Apple's agency model and the role of most‐favored‐nation clauses

• DOI: http://doi.org/10.1111/1756-2171.12195
• Published date: 01 August 2017
• Author: Øystein Foros,Greg Shaffer,Hans Jarle Kind
• Date: 01 August 2017

RAND Journal of Economics

Vol.48, No. 3, Fall 2017

pp. 673–703

Apple’s agency model and the role

of most-favored-nation clauses

Øystein Foros∗

Hans Jarle Kind ∗∗

and

Greg Shaffer∗∗∗

The agency model used by Apple and other digital platforms delegates retail-pricing decisions

to upstream content providers subject to a fixed revenue-sharing rule. Given competition both

upstream and downstream, weconsider how, under the agency model, retail prices depend on the

firms’ revenue-sharingsplits and the degrees to which consumers view the platforms and the goods

sold on the platforms to be substitutes. We show that the agency model may not be universally

adopted even if adoption would mean higher profits for all firms. Use of most-favored-nation

clauses in these settings can ensure industry-wide adoption and increase retail prices.

1. Introduction

Digital platforms (such as Apple and Google) often adopt the agency model in their dealings

with upstream content providers (such as e-book publishers and app developers). This business

model has two key ingredients: (i) retail pricing is delegated to the upstream content providers

and (ii) surplus is allocated between the parties according to a fixed revenue-sharing rule. Apple,

for example, has a 70/30 split, whereby 70% of the revenue goes to the upstream firm and 30%

goes to Apple.1The split is the same for all firms in each of Apple’s markets.

∗NHH Norwegian School of Economics; oystein.foros@nhh.no.

∗∗NHH Norwegian School of Economics and CESifo; hans.kind@nhh.no.

∗∗∗University of Rochester; shaffer@simon.rochester.edu.

Wethank YaronYehezkeland seminar par ticipants at the Economics of Platform Workshop,ESMT Berlin, April 2016; the

9th Annual Nordic IO Workshop,the University of Oslo, June 2014; the Norwegian School of Economics and BECCLE,

May 2014; the University of Toronto, April 2014; the National University of Ireland, April 2014; the University of

California at Berkeley,March 2014; Oslo Economics, March 2014; the University of Tubingen, December 2013; the 11th

Annual Workshopon Media Economics, in Tel Aviv, Israel, October 2013; the 40th Annual EARIE Conference in Evora,

Portugal, August 2013; and the 4th Annual Workshopon the Economics of ICT in Evora, Portugal, April 2013, for many

helpful comments. A previous version of this article was circulated under the title “Turningthe Page on Business Formats

for Digital Platforms: Does Apple’sAgency Model Soften Competition?”

1Apple has used the same split since they launched the iTunesmusic store in 2003 (see Isaacson, 2011).

C2017, The RAND Corporation. 673

674 / THE RAND JOURNAL OF ECONOMICS

The practice of delegating retail pricing decisions to the upstream content providers has

recently come under scrutiny.Although both Apple and Google adopted the agency model without

raising antitrust concerns when the first smartphones were introduced in 2008, this changed when

Apple entered the e-books market in 2010. The US Department of Justice took Apple to court,

and the judge in the case ruled that Apple’s use of the agency model in conjunction with most-

favored-nation (MFN) clauses (which prohibited publishers from selling their e-books at higher

retail prices on Apple’s iBookstore than they sold for elsewhere) had resulted in higher prices

for consumers. The judge also ruled that the decision by Amazon (the other major retailer of

e-books) to adopt the agency model soon after Apple entered the e-books market was the result

of a conspiracy between Apple and five of the largest publishers in the United States to pressure

Amazon into doing so.2

Little is known in general about the competitive effects of the agency model, or when

platforms would be expected to adopt it. Of particular interest in light of the increased scrutiny is

whether the agency model is intrinsically anticompetitive in the sense that it would be expected

to lead to higher retail prices even if it is not adopted in conjunction with MFN clauses (would,

e.g., a court-ordered restriction that prohibits a platform from using MFN clauses but not from

adopting the agency model be enough to protect consumers from higher prices?). Also of interest

is whether and under what conditions industry-wide adoption of the agency model would be

expected to arise even in the absence of a conspiracy among industry participants.

One might think that because pricing in the agency model resembles resale price maintenance

(RPM), one can simply borrow from the insights in the RPM literature when determining the

agency model’s competitive effects.3However, the usual justifications for RPM do not hold. For

example, it is often alleged that RPM can reduce free-riding on presale services (Telser, 1960),

stimulate interbrand competition by providing quality certification (Marvel and McCafferty,

1984), foster demand-enhancing activities (Mathewson and Winter, 1984; Winter, 1993), and

ensure that downstream firms have sufficient margins to maintain adequate supplies of inventory

(Deneckere, Marvel, and Peck, 1996; Krishnan and Winter, 2007). However, these justifications

implicitly assume that the upstream firms have the power, and have no swaywhen the downstream

firms are the ones deciding whether to adopt the agency model.4Moreover, the purpose of the

agency model cannot be to mitigate the well-known problem of double marginalization because

there is no double marginalization in the usual sense when the firms engage in revenue sharing.

In this article, we allow for both upstream and downstreamcompetition, with each upstream

firm selling through multiple downstream firms. We consider the downstream firms’ incentives

to delegate pricing control to the upstream firms, and we ask three questions:

1. When will the agency model lead to higher prices?

2. When will the agency model be adopted?

3. What is the role of MFN clauses?

In answer to our first question, we find that delegating pricing control to the upstream firms

leads to higher retail prices if and only if the willingness of consumers to substitute between

platforms is higher than their willingness to substitute between goods. By giving control of the

retail prices to the upstream firms, the downstream platforms in effect trade one type of pecuniary

externality (due to competition between platforms) for another (due to competition between

2See United States v. Apple Inc., 12 Civ. 2826 (DLC). The appeals court declined to reverse the ruling.

3They are economically similar in the sense that in both the agency model and under RPM, the upstream firms

control the retail prices. However, a major difference is that in the setting that we consider here, the agreements that

delegate retail pricing to the upstream firms are separately/individuallyentered into by different downstream firms, which

is in contrast to the classic case where RPM is imposed at the level of the market for anygiven good.

4To prevent free-riding on a retailer’s presale services, for example, RPM must be imposed on the downstream

firms that want to free-ride; it cannot be left to the free-riders to decide whether to abide by it.

C

The RAND Corporation 2017.