Side Effect,” Wall Street Journal, December 8, 2011.
Ubiquity has always been a top selling point for brand cards. This is clearly shown in card networks’
campaign slogans, such as Visa’s “It is everywhere you want to be,” and MasterCard’s “There are some
things money can’t buy. For everything else, there’s MasterCard.” Ubiquity externalities may arise from
various sources. First, in the presence of a fixed adoption cost, consumers are more likely to adopt
payment cards if the card is accepted by more merchants. Second, for consumers who have adopted
cards, universal card acceptance may allow them to carry less cash and as a result rely more on cards
for making payments. Third, universal card usage may allow card networks and issuers to collect more
complete information on consumer shopping patterns, so that they can design better services to encourage
further card usage (e.g. by offering more targeted card reward programs). All these ubiquity externalities,
regardless of their sources, are consistent with our following analysis.
4
3
Based on the model, I then study socially optimal card fees and alternative regulations.
The analysis shows that the social optimum generally would require lower interchange fees
than those chosen by the private market, but nevertheless it may maintain the differentiated fee structure by charging high (respectively, low) interchange rates to large-ticket
(respectively, small-ticket) merchants. This is because both the social and the private
optima seek to internalize the positive externalities of card usage across merchant sectors
by subsidizing small-ticket transactions. In the presence of card demand externalities, I
further show that capping the weighted average interchange fee, instead of the maximum
interchange fee, may help restore the social optimum.
The contribution of the paper is threefold. First, I address a puzzle of the debit card
interchange regulation by showing a “waterbed effect” may be at work, where regulating down the price of one sector may reduce the cross-subsidies that this sector provides
to another one. This provides a rational explanation for the unintended consequences
following the regulation. Second, I embed the analysis in an extended two-sided market model. In contrast to the existing payment card literature, the new model considers
endogenous issuer competition, heterogenous merchant sectors, and card demand externalities. Exploring these features yields a better understanding of both the structure
and the levels of socially optimal interchange fees. I show that the socially optimal fee
structure may allow price discrimination, and the fee levels are determined by multiple
factors, including merchant-and-consumer net benefits of card usage (which are subject
to ubiquity externalities), the competitiveness of issuers, and the acquirers’ cost. These
new results suggest that the popular interchange regulations adopted in various countries,
solely based on either issuer costs or merchant benefits, may have inadequate theoretical
foundation.5 Finally, I propose an alternative regulation that caps the weighted average
interchange fee. The alternative regulation is shown to provide incentives for card net5
Two types of interchange fee regulations are currently in practice. One is based on issuers’ costs,
first adopted by the Reserve Bank of Australia in early 2000. The Durbin regulation in the U.S. is a
recent example. The issuer-cost based regulation has been criticized for ignoring the two-sided nature of
payment card markets. Instead, Rochet and Tirole (2011) proposed regulating the interchange fee based
on merchant transaction benefit of card acceptance, which was adopted by the European Commission.
The merchant-benefit based regulation addresses the two-sided market concerns, but relies on a strong
assumption that issuers set a constant markup. Moreover, neither type of the regulations has considered
card demand externalities across merchant sectors.
4
works and issuers to internalize card demand externalities and hence avoid unintended
consequences on small-ticket merchants.
The paper is organized as follows. Section 2 provides the background of the payment
card industry and the debit interchange fee regulation. Section 3 lays out a two-sided
payment card market model with heterogenous merchant sectors and differentiated interchange fees. The model allows for card demand externalities across merchant sectors.
Section 4 characterizes the model equilibria with and without the interchange cap regulation. Section 5 discusses socially optimal interchange fees and an alternative cap
regulation. Section 6 provides concluding remarks.
2
Industry background
Credit and debit cards have become an increasingly important part of the U.S. payments
system. Recent data show that the share of their transactions in personal consumption
expenditures rose to 48 percent in 2011. Among those, credit cards were used in 26
billion transactions for a total value of $2.1 trillion, and debit cards were used in 49
billion transactions for a value of $1.8 trillion.6
Credit cards typically provide float or credit to cardholders, while debit cards directly
draw from the cardholder’s bank account right after each transaction. In practice, debit
card payments are authorized either by the cardholder’s signature or with a PIN number.
The former accounts for 60 percent of debit transactions and the latter accounts for 40
percent.
Visa and MasterCard are the two major card networks in the United States. They
provide card services through member financial institutions (issuers and acquirers) and
account for 85 percent of the U.S. consumer credit card market.7 Visa and MasterCard
are also the primary providers of debit card services. The two networks split the signature
debit market, with Visa holding 75 percent of the market share and MasterCard holding
6
Source: Nilson Report, December 2011. Prepaid cards are another type of general-purpose cards but
with much smaller volumes. They accounted for 2% of U.S. personal consumption expenditures in 2011.
7
American Express and Discover are the other two credit card networks holding the remaining market
share. They handle most card issuing and merchant acquiring by themselves and are called “three-party”
systems. For a “three-party” system, interchange fees are internal transfers.
5
25 percent.8 In contrast, PIN debit transactions are routed over PIN debit networks.
Interlink, Star, Pulse and NYCE are the top four networks, together holding 90 percent
of the PIN debit market share. The largest PIN network, Interlink, is operated by Visa.
2.1
Interchange controversy
Along with the development of payment card markets, there has been a long-running controversy about interchange fees. Merchants are critical of the fees that they pay to accept
cards. These fees are referred to as the “merchant discounts,” which are composed mainly
of interchange fees paid to card issuers (i.e., banks issuing cards and make payments on
behalf of cardholders) through merchant acquirers (i.e., banks collecting payments on behalf of merchants). Merchants believe that the card networks and issuers have wielded
their market power to set excessively high interchange fees. The card networks and issuers
counter that these interchange fees are necessary for covering issuers’ costs as well as providing rewards to cardholders, which may also benefit merchants by making consumers
more willing to use the cards.
In recent years, merchant groups launched a series of litigation against what they claim
is anticompetitive behavior by the card networks and their issuers. Some of the lawsuits
have been aimed directly at interchange fees of credit and debit cards. For example,
a group of class-action suits filed by merchants against Visa and MasterCard alleged
that the networks violated antitrust laws by engaging in price-fixing. As a result, Visa,
MasterCard and their major issuers reached a $5.7 billion settlement agreement with U.S.
retailers in December 2013, which is the largest antitrust settlement in U.S. history.
The heated debate on interchange fees has also attracted attention from researchers
and regulatory authorities. On the research side, a sizeable body of literature, called “twosided market theory,” has been developed to evaluate payment card market competition
and pricing issues.9 On the regulatory side, three bills restricting interchange fees were
8
Discover has recently entered the signature debit market, but its market share is small.
For example: Baxter (1983), Carlton and Frankel (1995), Katz (2001), Schmalensee (2002), Rochet
and Tirole (2002, 2006, 2011), Gans and King (2003), Wright (2003, 2004, 2010, 2012), Cabral (2005),
Armstrong (2006), Schwartz and Vincent (2006), Rysman (2007, 2009), Bolt and Chakravorti (2008),
Robin Prager et al. (2009), Rochet and Wright (2010), Wang (2010), Weyl (2010), Shy and Wang (2011),
McAndrews and Wang (2012), and Bedre-Defolie and Calvano (2013).
9
6
introduced in Congress shortly before the Durbin Amendment was passed.10 Similar
trends are also taking place in many other countries.11
2.2
Durbin regulation
In 2010, an amendment sponsored by Sen. Dick Durbin was added to the Dodd-Frank bill,
which was passed and signed into law in July 2010. The Durbin Amendment directs the
Federal Reserve Board of Governors to ensure that debit card interchange fees are “reasonable and proportional to the cost incurred by the issuer with respect to the transaction.”
The Federal Reserve Board thereafter issued Regulation II (Debit Card Interchange Fees
and Routing), which went into effect on October 1, 2011.
The new regulation establishes a cap on the debit interchange fees that banks with
more than $10 billion in assets can collect from merchants through merchant acquirers.
The permissible fees were set based on the Fed’s evaluation of issuers’ costs associated
with debit card processing, clearance and settlement. The resulting interchange cap is
composed of the following: A base fee of 21 cents per transaction to cover the issuer’s
processing costs, a five basis point adjustment to cover potential fraud losses, and an
additional 1 cent per transaction to cover fraud prevention costs if the issuer is eligible.
This cap applies to both Signature and PIN debit transactions.
The regulation has a major impact on card issuers’ interchange revenues. According
to a recent Federal Reserve study, the average debit card transaction in 2009 was approximately $40. Based on the regulation, the interchange fee applicable to a typical debit
card transaction would be capped at 24 cents (21 cents + ($40 × .05%) + 1 cent), which
is about half of its pre-regulation industry average level. As a result, card issuers were
expected to lose an estimated $8.5 billion annual interchange revenues.12
In response to the reduced interchange revenues, many card issuing banks have cut
back their debit reward programs and free checking services. A recent Pulse debit issuer
10
The three bills are a House version of the Credit Card Fair Fee Act of 2009, a Senate version of the
same act, and the Credit Card Interchange Fees Act of 2009.
11
Recent examples of interchange fee regulation include Argentina, Australia, Austria, Brazil, Canada,
Chile, China, Colombia, Denmark, European Union, France, Hungary, Israel, Mexico, New Zealand,
Norway, Panama, Poland, Portugal, South Africa, South Korea, Spain, Switzerland, Turkey, and United
Kingdom.
12
Wang (2012) provides some estimates of issuers’ lost interchange revenues using Call Report data.
7
study shows that 50 percent of regulated debit card issuers with a reward program ended
their programs in 2011, and another 18 percent planned to do so in 2012.13 Meanwhile,
the Bankrate’s 2012 Checking Survey shows that the average monthly fee of noninterest
checking accounts rose by 25 percent compared with the year before, and the minimum
balance for free-checking services rose by 23 percent.14 Several major banks including
Bank of America, Wells Fargo, and Chase attempted to charge a monthly debit card fee
to their customers, though they eventually backed out due to customer outrage.15
2.3
Small-ticket effect
Merchants as a whole may have greatly benefited from the reduced debit interchange
fees under the regulation.16 However, the distribution of the benefits is quite uneven.
Particularly, an unintended consequence quickly surfaced: Small-ticket merchants find
their interchange fees higher after the regulation.
Prior to the regulation, Visa, MasterCard, and most PIN networks offered discounted
debit interchange fees for small-ticket transactions as a way to encourage card acceptance
by merchants specializing in those transactions.17 For instance, Visa and MasterCard set
the small-ticket signature debit interchange rates at 1.55 percent of the transaction value
plus 4 cents for sales of $15 and below. As a result, a debit card would only charge a 7
cents interchange fee for a $2 sale or 11 cents for a $5 sale. However, in response to the
regulation, most card networks eliminated the small-ticket discounts, and all transactions
(except those on cards issued by exempt issuers) have to pay the maximum cap rate set
by the Durbin regulation.18 For merchants selling small-ticket items, this means that the
13
The 2012 Debit Issuer Study, commissioned by Pulse, is based on research with 57 banks and credit
unions that collectively represent approximately 87 million debit cards and 47,000 ATMs.
14
Bankrate surveyed banks in the top 25 U.S. cities to compare the average fees associated with checking
accounts in their annual Checking Account Survey.
15
See “Banks Adding Debit Card Fees,” The New York Times, September 29, 2011.
16
Depending on merchant comptition, some of the benefits may be passed along to consumers through
lower retail prices.
17
Visa and MasterCard introduced small-ticket discounted interchange fees in the early 2000s. The
rates were applied to merchant sectors specializing in small-ticket transactions, including Local Commuter
Transport, Taxicabs and Limousine, Fast Food Restaurants, Coffee Shops, Parking Lots and Garages,
Motion Picture Theaters, Video Rental Stores, Cashless Vending Machines and Kiosks, Bus Lines, Tolls
and Bridge Fees, News Dealers, Laundries, Dry Cleaners, Quick Copy, Car Wash and Service Stations,
etc. In October 2010, Visa expanded the program to include more merchant sectors.
18
Hayashi (2013) compares the increases of interchange fees for small-ticket transactions for Visa,
8
cost of accepting the same debit card doubled or even tripled after the regulation.
The increase of small-ticket interchange fees could affect a large number of transactions. According to the 2010 Federal Reserve Payments Study, debit cards were used for
4.9 billion transactions below $5, and 10.8 billion transactions between $5-$15 in 2009.
The former accounts for 8.3 percent of all payment card transactions (including credit,
debit, and prepaid cards), and the latter accounts for 18.3 percent. Depending on their
compositions of transaction sizes, merchants in different sectors could be affected differently by the post-regulation debit interchange fees.19 However, merchants who specialize
in small-ticket transactions would be most adversely affected.
In response, many small-ticket merchants have tried to find ways to offset their higher
interchange rates. Some raised prices, or chose to restrict or reject the use of debit cards.20
Some others offer customers incentives to consolidate transactions using prepaid cards or
online wallets.21 In the meantime, a lawsuit was filed in November 2011 by a group of
trade associations and retail companies against the Federal Reserve’s debit interchange
regulation. The lawsuit alleges that the Fed has set the interchange cap too high by
including costs that were barred by the law, and “forcing small businesses to pay three
times as much to the big banks on small purchases was clearly not the intent of the law
and is further evidence that the Fed got it wrong.”22
The unintended regulatory impact on small-ticket merchants calls for a further examination of the payment card market. According to the Federal Reserve’s evaluation, debit
card issuers incur a per-transaction cost around 21 cents, which exceeds the interchange
MasterCard, and most PIN debit networks.
19
E.g. Shy (2012) used the data from a diary study of consumer payment choices to identify the types
of merchants who are likely to pay higher or lower interchange fees under the debit regulation.
20
Notable examples in the press include: the DVD-rental company Redbox raised rental prices from
$1 to $1.20 to cover increased debit fees; USA Technologies and Apriva, two large payment facilitators
in the vending industry, stopped accepting MasterCard debit cards; the fast food restaurant chain Dairy
Queen asked customers to pay with cash for purchases under $10. See “Debit-Fee Cap Has Nasty Side
Effect,” Wall Street Journal, December 8, 2011.
21
Merchants are charged one transaction fee when a customer loads the prepaid card or online wallet
rather than multiple times each instance a user pays with a debit card. Notable examples in the press
include coffeehouse chain Starbucks promoting in-store prepaid cards and Washington, D.C. parking
operator Parkmobile offering discounts for customers who pay with an online wallet. E.g. see “SmallTicket Retailers Squeezed By High Transaction Fees,” U.S.News & World Report, October 26, 2012.
22
See “Merchants’ Lawsuit Says Fed Failed to Follow Law on Swipe Fee Reform,” Business Wire,
November 22, 2011.
9
fees that they charged for small-ticket transactions prior to the regulation. Considering
that issuers typically do not recover costs from the cardholder side (cardholders often
receive a reward rather than pay a fee for each card transaction), card issuers appeared
to have subsidized small-ticket transactions. This is also the reason that card networks
claim why they eliminate the small-ticket discounts under the regulation.23 While the
existing two-sided market theories have shed great light on the functioning of interchange
fees, they do not explain the subsidies on small-ticket transactions prior to the Durbin
regulation, nor do they explain why these subsidies were discontinued afterwards. In this
paper, I try to address this puzzle and draw some new implications.
3
Model environment
I consider a payment card system composed of five types of players: consumers, merchants,
acquirers, issuers, and the card network, as illustrated in Figure 1. The setup extends the
standard two-sided market model, such as Rochet and Tirole (2002, 2011), to allow for
card demand externalities across merchant sectors.
¥ Consumers: There is a continuum of measure one of consumers, who purchase
goods from two distinct merchant sectors and . In this setting, (respectively, ) refers
to the large-ticket (respectively, small-ticket) sector where merchants and consumers enjoy
high (respectively, low) transaction benefits of card acceptance and usage.24
Consumers have inelastic demand and buy one good per sector. Within each sector,
consumers need to decide which store to patronize. They know the stores’ price and card
acceptance policy before making the choice. Once in a store they then select a payment
method (a card or an alternative payment method such as cash), provided that the retailer
indeed offers a choice among payment means. I assume price coherence such that retailers
charge the same price for purchases made by card and by cash.25 Whenever a transaction
23
According to MasterCard, “the company decided that it couldn’t sustain the [small-ticket] discounts
under the new rate model because the old rates had essentially subsidized the small-ticket discounts.”
See “Debit-Fee Cap Has Nasty Side Effect,” Wall Street Journal, December 8, 2011.
24
For both merchants and consumers, replacing cash with cards may reduce their handling, safekeeping
and fraud expenses on payments, and the benefits typically increase with transaction values. Therefore,
it is natural to assume that merchants and consumers benefit more from card usage in large-ticket
transactions than in small-ticket transactions.
25
Price coherence is the key feature that defines a two-sided market. Rochet and Tirole (2006) show
10
C a rd N e tw o rk
(S ets in ter c ha ng e f ee a)
I s su e r
A cq u ire r
( ne t c o s t c B )
( ne t c os t c S )
P a ys p - a
( a: inte rc h an ge fee )
P ay s p - f S
(f S : me rc h an t
dis c o unt )
P a y s p + fB
(f B : c ons um er fe e)
Con s u m e r
M e rc h a n t
( n et b en efit b B )
( ne t b ene fit b S )
S e lls g oo d
a t pr ic e p
Figure 1: A Payment Card System
between a consumer (buyer) and a retailer (seller) is settled by card, the buyer pays a
fee to her card issuing bank (issuer) and the seller pays a merchant discount to
her merchant acquiring bank (acquirer). These fees, and , depend on the merchant
sector ∈ { }, and is allowed to be negative, in which case the cardholder receives
a reward. There are no annual fees and all consumers have a card.26
A consumer’s transaction benefit of purchasing good with a card instead of cash is
a random variable drawn from a cumulative distribution function on the support
[ ]. It is natural to think that the mean and variance of consumer card usage benefits
positively relate to the transaction value. Denoting as the mean of , this implies that
and ( ) ( ). For simplicity, I thereafter assume to be a uniform
distribution on the support [ ( ) − ( ) + ], while is a degenerate distribution
that the two-sided market pricing structure (e.g. interchange fees) would become irrelevant without
the price coherence condition. In reality, price coherence may result either from network rules or state
regulation, or from high transaction costs for merchants to price discriminate based on payment means.
In the U.S., while merchants are allowed to offer their customers discounts for paying with cash or checks,
few merchants choose to do so. On the other hand, card network rules and some state laws explicitly
prohibit surcharging on payment cards.
26
This model assumes a representative consumer framework developed by Wright (2004) and used in
the subsequent literature. Alternatively, the model could use the framework developed by Rochet and
Tirole (2002) and assume heterogenous consumers who differ systematically in their transaction benefits
of using cards. As Rochet and Tirole (2011) show, these two alternative frameworks deliver convergent
results, so the analysis and findings can be interpreted using either framework.
11
taking a single value . The latter is an innocuous assumption given that the variance
of is sufficiently small.27 Moreover, I assume that is positively affected by the
consumer’s card usage in the small-ticket sector , i.e. 0. (I define =
, where indicates whether -sector merchants accept cards and is a consumer’s
frequency of card usage in the sector conditional on cards being accepted).28 This
assumption captures the idea that ubiquity externalities shift up consumers’ valuation of
paying with cards in the sector.29
Cardholders are assumed to observe the realization of once in the store. This is a
standard assumption introduced by Wright (2004) and used in the subsequent literature.
Because the net benefit of paying by card is equal to the difference − , a card payment
is optimal for the consumer whenever ≤ . Hence, whenever ≤ ( ) + , the
proportion of card payments at an -sector (i.e. large-ticket) store that accepts cards is
( ) = Pr( ≥ ) =
( ) + −
2
(1)
and the average net consumer benefit of paying with a card is
( )
=
[
−
|
≥
]
( ) + −
=
2
(2)
Note that ( ) = ( ) = 0 if ( ) + .
Similarly, whenever ≤ , the proportion of card payments at an -sector (i.e.
small-ticket) store that accepts cards is
( ) = Pr( ≥ ) = 1
27
(3)
Intuitively, we can think that a consumer’s transaction benefit of paying by card relative to using cash
is a random variable = , where is a random factor and is the price of good largely determined by
the non-payment cost of the good. This implies that ( ) ( ) and ( ) ( ). Moreover,
given that is small, both ( ) and ( ) could be close to zero.
28
Under the assumption that is degenerate, consumer card usage in the sector becomes a simple
binary outcome, i.e. ∈ {0, 1}. This makes it easier to model card usage externalities between the
and sectors. Note that if is a non-degenerate distribution, we then need to specify how card
demand externalities vary by each of the multiple levels of card usage in the sector, which significantly
complicates the problem but does not provide greater intuition.
29
For ease of exposition, I assume that consumers’ transaction benefit of using cards in the sector is
fixed, unaffected by card usage in the sector. However, relaxing this assumption would not change the
qualitative findings.
12
and the average net consumer benefit of paying with a card is
( ) = −
(4)
Note that ( ) = ( ) = 0 if .
¥ Merchants: Merchants belong to one of the two sectors, and . A merchant in
a given sector ∈ {, } derives the transaction benefit of accepting payment cards
(relative to handling cash), and naturally . Moreover, the heterogeneity between
sectors is observable to the card network so that the card network can perfectly price
discriminate by charging differentiated interchange fee to the merchant sector .
By accepting cards, under the price coherence assumption, a merchant is able to offer
each of its card-holding customers an additional expected surplus of ( ) ( ), but faces
an additional expected net cost of ( )( − ) per cardholder from doing so. Here, is
the sector-specific merchant discount paid to the acquirer. Therefore, a merchant accepts
cards if and only if ≤ + ( ). Rochet and Tirole (2011) show this condition holds
for a variety of merchant competition setups, including monopoly, perfect competition and
Hotelling-Lerner-Salop differentiated products competition with any number of retailers.
Wright (2010) shows the same condition holds for Cournot competition.
I denote as an indicator function whether merchants in sector accept cards or not.
Accordingly,
⎧
⎨ 1 if ≤ + ( )
=
⎩ 0 otherwise
(5)
Note that merchants in the and sectors do not directly coordinate to internalize
card usage externalities. This is a realistic assumption given that there could be a large
number of merchants in each sector, which makes the coordination too costly. Moreover,
due to antitrust restrictions, merchants in reality can not engage in group bargaining
regarding interchange fees, so they typically face “take-it-or-leave-it” offers from card
networks.
¥ Acquirers: I assume acquirers incur a per-transaction cost and are perfectly
competitive. Thus, given the interchange fee , they charge a sector-specific merchant
13
discount such that
= +
(6)
Because acquirers are competitive, they play no role in the analysis except passing through
the interchange charge to merchants.
¥ Issuers: There are ≥ 1 issuers who have market power.30 Issuers incur a per-
transaction cost and receive an interchange payment of in a card transaction. I
consider a symmetric equilibrium at which all issuers charge the same consumer fee ,
which can be negative if cardholders receive a reward.
As pointed out in Rochet and Tirole (2002, 2011), there are various ways to model
issuer competition. To be concrete, I assume an explicit setting: Issuers coordinate on
their pricing in the sector where they make a loss (so that they internalize card demand externalities between the and sectors), but engage in a Cournot competition in
the sector where they make a profit. The former assumption simplifies the setting of
small-ticket card fees in order to focus on the card demand externalities, while the latter
assumption allows for endogenizing issuers’ markup for large-ticket transactions. Note
when = 1, the model reduces to a special case where there is a monopoly issuer.31
For small-ticket transactions, issuers take the interchange fee as given and set the
consumer fee to maximize their total profit conditional on merchants accepting cards
(i.e. = 1):
Π̂ = max( + − )
30
⎧
⎨ 1 if ≤
=
,
⎩ 0 otherwise
(7)
(8)
This is a standard assumption in the literature. As pointed out in Rochet and Tirole (2002), the
issuer market power may be due to marketing strategies, search costs, issuer reputation or the nature of
the card. Note that were the issuing side perfectly competitive, issuers and card networks would have no
preference over the interchange fee, and so the latter would be indeterminate.
31
The assumption that payment card issuers engage in Cournot competition is consistent with Rochet
and Tirole (2002). Alternatively, we could assume symmetric monopoly issuers, each making their own
pricing decisions in and sectors to internalize card demand externalities. The assumption of monopoly
issuers is likely to be true for the case of debit cards because a debit card holder typically has long-term
banking relationship with her card issuer. Our following analysis can equally apply to this alternative
setup (by simply setting = 1), in which case we show the welfare-maximizing interchange fees coincide
with the market-determined ones, but the user-surplus-maximizing interchange fees are lower.
14
where (8) follows (3). Whenever = 1, the highest possible consumer fee that issuers
choose is
=
(9)
and the corresponding total issuers’ profit in the sector is
Π̂ = + −
(10)
For large-ticket transactions, issuers engage in a Cournot competition if merchants
accept cards (i.e. = 1). Each issuer sets the output level taking the output by
competing issuers, −
= − , as given and maximizes profit:
̂ = max( + − )
= ( ) + − 2( + −
)
(11)
(12)
where (12) follows Eq (1). In a symmetric equilibrium, the total card usage and the
consumer fee are pinned down as follows:
= =
=
[ ( ) + + − ]
2( + 1)
1
[ ( ) + + ( − )]
+1
(13)
(14)
and the total issuers’ profit in the sector is
Π̂ =
[ ( ) + + − ]2
2( + 1)2
(15)
¥ Network: I consider a monopoly network that sets sector-specific interchange fees
and to maximize the total issuers’ profit, namely
Π = max (Π̂ + Π̂ )
where Π̂ and Π̂ are given by Eqs (10) and (15) above.
15
(16)
Because the network maximizes the issuers’ profit, it makes a decision consistent with
issuers on whether to provide card services to the sector. Therefore, = always holds
at equilibrium, so we can simply replace ( ) with ( ) in the following analysis.
In the welfare and policy analysis (Section 5), I will also consider an alternative regime
where the network is run by a social planner who maximizes social welfare or total user
surplus.
¥ Timing: I solve for a subgame perfect Nash equilibrium of the model. The timing
of the game can be summarized in the following four stages.
1. The card network sets sector-specific interchange fees .
2. Issuers and acquirers set fees and .
3. Depending on their value of , merchants decide whether to accept cards and set
retail prices.
4. Observing which merchants accept cards and their prices, consumers decide which
merchants to purchase from. Once in the store, consumers receive their draw of
and decide how to pay.
4
Model characterization
I first consider a monopoly network, which sets sector-specific interchange fees to maximize the total issuers’ profit. In the absence of regulation, the network solves the following
problem:
[ ( ) + + − ]2
+ ( + − )
2( + 1)2
Π = max
⎧
⎨ 1 if ≤ [ ( )++ − − ] + −
+2
=
⎩ 0 otherwise
⎧
⎨ 1 if ≤ −
=
⎩ 0 otherwise
16
(17)
(18)
(19)
The condition (18) is derived from (2), (5), (6) and (14), while the condition (19) is
derived from (4), (5), (6) and (9).
Once an interchange fee cap is introduced by regulation, the network then solves a
similar problem as above but with an additional constraint:
≤ for ∈ { }
(20)
To help characterize the model equilibrium, I make three basic assumptions on parameter values.
Assumption A1.
( ) = + ( ) + − − 0 for ∈ {0 1}
The first assumption states that the maximum merchant-and-consumer joint transaction benefit of using cards in the sector net of costs is always positive As will be shown,
this ensures that issuers earn a positive profit for serving card transactions in the sector.
Assumption A2.
= + − − 0
The second assumption states that the merchant-and-consumer joint transaction benefit of using cards in the sector net of costs is negative. As will be shown, this implies
that card issuers make a loss for serving card transactions in the sector per se.
Assumption A3.
(1) − (0) = (1) − (0)
( + 2)2 (− )
2[ (0) + (1)]
The third assumption states that card demand externalities are sufficiently large between the and sectors. As will be shown, this ensures that in the absence of regulation,
the card network would charge differentiated interchange fees to serve card transactions
in both the and sectors.
17
Under the above assumptions, I first characterize the model equilibrium in the absence
of regulation. The findings are shown by the following proposition.
Proposition 1 Under Assumptions A1-A3, an unregulated card network which maximizes total issuers’ profit sets differentiated interchange fees such that cards are used in
both the h and l sectors.
Proof. Consider three options for the card network. First, when only the sector is
served with card services (i.e. = 1 = 0), the card network maximizes the total
issuers’ profit (17) by setting the -sector interchange fee such that the constraint (18) is
binding
( = 0) = − +
(0)
+2
(21)
As a result, the total number of card transactions is
(0)
( + 2)
(22)
2
[ (0)]2
( + 2)2
(23)
=
and the total issuers’ profit is
Π =
Under Assumption A1, this implies that 0 and Π 0
Second, when only the sector is served with the card services (i.e. = 0 = 1),
the card network maximizes the issuers’ profit (17) by setting the -sector interchange fee
= −
(24)
Under Assumption A2, the total issuers’ profit is
Π = = + − − 0
(25)
Finally, when both the and sectors are served with card services (i.e. = = 1),
the card network maximizes the issuers’ profit (17) by charging differentiated interchange
18
fees to the two sectors:
( = 1) = − +
(1)
+2
= −
(26)
(27)
The resulting total issuers’ profit is
Π+ =
2
[ (1)]2 +
2
( + 2)
(28)
Comparing Eqs (21), (26) and (27), it is found that the interchange fee is always higher
in the sector than the sector, i.e.
( = 1) ( = 0)
(29)
given that and (1) (0) 0. Comparing (23) and (28), it is also verified
that Π+ Π iff Assumption A3 holds. Therefore, under Assumptions A1-A3, the card
network charges differentiated interchange fees given by (26) and (27) and serves card
transactions in both the and sectors.
In comparison, I now characterize the model equilibrium under the interchange cap
regulation. Under the regulation, the card network needs to solve the problem (17) subject
to the cap constraint (20) in addition to (18)-(19). The goal here is to derive conditions
that rationalize the card network’s pricing response to the cap regulation as seen in the
market. Namely, under the regulation, the card network charges a single interchange fee
exactly at the cap level . As a result, merchants in the sector continue to accept card,
but merchants in the sector do not.
Recall Eq (29) that ( = 1) ( = 0) . For the purpose stated, I consider
a cap level that satisfies ( = 0) ≥ . This ensures that the cap is binding for
the sector regardless of whether or not the sector is served with card services.32 I now
32
Note that if the cap value is set at a level such that ( = 1) ( = 0), the cap would
not be binding for the sector if the sector is dropped out of the card services. The case could be a
theoretical possibility, but is less relevant for explaining the market reality.
19
establish the following proposition.
Proposition 2 Given any interchange cap that satisfies ( = 0) ≥ , the card
network sets a single interchange fee at such that cards are used only in the sector if
the following condition holds
2( + 1)2 (− )
(1) − (0)
[ (1) + 3+2
(0)]
+2
(A4)
Proof. Given that ( = 0) ≥ , the cap is binding for the sector regardless
of whether or not the sector is served with card services. Therefore, Eqs (13) and (14)
imply that
=
[ ( ) + + − ]
2( + 1)
=
1
[ ( ) + + ( − )]
+1
If both the and sectors are served (i.e. = = 1), the total issuers’ profit is
Π+ = ( + − ) + =
[ (1) + + − ]2 +
2( + 1)2
In contrast, if only the sector is served (i.e. = 1, = 0), the total issuers’ profit is
Π = ( + − ) =
[ (0) + + − ]2
2
2( + 1)
Therefore, Π+ Π iff
(1) − (0)
2( + 1)2 (− )
[ (1) + (0) + 2 + 2 − 2 ]
(30)
Because ≤ ( = 0), a sufficient condition for (30) to hold is that
2( + 1)2 (− )
(1) − (0)
[ (1) + (0) + 2 + 2 ( = 0) − 2 ]
(31)
Inserting the expression of ( = 0) from Eq (21), the condition (31) can then be
rewritten as (A4).
20
Under Assumptions A1-A2, it is straightforward to verify that
2( + 1)2 (− )
( + 2)2 (− )
[ (1) + 3+2
(0)] 2[ (0) + (1)]
+2
Therefore, there exists a non-empty set of values that satisfy Assumption A3 and Condition A4. Hence, for any value of (1) − (0) within that set, the card network sets
differentiated interchange fees to serve both the and sectors in the absence of regulation, and sets a single interchange fee at such that only the sector is served with the
card services under the cap regulation.
5
Welfare and policy analysis
I have provided a model that rationalizes card networks’ interchange pricing before and
after the cap regulation introduced by the Durbin Amendment. The analysis suggests
that card demand externalities between the small-ticket and large-ticket sectors could
play an important role in explaining card networks’ response to the regulation. Based on
the model framework, I now take a step further to conduct welfare and policy analysis.
5.1
Welfare maximization
I first consider an alternative regime where the network is run by a social planner who
maximizes social welfare.33 Social welfare is generated whenever consumers use cards
for payment at retailers provided consumer-and-merchant combined transaction benefits
exceed the combined costs (i.e., + + ), which is shown as
X
∈{}
Ã
Z
!
[ + − − ] ( )
(32)
To be comparable with the analysis in the previous section, I assume that the social
planner can collectively set card fees ( ) for small-ticket transactions to internalize
33
In the welfare analysis, I abstract from the concern that social costs of alternative payment means may
deviate from private costs (e.g. the cash and check services are partially sponsored by the government,
so social costs of providing those services may diverge from private costs). Those are interesting but
separate issues, which are beyond the scope of this paper.
21
card demand externalities (I will show later that this outcome can indeed be implemented
by an alternative interchange cap regulation). Therefore, under the model’s distributional
assumptions of and , the social planner sets card fees to maximize social
welfare as follows,
= max
2
µ
¶
[( ( ) + ]2 − 2
[ − − ][ ( ) + − ] +
2
(33)
+ [ + − − ]
(9), (14), (18), (19).
The following proposition characterizes the solution to the welfare maximization problem (33). The results show that under Assumptions A1-A3, the social planner would also
set differentiated interchange fees to serve both the and sectors, but the fee level in
the sector tends to be lower than that set by the private network.
Proposition 3 The social planner who maximizes social welfare sets differentiated interchange fees to serve card transactions in both the and sectors. In addition, (i)
when issuer competition is high (i.e. 2), the h-sector interchange fee set by the social
planner is lower than that set by the private network; (ii) when issuer competition is low
(i.e. ≤ 2), the h-sector interchange fee set by the social planner coincides with that set
by the private network.
Proof. Consider that the card network is run by a social planner who maximizes social
welfare. In the case where issuer competition is high (i.e. 2), the constraint (18) does
not bind. The first order condition with regard to yields that
˜ = + −
(34)
Eqs (14) and (34) then determine the interchange fee in the sector
̃ = − +
22
( )
(35)
The social planner can also set
̃ = − and ˜ =
(36)
to serve the sector. Therefore, if the social planner sets a single interchange fee and only
serves the sector, the maximum welfare is determined by (33) as
=
[ (0)]2
4
(37)
In contrast, if the social planner sets differentiated interchange fees and serves both the
and sectors, the maximum welfare is
+ =
[ (1)]2
+
4
(38)
Under Assumption A3, + so the social planner prefers the latter.34 Comparing
Eqs (35) and (26) for = 1 shows that the -sector interchange fee set by the social
planner is lower than that set by the private network.
In the case where issuer competition is low (i.e. ≤ 2), the constraint (18) is binding.
Hence,
̃ = − +
( )
+2
(39)
Eqs (39) and (14) then determine the consumer fee in the sector
2( + − ) − ( − 2)[ ( ) + ]
˜ =
+2
(40)
Again, the social planner can also set card fees
̃ = − and
˜ =
(41)
to serve the sector. Therefore, if the social planner sets a single interchange fee and only
34
For issuers to participate in the card network, they need to make a non-negative profit. This can be
satisfied under plausible parameter values, i.e. [ (1)]2 2(− ), or the social planner is allowed to
conduct lump-sum transfers.
23
serves the sector, the maximum welfare is determined by (33) as
=
2[ (0)]2
( + 2)2
(42)
In contrast, if the social planner sets differentiated interchange fees and serves both the
and sectors, the maximum welfare is
+ =
2[ (1)]2
+
( + 2)2
(43)
Under Assumption A3, + so the social planner prefers the latter as well. Note
that in this case, the social planner’s decision is indeed equivalent to the private network’s
decision analyzed in Proposition 1. Therefore, the welfare-maximizing interchange fees
coincide with those set by the private network.
The welfare findings can be intuitively explained as follows. There are two counteracting distortions in the card payment system that we consider (particularly in the
sector where consumer transaction benefit of card usage follows a non-degenerate distribution).35 On the one hand, price coherence allows consumers to pay the same retail price
regardless of the payment method they use. As a result, merchants internalize consumers’
inframarginal card usage benefits when they decide whether to accept cards. This raises
the interchange fee that merchants are willing to accept.36 On the other hand, issuers
impose a markup when setting consumer fees, which drives down the inframarginal card
usage benefits and lowers the interchange fee that merchants are willing to accept. In
the case where the issuers’ market power is small (i.e. 2), the distortion due to
price coherence dominates, so the privately determined interchange fee in the sector
exceeds the socially optimal level. In the case where the issuers’ market power is large
(i.e. ≤ 2), the distortion due to issuer markup dominates. However, because the so35
Because of the assumption that consumers’ transaction benefit of card usage in the sector follows
a degenerate distribution, we abstract from distortion in the sector per se (which is supposed to be
sufficiently small anyway). However, the regulator and the private network still have different objectives
for internalizing cross-sector card demand externalities.
36
As mentioned before, the analysis in this paper can be carried over to the framework of Rochet and
Tirole (2002), where heterogenous consumers differ systematically in their transaction benefits of using
cards. In that framework, price coherence implies that cash-paying consumers are subsidizing those who
use cards.
24
cially optimal interchange fee is limited by the merchant card acceptance constraint, the
privately determined interchange fee coincides with the social optimum.37
In spite of the result that the privately determined interchange fee in the sector may
exceed the socially optimal level, we find that under the same set of assumptions, the
social planner behaves similar to the private network by setting differentiated interchange
fees to serve card transactions in both the and sectors. Essentially, both the social
planner and the private network treat the transactions in the sector as a loss leader. In
doing so, they subsidize the -sector card transactions in order to internalize the positive
externalities of card usage between the and sectors.
A similar analysis can be done if we assume that the social planner maximizes total
user surplus instead of social welfare. Total user surplus is the sum of consumer surplus
and merchants’ profit (but not issuers’ profit).38 Focusing on total user surplus is legitimate
when card issuers’ profit is dismissed by competition authorities. In this case, the results
turn out to be even stronger. Under plausible parameter values, I again find that the
social planner who maximizes total user surplus would set differentiated interchange fees
to serve card transactions in both the and sectors. Moreover, the resulting interchange
fees in both the and sectors are lower than those maximizing total issuers’ profit or
social welfare. The proof of the results can be found in the Appendix.
5.2
Alternative regulation
The analysis above shows that privately determined interchange fees tend to be too high
(based on the criterion of social welfare maximization or total user surplus maximization),
a finding consistent with previous studies. This implies that payment cards could be
overused at equilibrium. Therefore, lowering interchange fees may potentially improve
payments efficiency, which provides some justification for regulating interchange fees.
However, Eq (35) points out that the socially optimal -sector interchange fee is determined by multiple factors including merchant-and-consumer net benefits ( ), merchant
37
In reality, a card network typically has a large number of issuers. Therefore, it is likely that the
privately determined interchange fee exceeds the socially optimal level.
38
Maximizing total user surplus is the criterion Rochet and Tirole (2011) used to derive the optimal
interchange fee regulation based on the “merchant avoided-cost test.”
25
transaction benefit , issuer competition , and the acquirers’ cost . The finding suggests that the Durbin regulation that requests interchange fees to be capped by issuers’
marginal cost lacks theoretical foundation.
Our analysis also suggests that in the presence of card demand externalities, capping
the maximum interchange fee may not restore the social optimum because it could adversely affect small-ticket transactions, as shown by Proposition 2. On the other hand,
policymakers may not want to directly regulate card fees in all merchant sectors (including interchange fees and consumer fees) because that would risk being too heavy-handed.
Therefore, an interesting question is whether we could design an alternative interchange
cap regulation restoring the social optimum. The following discussion illustrate how this
can be done conceptually.
Proposition 4 In the presence of card demand externalities across merchant sectors,
capping the weighted average interchange fee, instead of the maximum interchange fee,
may restore the social optimum.
Proof. Consider a regulator who maximizes social welfare (a similar analysis can be
done using the criterion of total user surplus). When 2 the privately determined
interchange fee ( = 1) given by (26) exceeds the welfare-maximizing level ̃ ( = 1)
given by (35). Assume that Assumptions A1-A3 and Condition A4 hold, so both the
social planner and the private network would want to serve card transactions in the
and sectors. Following the analysis in Section 2, I focus on the scenario where ̃ ( =
1) ( = 0).39 In this case, as suggested by Proposition 2, capping the maximum
interchange fee at ̃ ( = 1) would not restore the social optimum because the card
network and issuers would stop subsidizing the sector.
Instead, the regulator could consider a weighted average cap (WAC) on interchange
fee as follows:
̃ + max(1 − 1 − ) ≤ ̄
(WAC)
where ̄ is the cap, 0 ≤ ≤ 1 is the weight chosen by the regulator, and ̃ = − is
the socially optimal -sector interchange fee. The WAC rule essentially requires ≤ ̄
39
Eqs (21) and (35) imply that ̃ ( = 1) ( = 0) whenever (1)
26
2
+2 (0)
when = = 0 but allows ̃ + (1 − ) ≤ ̄ when = 1, so it imposes a penalty on
the maximum permissible level of in case card networks and issuers raise fees to shut
off small-ticket transactions.
Recall that the welfare-maximizing interchange fees are given by (35) and (36) that
(1)
+ −
= −
̃ =
(44)
̃
(45)
Note that ̃ ̃ because and (1) 0. The corresponding total issuers’ profit
is determined by (16), (44) and (45) as
Π+ (̃ ̃ ) =
[ (1) + + ̃ − ]2
+
2( + 1)2
(46)
Denote Π (̃ ) as the total issuer’s profit for only serving the sector at the interchange
fee level ̃ = − , so
Π (̃ ) =
[ (0) + + ̃ − ]2
2( + 1)2
Under plausible parameter values, we have that40
Π+ (̃ ̃ ) Π (̃ ) 0
(47)
Given that ̃ ̃ , (47) implies that there exists a non-empty set of values that ̄ can
take, which satisfy ̃ ̄ ̃ and
Π+ (̃ ̃ ) Π (̄) Π (̃ ) 0
(48)
The regulator can then choose any cap value ̄ from this set and determine the corresponding weight by solving
̃ + (1 − )̃ = ̄
40
Note that the first inequality in (47) holds if
the second inequality holds if (0)
−
h
+1 (1)
27
i2
(49)
£
¤2
− (0) − ( − )
(− )
2(+1)2
and
Given ̄ and determined by (48)-(49), if the card network raises the -sector interchange fee above ̃ (or if issuers raise the -sector consumer fee so that ) and
hence only serves the sector, the highest total issuers’ profit that can be achieved is
Π (̄) under the WAC rule. In contrast, if the card network serves the -sector with the
interchange fee ̃ , it is allowed to set the -sector interchange fee as high as ̃ , so the
maximal profit is Π+ (̃ ̃ ). As (48) shows, the latter is more profitable. On the other
hand, the card network has no incentive to set the -sector interchange fee below ̃ under
the WAC rule because this will result in further losses in the sector but not help increase
the -sector interchange fee beyond the level of ̃ .
Proposition 4 shows that a regulation capping the weighted average interchange fee
may restore the social optimum. The analysis has intuitive implications for implementation. In principle, policymakers could design the interchange cap in such a way that
penalizes card networks and issuers in case they stop subsidizing small-ticket transactions. Under the WAC rule that we consider, card networks and issuers would have to
face a lower interchange cap for large-ticket transactions if they make fee adjustments
that adversely affect card usage for small-ticket transactions. As long as the penalty is
sufficiently large (cf conditions (48) and (49)), this will provide incentives for card networks and issuers to keep the small-ticket discounts and therefore internalize card demand
externalities across merchant sectors.
6
Conclusion
The recent U.S. debit card regulation introduced by the Durbin Amendment to the DoddFrank Act has generated some unintended consequences. While the regulation was intended to lower merchant card acceptance costs by capping the maximum interchange
fee, small-ticket merchants find their fees instead higher.
In this paper, I address this puzzle by introducing card demand externalities into
a two-sided market model. The findings rationalize the card networks’ response to the
cap regulation: Before the regulation, card networks were willing to offer discounted interchange fees to small-ticket merchants because their card acceptance boosts consumers’
28
card usage for large-ticket purchases from which card issuers can collect higher interchange
fees. After the regulation, however, card issuers profit less from this kind of externality,
so they discontinued the discounts. Based on the model, I then study socially optimal
interchange fees and alternative interchange regulation. The analysis suggests that the
social optimum generally would require lower interchange fees than those chosen by the
private market, but nevertheless it may maintain the differentiated fee structure in order
to internalize the positive externalities of card usage across merchant sectors. In this
case, simply capping the maximum interchange fee would not restore the social optimum
because of the side effect on small-ticket transactions. As an alternative, I propose a cap
regulation based on the weighted average interchange fee.
Overall, the takeaway of the paper is that interchange fees encompass more than
just the costs of processing payment card transactions. In the two-sided market, they
also serve to balance demand between consumers and merchants, as well coordinate acceptance and usage among different merchant sectors. The model shows that privately
determined interchange fees tend to be too high to maximize social welfare, so regulating
down interchange fees may help to improve the market outcome. But regulation that only
considers one-sided market logic (setting fees equal to issuers’ costs, for example) or one
sector of the market (ignoring card demand externalities between large- and small-ticket
merchants, for example) may result in unintended consequences.
There are several avenues for future research. First, it would be useful to quantify the
card demand externalities across merchant sectors. This can be an important step for
assessing the empirical impact of current as well as alternative interchange regulations.
Second, it would be useful to consider policy options other than price regulation. For
instance, in theory, if merchants can set different retail prices conditioning on payment
means, interchange fees may become less of an issue.41 Finally and more broadly, our
analysis can be extended beyond payment cards or two-sided markets. Policymakers may
always want to be alert to cross-sector externalities when designing regulatory policies so
that unintended consequences can be reduced or avoided.
41
However, those policy options may also have their own limitations, so some cautions need to be taken.
For example, in countries where card surcharging is allowed, few merchants choose to do so. Moreover, for
some merchants who are indeed surcharging, they are found surcharging excessively or in nontransparent
ways. See Hayashi (2012).
29
Appendix: Total user surplus maximization
Total user surplus is generated whenever consumers use cards for payment at retailers
provided consumer-and-merchant joint transaction benefits exceed the joint fees that they
pay, namely + + . In other words, total user surplus is the sum of consumer
surplus and merchants’ profit (but not issuers’ profit). The expression of total user surplus
can be derived from (17) and (33), as shown below. Again, I assume that the social
planner can collectively set card fees ( , ) for small-ticket transactions to internalize
card demand externalities. Therefore, the social planner sets card fees to
maximize total user surplus as follows:
= max
2
µ
¶
[( ( ) + ]2 − 2
[ − − ][ ( ) + − ] +
2
−
2
µ
[ ( ) + + − ]2
( + 1)2
¶
+ ( − − )
[ ( ) + + − ]2
+ ( + − ) ≥ 0
2( + 1)2
and
(50)
(51)
(9), (14), (18), (19).
The constraint (51) requires that card issuers make a non-negative profit.
The following proposition characterizes the solution to the problem. The results show
that under plausible parameter values, the social planner who maximizes total user surplus
would also set differentiated interchange fees to serve both the and sectors, but the
interchange fees in both sectors are lower than those maximizing total issuers’ profit or
social welfare.
Proposition 5 The social planner who maximizes the total user surplus sets differentiated interchange fees to serve card transactions in both the and sectors, and the
interchange fees in both sectors are lower than those maximizing total issuers’ profit or
social welfare if
[ (1)]2
2( + 2)2 (− )
30
(A5)
Proof. Consider that the card network is run by a social planner who maximizes total
user surplus. The constraint (18) never binds and the first order condition with regard to
yields
˜ = + − +
2
( )
+2
(52)
Eqs (14) and (52) then determine the interchange fee
̃ = − −
( )
+2
(53)
Therefore, if the social planner only provides card services to the sector but not the
sector, it is optimal to set
̃ ( = 0) = − −
(0)
+2
(54)
The total user surplus is
=
[ (0)]2
4( + 2)
(55)
and issuers make the total profit
Π =
[ (0)]2
2
2( + 2)
(56)
Alternatively, if the social planner also provides card services to the sector, it is
optimal to set
̃ ( = 1) = − −
(1)
+2
(57)
and set the lowest interchange fee ̃ that satisfies the constraint (51):
̃ =
− −
µ
¶
2
[ (1)] +
2( + 2)2
(58)
The total user surplus is
+ =
2 + 4
[ (1)]2 +
4( + 2)2
31
(59)
and the total issuers’ profit is
Π+ = 0
(60)
Under Condition A5, + , so the social planner achieves higher total user
surplus by setting differentiated interchange fees to serve card transactions in both the
and sectors. Moreover, under Condition A5, Eqs (57) and (58) confirm that the
interchange fees in both the and sectors are lower than those maximizing total issuers’
profit (given by (26) and (27)) or those maximizing social welfare (given by (35) and (36)
when 2 or (39) and (41) when ≤ 2).
References
[1] Armstrong, Mark (2006), “Competition in Two-Sided Markets,” RAND Journal of
Economics, 37(3): 668—681.
[2] Baxter, William (1983), “Bank Interchange of Transactional Paper: Legal Perspectives,” Journal of Law and Economics, 26: 541-588.
[3] Bedre-Defolie, Özlem, and Emilio Calvano (2013), “Pricing Payment Cards,” American Economic Journal: Microeconomics, 5(3): 206-31.
[4] Bolt, Wilko and Sujit Chakravorti (2008), “Consumer Choice and Merchant Acceptance of Payment Media,” DNB Working Papers 197, Netherlands Central Bank.
[5] Cabral, Luis (2005), “Market Power and Efficiency in Card Payment Systems: A
Comment,” Review of Network Economics, 5: 15-25.
[6] Carlton, Dennis and Alan Frankel (1995), “The Antitrust Economics of Credit Card
Networks,” Antitrust Law Journal, 63: 903-915.
[7] Foster, Kevin, Erik Meijer, Scott Schuh and Michael Zabek (2010), “The 2008 Survey of Consumer Payment Choice,” Public Policy Discussion Paper, No. 09-10,
Federal Reserve Bank of Boston.
[8] Gans, Joshua and Stephen King (2003), “The Neutrality of Interchange Fees in Payment Systems,” Topics in Economic Analysis & Policy, 3 (1): Article 1.
32
[9] Hayashi, Fumiko (2012), “Discounts and Surcharges: Implications for Consumer Payment Choice,” Payments System Research Briefings, June, Federal Reserve Bank
of Kansas City.
[10] Hayashi, Fumiko (2013), “The New Debit Card Regulations: Effects on Merchants,
Consumers, and Payments System Efficiency,” Federal Reserve Bank of Kansas
City Economic Review, Q1: 89-118.
[11] Katz, Michael (2001), “Reform of Credit Card Schemes in Australia II: Commissioned
Report” RBA Public Document. August.
[12] McAndrews, James and Zhu Wang (2012), “The Economics of Two-Sided Payment
Card Markets: Pricing, Adoption and Usage.” Research Working Paper 12-08,
Federal Reserve Bank of Richmond.
[13] Prager, Robin, Mark Manuszak, Elizabeth Kizer, and Ron Borzekowski (2009), “Interchange Fees and Payment Card Networks: Economics, Industry Developments,
and Policy Issues,” Finance and Economics Discussion Series 2009-23. Washington: Board of Governors of the Federal Reserve System.
[14] Rochet, Jean-Charles and Jean Tirole (2002), “Cooperation among Competitors:
Some Economics of Payment Card Associations,” RAND Journal of Economics,
33: 549-570.
[15] Rochet, Jean-Charles and Jean Tirole (2006), “Two-Sided Markets: A Progress Report,” RAND Journal of Economics, 35: 645-667.
[16] Rochet, Jean-Charles and Jean Tirole (2011), “Must-Take Cards: Merchant Discounts and Avoided Costs,” Journal of the European Economic Association, 9
(3): 462—495.
[17] Rochet, Jean-Charles and Julian Wright (2010), “Credit Card Interchange Fees,”
Journal of Banking and Finance, 34: 1788-1797.
[18] Rysman, Marc (2007), “An Empirical Analysis of Payment Card Usage,” Journal of
Industrial Economics, 60: 1-36.
[19] Rysman, Marc (2009), “The Economics of Two-Sided Markets,” Journal of Economic
33
Perspectives, 23: 125-144.
[20] Schmalensee, Richard (2002), “Payment Systems and Interchange Fees,” Journal of
Industrial Economics, 50: 103-122.
[21] Shy, Oz (2012), “Who Gains and Who Loses from the 2011 Debit Card Interchange
Fee Reform?” Public Policy Discussion Paper, No. 12-6, Federal Reserve Bank of
Boston.
[22] Schwartz, Marius and Daniel Vincent (2006), “The No Surcharge Rule and Card User
Rebates: Vertical Control by a Payment Network,” Review of Network Economics,
5: 72-102.
[23] Shy, Oz and Zhu Wang (2011), “Why Do Payment Card Networks Charge Proportional Fees?” American Economic Review, 101 (4): 1575-1590.
[24] Wang, Zhu (2010), “Market Structure and Payment Card Pricing: What Drives the
Interchange?” International Journal of Industrial Organization, 28: 86-98.
[25] Wang, Zhu (2012), “Debit Card Interchange Fee Regulation: Some Assessments and
Considerations,” Federal Reserve Bank of Richmond Economic Quarterly, 98 (3):
159-183.
[26] Weyl, Glen (2010), “A Price Theory of Multi-Sided Platforms,” American Economic
Review, 100 (4): 1642-1672.
[27] Wright, Julian (2003), “Optimal Card Payment Systems,” European Economic Review, 47: 587-612.
[28] Wright, Julian (2004), “Determinants of Optimal Interchange Fees in Payment Systems,” Journal of Industrial Economics, 52: 1-26.
[29] Wright, Julian (2010), “Why Do Merchants Accept Payment Cards?” Review of
Network Economics, 9, Article 1.
[30] Wright, Julian (2012), “Why Payment Card Fees Are Biased Against Retailers,”
RAND Journal of Economics, 43: 761-80.
34
@FedFRASER