Oxera Agenda November 2014
Agenda
Advancing economics in business
Another recent case concerns the pricing of e-books
available through online platforms, such as Amazon.
The US Department of Justice investigated six major
publishers of e-books as well as Apple (which is active in
e-book retailing), and found that they used MFNs as a tool
to implement a collusive agreement to increase prices to
consumers.
3
MFN clauses have also been scrutinised in the recent
investigation into the UK motor insurance market, in which
the CMA found that certain types of MFN between motor
insurance providers and price-comparison websites (PCWs)
restricted competition between PCWs.
4
How do MFNs operate under different
business models?
To understand the authorities’ concerns, it is useful to first
set out the two main business models under which MFNs
typically operate.
In the merchant model, the manufacturer offers the product
to a retailer at wholesale terms and the retailer sets the retail
price to consumers.
5
Figure 1 overleaf illustrates this model
with an example involving one hotel and two OTAs. Suppose
hotel H offers one room to OTA1 for £200/night (denoted by
W
1
in the figure). OTA1 can resell it at any price (e.g. £220).
A wholesale MFN agreement between hotel H and OTA1
in this case would typically stipulate that the hotel cannot
provide the same room to OTA2 at a ‘wholesale’ cost of less
than £200 (denoted by W
2
). However, OTA2 is free to set
the final price to consumers at any level. Such wholesale
MFNs, in isolation, therefore maintain the freedom of retailers
to set prices and hence are often not considered to be
problematic.
6
In the agency model, a ‘retailer’ is considered to be an ‘agent’
of the manufacturer and has no involvement in price-setting
MFN clauses, also known as most-favoured-customer
(MFC) clauses, are contractual terms agreed between firms
at different levels of the value chain. They usually stipulate
that a seller will offer its good or service to the counterparty
on terms that are as good as the best terms offered to third
parties.
The concept has its roots in trade agreements between
countries, ports or city-states going back as far as the
Middle Ages. According to these clauses, trade privileges
offered to a third party had to be extended to the counterparty
of an agreement so that the counterparty remained a
‘most-favoured nation’.
1
In the commercial world, MFN
clauses are quite common and are used in a wide range
of industries from books, movies and music, to hotels,
insurance and energy supply. They have come under the
spotlight following a number of investigations by competition
authorities.
For example, the online hotel bookings market is being
investigated by at least ten national competition authorities
(NCAs), including those of Austria, France, Germany,
Hungary, Italy, Switzerland and the UK. While not all of
these investigations are identical in scope, MFNs between
hotels and online travel agencies (OTAs)—through which
hotels guarantee certain OTAs the best room prices among
those offered to all OTAs—have been or are being explicitly
considered by most NCAs. In some cases, the clauses
also restrict the hotels themselves from offering better
deals through their own websites or to last-minute walk-in
customers. The main concern of the authorities is that these
clauses can dampen competition among OTAs and increase
the prices that consumers pay. The investigation by the
German NCA has culminated in a ban on MFNs by the hotel
booking service provider, HRS, while the investigation in the
UK is being reviewed following a successful appeal against
the earlier decision of the Office of Fair Trading (OFT, now
the Competition and Markets Authority, CMA) to settle with
the parties.
2
Most-favoured-nation clauses:
falling out of favour?
Competition authorities across the globe are investigating most-favoured-nation (MFN) clauses
in distribution contracts in industries as diverse as hotel bookings, books and insurance. Such
clauses guarantee to a distributor that no other distributor will receive a better deal. Although
there is concern that they may restrict competition and harm consumers, MFNs can also deliver
benefits. How should competition authorities strike a balance?
1
Oxera Agenda November 2014
decisions.
7
So, in the hotel example, both OTA1 and OTA2
would advertise rooms of hotel H at their respective prices,
P
1
and P
2
, which are set by hotel H (see Figure 2). For
every booking it makes, the OTA receives from hotel H a
commission (denoted by C
1
and C
2
in the figure), which
may or may not be dependent on the price of the room.
8
In this case, a retail MFN clause between OTA1 and hotel H
may stipulate that hotel H cannot advertise the same (or a
similar) room for a lower price through any alternative OTA
such as OTA2, or any other sales channel (e.g. hotel H’s
own website, where the price advertised is P
H
). Such an MFN
clause would be considered a ‘broad’ retail MFN—i.e. one
that specifies conditions on the retail price advertised on
all alternative channels. A ‘narrow’ retail MFN, on the other
hand, might be one that stipulates only that hotel H will not
advertise a lower price through its own website (i.e. P
H
≥ P
1
),
and leaves the option for hotel H to offer consumers a lower
price (P
2
) through OTA2.
The parties could also have a ‘wholesale MFN’ that
stipulates that the commission paid to OTA1—i.e. C
1
—
will not be lower than that paid to any other OTA (e.g. C
2
).
Synchronised shifting:
the authorities’ concerns
As described above, concerns have been raised across
jurisdictions and sectors. In the most recent cases, potential
restriction of retail competition through the use of retail MFN
clauses under agency agreements has been the primary
concern.
9
While the degree of concern is likely to differ depending
on how broad or narrow the retail MFN is, for illustrative
purposes suppose that in Figure 2 there is a retail MFN in
place which covers the prices advertised through both OTAs
(although similar concerns are relevant when the MFN also
restricts the pricing through the hotel’s own website). In the
absence of any MFN, hotel H would be able to set prices P
1
and P
2
freely. In particular, it has an incentive to encourage
sales—by lowering the retail price—at the OTA that charges
2
Most-favoured-nation clauses
it a lower commission level, thereby providing hotel H with
a higher margin while it competes more aggressively with
other hotels.
The retail MFN clause, for example with OTA1, therefore
prevents the seller from ‘rewarding’ other low-commission
distributors. This issue may raise the following concerns.
• If a specific agent has in place a sufficiently broad retail
MFN (i.e. one covering a large proportion of the sales
channels), the agent, when thinking about negotiating an
increase in its commission, will not be concerned about
its competitiveness relative to other agents (because the
manufacturer will be bound by the MFN not to reflect the
higher commission in the price advertised through this
agent). Equally, agents that wish to increase their market
share by lowering their commission will not be able to
advertise lower prices, as these will be constrained by
the MFN. Thus, there is likely to be a lower incentive for
agents to reduce commission rates, although elements
of non-price retail competition may remain.
This, in turn, raises the likelihood of consumer harm,
assuming that higher commissions paid by a supplier
ultimately feed through into higher retail prices.
• MFNs may also undermine market entry. For example,
a new agent that wishes to advertise lower prices to
consumers will not be able to do so in the presence
of MFN clauses between existing agents and sellers,
thereby preventing low-cost/low-price entry.
• The incentives for agents to reduce costs may be
undermined. In a well-functioning market, agents would
have an incentive to cut their costs and pass some of the
savings on in the form of lower commissions, in order to
incentivise the seller to lower the advertised price. With
MFNs, this incentive is weakened.
Other concerns might also arise, depending on the case.
For example, MFNs can be used as a tool to support
collusive actions (as in the Apple e-books case, where
Figure 2 Stylised example of MFNs in an
agency model
Source: Oxera.
Figure 1 Stylised example of MFNs in a
merchant model
Source: Oxera.
Oxera Agenda November 2014
3
Most-favoured-nation clauses
Are there any benefits of MFNs?
In short, yes. Indeed, although MFN clauses have been
increasingly scrutinised in recent years, they are not
regarded as anticompetitive per se, as they can deliver
benefits. The scope of these benefits is well accepted by
academics and competition authorities, although detailed
assessment has so far been limited.
First, MFN clauses can assist in preventing hold-up of
investments by retailers. This is because such clauses,
by preventing the supplier from offering better terms to other
retailers and/or undercutting via the supplier’s own sales
channels, can prevent ‘free-riding’ on a specific retailer’s
investment.
For example, consider an online platform such as eBay,
which invests in offering a high quality of service (e.g. in
terms of customer reviews and purchase advice), and also
invests in advertising. Despite this, if consumers find that
the product is less expensive elsewhere, they may use eBay
to obtain necessary information, but purchase through the
chosen seller’s own website. Indeed, without an MFN in
place, the price on the seller’s website is likely to be lower
in order to attract the customer (as the seller avoids paying
a commission to eBay), and eBay’s investments in service
quality would therefore not be protected.
12
This, in turn, could
lead to sub-optimal investments by retailers, and potentially
to market exit.
13
MFN clauses can also benefit consumers directly. For
example, they allow retailers to advertise ‘lowest-price
offers’, which offer consumers the assurance that they will
find the best available price on a single platform. This in turn
can significantly reduce search costs and may even lead
to a higher switching rate (for example, a consumer can
switch mobile contracts more easily than they otherwise
could).
14
Furthermore, although broad MFNs may limit price
variation in the market and lessen price competition, they
might enhance competition on other aspects of the offering.
For example, retailers might differentiate themselves by
investing in pre-sales advice or after-sales support, which
consumers may welcome.
Concluding remarks
The growth of e-commerce and the use of agency
agreements by online platforms have made MFNs a hot
topic in competition law. While some academics and
practitioners in Europe have questioned whether certain
types of retail price MFNs should be treated as equivalent
to RPM and as infringements per se, others support a
case-by-case assessment of their potential harm and
efficiency benefits. While this debate is unlikely to be
resolved in the near future, the decisions of the numerous
NCAs in the hotel investigations, as well as potential
investigations in the online book retailing market,
15
should shed some light on the future of MFNs.
MFNs were not considered problematic as such); a pervasive
network of broad MFNs across the market can also reduce
inter-brand price competition between sellers; and the lack
of price competition can incentivise excessive advertising
by retailers (which may have caused the deluge of TV
advertisements in the UK from PCWs involving robots, opera
singers and ‘meerkats’). Box 1 below provides further details
on how the CMA analysed MFNs in the motor insurance
market.
The extent to which an MFN clause is likely to give rise to
harm will depend on a combination of factors including the
type and scope of the clause (retail versus wholesale, broad
versus narrow); the extent of use of similar clauses by other
suppliers and agents; the payment arrangements in place
(e.g. the structure of the commission rates); and, importantly,
the market structure and positions of respective firms.
For example, use of MFNs by smaller retailers is likely to be
less of a concern, as these clauses would help such firms
to become, or remain, competitive. Recent findings in the
academic literature also show that MFNs can assist market
entry under certain circumstances (e.g. when the payment
to the retailer is based on profit-sharing rather than
revenue-sharing arrangements; and when the potential
entrant has a less-differentiated, higher-cost business model
than the incumbent).
10
The extent of retail competition on
parameters not covered by MFNs (e.g. customer service
and other add-on services) is also a relevant factor in the
assessment.
11
MFNs in the UK motor insurance market
The CMA analysed MFNs between insurance providers and
PCWs. Two types of MFN occur in this market: ‘wide’ MFNs,
where the insurance provider has to offer prices to a PCW
that are not worse than prices offered through any other
PCWs or through the provider’s own website; and ‘narrow’
MFNs, which stipulate only that the price advertised through
a specific PCW cannot be higher than that offered through
the provider’s website. The CMA found that the wide MFNs
restrict competition between PCWs, due to the reasons
discussed above.
The CMA found that the narrow MFNs did not raise
substantial concerns, mainly because they do not restrict
prices advertised through other PCWs. However, narrow
MFNs may be problematic in the motor insurance industry.
This is because if the PCW with the MFN increases its
commissions, and the insurer reflects this in the prices
for that PCW only, it also has to increase the prices on its
own website. This is likely to lower sales through the direct
channel relative to third-party PCWs, where prices are lower.
If the direct channel is more profitable, the insurer may prefer
to pass on the increased commission to prices on all PCWs
so as not to disadvantage the direct channel. This, in turn,
means that the PCW with the MFN remains competitive, and
would therefore have the incentive to increase commissions
in the first place. The argument discussed above about
softening of competition may therefore also hold in this case.
Oxera Agenda November 2014 4
Most-favoured-nation clauses
1
For a discussion of MFNs in the context of international trade agreements, see, for example, United Nations (1980), ‘Yearbook of the International
Law Commission 1978’.
2
For the German decision, see Bundeskartellamt (2013), ‘Decision in the administrative proceedings HRS-Hotel Reservation Service, Hotelverband
Deutschland (IHA) e.V., JBM JustBook Mobile GmbH, Unister GmbH, for examination of a violation of section 1 of the Act Against Restraints of
Competition (
Gesetz gegen Wettbewerbsbeschränkungen
(
GWB
))/Art. 101 (1) TFEU and section 20 (1) in conjunction with section 19 (1) and (2)
No. 1 GWB’, 20 December. For the OFT’s decision and subsequent developments, see Competition and Markets Authority (2014), ‘CMA statement
on CAT judgment in hotel online booking case’, 26 September. Oxera advised Skyscanner, the price-comparison website that lodged the appeal.
3
For the US order, see United States District Court Southern District of New York (2013),
United States of America v Apple Inc. et al., 12 Civ. 2826
(DLC); The State of Texas et al. v Penguin Group (USA) Inc. et al., 12 Civ. 3394 (DLC)
, 10 July. The European Commission also investigated the
parties but decided to settle. See European Commission (2013), ‘Summary of Commission Decision of 25 July 2013 relating to a proceeding under
Article 101 of the Treaty on the Functioning of the European Union and Article 53 of the EEA Agreement (Case COMP/39.847/E-BOOKS) (notified
under document C(2013) 4750), 2013/C 378/14’. In both jurisdictions, these clauses were prohibited from the parties’ contracts for a number of years.
4
Competition and Markets Authority (2014), ‘Private Motor Insurance market investigation’. Oxera advised a major insurance company in this
investigation.
5
The retailer may or may not purchase the product before reselling; if it does, this will be equivalent to a wholesaler/retailer setting. In any event, the
retailer assumes some financial risk (for example, due to low margins if there is low demand or strong competition downstream).
6
However, potential concerns could be identified on a case-by-case basis. For example, MFNs may prevent retailers exercising buyer power to
negotiate lower deals with suppliers that would, to a greater or lesser extent, have been passed on to final consumers.
7
In the context of this article, an agent is simply a distributor in the economic sense. A stricter legal assessment of what constitutes a genuine agency
agreement is required in order to determine whether an agreement can be exempt from Article 101 (1) TFEU under the Block Exemption Regulation.
For further details, see European Commission (2010), ‘Guidelines on Vertical Restraints’, paras 12–21.
8
The merchant model and the agency model can also co-exist in the same market (e.g. in the UK online hotel booking market).
9
The focus here is on MFNs that relate to the terms offered by a supplier across different retailers—i.e. terms that affect intra-brand competition
directly. Similar price-parity agreements could occur where a supplier imposes conditions on a specific retailer regarding the retail prices of its own
product vis-à-vis a competitor’s product.
10
Johnson, J.P. (2014), ‘The Agency Model and MFN Clauses’; Boik, A. and Corts, K.S. (2013), ‘The Effects of Platform MFNs on Competition and
Entry’.
11
The types of business model in the market are also relevant. For example, the academic literature suggests that, in the presence of mixed business
models, even retail MFNs for one retailer can raise significant concerns and eliminate retail competition (some suggest that this could lead to the
same effect as industry-wide resale price maintenance, which is considered an infringement per se under EU competition law). The US Apple e-book
case, although categorised as a price-fixing case, had this aspect. See Foros, Ø., Kind, H.J. and Shaffer, G. (2013), ‘Turning the Page on Business
Formats for Digital Platforms: Does Apple’s Agency Model Soften Competition?’, CESifo Working Paper Series.
12
The argument that MFN clauses can limit free-riding has been put forward in most of the recent cases. See, for example, Competition and Markets
Authority (2013), ‘Private Motor Insurance Market Investigation: Provisional findings report’, December.
13
See, for example, Baker, J.B. (1996), ‘Vertical restraints with horizontal consequences: Competitive Effects of “Most-Favored-Customer” clauses’,
Antitrust Law Journal
, 64:3, pp. 517–34, Spring.
14
In certain markets, MFNs may not be necessary to reduce search costs—for example, where PCWs or meta-search engines are able to work
effectively and are widely used by consumers.
15
Publishers in Germany and the UK recently called on competition authorities to open an investigation into the market for book retailing.
See Mance, H. (2014), ‘Publishers call for UK antitrust inquiry into Amazon’,
Financial Times
, 18 September.
© Oxera, 2014. All rights reserved. Except for the quotation of short passages for the purposes of criticism or review, no part may be used or
reproduced without permission.
