Do slo tt in g a llowa n ce s ha r m re ta il co mpet itio n ?
1
Øystein Foros
Norwegian School of Econom ics and Busin ess Ad m inistration
oystein.foros@nhh.no
Hans Jarle Kind
Norwegian School of Econom ics and Busin ess Ad m inistration
hans.kind@nhh.no
JEL classification: L13, L42, L81
Keyw ords: Slotting allowances, retail com petition, an ti-trust policy
Abstract: Slotting allow a nces ar e fees paid by manufacturers to get access to
retailers’ shelf space. Both in the USA and Eu rope, the use of slotting allo wances
has attracted attention in the general press as well as amo ng policy makers and
economists. O ne sch ool of thought claim s that slotting allowances are efficiency
enhan cin g, wh ile another sc hool of thought maintains that slotting allowances are
used in an an ti-competitiv e m ann er. In this paper, w e argue that this controversy is
partially caused b y inad equate assum ption s of ho w the retail market is structured
and organized. Using a formal model, w e show that there are good reasons to
expect anti-competitive effects of slotting allowances. We further point out that
competition authorities tend to use an unsatisfactory basis for com parison when
analyzing welfare consequences of slotting allo wances.
1
We would like to thank Victor D. Norman, Kåre P. Hagen, Lars Sørgard and Einar Hope for
stimulating discussions and comments.
1 Introduction
Slotting allo w a nces are fixed fees that manufacturers pa y to retailers in order to get
access to their shelf space. While slotting allo wances w ere hardly kno w n before the
late 80’s, they are no w widely used, not least in the grocery industry. The boost of
slotting allowances has coincided with a trend tow ards higher retail concentration.
In Europe, in particu lar, the grocery reta iling sector has becom e striking ly m ore
concentrated ov er the last decades (Dobson and Wa terson, 1999, and Clark e et al.,
2002). T her eby retailers’ market pow er ov er m anufacturers has increased, and there
is a broad consensus that this is the major reason why the use of slotting allo wances
hasbecomemorewidespread.
2
How ever, economists (and policy makers) disagree
as to wh ether slotting allo wances tend to mitigate retail com petition. T he main
purpose of the presen t paper is to help resolve this controv ersy, and draw some
po licy implica tion s.
Tw o schools of thought dominate the debate o v er w elfare effects of slotting al-
lo wances. The so-ca lled ma rket pow er school argues that slotting allowances m ay
have anti-competitive rationales (Shaffer, 1991). To see why, suppose that a retailer
can c hoose bet ween contract A, with no slotting allo wances an d a low wholesale
price, and con tract B, where the retailer receives a slotting allowance from the man-
ufacturer but in return pa ys a higher wholesale price. Since slotting allowances are
up-front payments, margin al costs for the retailer are th us relatively high in con tract
B. By signing this contract, the retailer sends a signal to her rivals that she will be
a soft com petitor and set a relatively high end -user price. This in turn ind uces the
rivals to raise their prices too. Sha ffer sho w s that this mechanism ma y lead us to an
equilibrium where retailers use slotting allowances as a device to increase end-use r
2
According to a US-based survey by Bloom et al. (2000), retailers and manufactures agree that
greater retail power has contributed to more use of slotting allowances also in this country. This is
true even though retail concentration is much lower in the USA than in Europe. See also Rey, Thal
and Vergé (2005), who further note that in the UK even the products of leading manufacturers
typically ”represent a very small proportion of the total business for each of the major suppliers”.
In contrast, also the largest manufacturers are highly depend on their major buyers (Rey et al
2005, p. 3).
1
prices. This has a negative w elfare effect.
3
In contra st, the efficiency sc hool argues that slotting allo wances ha ve positiv e
welfare effects, for instance b y solving problem s connected to uncertain ty and/o r
asymm etric information, and by allocating scarce shelf space.
4
The efficiency sc hool
dismisses Shaffer’s hypothesis that slotting allowances are used as a tool to soften
retail competition, one of their main arguments being that retailers and manufac-
turers ty pically enter secret con tracts. Thereby wholesale prices are uno bservable,
and cannot be used strategically to increase end-user prices.
In our view, a m ain problem with both th e efficiency and the mar ket po wer
sc h ool is their assum ption on how the m arket is structured and organized. First,
both sch ools of though t presuppose that there are only two layers; manufacturers
at the upstream lev el and retailers at the do wnstrea m level. A second presumption
they ha ve in common, is that each retailer behav es lik e a vertically integrated firm
in its de cision on p r ocurement contracts and retail pricing. How ever, this is not a
proper description of the grocery market, especially not in Euro pe. Indeed, what we
ha ve observed is that large retail chains have formed procurement alliances (buy er
groups), suc h that the level of concentration is high er for procurement than for re-
tailing (see Dobson and Waterson, 1999).
5
In these constellations, the headquarter s
of eac h buy er group typically deals with procuremen t, while the retail sub-c hains
take care of retailing (e.g. end-user pricing). Even when sub-c h a ins are fully ow n ed
3
Slotting allowances may also reduce product variety through foreclosure of smaller suppliers.
Shaffer (2005) shows specific market structures where such practice may be optimal. Marx and
Shaffer (2004) demonstrate that retailers may also benefit from foreclosure of suppliers, since this
may shift profit from the manufacture-level to the retail lev el.
4
See further discussion in Section 3.
5
The largest food buyer in Germany is the buyer group Markant Handels. The buyer groups
Euromadi and IFA Espanñola are the two largest food buyers in Spain, and Intermarché dominates
in France (Dobson and Waterson, 1999). In Norway, the largest retailer group, NorgesGruppen
(NG), was formed as a buyer group in 1994. Even though there has been a process of closer
integration, NG may still be considered as a buyer group where the headquarters takes care of
procurement, and each store brand decides end-user prices autonomously. Several of the retail
formats within NG are also independently owned by the retailers themselves. An overview of the
Nordic markets is giv en by the Danish Competition Authorit y (2005).
2
b y the procurement headquarters, they are typically organized as divisionalized
firm s.
6
We show that in this context each buyer group will use slotting allowances
to dampen in tr a-retailer competition even if rival retail chains cannot observe the
wholesale contracts. As long as the procurement con tracts can be observed within
each buy er group, which is a plausible assumption, they can transfer their buying
power into the retail market b y using slotting allowa nces.
This paper contain s a relativ ely broad discussion of antitrust issues. First, we
sho w that our findings are supported b y severa l an titrust in vestigations of the gro-
cery industry in Europe (European Commission, 1996, Competition Commission,
2000, Danish Competition Authority, 2005, and The Norwegian Competition Au-
thority, 2005). Second, we emphasize the importance of recognizing the degree of
substitutabilit y that exists betw een different kinds of v ertical restrain ts. C on sider a
buy er group of independent retailers. They cooperate in the procuremen t market,
but compete at the retail level. Building on ou r forma l model, w e argue that b y
using a vertical restraint like a slotting allowance, the group can achiev e the same
outcome as they w ould with v ertical integration. In the latter case, decisions on pro-
curement and end-user pricing are tak en by the group’s headqua rters. Consequently,
it is pointless to outlaw slotting allowances if the competition authorities w ou ld not
ban a merger among alliance members. Vice versa, if a potential merger betwe en
the firms raises serious doubts by the competition authorities, slotting allowances
should raise the same concerns.
1.1 Related Literature
The present paper is an extension of Shaffer (1991), who considers competition
betw een two retailers in the end-user mark et. By assuming that the retailers hav e
6
Thus, while the headquarters decide procurement contracts cent rally, each sub-chain is rela-
tively autonomous with respect to end-user pricing. The leading Finnish retailer groups, Kesko and
Tuko , are orgainized in this a way (The European Commission,1996), and the same holds for ICA’s
retailing operations in Norway (NCA, 2005). Just one of the four dominating Norwegian retailer
groups operates a completely vertically integrated firm with respect to procurement contracts and
retail pricing (NCA, 2005).
3
complete bargaining po wer ov er manufacturers, Shaffer (1991) sho ws that it is in
the in t erest of each retailer to set a high wholesale price in the contra ct with the
manufacturer. W hen wholesale prices rise, retail competition softens. Thus the
total profit made by the vertical c h ain increases, and this profitiscapturedbythe
retailers through slottin g allowances.
Shaffer’s idea is b ased on the strategic delegation literature, wh ere Fersh tm an
and Judd (1987) is the seminal paper. Gal-Or (1991), Bonanno and V ic kers (1988)
and Rey and Stiglitz (1988, 1995) build on the same framework, but they assume
that the bargaining po wer is in the hands of the suppliers. Irmen (1998) shows that
the outcom e in this case resem bles the one found b y Sha ffer (1991). The differen ce is
the sign of the fixed fee.
7
Consequently, the strategic delegation theory is consistent
with the observation that the use of slotting allowances has increased as bargaining
power has been transformed from the manufactur ing lev el to the retailing lev el.
A critical assum p tion within the strategic delegation literature is that the con-
tract bet ween a manufacturer and a retailer is irreversible, so that wholesale prices
are determined prior to the price gam e betw een retailers. We agree with Rey and
Stiglitz (1995) that this is lik ely to hold. The reason is that retailers rarely hav e
long-term con tr acts with their customers, while the w holesale contra ctual arrange-
ments often are set for no less than a year. Moreov er, the type of wholesale contracts
(e.g. slotting allo wances or not) will t yp ica lly be specified in long - term co ntracts
(see e.g. discussion by Rey and Stiglitz, 1995).
The second critical assumption emp loy ed by Shaffer and other papers in the
same tradition, is that wholesale tariffs are observed by rival retailers. This assump-
tion is more dubiou s than that of irreversibilit y.
8
However, we show that slotting
7
If the retailers have the bargaining power, a fixed fee is paid by the suppliers, defined as a
slotting allowance. If the suppliers have the bargaining po wer, the fixed fee is paid by the retailers,
denoted as a franchisee fee.
8
The Norwegian Competition Authority (NCA) doubts that slotting allowances can be used
by retailers as a facilitating practice a la Shaffer (as claimed in the press), since they find the
assumption of contract observability unrealistic (NCA, 2005, and Gabrielsen and Sørgard, 2005).
However, NCA has initiated an investigation to clarify the extent of information exchange betw een
retailer groups.
4
