The Value of Switching Costs
The Value of Switching Costs
The Value of Switching Costs
Gary Biglaiser
University of North Carolina, Chapel Hill
Jacques Cremer
Toulouse School of Economics
(GREMAQ, CNRS and IDEI)
Gergely Dobos
Gazdas agi Versenyhivatal (GVH)
June 13, 2011
We would like to thank Andrew Clausen, Philipp Kircher, George Mailath, Curt Taylor,
Paul Klemperer, and the participants at seminars where this paper has been presented, especially
those of the Economics department at the University of Cambridge.
We study the consequences of heterogeneity of switching costs in a dynamic model
with free entry and an incumbent monopolist. We prove that even low switching cost
customers have value for the incumbent: when there are more of them its prots increase.
Indeed, their presence hinders entrants who nd it more costly to attract high switching
cost customers. This leads to dierent comparative statics from the standard static
model with free entry and homogeneous switching costs: for instance, an increase in the
switching costs of all consumers can lead to a decrease in the prots of the incumbent.
In many industries, the market power of incumbents is protected by the switching
costs that consumers have to incur when they purchase from an entrant. This paper
focuses on the consequences of the presence of low switching cost customers; we point out
that their presence hinders entrants, who nd it more costly to attract high switching
cost customers.
As we will discuss in Section 3, in the simplest static economic model of switching costs,
with one incumbent and free entry, heterogeneity of switching costs does not matter.
If a proportion > 0 of the agents have switching cost > 0, while the others have
no switching cost, the incumbent will charge and its prots will be equal to , the
average switching cost of all the consumers, multiplied by the number of agents. We
show that this result changes drastically in a dynamic model in which there are new
potential competitors in every period; the more skewed the distribution of switching
costs, the greater the prots of the incumbent. To the best of our knowledge, this fact
and the importance of the distribution of switching costs has not been recognized in
the literature, despite the existence of a signicant body of theory which explores the
consequences of consumer switching costs. (We discuss the literature below in Section 1.)
Our results have implications for managerial practice and for public policy. Entrants
should beware of not pricing aggressively while attracting footloose consumers who
will not stick with them when they increase prices. Antitrust authorities should take
into account the whole distribution of switching costs (including the switching costs of
consumers who decide to purchase from entrants) when determining whether incumbent
rms are behaving anticompetitively.
We conduct our analysis by constructing a series of models that share the following
features: a) the switching costs of consumers are invariant over time; b) at the start
of the game there is a single incumbent rm; and c) there is free entry by competing
rms in every period. Following much of the literature, we assume that only short term
contracts are used and that a consumers switching cost does not depend on the rm
from which it is purchasing.
In Section 2, we introduce our analysis by considering the case where all consumers
have the same switching cost . In a one period model, the incumbent would charge ,
and, assuming that the mass of consumers is equal to 1, its prot would also be equal
to . We embed this static model in a dynamic framework and show that in equilibrium
aggregate discounted prot over all periods is also equal to , whether the number of
periods is nite or, subject to stationarity assumptions, innite. In the latter case, this
implies that the prot of the incumbent is equal to the value of a ow of per period
payments equal to (1 ), not to ! Although this result is very easy to prove, and
is implicit in some of the literature, we feel that it is worth stressing as it shows that
switching costs are a leaner cash cow than sometimes assumed.
We begin our analysis of the heterogeneity of switching costs in Section 3. A proportion
(0, 1) of consumers have a switching cost equal to > 0, while the others have no
switching cost. We identify the (stationary) equilibrium of the innite horizon model. As
opposed to the case where all consumers have the same switching cost, the intertemporal
prot of the incumbent is greater than the one period prot, although it is smaller than
the value of an innite stream of one period prots. We prove that even zero switching
3
customers have value for the incumbent, despite the fact that they never purchase from
him again: when, keeping xed the number of high switching cost consumers, there
are more of them the prots of the incumbent increase. Indeed, their presence hinders
entrants who nd it more costly to attract high switching cost customers.
1
In order to conduct more complete comparative statics, in Section 4 we generalize
the model of Section 3 by assuming that the low switching cost consumers have strictly
positive switching costs. For technical reasons, we turn to a two period model. We are
able to conrm the economic insights highlighted in Section 3. Increasing the number
of low switching cost consumers increases the prot of the incumbent, even though in
equilibrium they always choose to purchase from one of the entrants. Furthermore, we are
able to show that these consumers are all the more valuable to the incumbent that their
switching costs are lower: a decrease in their switching costs increases their eagerness
to change supplier, makes entrants less aggressive and, as a consequence, leads to an
increase in the prot of the supplier. This is true despite the fact that the low switching
cost customers never purchase from the incumbent! This eect is strong enough that for
a large class of parameters it overwhelms the negative consequences for the incumbent
of a decrease in the switching costs of high switching cost consumers, so that an equal
decrease in the switching cost of all consumers increases its prots.
In most of the paper, we assume that price discrimination based on purchasing histories
is not possible. Due to free entry, all our results still hold if we allow rms to price
discriminate, as we discuss in see Section 4.5.
The conclusion discusses further research as well as policy implications.
1. Literature
The literature has distinguished switching cost models proper and subscription models:
in switching cost models, rms must charge the same price to both current and new
consumers, while in subscription models they can oer dierent prices to consumers with
dierent purchase histories. Switching cost models were introduced in the economics
literature by
2
Klemperer (1987b) (see the surveys of the theoretical literature in Klemperer
(1995), Annex A of National Economic Research Associates (2003), and Farrell and
Klemperer (2007), and the discussion of policy implications in National Economic
Research Associates (2003, especially Annex C)). Chen (1997) initiated the investigation
of subscription models. We present our model as a switching cost model, but, as discussed
in the introduction and in Section 4.5, our results hold if rms can discriminate between
consumers.
Much of the switching cost literature focuses on two-period duopsony models in which
rms choose between charging a high price in order to extract rents from their customers
and charging a low price in order to attract customers from their rivals. Farrell and
1
However, if the proportion of zero switching cost consumers increase keeping xed the total number
of consumers (and therefore decreasing the number of high switching cost consumers) the prots of the
incumbent decrease high switching cost consumers are still more valuable than zero switching cost
consumers.
2
See also Klemperer (1983) and Klemperer (1986).
4
Shapiro (1988), Beggs and Klemperer (1992), Padilla (1995), and Anderson, Kumar, and
Rajiv (2004) study innite horizon switching cost models, in each of these cases with two
rms and homogenous switching costs;
3
they focus on the evolution of market shares and
on the eect of switching costs on prices. Klemperer (1986) studies an innite horizon
model with homogeneous switching cost and free entry by rms.
In this framework, Klemperer (1987a) shows that higher switching costs may make
entry more likely, by inducing incumbents to abandon the hope of attracting the customers
of other incumbents and therefore choosing higher prices. Farrell and Klemperer (2007,
p. 1997) explain that
the rm must balance the incentive to charge high prices to harvest greater
current prots . . . against the incentive for low prices that invest in market
share and hence increase future prots.
Recently, Dube, Hitsch, and Rossi (2009) have studied the interaction these two eects
in an innite horizon model where a single consumer has random utility and rms have
dierentiated products; their focus is on empirics, and, through the use of simulation
methods, they provide numerical examples where prices fall when switching costs increase.
In a theoretical investigation of a simplied version of this model, Cabral (2008, 2009) has
shown that for low switching costs, the incentives of rms to invest in the acquisition of
new customers outweigh their harvesting incentives; as a result an increase in switching
costs leads to lower prices and to lower prots.
4,5
We also nd conditions under which higher switching costs lead to lower prots for
the incumbent, but the reasons are very dierent from those stressed in the literature.
Whereas previous authors have assumed a xed number of rms, we have free entry, and
therefore competition is more intense. Thus, the incumbent has no incentive to invest
in the acquisition of new customers, on which it can only make zero prots indeed,
in equilibrium, the incumbent does not try to recover the consumers that it has lost
to other rms. In other words, our comparative statics are entirely the consequence of
the heterogeneity of switching costs, and of the fact that low switching cost customers
protect the incumbent from entry.
The paper that is closest to ours is Taylor (2003). It analyzes a nite horizon
subscription model where consumers have dierent switching costs and where there is
free entry. In his primary model, consumers draw new switching costs from identical,
independent distributions in each period. He shows that free entry limits the advantages
of incumbency and that a rm makes zero expected prots from the consumers that
it attracts from its rivals. In an extension, Taylor examines a two period model with
two types of consumers who draw their switching cost (as before, independently in each
period) from dierent distributions. His focus is on the incentives of consumers to build
3
Beggs and Klemperer assume that consumers are horizontally dierentiated, but that, once they
have purchased from a rm, they never buy from another rm.
4
Arie and Grieco (2010) also provide a theoretical analysis of Dube et al. (2009) but, unlike Cabral
(and like Dube et al.), they assume that the consumers are myopic. (Our consumers are forward looking.)
5
In a subscription model based on Chen (1997), Bouckaert, Degryse, and Provoost (2008) also show
that higher swithching costs can lead to lower prots.
5
a reputation of having low switching cost in order to get better oers in the future, while
consumers in our model do not have an incentive to build a reputation.
In our model, switching costs are constant over time and this implies that it is harder
for an entrant to attract the more valuable consumers, those with higher switching costs,
than to attract the less valuable customers. As in Taylor, the presence of low switching
cost consumers hurts the high switching cost consumers, but in our model, we show that
it can also increase the incumbents prot.
Finally, in our model, because of free entry, incumbent rms nd it just as dicult
as entrants to attract customers of other rms. Therefore, incumbent rms, just like
entrants, make zero prots on customers of other rms, and in equilibrium they ignore
them when choosing the price they charge. As a consequence, our model would generate
exactly the same results if we transformed it into a subscription model and allowed for
the possibility for them to price discriminate based on consumer purchasing histories.
2. When all consumers have the same switching cost:
You cannot get rich on switching costs alone
In this section, we consider a repeated version of the most standard textbook model of
switching cost, with one incumbent and free entry. We show that, in equilibrium, the
prot of the incumbent is equal to its prot in the one period version of the game. This
is true for all equilibria when there are a nite number of periods, and for stationary
equilibria when there are an innite number of periods. We begin by presenting the one
period version of the model and then turn to the repeated game with a nite number of
periods.
There is a continuum of consumers with mass normalized to 1, and a good which can be
supplied by a number of rms, as we will describe below. Each consumer has a perfectly
inelastic demand for one unit of the good, and therefore always buys one unit either from
the incumbent or from one of the entrants. In this section only, all consumers have the
same switching cost . This switching cost is incurred every time a consumer changes
from one supplier to another. It reects industry wide similarities or compatibilities
between products, rather than idiosyncrasies of specic sellers. This implies, for instance,
that our comparative statics results which describe the consequences in changes of the
switching costs bear on circumstances where the cost of changing between any pair or
products increase or decrease.
In previous periods, the consumers have bought from the incumbent
6
rm I. We do not
study the process by which rm I became the incumbent, but only the continuation game
after entry becomes possible. In general, at least some of the incumbency rents which we
identify would have been dissipated in the competition to become the incumbent.
Let us consider rst a one period model with a denumerable number of entrants who
can enter the market at zero cost in each period. The main focus of our study is the
following Bertrand game:
6
In the dynamic version of the model, there could be, in some periods t > 1, several incumbents,
i.e., rms who have sold goods to a positive mass of consumers in previous period.
6
Stage 1: The incumbent and the entrants set prices;
Stage 2: The consumers choose from which rm to buy.
All of our qualitative results also hold true, and are sometimes easier to establish, in
the Stackelberg version of this game:
Stage 1: The incumbent sets a price;
Stage 2: The entrants set their prices;
Stage 3: The consumers choose from which rm to buy.
Assuming, as we will throughout this paper that all rms have zero marginal cost, it is
easy to prove that, in both the Bertrand and Stackelberg versions of the game, there is
(essentially) only one equilibrium, where the incumbent
7
charges , the entrants 0, and
all consumers buy from the incumbent.
We will now show that in the repeated version of the game, the discounted intertemporal
prot of the incumbent is not increased: it is still equal to . One can only pocket the
switching cost once.
8
This is easy to prove when there are two periods. Formally, we expand the game above
by assuming that every entrant that has sold to a positive measure of consumers in the
rst period becomes a second period incumbent, and that there are new entrants, again
in denumerable number,
9
in the second period. We assume that rms cannot commit to
prices beyond the current period. We also assume that they cannot discriminate between
consumers but this is not essential this second assumption only serves expository
purposes, and, as discussed in the introduction and in Section 4.5, is not essential.
In equilibrium, whether in the Bertrand or Stackelberg model, all second period
incumbents charge , and make prots equal to multiplied by the number of their
rst period customers. Therefore, competition between rst-period entrants pushes the
price that they charge down to , where (0, 1] is the discount rate.
10
Consumers
know that all incumbents will charge in the second period. Hence, rm I will be able
to keep its customers only by charging a price less than or equal to + . It is
straightforward to show that it indeed charges this price, and therefore that it keeps
all its customers. Hence its discounted prot is
( + ) + = .
An easy proof by induction shows that the same result holds with any nite number of
7
The results would be the same with several incumbents.
8
Although the model we use is a trivial extension of the most elementary model of switching costs,
we have not found in the literature a clear statement of what happens when this game is repeated, with
new entrants in every period; almost all of the literature focuses on the case of duopsony, where the
same two rms compete again each other period after period.
9
For the two period model, two entrants would be enough, but we need more in the innite horizon
models that we will discuss later.
10
This negative price should be interpreted as a discount below marginal cost. Thus, like much of the
literature, we assume that entrants can charge prices below marginal cost, and this is often the case in
commercial practice.
In some cases could be greater than the marginal cost, and, in practice, the entrant may not be
able to oer the full discount that we assume, for instance because consumers have free disposal. The
constraint which this imposes on the strategies of the rms have been studied recently by Chen and Xie
(2007). Throughout the paper, we assume that this constraint does not bind.
7
periods.
We now show that this result also holds true in the innite horizon version of this
model (we now assume < 1). In each period, we assume only a nite number of active
entrants oer the good. We look for subgame perfect stationary equilibria where rms
play undominated strategies that satisfy three other restrictions.
11
Entry and exit We assume that if a rm does not sell to a positive measure of
consumers in any period t, then it is not active in any period t
makes a much better oer than the incumbent, taking into account the fact that
the consumers have to pay the switching cost . However, every consumer thinks that
the others will refuse the oer, and therefore, by our entry and exit assumption, he
will have to pay the switching cost once again next period.
Formally, we allow small groups of consumers to coordinate on a strategy.
12
To
introduce this concept, we must introduce some notation. For consumers, a strategy
t
c
in period t is a function from the concatenation of the history of the game h
t1
and
the moves of the rms in that period t into consumer decisions in period t. We will
designate by
t
(c) the rm from which consumer c buys in period t and call
t
(f) the
set of consumers who buy from f in period t:
t
(f) = {c | f =
t
(c)}. For all periods
t N, all possible h
t1
and all (feasible) prices p
t
f
by active rms (that is for all possible
histories at the time the consumers choose their moves) it is not true that for all > 0
there exists a set C
;
b)
t
c
(h
t1
, p
t
f
) is independent of c over C
;
c) there exists a rm f
than from
t
c
(h
t1
, p
t
f
) (in the sense that their discounted disutility
would be lower after this deviation from the equilibrium).
Stationarity We state the stationarity assumption separately for incumbents and
entrants:
a) Incumbents use the same mixed strategy, if they sold to a positive measure of customers
in the last period, whatever the history.
13
11
The formal denitions of history, strategies and other elements of the game are in the web Appendix.
12
Of course, with a continuum of consumers, they cannot literally have mass; we use this terminology
to stress the fact that we are approximating a market with a large but nite number of consumers.
In many models of network externalities, it is assumed that the consumers coordinate on the purchasing
decision which maximize their utility. We do not make this assumption. In a dynamic model, either
we would have to assume that they are able to coordinate on a, potentially innite, sequence of moves,
which requires very strong coordination, or that this coordination has a myopic component, which is
not very attractive. Furthermore, as the game progresses even similar consumers can nd themselves in
situations where they face dierent payos moving forward; their interest might diverge.
13
This assumption needs to be rened when they are dierent types of consumers. For the results of
8
b) In or out of equilibrium, the distribution, F
E
of the minimum of the prices charged
by entrants is independent of history.
14
We now state the main result
15
of this section a sketch of the proof follows. (See
Appendix A for a formal proof.)
Proposition 1. In both the Stackelberg and the Bertrand models, when all consumers
have the same switching costs , the intertemporal discounted prot of the incumbent is
equal to , whatever the number of periods.
Therefore, as in the two period model, the incumbent can only collect the switching
cost once: he only gets one bite at the apple.
The result has been proved with a nite number of periods. When their number is
innite, let be the present discounted prot of an incumbent rm which supplied
all consumers in the previous period. By the stationarity assumption, this prot is
independent of the rms name and of the date. Entrants are willing to charge to
attract all the buyers. As in the two period model, consumers know that their welfare in
subsequent periods does not depend on the identity of rm they choose to purchase from
in the current period, and the incumbent will have to set a price equal to + in
order to keep its customers.
16
Hence, the equilibrium prot of the incumbent satises
= ( + ) + = .
In every period the entrants charge , while the incumbent charges (1 ), which
does yield a discounted prot equal to .
3. Heterogeneity of switching costs
We now turn to the main theme of the article: the consequences of heterogenous switching
costs. In this section, we study a model with two types of consumers: high switching cost
(hsc) consumers, who are a fraction (0, 1) of the population, have a switching cost
equal to > 0, while low switching cost (lsc) consumers, who form a fraction 1 of
the population, have a switching cost equal to 0. In the one period model, competition
drives the prices of entrants to 0, while the incumbent charges a price of , and obtains
a prot equal to : under the assumptions of this section, its prots are the average
switching cost of consumers multiplied by their mass. We analyze the innite horizon
version of this game.
Section 3, we simply need to assume that rms who sold to a positive measure of hsc customers in the
last period, always charge the same price.
14
This assumption needs to be adapted when they are several types of consumers. In particular, rms
who sold to consumers with a zero switching costs, are considered as entrants for the purpose of this
condition.
15
In a companion paper, Biglaiser and Cremer (2011) prove that there exist many other equilibria of
this game, which satisfy a weaker version of stationarity: although the outcome of the game is stationary
(with prices in each period as low as 0 or as high as ), after a deviation incumbents may charge prices
dierent from the prices along the equilibrium path.
16
Technically, the incumbent will charge + , the entrants charge and in the continuation
equilibrium, all the consumers buy from the incumbent.
9
3.1. Results
As in the model where consumers have the same switching costs, we restrict attention to
equilibria that satisfy the consumers have mass condition, the stationarity conditions,
and where players do not use weakly dominated strategies.
The following proposition summarizes our results.
Proposition 2. In the innite horizon model, where consumers have switching costs
equal to > 0, while the remaining consumers have 0 switching costs, under either
Stackelberg or Bertrand competition
i. the expected prot of the incumbent is
=
1 +
. (1)
ii. is greater than the prot of the incumbent in the one period model, , but
smaller than the value of an innite stream of one period prots, /(1 ).
iii. is strictly smaller than , but lim
1
= for all .
Parts i and ii of the proposition show that, contrary to what happens when all
consumers have the same switching costs, the intertemporal prot is not equal to the
one period prot, but is greater; however the per period prot is smaller in the innite
horizon model than in the one period model. Finally, part iii shows that when the agents
are very patient, the prot of the incumbent is independent of the proportion of hsc
consumers, whereas in the one period model prots are proportional to . As we will
explain below, lsc consumers, who always purchase from the lowest price entrant, make
it more costly to attract protable hsc customers away from the incumbent.
Proposition 2 yields interesting comparative statics, which we summarize in the
following corollary.
Corollary 1. Under the conditions of Proposition 2:
i. is increasing in , and ;
ii. for a given average level of consumer switching costs, , the prot of the incumbent,
, is decreasing in ;
iii. adding lsc consumers without changing the number of hsc consumers increases ;
iv. under Stackelberg competition, the utility of hsc consumers is an increasing function
of .
Parts i and ii of the corollary are obvious from equation (1). Part iii is easy to prove.
Assume that we add a mass > 0 of lsc consumers; the total mass of consumers becomes
= (1 + )
1 +
=
1 +
1+
,
which is increasing in . Lsc consumers are valuable to the incumbent, even though
they never buy its product, as they make it more costly for entrants to make aggressive
discounts in order to attract hsc customers.
10
Although they lead to the same prots for the incumbent, the equilibria under Bertrand
and Stackelberg competition are very dierent. Under Stackelberg competition, the
incumbent oers the same price in every period, and hsc consumers never change
suppliers. On the other hand, in Bertrand competition, the incumbent and the entrants
play mixed strategies, and in each period there is a strictly positive probability that all
the hsc consumers change suppliers. As switching is socially wasteful and as the prots
of the incumbent are the same in these two models, consumer surplus and social welfare
is lower under Bertrand than under Stackelberg competition. We summarize this result
in the following proposition.
Corollary 2. In the innite horizon model, where a proportion of the consumers have
switching costs equal to > 0, while the others have zero switching costs, consumer surplus
and welfare is lower under Bertrand competition than under Stackelberg competition.
Before proceeding, remember that, as we have discussed in Section 2, our comparative
statics results assume that the changes in switching costs that apply to all changes from
one supplier to another. A consumer who chooses to switch in the rst period from the
incumbent to an entrant would have to nd that his cost of switching once again, to a
future entrant, has also increased. On the other hand, the result does not apply if the
increase in switching costs applies only to a switch from the incumbent to a period 1
entrant. From a policy point of view, this implies that our theory can illuminate changes
which aect the whole industry, for instance changes in standards or new regulations
such as number or bank account portability.
We now present an informal proof of Proposition 2, starting with the Stackelberg case,
which is easier to analyze. Complete proofs are presented in Appendices B and C.
3.2. Analysis of Stackelberg competition
By stationarity,
17
hsc consumers know that the price that they will face in future periods
is independent of the rm they choose in the current period. Hence, they will denitely
switch suppliers if the dierence of price is strictly greater than and denitely not
switch if this dierence is strictly smaller than . In the rst period, entrants will be
willing to underbid the incumbent by (slightly more than) as long as the price it charges
is (strictly) greater than + . Hence, the incumbent will charge + and sell
17
The stationary requirements which we impose on the pricing strategy of the rms implies that we
search for equilibria where the pricing strategies of active rms only depend on whether consumers with
positive switching costs (that is, the protable consumers) purchased from them in the previous period.
In particular, the incumbents who have a positive measure of protable customers always choose the
same price (or the same distribution of prices) whatever the history, and the lowest price charged by an
entrant is the same (or has the same distribution) in all periods.
11
to the high cost customers at this price.
18
Therefore,
= ( + ) + = =
1 +
. (2)
This implies that the price charged, in every period, by the incumbent, and paid by the
hsc consumers, is equal to
p
S
1
1 +
. (3)
In the next stage of the game one or several entrants charges and attract all the lsc
consumers.
3.3. Analysis of Bertrand competition
In the Bertrand game, there is no equilibrium in which the incumbent charges p
S
=
+ and at least one entrant charges p
S
= . Indeed, if the incumbent did
not retain all the hsc customers, he would have incentives to decrease its price; if it
retained them, the entrant would attract only the lsc consumers, who generate no prot
in future periods, at a negative price. More generally, it is easy to show that there is no
pure strategy equilibrium of the game, but we will still be able to show that the prots
of the incumbent are equal to the prots in Stackelberg competition.
We do this by proving that, if is the (expected) prot of the incumbent, then +
belongs to the support of the distribution of prices that it announces; furthermore when
it chooses this price, its hsc customers purchase its product with probability 1. This will
imply that equation (2) holds. (More precisely, we will show that + is the lower
bound on the support of prices charged by the incumbent, and that when it chooses a
price arbitrarily close to this lower bound, it keeps the hsc customers with probability
arbitrarily close to 1.)
Lsc consumers always purchase from one of the lowest price sellers. By the stationarity
hypothesis, hsc consumers who change suppliers can never gain from purchasing from
an entrant which does not charge the lowest price: in the next period, any entrant
who has attracted customers and become an incumbent will charge the same price.
Hence, calling p
E
the lowest price charged by an entrant and p
I
the price charged by the
incumbent, hsc consumers buy from the incumbent if p
I
< p
E
+ and from one of the
lowest price entrants if p
I
> p
E
+ .
In the current period, the aggregate revenues of all the entrants who charge p
E
is equal
to p
E
times the mass of (lsc and hsc) customers that they attract. By stationarity, their
total future prots discounted to the next period are smaller than or equal to . Any
p
E
< would generate strictly negative discounted prots in the aggregate for the
lowest price entrants, and therefore b
E
, the lower bound of the support of the strategies
of entrants, is greater than or equal to .
18
More precisely, along the equilibrium path the incumbent charges + and the entrants charge 0
(so as not to subsidize lsc consumers who would bring them no future prots). In any continuation
equilibrium after one or several entrants charge , at least some hsc consumers buy from the
incumbent. In any continuation equilibrium after the incumbent charges more than + , some
entrants charge .
12
Clearly, the incumbent never charges less than b
E
+ . Furthermore, b
E
cannot be
strictly greater than : otherwise, an entrant could attract all the consumers and
make strictly positive discounted prots by charging a price in the interval (, b
E
).
Thus, b
E
= , and it is possible to show that the distribution of the lowest prices
charged by the entrants does not have a mass point at this price.
19
Therefore, when the
incumbent charges a price close to + (which is the lower bound of the prices it
charges), it keeps all the hsc customers, and in all stationary equilibria equation (2)
must hold.
4. Comparative statics
In Section 3, the switching cost of lsc consumers is equal to zero; with an innite horizon,
we have not been able to extend the analysis
20
to the case where all consumers have
strictly positive switching costs. This prevents us from examining questions such as the
consequences of an increase in the switching cost of the lsc consumers (and, a fortiori,
of an increase in the switching costs of all consumers). Therefore, in this section we
consider a two period model where lsc consumers can have strictly positive switching
costs. This leads to new economic insights and to unexpected comparative statics, which
are presented in Proposition 3.
4.1. Results and intuition
There are two
21
types of consumers: a mass of hsc buyers, with a switching cost equal
to
H
, and a mass (1 ) of lsc buyers with a switching cost equal to
L
[0,
H
). We
assume that
L
is small, more precisely,
L
<
1 +
H
, (4)
which implies
L
<
H
. Thus, in the one period model the incumbent would charge
H
,
sell to all the hsc consumers and to no lsc consumer, and make a prot equal to
H
.
(In subsection 4.4, we study environments where inequality (4) does not hold.)
The following proposition states our main result.
19
If there was such a mass point, for some > 0 the incumbent would never choose a price in
(b
E
+ , b
E
+ + ]: it could increase its prot by choosing a price slightly smaller than b
E
+ and
selling to its hsc customers with probability 1. Then, entrants who make at best zero prots by charging
would obtain higher prots by charging any price in the interval (b
E
, b
E
+ ) than by charging b
E
,
which establishes the contradiction.
20
When lsc consumers have zero switching costs, in every period they purchase from one of the rms
which charges the lowest price and that price is negative in equilibrium. Therefore, attracting them
generates no prot; this fact greatly simplies the analysis of the innite horizon model.
21
We have studied a model with a continuum of switching costs, which we do not discuss in this
paper. It leads to the same economic insights and more technical complications.
13
Proposition 3. In the two period model, where a proportion of consumers have
switching costs equal to
H
while the others have switching costs equal to
L
with
inequality (4) satised, the equilibrium prot of the incumbent is
=
H
_
H
L
H
L
(1 + )
_
(5)
under either Stackelberg or Bertrand competition. is greater than the one period prot,
H
, and smaller than the discounted value of a ow of one period prot,
H
(1 + ).
We discuss the proof of Proposition 3 in Sections 4.2 and 4.3. Before doing so, we
comment on its economic signicance.
As in the innite horizon model of Section 3, the presence of lsc buyers enables the
incumbent to generate higher prots than it would receive in the one period model.
Furthermore, when specialized to the case
L
= 0, equation (5) is consistent with
equation (1): it yields =
H
(1 + (1 )), which is equal to the value of a ow of
one period prots discounted at the rate of (1 ), as in equation (2). It is also worth
noticing that as converges to 1, converges to
H
, the one period prot, as we would
expect from point iii of Proposition 2.
Corollary 3. Under the hypotheses of Proposition 3
i. is increasing in and
H
and decreasing in
L
;
ii. If < (
L
+
H
)/2
H
, which is always satised if < 1/2, then an equal increase
in
H
and
L
leads to a decrease in (/
L
+ /
H
<0).
iii. If
L
<
2
H
/(1 + ), then a small increase in the number of lsc consumers
increases the prots of the incumbent.
iv. If
H
is increased by and
L
reduced by (1)/, so that the average switching
cost does not change, then increases.
Without surprise, when or
H
increase, the prot of the incumbent increases. To
understand why an increase in
L
decreases prots, we note rst that the incumbent
will always price in such a way that it sells to no lsc consumer. Let us assume, only
for expository purposes, that only one entrant attracted customers in the rst period,
and let
H
if it charges
H
, and
(
+ (1 ))
L
if it charges
L
. If it has attracted the proportion of hsc customers
such that
H
= ( + (1 ))
L
1
H
L
=
1
_
H
H
L
1
_
, (6)
it will be indierent between charging
L
and
H
. From (6), it is straightforward that
an increase in
L
leads to an increase in : the benets of keeping the lsc customers
increases, thus the number of hsc consumers attracted in the rst period must increase
14
if the entrant is to be kept indierent between its two plausible second period strategies.
In equilibrium, a proportion of hsc consumers purchase from the entrant in the rst
period: if fewer than this proportion did so, the entrant would charge a low price in the
second period, and be very attractive to hsc customers.
22
Therefore when
L
increases,
the rst period incumbent loses more customers, which explains the result.
Whether an equal increase in both
H
and
L
will increase or decrease the prot of
the incumbent will therefore depend on the relative strengths of two opposing eects,
which, by (5), can be determined by evaluating the change in
H
(
H
L
). Adding
to both
H
and
L
and taking the derivative for = 0, we obtain result ii) in Corollary 3:
the negative consequences for the incumbent of an increase in
L
swamps the positive
consequences of an equal increase in
H
when is small enough. (As explained in
section 1, we obtain these counterintuitive comparative statics for reasons dierent
than the rest of the literature.)
Part iii of the corollary is similar to part iii of Corollary 1.
23
Note that it requires a
L
smaller than the upper bound authorized by equation (4). Indeed, when
L
is small, the
same reasoning as in Section 3 holds: entrants do not want to attract lsc customers, and
an increase in their number makes them less aggressive. On the other hand, when
L
is larger, lsc customers become valuable enough to entrants that an increase in their
number makes them more aggressive.
Much empirical research is focussed on changes the average switching cost in the
market. Routine calculations demonstrate that if
H
is increased by and
L
is reduced
by (1 )/, so as to maintain the average switching cost but increase the variance of
the switching costs, then the incumbents prot is increasing in .
We now turn to the proof of the Proposition 3.
4.2. Proof of Proposition 3 for the Stackelberg model
In period 2, all the rms which sold strictly positive amounts in period 1 announce their
prices rst, followed by the entrants, who in equilibrium charge 0. The incumbents charge
H
if the proportion of hsc buyers among their period 1 consumers is strictly greater
than
L
/
H
and
L
if it is strictly less than
H
/
L
; if this proportion is exactly equal to
H
/
L
, they will be indierent between
L
and
H
, and charge either one of these two
prices with probability 1, or mix between the two.
If the rm from which it purchased in period 1 charges
H
in period 2, a lsc consumer
will choose to purchase from a period 2 entrant at a price of 0. Hence, his total period 2
cost will always be exactly
L
, whatever he does in period 1. As a consequence, if, as
above, we denote by p
E
the lowest price charged by any entrant in period 1 in response
to the period 1 price p
I
charged by the incumbent, lsc consumers will purchase from
22
As we will see shortly, the entrant mixes between
H
and
L
in the second period.
23
It is easy to prove by computing the value of the derivative of
(1 + )
H
_
1+
H
L
H
L
(1 +
1 +
)
_
=
H
H
L
(
H
(1 + )
L
)(1 +
1 +
).
with respect to for = 0.
15
one of the lowest price entrants if p
E
+
L
< p
I
, from the incumbent if p
E
+
L
> p
I
, and
from one or the other if p
E
+
L
= p
I
. Eectively, lsc customers minimize their cost in
each period. Therefore, in equilibrium, the expected value of the second period price
of all the entrants who attract lsc customers in the rst period must be equal to each
other.
Because second period prices are increasing functions of the proportion of hsc customers
in the client`ele of a rm, the hsc customers who purchase from an entrant will also
allocate themselves among the lowest cost rst period entrants, and it cannot be an
equilibrium for these entrants to charge dierent prices in the second period. Therefore,
the pricing strategy of the successful entrants will only depend on whether or not in
the aggregate they attracted a proportion of the hsc consumers smaller than, equal to,
or greater than , as dened in (6).
This enables us to prove the following lemma, which describes the continuation payo
of the incumbent as a function of the price it charges in the rst period. (The proof is in
Appendix D.)
Lemma 1. For a given price, p
I
, charged by the incumbent in the rst period:
1. if p
I
< (1 )
L
, the incumbent sells to all consumers in period 1 and to all hsc
consumers (at price
H
) in period 2. Its prot is p
I
+
H
.
2. if p
I
((1 )
L
, (1 )
H
), the incumbent sells to all hsc consumers in both
periods and to no lsc consumer in either period. Its prot is (p
I
+
H
).
3. if p
I
_
(1 )
H
, (1 )
H
_
, the incumbent sells to (1 ) hsc consumers at
price p
I
in period 1 and at price
H
in period 2, while its sales to lsc consumers are
equal to 0 in both periods. Its prot is (1 )(p
I
+
H
).
4. if p
I
> (1 )
H
, the incumbent has zero sales in both periods.
From Lemma 1, the prots of the incumbent are increasing on the intervals (, (1
)
L
), ((1 )
L
, (1 )
H
) and
_
(1 )
H
, (1 )
H
_
. Given the restrictions that
we have imposed on
H
/
L
, it is easy to check that it is maximized on the union of
these intervals for p
I
smaller than and very close to (1 )
H
. Therefore, the only
equilibrium of the game has the incumbent charging (1 )
H
in the rst period
with the continuation equilibrium described in point 3, yielding the prots described by
equation (5). This proves Proposition 3 for the Stackelberg model.
24
4.3. Sketch of the proof of Proposition 3 in the Bertrand model
We divide this subsection in two parts. In the rst, we provide a short sketch of the
proof of equation (5), which is derived formally in Appendix E; in the second part, we
describe in detail one of the payo equivalent equilibria of the game.
24
The identication of the equilibrium can be easily extended to the case where the cost of shifting
from the incumbent to an entrant is
L
, greater but close to
L
which now is the cost of shifting from a
period 1 to a period 2 entrant. The entrants will still charge
L
in the period 1, and the total cost
of the lsc consumers will be
L
. Apart from this, the equilibrium, and in particular the prot of the
incumbent, will not be aected.
16
4.3.1. Proving equation (5)
There are only mixed strategy equilibria in the Bertrand model, and we use a proof
similar to the proof in Section 3.3 to show that the prots of the incumbent are the same
as in the Stackelberg model.
In period 1, entrants never charge strictly less than
H
: at this price, they make
zero prot even if they attract all the buyers. By exactly the same reasoning as in the
innite horizon case, this price must be in the support of the lowest price charged by the
entrants and
H
+
H
must be in the support of the period 1 price charged by the
incumbent. Because we show that the incumbent never sells to a lsc customer in the
period 1 and sells only to a proportion 1 of the hsc consumers, its prot when it
charges
H
+
H
is
(1 ) [
H
(1 ) +
H
)] ,
(
H
(1 ) +
H
is the discounted prot per customer of the incumbent). It is easy to
check that this is indeed equal to the of equation (5).
4.3.2. What do equilibria look like?
The reasoning above is sucient to prove equation (5), but does not provide much
intuition about the equilibrium strategies of the agents. To help the reader build this
intuition we now describe explicitly one equilibrium of the Bertrand game.
As discussed in section 3.3, in all equilibria the incumbent and the entrants use mixed
strategies in period 1. For simplicity, we present an equilibrium where there is only
one
25
active entrant, who chooses its price p
E
in [
H
,
L
], while the incumbent
chooses p
I
in [
H
+
H
,
L
+
H
) and at least one other entrant charges
L
with probability 1. Then, all lsc customers buy from the entrant, and, depending on
the dierence between p
I
and p
E
, either all or a fraction of hsc customers purchase
from the entrant:
if p
I
p
E
H
, then all hsc consumers buy from the entrant, who therefore
charges
H
in the second period its second period prot is
H
;
if p
I
p
E
<
H
, a proportion purchases from the entrant, who in the second
period uses a mixed strategy: he chooses prices
L
and
H
with probabilities
such that the hsc customers are indierent between switching and not switching
suppliers in period 1 its second period prot is ( +(1 ))
L
=
H
(in the
states of nature where its second period price is
H
, all the lsc customers switch
to a period 2 entrant).
Therefore, in equilibrium, a proportion at least equal to of the hsc consumers
purchase from the entrant in period 1.
25
Our equilibrium is also an equilibrium if there are several active entrants and they each choose a
mixed strategy such that the distribution of the minimum of the prices they charge is the function G
E
dened below. In other words, there are many dierent payo equivalent equilibria. The key to all
equilibria is that there is free entry and that all entrants make 0 expected prots in equilibrium.
17
The entrant chooses p
E
according to the following distribution G
E
, which has a mass
point at
L
:
G
E
(p
E
) =
_
_
_
p
E
+
H
p
E
+ (1 + )
H
if p
E
[
H
,
L
),
1 if p
E
=
L
.
(7)
Then, if the incumbent chooses any p
I
[
H
(1 ),
H
L
), its expected discounted
prot is
G
E
(p
I
H
) 0 +(1 G
E
(p
I
H
)) (1 )(p
I
+
H
) = (1 )
H
(1 + ). (8)
To see why the incumbent must choose a price in the interval [
H
(1 ),
H
L
),
we check for possible deviations. A) It is not protable for the incumbent to choose a
price greater than or equal to
H
L
, as this implies p
I
p
E
H
with probability 1,
and no sales! B) To show that it is not protable to decrease prices below
H
(1 ),
we proceed in two steps. a) First, note that by charging
H
(1 ), the incumbent
sells to a proportion 1 of hsc customers. Claim E2 in Appendix E shows that to
increase its sales above this number the incumbent needs to choose p
I
H
(1), which
implies that as long as it does not sell to lsc customers, its prot, (p
I
+
H
), is at
most
H
which is smaller than (1 )
H
(1 + ), by (4). b) Second, in order to
sell to lsc customers, the incumbent needs to make their total costs, over both periods,
less than
L
, which is the upper bound of their cost if they switch to the entrant in the
rst period. Given that they will switch in period 2 when it charges
H
, this necessitates
p
I
(1 )
L
, which leads to prots p
I
+
H
smaller than the prots when using the
equilibrium strategy.
Similarly, in our equilibrium the incumbent chooses p
I
according to the distribution
G
I
(p
I
) =
p
I
H
(1 )
p
I
H
(1 ) + ((1 ) + )(
H
p
I
L
)
.
Then, the prot of the active entrant is
G
I
(p
E
+
H
) (p
E
+
L
)(1 + ) + (1 G
I
(p
E
+
H
))(p
E
+
H
) = 0
when it chooses a price in [
H
,
L
], and smaller than or equal to 0 when it
chooses a price outside of this interval (the presence of another inactive entrant who
charges
L
is crucial for this last point).
In all equilibria p
I
will be distributed according to G
I
and p
E
, interpreted as the lower
bound of the prices of the active entrants, will be distributed according to G
E
. We will
let the interested reader convince himself of this fact.
Figure 1 shows the equilibrium strategies with
H
= 1,
L
= .2, = .4 and = 1.
4.4. Equilibrium with large
L
For completeness, we now turn to a discussion of the equilibrium when equation (4) does
not hold. Proofs and more details can be found in the web Appendix.
18
0.4 0.35 0.3 0.25 0.2
.03
.2
.31
.4
.6
.8
1
entrant
p
e
0.6 0.65 0.7 0.75 0.8
.03
.2
.31
.4
.6
.8
1
incumbent
p
I
Figure 1: This gure represents the probability distributions in the mixed strategies
of the incumbent and the entrant with
H
= 1,
L
= .2, = .4 and = 1,
which implies = 37.5%. For instance, reading along the vertical dashed line,
if p
E
= 0.35, we obtain G(p
E
) 0.03, which implies that if the incumbent
chooses p
I
= 0.65 = 0.35 +
H
, then it looses all its hsc customers with
probability 3% and sells to a proportion 1 of them with probability 97%.
Similarly, if the entrant chooses p
E
= 0.35, it sells to a proportion of hsc
customers with a probability 31% and to all of them with probability 69%.
19
If
L
is very large, i.e., greater than
H
, then everything happens as if all the
consumers were lsc consumers: the incumbent charges
L
(1 ) in period 1,
L
in
period 2, and sells to all consumers. Its prot is
L
.
If
L
/
H
is smaller than but close enough to , then there is a pure strategy equilibrium
where in period 1 the incumbent sells to all the hsc consumers at a price
H
(1) and the
entrants sell to all the lsc consumers at a price
L
. In period 2, the incumbent charges
H
and keeps all the hsc customers its prot over both periods is therefore
H
, and
as when
L
is small, prots increase when there is a mean preserving spread in switching
costs . The best alternative strategy for the incumbent would be to charge
H
L
in
period 1, and sell to a proportion 1 of the hsc consumers. This strategy becomes more
attractive as
L
decreases, and dominates when
L
/
H
< x
C
, where x
C
(/(1 +), )
solves
x
C
[1 + + ] =
_
+ x
C
2
. (9)
When
L
/
H
(/(1 + ), x
C
), we show that there exists the same mixed strategy
equilibrium as when
L
<
1+
H
. The dierence is that now the one period game prot
is larger than the equilibrium prot and thus we need to check for deviations where the
incumbent could retain all the hsc customers.
26
4.5. Price discrimination
For expositional convenience, we have assumed throughout the paper that rms could
not price discriminate on the basis of consumers past purchases history; this is not
essential for our results. Typically, when considering such discrimination, the literature
has assumed that incumbents could oer to new customers prices lower than those they
oer to consumers which have bought their products in previous periods. We show that
allowing such behavior would not aect our results.
Let us allow an incumbent rm to oer in period 2 dierent prices to consumers who
have purchased its product in period 1 and to consumers who purchased from another
rm. Due to free entry, a period 2 entrant, who by denition has no customers from
period 1, cannot make positive expected prots. When choosing the price it oers new
customers, an incumbent nds itself in exactly the same position as a period 2 entrant.
That is, given the the price distribution of the lowest priced entrant, there exists no price
that it can charge (only) to new customers and improve its prots. Thus, enlarging the
strategy space for the incumbent and by allowing for price discrimination cannot lead to
an increase in its prot.
Similarly, redening the notion of stationarity and allowing incumbents to price
discriminate in the models of Sections 2 and 3 would also not change our results in any
way.
26
In the web appendix we only prove that there exists an equilibrium which satisfy these properties,
not that all equilibria do although we believe that this may well be the case.
20
5. Conclusion
A signicant body of theory explores the consequences of consumer switching costs: it
highlights the role of bargain then rip-o pricing patterns, where a rm makes very
protable introductory oers and raises its price in subsequent periods. To the best of our
knowledge, the fact that the distribution of switching costs changes considerably the way
in which these strategies play out has not been pointed out. We hope the present paper
will contribute to close this gap. Taking into account the heterogeneity of switching costs
has enabled us to identify very rich strategic interactions between the incumbent and the
entrants and led to surprising comparative statics.
The following story illustrates how dierences in switching costs play in the real
world. On February 6, 2007, in the same well-known letter in which he called for an end
to DRM (Digital Rights Management) for music distributed in electronic form, Steve Jobs
discussed the incumbency benets that the iPod enjoyed thanks to iTunes proprietary
format (Jobs, 2007). He noticed that
[s]ome have argued that once a consumer purchases a body of music from
one of the proprietary music stores, they are forever locked into only using
music players from that one company. Or, if they buy a specic player, they
are locked into buying music only from that companys music store.
He argued that on average there are 22 songs purchased from the iTunes store for
each iPod ever sold, and that this implied that under 3% of the music on the average
iPod is purchased from the iTunes store and protected with a DRM. He concluded that
there was no lock-in as it is hard to believe that just 3% of the music on the average
iPod is enough to lock users.
In a response to Jobs statement, Jon Lech Johansen
27
made the following interesting
points:
Many iPod owners have never bought anything from the iTunes Store. Some
have bought hundreds of songs. Some have bought thousands. At the 2004
Macworld Expo, Steve revealed that one customer had bought $29,500 worth
of music.
Therefore, the lock-in is non negligible as
its the customers who would be the most valuable to an Apple competitor
that get locked in. The kind of customers who would spend $300 on a set-top
box.
In essence, Johansen argued that the consumers that matter, those who buy lots of
online music, have high switching costs, and therefore that an entrant in the market will
face large obstacles attracting them.
27
See Johansen (2007). Jon Lech Johansen, also know as DVD Jon is a hacker made famous by
his work on reverse engineering of data formats, and in particular on the DVD licensing enforcement
software (see http://nanocr.eu/, last visited on 31 January 2010).
21
There are dierences between our very stark setup and the market for music players,
due in particular to the fact that switching costs are correlated with the amount that
consumers spend on music. However, our analysis supports Johansens insight that the
distribution of switching costs might be important, but our interpretation is dierent from
his: the heterogeneity of switching costs could be benecial to Apple not so much because
it implies that there exist a subset of consumers with very high switching costs, but
rather because the presence of customers with low switching costs makes an aggressive
pricing strategy potentially very costly for an entrant.
Our results should aect policy analysis. For instance, the liberalized UK domestic gas
and electricity markets analyzed by NERA in National Economic Research Associates
(2003) appears to broadly t the context we consider: the product is homogeneous,
discrimination between old and new customers was not an option, and entrants had
to attract customers away from the historical incumbent (British Gas and the public
electricity suppliers) as there were practically no unattached customers. Entrants oered
prices below cost, and a fortiori below those of the incumbent(s), which saw their market
share decrease. Our analysis shows that information on the distribution of switching
costs, for which no data is given, should have been gathered and that its consequences
for the strategy of the entrants should have been considered.
Our results should also have consequences for the empirical work which tries to estimate
the consequences of changes of switching costs on prices. For instance, Viard (2007)
examines the eect introducing number portability (if consumers change phone companies
they can keep their same number) for toll free calling. This reduces the switching costs
of buyers of toll free services. If consumers have dierent switching costs and these
switching costs are aected dierentially by the change to number portability, then
not only do average switching cost fall, but the distribution of switching cost changes.
Similarly, Kim, Kliger, and Vale (2003) examine a switching cost model in banking. Our
work demonstrates that when the empirical work does not take into account how the
distribution of switching cost changes, then the estimated model maybe misspecied.
On the theoretical side, we have used a very stark model, with free entry and many
entrants in every period whereas much of the literature on switching costs has emphasized
models where a limited number of incumbents compete over time, trying to vie for each
others consumers. In the rest of this conclusion, we briey discuss some extensions of
the model.
The rst extension would be to allow for the presence of xed costs for entrants.
Assume that the xed cost are incurred before any rm sets its price, and that it is small
enough that there is entry. Then, as in the standard Bertrand model, the only equilibria
are mixed strategy equilibria. When at least two potential entrants actually enter, the
pricing decisions in the continuation game will be exactly the same as without entry
costs. With zero or one entrant, on the other hand, prices will be higher. We believe
that, in general, the presence of switching costs would mitigate the eects studied in this
paper. Indeed, an increase in the number of lsc consumers or an increase in
L
will lead
to a higher probability that there will be at least two entrants, and therefore to lower
prices.
Another interesting avenue for future research is to assume that in each period some
22
consumers are replaced with new consumers who are initially not tied to any rm. To get
the avor of what the equilibrium would look like, we make the following observations,
assuming that the proportion of new consumers is small. First, an incumbent with some
consumers who have purchased in the past will not try to attract new consumers, and, as
in the main body of the paper, only the lowest priced entrants will attract new consumers.
The number of incumbents will therefore grow from period to period, and their market
shares will shrink. Furthermore, entrants should price as aggressively as when there were
no new consumers. This will lead to lower incumbent prices and prots.
We have done some very preliminary work on a model where the distribution of
switching costs is drawn from a continuous distribution. Additional complications occur
when doing comparative statics in this model, since entrants in period 1 will become
incumbents in period 2 and will lose some consumers in period 2 if the lowest bound on
switching cost is small enough. This may be an interesting avenue to pursue more fully
in the future.
We have not been able to identify the equilibria in a innite horizon model, except
in the case where the switching cost of the lsc consumers is equal to 0. Solving this
problem raises interesting, but dicult, questions; in particular, we are not sure that a
stationary equilibrium exists, or we do not even know what would be the appropriate
denition of stationarity for that case.
Finally, network externalities often play a role similar to switching costs they have
sometimes been called collective switching costs. In future work, we plan to study
models where agents have dierent trade-os between size of network and prices; we
believe that phenomena similar to those analyzed in the current paper can be identied.
23
References
Anderson, Eric T., Nanda Kumar, and Surendra Rajiv (2004), A comment on: Revisiting
dynamic duopoly with consumer switching costs. Journal of Economic Theory, 116,
177186.
Arie, Guy and Paul L. E. Grieco (2010), Who pays for switching costs? Unpublished
manuscript, Northwestern University.
Beggs, Alan and Paul Klemperer (1992), Multiperiod Competition and Switching Costs.
Econometrica, 60, 651666.
Biglaiser, Gary and Jacques Cremer (2011), Equilibria in an innite horizon game with
an incumbent, entry and switching costs. International Journal of Economic Theory,
7, 6576.
Bouckaert, Jan, Hans Degryse, and Thomas Provoost (2008), Enhancing market power
by reducing switching costs. CESifo Working Paper 2449, Munich.
Cabral, Lus (2008), Switching costs and equilibrium prices. Unpublished manusrcript,
available at http://luiscabral.org.
Cabral, Lus (2009), Small Switching Costs Lead to Lower Prices. Journal of Marketing
Research, 46, 449451.
Chen, Yongmin (1997), Paying Customers to Switch. Journal of Economics and
Management Strategy, 6, 877897.
Chen, Yuxin and Jinghong Xie (2007), Cross-market network eect with asymmetric
customer loyalty: Implications for competitive advantage. Marketing Science, 26,
5266.
Dube, Jean-Pierre, Guenter J. Hitsch, and Peter Rossi (2009), Do Switching Costs Make
Markets Less Competitive? Journal of Marketing Research, 46, 435445.
Farrell, Joseph and Paul Klemperer (2007), Coordination and Lock-In: Competition
with Switching Costs and Network Eects. In Handbook of Industrial Organization
(Mark Armstrong and Rob Porter, eds.), volume 3, 19672072, North-Holland.
Farrell, Joseph and Carl Shapiro (1988), Dynamic Competition with Switching Costs.
The RAND Journal of Economics, 19, 123137.
Jobs, Steve (2007), Thoughts on Music. Statement of 6 February 2006, URL http:
//bit.ly/iUeQJY.
Johansen, John L. (2007), Steves misleading statistics. URL http://bit.ly/mLSUZ5.
Kim, Moshe, Doron Kliger, and Bent Vale (2003), Estimating switching costs: the case
of banking. Journal of Financial Intermediation, 12, 2556.
24
Klemperer, Paul D. (1983), Notes on Consumer Switching Costs and Price Wars.
Klemperer, Paul D. (1986), Markets with consumer switching costs. Ph.D. thesis, Graduate
School of Business, Stanford University.
Klemperer, Paul D. (1987a), Entry Deterrence in Markets with Switching Costs.
Economic Journal, 97, 99117.
Klemperer, Paul D. (1987b), Markets with Consumer Switching Costs. Quarterly
Journal of Economics, 102, 375394.
Klemperer, Paul D. (1995), Competition When Consumers Have Switching Costs.
Review of Economic Studies, 62, 515539.
National Economic Research Associates (2003), Switching Costs. Report prepared for
the Oce of Fair Trading and the Department of Trade, UK.
Padilla, A. Jorge (1995), Revisiting Dynamic Duopoly with Consumer Switching Costs.
Journal of Economic Theory, 67, 520530.
Taylor, Curtis R. (2003), Supplier Surng: Competition and Consumer Behavior in
Subscription Markets. The RAND Journal of Economics, 34, 223246.
Viard, V. Brian (2007), Do switching costs make markets more or less competitive?
The case of 800-number portability. The Rand Journal of Economics, 38, 146163.
25
Appendices
A. Equilibrium in the innite horizon model when all
consumers have the same switching cost
We prove Proposition 1 for the Bertrand case. The Stackelberg case is very similar and
somewhat easier to prove.
We rst establish that (1 ) is an upper bound on p
. By stationarity, whether
consumers purchase from an incumbent or an entrant, in future periods they will incur
the same cost p
if p
+
p
1
.
Writing that this expression is negative for all > 0 yields p
(1 ).
We now show that (1 ) is also a lower bound on p
: otherwise the
incumbent(s) could increase its (their) price(s) without loosing customers. If the entrant
attracted customers at this price it would make prots equal to the mass of these
customers multiplied by p
+ p
to
be undominated.
The proposition is an immediate consequence of p
= (1 ).
B. Stackelberg equilibrium in the innite horizon model
with two levels of switching costs
In this appendix, we present a formal version of the proof sketched in 3.2. We will do so
by proving that in all equilibria the equilibrium price charged by an incumbent, p
, is
equal to p
S
as dened in (3).
We rst show that p
S
is an upper bound on p
. Let p
E
the minimum of the prices
charged by any entrant (this minimum exists as we are identifying equilibria with a nite
number of active entrants in each period). By stationarity, if p
E
< p
all consumers
purchase from one of the lowest price entrants. The sum of the prots of these entrants
is p
E
+ p
+
p
1
0 p
1
1 +
= p
S
26
To show that p
S
is a lower bound on p
, we show that if p
< p
S
a deviation by
the incumbent to any p
(p
, p
S
) would be protable. Indeed, entrant(s) who would
respond by charging p
+ p
/(1 ) . (This is their prot if they attract all the hsc consumers.) If
both p
and p
= p
S
, which implies the claims of the theorem in the
Stackelberg case.
C. Bertrand equilibrium in the innite horizon model
with two levels of switching costs
In this appendix, we begin by proving that equation (1) must hold for any Bertrand
equilibrium. We then show that there indeed exists an equilibrium by exhibiting one.
C.1. Proof of equation (1)
Most of the work consists in computing bounds on the distribution of prices charged
by the rms: b
I
and b
I
, respectively the lower and upper bounds of the support of the
prices charged by incumbent(s) as well as b
E
and b
E
the lower and upper bounds of the
distribution of the lowest price charged by the entrants.
Claim C1. Incumbents have strictly positive prots, which implies b
I
> 0.
Proof. We assume that incumbents have prots equal to 0, and show that this lead
to a contradiction. Because we are looking for stationary equilibria, any entrant who
would attract consumers would become an incumbent in the next period, and therefore
make 0 prots. It is therefore a dominated strategy for entrants to oer a negative price.
Now assume that an incumbent deviates in period t and charges p
(0, (1 )). An
upper bound for the total cost that an hsc consumer can incur by purchasing from the
incumbent in period t and switching in period t +1 is p
G
I
(p
I
) = 1, the function G
I
has no mass point.
Similarly, the distribution G
E
of p
E
is determined by the fact that the prots of the
incumbent are equal to for all prices in [b
I
, b
I
], and therefore
G
E
(p
E
) = 1
(p
E
+ ) +
for p
E
(b
E
, b
E
). (C2)
Because
lim
p
E
0
G
E
(0) =
(1 )
+
< 1;
the distribution G
E
has a mass point at p
E
= 0.
C.3. Existence of an equilibrium
We have proved that if there exists an equilibrium that satises our assumptions, the
distribution of prices must satisfy equations (C1) and (C2). We now prove that there
does indeed exist such an equilibrium; this is a proof by construction: we exhibit the
strategies followed by the agents.
28
If some of the hsc consumers do not purchase from the incumbent, none of the lsc consumers will.
29
In this equilibrium the consumers who buy from an entrant always buy from the same
(lowest price) entrant. The analysis which we have conducted to derive (C1) and (C2),
shows that these strategies are best responses for all the agents when there is only one
incumbent.
We need to examine the continuation equilibrium when there are several incumbents,
for two reasons: a) for the consumers to nd it optimal to coordinate on buying from
one incumbent, it must be the case that it is not a protable deviation for a small mass
of hsc consumers to purchase from another rm than the other hsc consumers; b) we
have imposed the requirement that all incumbents, i.e., all rms that have sold to hsc
consumers in the previous period, use the same pricing strategy. As we will see, the fact
that the strategies of the rms satisfy b) provides an easy proof of point a).
Let us therefore assume that in one period there are n 2 incumbents. It is straight-
forward to see that if the distribution of p
E
is G
E
, then all the incumbents are indierent
between all prices in [b
I
, b
I
]. We now show that the prots of the lowest price entrants
are equal to 0 if all the incumbents choose the strategy described by (C1).
Let
i
be the mass of hsc consumers of incumbent i = 1 . . . , n in the previous period.
The lowest price entrant sells to all the lsc consumers and to the mass of hsc consumers
who were in the previous period clients of rms who choose in the current period a
price p
i
p
E
+ . Because it will follow the same strategy as a unique incumbent,
and because the distribution of prices of the entrant is independent of the number of
incumbents, its prots discounted to the future of next period will be /.
Therefore, for given prices by the incumbents, the prot of the entrant is
(1 )p
E
+
{i|p
i
>p
E
+}
(
i
p
E
+
i
) = (1 )p
E
+
{i|p
i
>p
E
+}
_
p
E
+
_
= (1 )p
E
+
i
s
i
(p
i
)
_
p
E
+
_
,
where s
i
is the random variable, of expected value
i
(1 G
I
(p
E
+ )), that takes the
value
i
for p
i
p
E
+ and 0 otherwise. The p
i
s are independently distributed, and
therefore the expected value of
s
i
(p
i
) is (1 G
I
(p
E
+ )), and the expected prot of
the lowest price entrant, conditional of the fact that it has chosen a price of p
E
, is
(1 )p
E
+ (p
E
+ )(1 G
I
(p
E
+ )),
which, by equation (C1), is equal to 0.
Because all incumbents use the same pricing strategy hsc consumers have no incentive
to deviate from the focal strategy described above: in subsequent periods, they would
face the same distribution of prices both from the rm they purchased from in previous
periods and from the entrants.
D. Equilibrium in the two period Stackelberg model
In this Appendix, we prove Lemma 1. We begin by establishing three claims; the rst
one is part 4 of the lemma.
30
Claim D1. If p
I
> (1 )
H
, the incumbent has zero sales in both periods.
Proof. If p
I
> (1 )
H
, a unique lowest price entrant who would charge p
E
(
H
, p
I
H
) would make strictly positive prots equal to p
E
+
H
, as it would
attract all consumers in period 1. Free entry prevents this, and therefore in the continua-
tion game, one or several entrants must charge
H
, and attract all the consumers
while making zero prots.
Claim D2. If p
I
< (1 )
H
, then no entrant attracts enough hsc consumers in
period 1 that it nds it optimal to charge
H
with probability 1 in period 2.
Proof. Assume that entrant e attracted a large enough proportion of hsc customers
that it found it optimal to charge
H
in period 2. Because lsc consumers always nd
it strictly more protable to switch suppliers than do hsc consumers, the incumbent
would have no lsc customers and, therefore, hsc customers can guarantee themselves a
second price of
H
by staying with the incumbent. Therefore, entrant e must have
chosen a period 1 price p
e
p
I
H
<
H
<
L
. This implies that no entrant
nds it protable to attract period 1 customers and charge
L
in period 2: all entrants
that attract customers must choose the same strategy as e, and in the aggregate their
prots are smaller than p
e
+
H
< 0, which establishes the contradiction.
Claim D3. If p
I
< (1)
H
, all hsc consumers purchase from the incumbent in period
1.
Proof. By Claim D2, any period 1 entrant who has attracted consumers in period 1
charges
L
with positive probability in period 2. Therefore, its second period prot will
be
L
times the mass of consumers it attracted in the rst period and, by free entry, its
period 1 price must be
L
. The total discounted cost for a hsc consumer who would
purchase from a period 1 entrant would therefore be at least (
L
+
H
) +
L
=
H
(it would be greater if in period 2 the entrant charged
H
with a probability in (0, 1)). If
he purchases from the incumbent, his total cost is p
I
+
H
, and therefore strictly lower,
which establishes the claim.
Parts 1 and 2 of the lemma follow immediately from Claim D3.
If p
I
_
(1 )
H
, (1 )
H
_
, hsc consumers prefer to purchase from an entrant
if its period 2 price is
L
and from the incumbent if the entrants period 2 price is
H
.
Therefore, there can be an equilibrium only if the entrants play a mixed strategy in
period 2, which is feasible only if in period 1 they attract a proportion of the hsc
consumers. This establishes part 3 of the lemma and completes the proof.
E. Equilibrium in the two period Bertrand model
In subsection E.1, we begin by deriving some properties of the equilibrium of the two
period Bertrand model which hold for any values of the parameters. Then, in E.2, we
specialize the model to the case where equation (4) holds and prove Proposition 3. Finally,
in F, we conduct the analysis which leads to the large
L
results discussed in 4.4.
31
E.1. Some properties of the equilibrium in the two period Bertrand
model
Because of free entry, period 2 entrants choose a price equal to 0. As in the Stackelberg
case, a period 2 incumbent charges
L
or
H
depending on whether the proportion of
its hsc customers in period 1 was less or greater than
L
/
H
, and, clearly, the period 1
incumbent will charge
H
in period 2.
This implies that, as in the Stackelberg case, in period 1 lsc consumers will optimally
behave as if they were myopic, switching to one of the lowest price entrants if the dierence
between its price and the incumbents price is greater than
L
and not switching if this
dierence is smaller than
L
. It also implies that any hsc consumer who does not buy
from the incumbent in period 1 also buys from one of the lowest priced entrants. Indeed,
any other entrant would attract only hsc customers, and hence charge
H
in period 2.
We are now ready to study the pricing behavior of the rms in period 1. We begin
by Claim E1 which describes the behavior of entrants. Next, in Claims E2 and E3
we describe properties of the incumbents rst period demand function. They help us
characterize the strategy of the incumbent, which enables us to compute the lower bound
on its prices of Claim E6.
Claim E1. In period 1, any active entrant charges a price in [
H
,
L
].
Proof. Any entrant who has attracted consumers in period 1 will charge at least
L
in
period 2. Therefore, competition and free entry will ensure that in period 1 no entrant
which charges more than
L
attracts a positive measure of customers with positive
probability. If the lower priced entrants charge prices strictly smaller than
H
, their
aggregate prot is negative by the same line of reasoning as in the proof of Claim D2.
Claim E2. If the incumbent charges a price strictly greater than
H
(1 ) in period 1,
then it sells to at most (1 ) hsc customers.
Proof. Assume p
I
> (1 )
H
. Because
L
+
L
<
H
(1 ), by Claim E1 all the
lsc consumers, who act myopically in the rst period, purchase from entrants. If on the
aggregate the entrants attract a proportion of hsc customers smaller than , at least one
of them will have a proportion of period 1 hsc customers strictly smaller than
L
/
H
and therefore charge
L
with probability 1 in period 2. Hsc customers would nd this
entrant more attractive than the incumbent as (
L
+
H
) +
L
<
H
(1 ) +
H
,
which establishes the contradiction.
Claim E3. If in period 1 the incumbent charges a price strictly smaller than
H
(1 ),
then it sells to at least (1 ) hsc customers.
Proof. If this were not the case, at least one of the entrants would attract enough hsc
customers in the rst period to charge
H
in the second period; by Claim E1, these hsc
customers would incur total discounted costs equal to at least
H
+
H
+
H
, which
is strictly larger than the total discounted costs that they would incur from buying from
the incumbent in both periods.
32
Claims E2 and E3 show that for p
I
(
H
(1 ),
H
(1 )), the incumbent sells
to (1 ) customers. This implies the following claim.
Claim E4. The incumbent will never choose a rst period price in (
H
(1),
H
(1)).
Claim E5. By choosing p
I
below but close to
H
(1 ), the incumbent can guarantee
itself discounted prots arbitrarily close to (1 )
H
(1 + ).
Proof. It sells to at least (1 ) hsc consumers at price (1 )
H
in period 1 and at
price
H
in period 2.
Claim E6. b
I
H
(1 ).
Proof. The incumbent makes strictly positive prots. This implies that p
E
is not strictly
smaller than p
I
H
with probability 1. However, if b
I
>
H
(1 ) an entrant could
charge a price in (
H
, b
I
+
H
) and obtain strictly positive expected prots. In the
states of nature where it is not the lowest priced entrant, it would attract no consumers
and make a prot equal to 0. When it is the lowest price entrant, which would happen
with strictly positive probability by Claim E1, it would undercut the other entrants and
also undercut the incumbent by more than
H
; its discounted prot would be strictly
positive, which establishes the contradiction.
E.2. Equilibrium in the two period Bertrand model for small
L
In subsection E.1, we have not used any restrictions on ratio of switching costs,
L
/
H
.
We now restrict the analysis of the cases where equation (4) (
L
<
H
/(1 + )) holds,
which will enable us to prove Proposition 3.
Claim E7. If equation (4) holds, then at equilibrium b
I
>
H
(1 ).
Proof. From (4) and Claim E6, we have b
I
<
L
(1 ). By Claim E1, this implies
that the incumbent never sells to any lsc consumers. By Claim E2 if the incumbent
chooses p
I
>
H
(1 ), at least hsc consumers buy from a period 1 entrant. Thus,
the highest prot the incumbent could make while selling to all hsc consumers in
period 1 is
H
(1 ) +
H
=
H
. Using Claim E5, the incumbent can improve its
prot by charging a price larger than
H
(1 ), since equation (4) is equivalent to
H
< (1 )
H
(1 + ).
Claims E4, E6 and E7 imply b
I
=
H
(1 ) if
L
/
H
< /(1 + ). By Claim E2
this implies that the discounted prot of the incumbent is bounded above by
(1 )
H
(1 +
H
) = (1 )(b
I
+
H
).
By Claim E5, this quantity is also an lower bound on the prot, and this proves
Proposition 3.
33
Not for publication
The material presented in the following pages will be included in the web appendix of
the paper. It is composed of three sections. In the rst, we derive the equilibrium in
the two period model for large
L
, that is
L
that satises
L
/(1 + ). Next, we
provide formal denitions for the innite horizon model, including descriptions of the
game and of strategies of the players. Finally, we formally dene the constraints that we
impose on the equilibria.
For compactness, we refer to the uniform model when discussing the model where all
consumers have the same switching cost.
F. Equilibrium in the two period Bertrand model when
L
/
H
is greater than or equal to /(1 + )
In this section, we prove the results discussed in subsection 4.4. Note that we are less
ambitious than in E.2: we are only trying to identify one equilibrium for each value of
H
/
L
, not to characterize all the equilibria. We present the results under the form of
three claims, starting with the largest value of
L
/
H
.
Claim 8. If
L
/
H
> , then the two period Bertrand game has a unique equilibrium
in which the incumbent charges
L
(1 ) in period 1 and
L
in period 2. It sells to all
consumers and its prots are
L
.
As in the one period model, when
L
>
H
the incumbent and the entrant act as if
there were only lsc customers in the economy. We leave the proof of the claim to the
reader.
For
L
/
H
(x
C
, ), with x
C
dened by (9), we establish the following claim:
Claim 9. If
L
/
H
[x
C
, ], there exists a pure strategy equilibrium in which the
incumbent, whose prots are
H
, sells to the hsc consumers in both periods, at prices
respectively equal to
H
(1 ) and
H
. All lsc customers purchase from entrants at
price
L
in period 1 and at price
L
in period 2.
Proof. We show that the strategies described in the claim form an equilibrium. The lsc
customers are clearly better o switching in period 1. The strategy of the hsc customers
is a best response to the strategy of the other agents as they are indierent between
purchasing from the incumbent in both periods and switching to an entrant in the rst
period in both cases their total discounted costs are equal to
H
.
This indierence of hsc consumers implies that the incumbent would loose at least a
proportion of its customers if it increased its period 1 price. It is straightforward to
see that, under these circumstances, its most protable increase in price is to
H
L
.
This deviation is unprotable as long as
H
(1 )(
H
L
+
H
) = [
H
L
+
H
]
_
H
L
(
H
L
)
_
L
/
H
[1 + + ]
_
+ (
L
/
H
)
2
,
34
and therefore, by (9), holds on [x
C
, ]. A small decrease in period 1 price obviously
decreases the prots of the incumbent. A decrease to
L
(1 ) allows it to sell to all
consumers in period 1, but decreases its prots.
Finally, it is easy to show that the entrants strategy is indeed a best response to the
strategies of the other agents.
Claim 10. The equilibrium described in 4.3.2 is also an equilibrium when
L
/
H
[/(1 + ), x
C
].
Proof. The proof is exactly the same as in 4.3.2, except that we need to be a bit
more careful when showing that the incumbent does not gain by deviating to p
I
in
(
L
(1 ),
H
(1 )]. The incumbent sells to all the hsc customers if
p
I
+
H
< p
E
+
H
+
L
p
I
< p
E
+ (1 )
H
+
L
,
This implies that if p
I
[(1 )
H
+
L
, (1 )
H
), the incumbent sells to all the
hsc consumers with probability strictly between 0 and 1 in the other states of nature,
it sells to a proportion 1 of them. From (7), one can show that this implies that the
prots of the incumbent are increasing in this range, and that, because <
L
/
H
, they
are always smaller than the putative equilibrium prot, which establishes the claim.
G. The innite horizon games: denitions
G.1. The players
The players are:
The incumbent I;
Potential entrants E = {(i, t)}
iN
,tN
, where N
E.
Consumers are represented by a number from c [0, 1]. In the uniform switching
cost model, all consumers have a switching cost > 0. In the Stackelberg and
Bertrand models, consumers c [0, ] have switching cost , while consumers
c (, 1] have a switching cost equal to 0.
G.2. First period moves
Let us denote by =
E
1
oer their price in ,
35
while in the Stackelberg model in period 1, the incumbent {I} chooses its prices rst,
then all entrants in E
1
oer their price in . It will be notationally convenient to assume
that the other rms (i.e., the potential entrants in subsequent periods) also make an
oer, which is restricted to +. The set of active rms, those who make an oer is
F
1
=
_
f F | p
1
f
= +
_
. Of course, F
1
{I}
E
1
. If F
1
= , the game ends.
Consumers The consumers choose to purchase from one rm or another. We will
designate by
1
(c) the rm from which consumer c buys in period 1, and call
1
(f)
the set of consumers who buy from f in period 1:
1
(f) = {c | f =
1
(c)}.
G.3. First period payos
Firms The period 1 prot of rms which do not belong to F
1
is equal to 0.
If
1
(f) is measurable, let Q
1
f
= (
1
(f)) be the total quantity sold by rm f F
1
.
The period 1 prot of rm f,
1
(f) is equal to p
1
f
Q
1
f
. If
1
(f) is not measurable,
the prot of rm f is 0, and the game ends.
Consumers If F
1
= , the disutility of consumer c is p
1
1
(c)
if
1
(c) = I or if the consumer
is a lsc consumer (c > in the Stackelberg and Bertrand models), while it is
p
1
1
(c)
+ if
1
(c) = I and the consumer is a hsc consumer (c in the Stackelberg
and Bertrand models and all consumers in the uniform model). If F
1
= or if for
some rm f
1
(f) is not measurable, the payo of all consumers is, by harmless
abuse of language, +.
G.4. Moves for periods t > 1
We assume that in period t 1, a non-empty set F
t1
of active rms; functions
t1
(c)
and
t1
(f) which describe the allocation of customers to rms in the previous period.
The total sales of the rms and such quantities are also dened as in the case of period 1.
Firms We assume that if Q
t1
f
= 0 for some rm f active in period t 1, then it
drops out of the game. So that the set of potential rms in period t is
F
t
=
_
f F
t1
| Q
t1
f
> 0
_
E
t
.
The moves that the rms can make are for f
F
t
; the others are restricted to
playing +. The active rms are the rms which make an oer; they belong to the
set F
t
=
_
f | p
t
f
= +
_
. In the Stackelberg model, in every period, incumbent
rms choose prices rst and, after observing these prices, entrants choose their
prices. If F
t
= , the game ends.
Consumers If F
t
= consumer c must choose a rm
t
(c) from which to buy from. The
set
t
(f) is the set of consumers such that
t
(c) = f.
36
G.5. Period payos for t > 1
Firms The prot of a rm which does not belong to F
t
is equal to 0; for the rms in F
t
the prot is p
t
f
Q
t
f
where Q
t
f
= (
t
(f)), when
t
(f) is measurable and 0 otherwise.
Consumers If F
t
= , the disutility of consumer c is p
t
t
(c)
if
t
(c) =
t1
(c) or if the
consumer is an lsc consumer (c > in the Stackelberg and Bertrand models) and
p
t
t
(c)
+ otherwise. If F
t
= or if for some rm f
t
(f) is not measurable, the
payo of all consumers is + and the game stops.
G.6. Intertemporal Payos
For both rms and consumers the intertemporal payos is computed by adding each
period payo weighted by the discount factor (0, 1) over t N (except that in all
cases where the game ends in nite time, the disutility of the consumers is equal to +).
G.7. Strategies
Histories are well dened. Dene h
0
= {} and for t 1, h
t
is constructed in the usual
way by the concatenation of h
t1
and the moves in period t.
The strategies of rms in the Stackelberg Model For the Stackelberg model, a
strategy for the incumbent rm in period 1,
1
I
, is a period 1 move, with the constraints
described above. In period 1, a strategy for an entrant f is a function from the incumbents
rst period price, p
I
, to a price p
1
f
, with the constraints described in G.2. A strategy
in period t > 1 for an incumbent, is a function from h
t1
into moves at period t with
the constraints described above, while a strategy for period t entrants is a function from
h
t1
and the prices set by the incumbents in period t, p
t
I
. A strategy
f
is a strategy for
every period
_
t
f
_
tN
.
Firm Strategies for Bertrand and Uniform Models As we will see, in the Bertrand
model there exist no pure strategy stationary equilibrium which satises our assumptions.
Therefore, we need to expand our denitions to mixed strategies for rms.
For all rms f, a strategy in period 1,
1
f
is a probability distribution over period 1
prices, with the constraints described above. In period t > 1, a strategy is a function
from h
t1
into a distribution over prices at period t with the constraints described above.
A strategy
f
is a strategy for every period
_
t
f
_
tN
.
Strategies for Consumers For consumers, a strategy
t
c
in period t is a function from
the concatenation of h
t1
and the moves of the rms in that period t into moves in
period t, with the constraints imposed above. A strategy
c
is a strategy for every period
{
t
c
}
tN
.(Because we will assume that consumers coordinate to purchase from the same
rm, we need to assume that they have a pure strategy.)
The outcomes are dened as usual from the strategies, and the payos associated with
strategies are the discounted payos of the outcomes.
37