Quality differentiation and entry choice

between online and offline markets

Yijuan Chen
Australian National University

Xiangting Hu
Renmin University of China

Sanxi Li
Renmin University of China

ANU Working Papers in Economics and Econometrics

# 620

November 2014 JEL: L13

ISBN: 0 86831 620 2

 Quality dierentiation and entry choice between
online and o- ine markets
Yijuan Cheny, Xiangting Huz, Sanxi Lix

November 2, 2014

Abstract

We study a model where an entrant chooses between online and o- ine markets to compete
with an o- ine-market incumbent. When consumers buy a product from the online market,
they cannot inspect the products quality prior to purchase. Conventional wisdom and some
literature suggest that this feature drives low-quality products to hide themselves in the online
market. However, the literature on vertical product dierentiation indicates that a rm may
prefer to reveal its product quality in the o- ine market, because quality dierentiation helps
alleviate price competition. We show that under fairly general conditions the entrant will
choose the o- ine market for not only the highest qualities but also the lowest ones, and
choose the online market for intermediate qualities. While the average quality of the online
good is lower than the incumbents quality, the actual quality of the online good may be
higher than that.

Keywords: online vs. o- ine competition; market choice by entrant; quality dierentia-
tion
JEL Classication: L13

1 Introduction
When consumers buy goods such as toys, clothing, and furniture from an online market, they
can hardly inspect a products quality prior to purchase. In such cases what is usually regarded
as a search good in an o- ine market is, at least partially, turned into an experience good in the
online market. Conventional wisdom and some literature suggest that the online market has a
"pooling eect," which attracts rms with low-quality products, because it provides those rms
with the opportunity to pool with higher-quality products. For example, Jin and Cato (2007)
show that ungraded sport cards sold in the online market are of lower quality than those in
the o- ine market.1 This is akin to a standard result in the literature of information disclosure
We are grateful for comments from Martin Richardson, and seminar participants at the La Trobe University
and Huazhong University of Science and Technology.
y
Australian National University; E-mail: yijuan.chen@anu.edu.au
z
Renmin University of China; E-mail: xiangthu@ruc.edu.cn
x
Renmin University of China; E-mail: lisanxi@gmail.com
1
For a review of recent literature on online versus o- ine competition, see Lieber and Syverson (2011).

1
(Grossman and Hart 1980, Jovanovic 1982), which shows that rms with higher qualities have a
stronger incentive to disclose their qualities, and if disclosure is costly, rms with lower qualities
will pool with each other by not disclosing.2
However, the literature on vertical product dierentiation (Shaked and Sutton 1982) indicates
another driving force that may also aect a rms choice of marketplace. The literature shows
that vertical product dierentiation may help alleviate price competition, and thus a rm may
voluntarily choose to provide a low-quality product if the competitor produces a high-quality
good. In the context of market choice, this means a low-quality rm may want to choose the
o- ine market if the competitors quality is high. In other words, one can argue that the o- ine
market has a "dierentiation eect" that may also attract low-quality rms.
Since the pooling eect and the dierentiation eect are in opposite directions, it is not clear
overall how they aect a rms market choice. In this paper, we investigate a situation where an
entrant chooses between online and o- ine markets to compete with an o- ine incumbent. We
develop a signalling-game model that draws upon the literature of quality disclosure and that
of vertical product dierentiation. We show that, due to the dierentiation eect, in the o- ine
market the entrants prot increases with the dierence between her products quality and the
incumbents quality. This drives the entrant with high-dierentiated product qualities, which
consist of not only the highest qualities but also the lowest ones, to choose the o- ine market.
On the other hand, the dierentiation eect diminishes as the entrants product quality be-
comes closer to the incumbents. We assume that the o- ine market is more costly than the
online market. Consequently, there is an interval of qualities such that at the two boundary
points the entrant will be indierent between the two markets. For medium-dierentiated quali-
ties, which are slightly above the lower bound or below the upper bound, the entrant will choose
the online market mainly to save the cost of the o- ine market. The pooling eect then drives
the low-dierentiated qualities, namely those near the incumbents quality, to pool with the
medium-dierentiated qualities in the online market.
The dierentiation eect implies that, all else equal, the o- ine market will be more attractive
to the entrant with higher qualities. As a result, the average quality of the online good is lower
than the quality of the incumbents. However, the upper bound of the aforementioned quality
interval is above the incumbents quality. Hence the actual quality of the online good may be
higher than the incumbents.
In this paper we choose a simple model that allows us to illustrate the pooling and dieren-
tiation eects in an easy way. The results can also serve as a benchmark for investigating more
features of online / o- ine business in the future. For example, quality of online products can
be fully or partially revealed by consumer review mechanisms and / or third-party web sites.
Thus in a more dynamic setting one can study if review mechanisms and third-party web sites
attenuate or amplify the pooling and dierentiation eects, and consequently how they aect
a rms market choice. Moreover, for tractability of the analysis we have assumed exogenous
product qualities, an assumption that remains to be relaxed in future studies.
We discuss the results in more detail in Section 2 after we present the model, and Section 3
2
See Dranove and Jin (2010) for a recent review of literature on quality disclosure.

2
concludes. The Appendix contains proofs of a lemma and a proposition.

2 Model
2.1 Set-up
There are two prot-maximizing rms i = A; B: Firm A is an entrant who chooses between an
o- ine market and an online market, and rm B is an incumbent who sells in the o- ine market.
Each rms product quality qi is an independent draw from a uniform distribution on [q; q] with
0 < q < q. A rms product quality is observable to the public in the o- ine market. But if
rm A sells online, then qA will be unobservable to consumers prior to consumption and to the
incumbent as well.3 After rm A chooses the market, the rms simultaneously post their prices
pi , which are publicly observable.
We assume that the xed cost of selling on the o- ine market, including rent and utility
expenses, is larger than the xed cost of selling online. For simplicity, we assume the xed cost
for the o- ine market is F and the xed cost for the online market is zero. The marginal cost of
production is constant and normalized to 0.
The rms compete for a continuum of utility-maximizing consumers, indexed by j. Each
consumer has a unitary demand and is characterized by a type j, which measures the consumers
marginal utility of quality and is an independent draw from a uniform distribution on [ ; ] with
0< < . Consumers are of mass M = . For ease of illustrating the dierentiation eect,
we adopt a standard specication of consumer utility as in Belleamme and Peitz (2010): A
consumers (indirect) utility from buying rm is product is u(qi ; pi ; j) =r+ j qi pi , where
r > 0 represents the basic willingness to pay for the product, and is assumed to be su ciently
large so that all consumers will buy from one rm in equilibrium.
Denote rm is prot by i. If a mass of m consumers buys from rm i, then i = mpi F
if the rm sells in the o- ine market, and i = mpi if it sells online.

2.2 Analysis
A strategy of the entrant (rm A) is to choose a market and then a price, and a strategy of
the incumbent (rm B) is to choose a price given the entrants market choice. A consumers
strategy is to decide on which rm to buy from given the prices and the observability of product
quality. Because the entrants product quality will be unobservable in the online market prior to
purchase, the game is essentially a dynamic game with incomplete information, and we use the
perfect Bayesian equilibrium (PBE), which consists of a sequentially rational strategy prole and
a consistent belief system, as our solution concept.
A PBE is separating if the entrant always chooses the o- ine market, as this will lead to
perfect revelation of its product quality. An equilibrium is pooling if the entrant always chooses
the online market, since the market choice will not reveal extra information about its quality. An
3
We model uncertainty in the online market as consumers uncertainty about a rms product quality. In
contrast, Loginova (2009) assumes that consumers are uncertain about their own types if they buy online. Therefore
quality signalling by the rms market choice is abstracted from her model.

3
equilibrium is partially pooling (partially separating) if for some qualities the entrant chooses the
o- ine market while for others it chooses the online market.
In equilibrium the entrant will choose a market if the prot from that market is higher than
that from the other. We will rst study the rms equilibrium behaviors in two situations: (i)
the entrant has chosen the o- ine market, and (ii) the entrant has chosen the online market.
Lemma 1 below shows the prots of the two rms if they compete in the o- ine market.

Lemma 1 Suppose both rms sell in the o- ine market. Then, given (qi ; q i ) with qi > q i ,
in equilibrium the rms prots are i = H (qi q i) F and i = L (qi q i) F , where
1 1
L = 9( 2 )2 and H = 9 (2
2
) .

Proof. In the appendix.

Lemma 1 is a standard result in the literature of vertical product dierentiation (Belleamme

and Peitz 2010): In the o- ine market, if the rms have the same quality, they will essentially be
engaged in a cut-throat Bertrand competition. But when their qualities dier, their equilibrium
prots both increase with the quality dierence (qi q i ). Hence, quality dierentiation helps
alleviate price competition and leads to higher prots of both rms, with the higher-quality rm
making more prot than the other.
Next, if rm A chooses the online market, then consumers have to base their purchase decision
on the expected quality of the rms product. Suppose there is a set QA [q; q] such that rm
A will choose the online market if qA 2 QA . Denote QA = E[qA jqA 2 QA ], which is the expected
quality of rm As product from consumers perspective. The next corollary characterizes the
rmsequilibrium prots if the entrant chooses the online market:

Corollary 1 If rm A chooses the online market, then in equilibrium A = H (QA qB ) and

B = L (QA qB ) F if QA > qB ; A = L (qB QA ) and B = H (qB QA ) F if QA < qB .

Proof. The proof essentially replicates that of Lemma 1 by replacing qi and q i with QA and
qB , and thus is skipped.

The proposition below characterizes a partial pooling equilibrium where the entrant chooses
the o- ine market if its product quality is su ciently high or su ciently low, while for intermediate
qualities it chooses the online market.

2 2
Proposition 1 Let q A = qB H+ L
H
L
F and q A = qB + H+ L
F . If q < q A and q A < q,
then there exists an equilibrium where rm A will choose the o- ine market if qA 2 [q; q A )[(q A ; q]
and the online market if qA 2 [q A ; q A ].

Proof. In the appendix.

4
 Figure 1: Entrants market choice and average quality of online good

The dierentiation eect and the pooling eect together give rise to the entrants market
choice in Proposition 1, the intuition of which has several layers. First, as Lemma 1 shows,
the dierentiation eect increases with the dierence in the rmsproduct qualities. Hence the
entrant with a quality close to the high-end q or the low-end q has a stronger incentive to choose
the o- ine market than that with a quality near the incumbents. For the highest qualities [q A ; q]
as well as the lowest ones [q; q A ], the entrant has no incentive to pool with the qualities in the
online market because if going online, it will be regarded as having an intermediate quality,
which, due to competition with the incumbent, will lead to a less protable price than in the
o- ine market. Consequently, as shown in Figure 1, the entrant will choose the o- ine market
despite its higher xed cost.
Secondly, the dierentiation eect diminishes as the entrants quality gets closer to the in-
cumbents. For the entrants quality that is slightly above q A or below q A , were the two markets
equally costly the entrant would still have made a higher prot from the o- ine market than going
online. However, the higher xed cost F of the o- ine market now drives the entrant to choose
the online market. Those medium-dierentiated qualities then become the target for qualities
around qB to pool with in the online market. It is worth noting that, in contrast to conventional
wisdom, in this case the higher qualities (those close to qB ) want to pool with the lower qualities
(those close to q A ), not the other way around. Moreover, qA > qB implies that the actual quality
of the online product may be higher than the the incumbents.4
Thirdly, Proposition 1 implies that in equilibrium the expected quality of the online product
is QA = qB H L
L( H + L)
F < qB , as shown in Figure 1. That is, on average the online product
has a lower quality than the o- ine product. To see the intuition of this result, note that because
00
0 and q such that q 0 = q 00 00
0 and q equally distant from q ),
H > L, for qA A B qA A qB , (i.e. qA A B
00 s prot from choosing the o- ine market is higher than that of the lower
the higher quality qA
0 . So the o- ine market is more attractive to q 00 than to q 0 . Hence, there are more
quality qA A A
qA < qB to choose the online market than qA > qB . Consequently, the expected quality of the
online product falls below the incumbents product in the o- ine market.
4
This is contrasted with the result in Board (2009), where two competing rms simultaneously choose to disclose
product qualities. He is focused on an equilibrium where the higher-quality rm always discloses. This corresponds
to a hypothesis in our paper that the incumbents product quality is always higher than the online product, which,
as Proposition 1 shows, may not be true.

5
 It is worth noting that in Proposition 1 q A and q A are independent of q and q. Hence, the
condition that q < q A and q A < q, under which the partial-pooling equilibrium exists, is fairly
general. Below is a numerical example where [q A ; q A ] is characterized given qB , , , and F:

Numerical Example: Suppose qB = 10, = 1, = 20, and F = 5. Then Proposition

1 implies that q A = 9:77, q A = 10:05, and thus QA = 9:91. When the entrant chooses the
online market, i.e., qA 2 [q A ; q A ], the prots of the entrant and the incumbent are A = 3:24 and
B = 10:23.

3 Conclusion
To summarize, by taking into account the eect of vertical product dierentiation on a rms
market choice, we show that an entrant with the lowest qualities may prefer the o- ine market. In
the online market, higher qualities may have the incentive to pool with lower qualities. Moreover,
the actual quality of the online product may be higher than the o- ine incumbents. On the other
hand, consistent with prior studies, we show that the average quality of the online product is
lower than the o- ine incumbents. These ndings can be a starting point for more comprehensive
studies in the future. For example, one may endogenize the entrants quality choice prior to its
market choice. It will also be interesting to see how other features of the online market, such
as market size and consumer review systems, aect a rms market choice. Moreover, one may
explore the possibility that the incumbent also has the incentive to enter the online market.

4 Appendix
4.1 Proof of Lemma 1
Proof. The proof is similar to that in Belleamme and Peitz (2010) for a standard model of
vertical product dierentiation in an o- ine market: When both rms choose to sell in the o- ine
market, (qB ; pB ) and (qA ; pA ) are common knowledge. Without loss of generality, given (qB ; pB )
and (qA ; pA ) with qB < qA and pB < pA , there will be a consumer with ^ who will be indierent
between the two rms:

^ qB pB = ^ q A p A ,
pA pB
)^ = .
qA qB

Hence, rm Bs prot is B = (^ )pB F = ( pqA

A
pB
qB )pB F , while rm As prot is
=( ^ )pB F = ( pA pB
A qA qB )pA F.
Given qB and qA , each rm chooses the price to maximize its prot. The rst-order condition
pA pB pB
for rm B implies qA qB qA qB = 0; which leads to

1
pB = [pA (qA qB )] :
2

6
 pA pB pA
The rst-order condition for rm A implies qA qB qA qB = 0, which implies

1
pA = (qA q B ) + pB :
2

Solving the two rst-order conditions for pB and pA , we have

1
pB = ( 2 )(qA qB );
3
1
pA = (2 ) (qA qB ) :
3

Thus given qB < qA , in equilibrium

pA p B
B (qB ; qA ) = ( )pB F
qA qB
1
(2 ) (qA qB ) 13 ( 2 )(qA qB ) 1
= (3 ) ( 2 )(qA qB ) F
qA qB 3
1
= ( 2 )2 (qA qB ) F
9

pA pB
A (qB ; qA ) = ( )pA F
qA qB
1 1
3 (2 ) (qA qB ) 3( 2 )(qA qB ) 1
= ( ) (2 ) (qA qB ) F
qA qB 3
1
= (2 )2 (qA qB ) F
9

4.2 Proof of Proposition 1

Proof. First suppose QA < qB . We consider two possible cases.
Case 1: qA 2 [q; qB ].
If rm A chooses the o- ine market, then by Lemma 1 A = L (qB qA ) F . If rm A
chooses the online market, then by Corollary 1 A = L (qB QA ). Firm A will choose the online
market if L (qB QA ) L (qB qA ) F , that is,

F
qA QA . (1)
L

Case 2: qA 2 (qB ; q].

If rm A chooses the o- ine market, then A = H (qA qB ) F . If rm A chooses the online
market, then A = L (qB QA ). Firm A will choose the online market when L (qB QA )
H (qA qB ) F , that is,
1
qA qB + [F + L (qB QA )]. (2)
H

h i
1 F 1
(1) and (2) imply that QA = 2 QA L
+ H
[F + L (qB QA )] . Solving this equation

7
for QA , we have
H L
QA = qB F < qB . (3)
L( H + L)

Note that QA < qB conrms our assumption.

F 1
Since rm A will choose to sell online when qA 2 [QA L
; qB + H
[F + L (qB QA )]], plugging
(3) into the lower bound and the upper bound of the set, we get

2 H
q A = qB F,
H + L L
2
q A = qB + F:
H + L

To complete the description of the PBE, we need to construct the consumersbelief system. On
the equilibrium path, the belief system is straightforward: By assumption, if the entrant chooses
the o- ine market, consumers can verify her quality. On the other hand, if the entrant chooses
the online market, consumers believe that her quality is QA . Then the characterization of q A and
q A implies that an o- ine entrant has no incentive to deviate to the online market, and vice versa.
A subtle issue remains in the o-equilibrium-path belief: Since the entrant chooses a marketplace
and then sets the price, a quality in the online market may want to post a dierent price in order
to separate itself from other online qualities. In particular, q A or q A has the strongest incentive to
deviate in this way, because the deviation will allow them to reveal their types while not paying
the o- ine-market cost F . To deter this kind of deviation, it su ces to let consumers hold the
o-equilibrium-path belief that any online price other than the one in equilibrium comes from an
entrant with quality QA , because then the most protable price-deviation in the online market
will lead to the same prot as staying on the equilibrium path.
Next, we will show that it cannot be QA > qB in equilibrium: Suppose, toward a contradiction,
QA > qB . There are two cases:
Case 1: qA 2 [q; qB ].
If rm A chooses the o- ine market, then A = L (qB qA ) F . If rm A chooses the
online market, then A = H (QA qB ). Firm A will choose the online market if H (QA qB )
1
L (qB qA ) F , that is, qA qB L
[F + H (QA qB )].
Case 2: qA 2 (qB ; q].
If rm A chooses the o- ine market, then A = H (qA qB ) F . If rm A chooses the
online market, then A = H (QA qB ). Firm A will choose the online market if H (QA qB )
F
H (qA qB )F , that is, qA QA + H .
h i
Hence, QA = 12 QA + FH + qB 1
L
[F + H (QA qB )] . Solving the equation for QA ,
we have QA = qB H
L H
L
F < qB , contradicting QA > qB . Therefore it cannot be QA > qB in
equilibrium.

References
[1] Belleamme, Paul and Martin Peitz, 2010, "Industrial Organization: Markets and Strategies",
Cambridge University Press.

8
[2] Board, Oliver., 2009, "Competition and Disclosure", Journal of Industrial Economics, Vol 57,
197-213.

[3] Dranove, D., and Jin, G.Z., 2010, Quality Disclosure and Certication: Theory and Practice.
Journal of Economic Literature, Vol 48, 935-963.

[4] Grossman, Sanford J., and Oliver D. Hart, 1980, "Disclosure Laws and Takeover Bids",
Journal of Finance, Vol 35(2): 323-34.

[5] Jin, Ginger Zhe and Kato, Andrew, 2007, Dividing Online and O- ine: A Case Study.,
Review of Economic Studies, Vol 74(3), 981-1004.

[6] Jovanovic, Boyan, 1982, "Truthful Disclosure of Information", Bell Journal of Economics, Vol
13 (1): 36-44.

[7] Lieber, Ethan and Syverson, Chad, 2011, "Online vs. O- ine Competition", Oxford Handbook
of the Digital Economy

[8] Loginova, Oksana, 2009, Real and Virtual Competition, Journal of Industrial Economics,
Vol 57 (2), 319-342.

[9] Shaked, Avner, and Sutton, John, 1982, "Relaxing Price Competition Through Product Dif-
ferentiation", Review of Economic Studies, Vol 49, 3-13.