This article was downloaded by: [139.80.123.

44] On: 26 July 2015, At: 00:01

Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA

Management Science
Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org

Selling and Leasing Strategies for Durable Goods with

Complementary Products
Sreekumar R. Bhaskaran, Stephen M. Gilbert,

To cite this article:

Sreekumar R. Bhaskaran, Stephen M. Gilbert, (2005) Selling and Leasing Strategies for Durable Goods with Complementary
Products. Management Science 51(8):1278-1290. http://dx.doi.org/10.1287/mnsc.1050.0421

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.

© 2005 INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
 MANAGEMENT SCIENCE informs ®

Vol. 51, No. 8, August 2005, pp. 1278–1290 doi 10.1287/mnsc.1050.0421

issn 0025-1909
eissn 1526-5501
05
5108
1278 © 2005 INFORMS

Selling and Leasing Strategies for Durable

Goods with Complementary Products
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.

Sreekumar R. Bhaskaran, Stephen M. Gilbert

Management Department, University of Texas at Austin, Texas 78712
{sreekumar.bhaskaran-nair@phd.mccombs.utexas.edu, steve.gilbert@mccombs.utexas.edu}

I t has been recognized that when a durable goods manufacturer sells its output, it has an incentive to produce
at a rate that will drive down the market price of the product over time. Because anticipation of declining
prices makes consumers less willing to invest in owning the durable good, selling can be self-defeating for
the manufacturer. If the manufacturer instead leases the product, it can eliminate its own incentive to decrease
the price over time, which allows it to extract larger rents from consumers. In this paper, we investigate how
a durable goods manufacturer’s choice between leasing and selling is affected by a complementary product
that is produced by an independent firm. We show that a durable goods manufacturer that leases its product
has an incentive to increase prices (by limiting the availability of the product) in response to the availability
of a complement. Because this potential for opportunistic behavior discourages output of the complement,
leasing can also be problematic. As a result, the durable goods manufacturer faces a trade-off between leasing,
which commits the manufacturer to not overproduce, and selling, which commits it to not underproduce. Our
contribution is to identify this trade-off and show how a durable goods manufacturer can use a combination of
leasing and selling to balance its strategic commitment across both its own market as well as the complementary
market.
Key words: complementary markets; selling versus leasing; leasing with an option to buy
History: Accepted by Jagmohan S. Raju, marketing; received October 27, 2003. This paper was with the
authors 7 months for 3 revisions.

1. Introduction the overall cost of operating the vehicle. Fuel cell

Many durable or semidurable products are heav-
ily influenced by the availability of complementary
goods or services that can increase the value of
the durable product. In many instances the lack
of sufficient availability of complementary products
can impede the success of a durable product. Com-
plementary relationships frequently arise in con-
sumer electronics. For example, the demand for high-
definition television (HDTV) sets is highly dependent
upon the availability of high-definition programming.
In fact, although the scarcity of such programming
has taken much of the blame for the disappointing
rate of adoption of HDTV, it is widely anticipated that
the recent launch of ESPN’s high-definition channel
will accelerate the rate of adoption of HDTV (Business
Week 2003). Another example of a complementary
relationship in consumer electronics is that between
video game stations and video games.
In the automotive industry, complementarities
affect the adoption of alternative fuel technologies
such as hybrid and fuel cell vehicles, both of which
depend on the availability of complements. For
hybrids, batteries need to be replaced every couple of
years, and greater availability of batteries decreases
1278
vehicles have an even stronger complementary
dependence on the availability of hydrogen fuel
because they cannot be operated without near-daily
access to fuel. One reason fuel cell cars have not
been widely sold is the lack of widespread availability
of hydrogen fuel. Energy companies and automotive
firms are trapped in a chicken-and-egg problem, in
which each industry could benefit from increases in
the other’s output (Wall Street Journal 2003), but nei-
ther has an incentive to unilaterally increase its own
production.
It has long been recognized that one of the strate-
gic issues that a durable goods manufacturer (DGM)
faces is that of mitigating the potential for its own
opportunistic behavior with respect to consumers.
This issue arises because, when a durable goods
monopolist sells a product, it has an incentive to skim
the demand curve by producing at a rate that, over
time, drives down prices and decreases the market
value of owning the product. In anticipation of this
opportunistic behavior, consumers are less willing to
“invest” in ownership. This issue has been referred
to as time inconsistency in reference to the fact that a
monopolist’s ability to sell a durable good at a price
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
above marginal cost is inconsistent with its own incen-
tives to produce at a rate that causes the price of
the product to decrease. If the DGM could make
a credible commitment to its future output, then it
could avoid the problem created by time inconsis-
tency. Unfortunately, in practice it is often impossible
for a DGM that sells its output to make such a com-
mitment credible, since there is a strong incentive to
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
renege.
However, one well-known way for a DGM to elim-
inate the problem of time inconsistency is to lease the
product to the consumer instead of selling it. Leas-
ing gives a monopolistic DGM control over the mar-
ket for used goods, eliminating its own incentive to
produce at a rate that drives down prices. However,
as much as leasing eliminates the time inconsistency
problem, it can create another problem in situations
where consumers’ valuations for the product depend
on the availability of complementary products. Gen-
erally, when two products complement one another,
the market price of each product is decreasing in its
own quantity and increasing in the quantity of the
other. As a result, if the two products are produced
by two independent firms, there is an externality that
gives each firm an incentive to produce less than
would be necessary to achieve the first-best outcome.
Because leasing provides the DGM with more flexi-
bility to withhold the availability of a product, it can
discourage output of the complement and can be self-
defeating.
However, we demonstrate that, by selling at least
some portion of the output, the DGM can credi-
bly commit to a greater availability of the product,
thereby encouraging the output of a complement. In
addition, we show how the DGM can use a mixture of
selling and leasing to balance the need to manage the
time inconsistency issue against the need to encour-
age the output of a complementary product.
The remainder of the paper is organized as follows.
In §2, we review and analyze the relevant literature.
In §3, we develop a model to capture the interac-
tion between the manufacturer of a durable product
and a firm that produces a complementary product.
We analyze how the strategic choice between leas-
ing and selling can be used to the durable goods
firm’s advantage and extend our analysis to the case
where the complementary market opens subsequent
to the durable goods market. Finally in §4, we sum-
marize the managerial implications through a con-
ceptual framework and point to directions for future
research.
2. Literature Review
The study of durable goods has long been a central
subject in the industrial organization literature. It is
1279
well established that the durability of a product can
interfere with a monopolistic manufacturer’s extrac-
tion of rents from consumers. The reason for this is
that after a manufacturer sells its durable product to
some subset of the market, it has an incentive to con-
tinue production, selling its output at lower and lower
prices. Coase (1972) conjectures that if rational con-
sumers anticipate this behavior, then prices should
fall down to competitive levels. This issue has been
referred to in the literature as time inconsistency to
reflect the fact that for a durable product, the tra-
ditional monopolistic model is inconsistent with the
passage of time. This line of thought is formalized by
Bulow (1982) and Stokey (1981), who propose that, by
leasing its product, a durable goods manufacturer can
avoid the time inconsistency problem. Bulow (1982)
also suggests that while leasing helps the firm to
retain its monopoly power, selling also might have
socially undesirable effects.
Later research identifies conditions under which
the Coase conjecture does not apply, resulting in
prices that are above marginal costs. Notable among
these are the works by Conlisk et al. (1984), who
model constant inflow of new consumers; Bond and
Samuelson (1984), who incorporate replacement sales;
Kahn (1986), who studies increasing marginal produc-
tion costs; and Bagnoli et al. (1980), who use discrete
demand. In contrast to these papers, which consider a
single durable product in isolation, Kuhn and Padilla
(1996) show that when a monopolist manufacturer
of a durable product also sells nondurable substi-
tutes, the price of the durable product will be above
marginal cost.
Research on durable goods has been pursued in
two main directions. The first of these examines the
relationship between durability and a firm’s incentive
to innovate. Prominent among these are the works
by Levinthal and Purohit (1989) and Waldman (1993),
who argue that the incentive for a durable goods
manufacturer to make the existing product obsolete
by introducing a newer version is high and that this
can intensify the time inconsistency effect. In a related
work, Dhebar (1994) analyzes the case wherein a
monopolist supplies a series of durable products of
increasing quality to a heterogeneous customer base
and shows that intertemporal price discrimination
issues in such circumstances could prevent a producer
from credibly committing to future prices and quali-
ties. This can result in a situation where there is no
equilibrium strategy. Kornish (2001) points out that
an equilibrium strategy exists if the monopolist does
not offer upgrade pricing. Also, Subramaniam and
Srinivasan (1998) demonstrate how an introductory
product strategy can be used to signal the trajectory
of its cost curves to consumers. They show that firms
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
1280
can use a high introductory price to signal to con-
sumers that cost reduction through learning experi-
ences is low.
The other main direction in durable goods research
explores the interaction between a firm’s strategy for
leasing-selling and the durability of its product. How-
ever, this line of research tends to focus on compet-
itive rather than complementary interactions. Bulow
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
(1986) models the durability choice in an oligopoly
and finds that the lease-sales ratio is dependent on
the number of firms in the market. Bucovetsky and
Chilton (1986) demonstrate that selling can dominate
leasing when there is a threat of entry from compet-
itive firms or actual competition among firms. Com-
petition is also examined by Desai and Purohit (1999),
who show that in a duopoly, firms tend to prefer sell-
ing over leasing when their products are sufficiently
substitutable. They also demonstrate that a combina-
tion of leasing and selling might be the optimal strat-
egy. In a separate work, Desai and Purohit (1998) also
show that firms can employ leasing and selling as a
mechanism to differentiate among consumers.
Our work also draws on the literature on markets
that involve either direct or indirect network effects.
Katz and Shapiro (1994) categorize such markets and
identify the issues firms and consumers face while
dealing with these markets. Farrell and Saloner (1986)
investigate how the installed base for such products
interacts with a firm’s incentive to innovate and eval-
uate the welfare implications of certain strategies that
firms might adopt. Similar issues are also considered
by Katz and Shapiro (1985), Farrell and Saloner (1985),
and Choi (1994). In another related work, Conner
(1995) analyzes a market with direct network effects
and suggests how a manufacturer can derive external-
ity benefits from a lower-quality version of its prod-
uct that is produced by an imitator. Recent work by
Parker and Van Alstyne (2001a, b) and Rochet and
Tirole (2001) also examines how complementary inter-
actions between products could be used to stimu-
late demand and handle competitive influences from
other firms and thus provide a rationale for cross-
subsidization of products. However, most of this lit-
erature focuses on network externality issues and
ignores the effect of the durability of a product under
such situations.
Thus, while the effect of durability on the profit-
ability of firms has been studied before, little atten-
tion has been paid to the interactions between durable
goods and complementary products. This aspect is
important because the availability of such comple-
ments can stimulate demand for the durable good,
enhancing the manufacturer’s profits. The contribu-
tion of this article is to identify the trade-off that
exists when durable goods interact with complemen-
tary products and to show how a durable goods man-
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
ufacturer can manage this trade-off with its strategy
for leasing and/or selling its product.
3. The Model
Consider two complementary products, one of which
is a durable good (DG). For simplicity, we assume
that the other is a nondurable product or service and
refer to it as the complementary product (CP). We
assume that these two products are produced by dif-
ferent firms, which we refer to as the DGM and the
complementary product provider (CPP). To represent
durability, we adopt a variation of the two-period lin-
ear demand model that Bulow (1982) proposed for
durable goods. In this model, it is assumed that the
DG lasts for exactly two periods. Although there is
no depreciation of its value between periods 1 and 2,
a discount factor of
is applied to revenues or cash
flows received in period 2. Assuming that the durable
good lasts for two periods is not critical; it is only
important to assume that it lasts for a finite amount
of time, so that optimal decisions can be calculated in
a recursive fashion. The assumption that there is no
depreciation simplifies the presentation of our anal-
ysis. Although our model can easily be generalized
to allow for depreciation, this blurs the distinction
between durable and nondurable goods. Note that the
assumption of no depreciation is consistent with that
of Bulow (1982) and Bucovetsky and Chilton (1986).
As in the Bulow (1982) model, we assume that con-
sumers’ utility for the DG is defined by the value of
the service it provides. In each period, there are M
potential consumers, each of whom has valuation vd
for each period of service from the DG and can con-
sume at most one unit of it. In the absence of the CP,
the valuation vd for the service of the DG has a dis-
tribution that is uniform over this population on the
interval ad − M ad , where 0 ≤ ad ≤ M. Note that if
ad < M, then some consumers have negative valua-
tions for the DG. As we will soon see, this ensures
that there are some consumers who derive a positive
utility from the DG only when it is used in combi-
nation with the complement. Note that a product can
yield a negative utility if the costs of shopping for
and obtaining it outweigh the benefits of ownership.
Although our results do not necessarily require neg-
ative valuations, we do require that, at equilibrium,
some consumers will not pay for access to DG; that is,
the market is not covered. As long as the market is not
covered, a change in the price of the DG will affect
the number of consumers who use it.
We also require that either the same set of M con-
sumers be present in both periods, or that the DG
can be bought and sold among consumers with no
transaction costs; that is, there is a perfect second-hand
market. As discussed in Bulow (1982), the existence
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
of a perfect second-hand market allows for the pos-
sibility that a different set of M consumers may be
doing the demanding in each period. Bulow provides
a compelling illustration of this possibility with an
example about baby strollers. Although the number
of consumers demanding baby strollers is stable, the
specific set of consumers who demand a baby stroller
this year is not composed of the same individuals
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
who demanded baby strollers last year. The existence
of a second-hand market ensures that, in period 2, all
of the units that are available to the market, including
those sold in period 1, are allocated to the consumers
with the highest valuations.
If either the same set of consumers are present in
both periods or there is a perfect second-hand mar-
ket, then the price that a consumer is willing to pay
for ownership of the DG in the first period will be
equal to the sum of his valuation vd and the antici-
pated second period market price, discounted by
. It
follows that if consumers anticipate that a DGM will
have an incentive to produce at a rate that decreases
the market price for a product in the second period,
they will be less willing to pay for ownership of the
product in the first period.
For the CP, we assume that each consumer pur-
chases a scalar amount and that each consumer’s
marginal valuation for additional units of the CP is
linearly decreasing. In addition, we allow for a con-
sumer’s marginal utility for the complement to be
related to the valuation vd for the use of the DG.
Specifically, we assume that if a consumer does not
have the use of the DG, his or her marginal util-
ity for the yth unit of the CP can be expressed as
ac + vd − y/ where ac ,  ≥ 0 and ∈ 0 2 are
constants. Note that if > 0, then consumers with the
highest valuations for the use of the DG will have the
highest marginal utilities for the complement. Alter-
natively, if = 0, then all consumers are homoge-
neous with respect to their marginal utilities for the
complement. Many of our results can be extended
to cover cases in which < 0. However, this would
imply that consumers with the highest valuations for
the use of the DG have the lowest marginal utilities
for the CP. Although such scenarios may be possible,
they are likely to be rare.
Thus far, we have described a model of consumers’
independent valuations for the DG and the CP, that is,
how they value each product in the absence of the
other. However, to capture the complementary nature
of the two products, we must also define how a con-
sumer’s use of one product affects his or her util-
ity for the other. Complementary products are those
for which a consumer’s utility from using both of
them together is greater than the sum of the utilities
that he or she would have received from using each
product separately. For example, a consumer who has
1281
the use of an HDTV can obtain some utility from it
by watching non-high-definition (HD) programming.
This same consumer would derive some utility for
HD programming, even if he or she had to view it
on a non-HD television. However, the total value that
this consumer derives from having both an HDTV
and HD programming is greater than the sum of the
independent valuations that he or she would have
for these products separately. To represent this com-
plementary effect, let us introduce k as a parameter
of complementarity. We propose the following utility
function for a consumer whose independent valua-
tion for the DG is vd when the per unit prices for the
DG and the CP are pd and pc respectively:


ac + k + vd − x
y
U y  =  vd − pd  + dx − ypc 
0 
(3.1)
where y is the amount of the complement consumed,
 = 1 if the consumer has the use of the DG, and  = 0
otherwise. From (3.1), it can be shown that at price pc ,
a utility maximizing individual consumer with valu-
ation vd for the DG would consume
yi pc  vd   = ac + k + vd − pc (3.2)
units of the CP. Thus, at a given price pc , access to the
use of the DG increases the amount of the CP that the
consumer consumes by k units.
To determine how the CP affects a consumer’s will-
ingness to pay for the DG, we must consider his or
her total utility as a function of the price of the CP. If
the price of the CP is pc and the consumer’s indepen-
dent valuation for the DG is vd , then the use of the
DG increases his or her total utility by the following
amount:
U y pc  vd  1 1 − U y pc  vd  0 0
ac + k + vd − pc 2 a + vd − pc 2
= vd − pd + − c
2 2
k2 + 2k ac + vd − pc 
= vd − pd +  (3.3)
2
Note that the latter term in this expression repre-
sents the amount by which the availability of the CP
increases a consumer’s willingness to pay for the DG.
Thus, if the price charged for use of the DG for a
single period is pd , then all consumers with an inde-
pendent valuation of more than pd − k2 + 2k ac +
pd − pc / 2 will pay for access to the DG, and
the total number of consumers who pay to use the
service of the DG will be
  +
k2 + 2k ac + pd − pc 
Q = Min M ad + − pd 
2
(3.4)

1282
To facilitate the analysis, we will introduce several
restrictions on our parameters:
2 M − ad  − k2
ac ≤
2k
ac ≥ k + 2M  + k  3M − 2ad  + k2
3
2
+ 2kad  · 2 k + 2M + 2kM  
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
The first of these restrictions is sufficient to guarantee
that at equilibrium, Q ≤ M; that is, some consumers
will not have the use of the DG. This plays a major
role in our results because it implies that a decrease
in the price of the DG leads to more consumers hav-
ing access to it, which increases demand for the CP.
Note that because we have assumed that ac ≥ 0,
Assumption (3.5) also implies that k2 ≤ 2 M − ad .
The restriction shown in (3.6) implies that, at equi-
librium, all consumers purchase a positive amount
of the CP, even those who lack access to the DG.
Although this restriction simplifies the mathematical
analysis, our results do not depend on it qualitatively.
Thus, even though (3.6) precludes the case in which
ac = 0, by sacrificing some clarity of exposition, our
results can be extended to include this case as well.
Note that the assumption that ≥ 0 implies that the
right-hand side of (3.5) is larger than the right-hand
side of (3.6).
Under the assumptions above, Q ≤ M, and in (3.4)
the term k2 + 2k ac + pd − pc / 2 represents the
number of additional consumers who purchase the
DG at price pd as a result of the availability of the CP
at price pc . Moreover, when Q ≤ M, there exists vd∗ ∈
ad − M ad  such that all consumers with valuation
vd ≥ vd∗ will acquire access to the durable product,
where
−2ac k − k2 + 2kpc  + 2pd 
vd∗ =  nondecreasing.
2 k + 
Because each consumer’s consumption of the CP
depends on whether he or she has the use of the DG,
the total consumption of the CP at price pc can be
expressed as follows:

ad
vd∗
yT pc  pd  = y pc  v 1 dv +
vd∗ ad −M
M2
= M ac − pc  −
2 tinue to use their product or sell it at the market price.
2
k k +2kac −2 kpc +pd 
+ +ad k+M 
2  +k 
(3.7)
Note that in the case where = 0, each consumer’s
consumption of the CP depends only on pc and
whether he or she has access to the DG. Each con-
sumer who does not have access to the DG consumes
ac −pc units of the CP, while each one who has access
Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
consumes ac + k − pc units of the CP. As a result,
when = 0, the total consumption of the CP would be
M ac − pc  + kQ. From (3.4) and (3.7), we can derive
the following inverse demand functions for the DG
(3.5)
and the CP, respectively:
5M − 2ad 
k2 + Mk  M − 2Q + 2M ad − Q + 2ky
−1
(3.6) pd Q y =
2M
(3.8)
2Mac + 2kQ − 2y + M 2ad − M
pc Q y =  (3.9)
2M
where pd Q y is the implicit single-period rental
price for the DG when Q units of the DG and y units
of the CP are available. The inverse demand func-
tion for each of these products is decreasing in its
own quantity and increasing in the quantity of the
other product. This represents an externality so that,
when the two firms act simultaneously and indepen-
dently, they will set quantities too low (prices too
high) to achieve the first-best outcome. This external-
ity could be partially mitigated if one of the two firms
could move first to provide a credible commitment
to its output quantity. However, it is often difficult to
identify a mechanism for making such a commitment
credible. We will argue that selling can provide the
DGM with such a mechanism.
In all of our analysis, we assume that the marginal
costs of production for both the DGM and the CPP
are constant, and for ease of exposition we normalize
these costs to zero. Although the assumption of linear
production cost is simplistic, it is not unreasonable
in a wide variety of settings. Moreover, the funda-
mental interaction between the selling versus leasing
strategy of the DGM and the complementary product
requires only that the marginal production costs be
3.1. The Sales and Leasing Strategy
We can now begin to explore how the profits of both
firms are influenced by the DGM’s decision regarding
the extent to which it will lease or sell its product.
As shown in Figure 1, consumers have three options
y pc  v 0 dv with respect to the DG in period 1: They can pur-
chase, lease, or go without the use of it. In period 2,
consumers who purchased in period 1 can either con-
Consumers who leased or did without the DG in
period 1 can either buy it or do without it in period 2.
Note that in period 2, leasing and buying are indis-
tinguishable, since either one provides one period of
use of the DG. In each period, each consumer also
determines the amount of the CP that he or she will
consume.
Let Q1 be the total quantity that the DGM dis-
tributes in the first period, and let  ∈ 0 1 be the
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
Figure 1 Consumer Decision Tree
Lease/buy
Lease
Do not use
DG
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
Lease/buy
Do not use
DG
Do not use
DG
Sell & don’t
use DG
Buy
Hold
fraction of this quantity that it leases. Let Q2 be the
quantity that the DGM introduces to the market in
period 2, and let Q2T be the total number of units of the
DG that are in use in period 2. Let y1 and y2 denote
the CPP’s first and second period output quantities.
To proceed with the analysis for the DGM’s strat-
egy for leasing or selling, we must first establish the
role that is played by the second-hand market. Recall
that we have assumed that there is a perfect second-
hand market that allows the DG to be bought and
sold without transaction costs. This has two important
implications. First, it implies that all of the 1 − Q1
units that the DGM sells in period 1 are potentially
available in the second-hand market in period 2. As a
result, these units compete with any additional Q2 
units that the DGM sells/leases in period 2, and the
total quantity of the DG that will be available to the
market in period 2 will be Q2T = Q2 + 1 − Q1 . Note
that, although Q2 can include units that were leased
in period 1, the DGM has no obligation to make the
leased units from period 1 available to the market in
period 2. However, it can be verified that, for any frac-
tion  ∈ 0 1 of leasing in the first period, if the DGM
leases a conditionally optimal quantity Q1∗  in the
first period, then it will never be in the DGM’s interest
to withhold these units from the market in period 2.
The other important implication of the perfect
second-hand market is that it provides a mechanism
for guaranteeing that the Q2T = Q2 + 1 − Q1 units
of the DG that are available in period 2 will be allo-
cated to the consumers with the highest valuations
for it. Otherwise, there would be opportunities for
Pareto-improving trade between low-valuation and
high-valuation consumers. Thus, the market prices for
the DG and the CP in period 2 will be pd Q2T  y2  and
pd Q2T  y2 , respectively.
1283
We can now proceed with the standard backward
induction approach to determine the DGM’s optimal
selling/leasing strategy. In the second (final) period,
the profits of the CPP can be expressed as
2c Q1   Q2  y
 
= ypc Q2T  y
y    
= 2 ac M + ad M + kQ2T − y − M 2 (3.10)
2M
and the DGM’s profits can be expressed as
 
2d Q1   Q2  y = Q2 pd Q2T  y
Q2
= 2 ad M + ky + kM k + M 
2M
−2Q2T k2 +M +kM  (3.11)
As is evident from these expressions, each of the
two firms’ second-period profits are increasing in the
amount of the other firm’s product that is available.
Taking first-order conditions, we obtain the following
second-period response functions for the DGM and
the CPP, respectively:
Q2 Q1   y

1 M 2ad − M + 2ky
= M+ − 2 1 − Q1 (3.12)
4 k2 + M + kM
y2 Q1   Q2 
1
= 2a M + 2ad − MM + 2k 1 − Q1 + Q2 
4 c
(3.13)
From these response functions, we can confirm that
each firm’s optimal output is increasing in the quan-
tity of the other firm’s available product. These ex-
pressions provide insight about why the DGM might
want to sell its product in period 1: By doing so,
the DGM can make a credible commitment about the
amount of product that will be available in the second
period. From (3.12), it can be seen that the amount
of the DG that is released in period 2 Q2 Q1   y
is decreasing in the amount that is sold in period 1,
1 − Q1 . However, the total amount of the DG that
will be in use in the second period, Q2 Q1   y +
1 − Q1 , is increasing in the amount sold in period 1.
If the DGM could somehow make a credible commit-
ment to make a given amount of product available in
period 2 without selling it in the first period, then it
would no doubt benefit from doing so to encourage
a higher level of output of the CP. However, making
such a commitment credible would be a major obsta-
cle because the DGM would have a strong incentive
to renege.
It follows that the strategic motivation for selling
is that it allows the DGM to demonstrate a higher

1284
propensity to make its product available in the future,
thereby encouraging output of the complement. As a
consequence, the question of selling versus leasing is
most interesting when the DGM has no other mech-
anism for making such a credible commitment. Our
purpose is to recognize the strategic benefit of selling
as a means of encouraging output of a complemen-
tary product rather than to compare different mech-
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
anisms for making quantity commitments. Therefore,
we assume that within each of the two periods, nei-
ther the DGM nor the CPP is able to credibly commit
to the amount of product that it will produce in that
period.
By simultaneously solving (3.12) and (3.13), we
have that the second-period quantities will be the fol-
lowing function of the amount of the durable good
that was available in the first period and the fraction
of these units that was leased:
y2∗ Q1   = f M ad  ac  k  
+ 2kQ1 1 −  k3 + kM + M 
· 6k2 + 8M + 8kM −1
Q2∗ Q1   = g M ad  ac  k  
−2Q1 1 −  k2 + 2M + 2kM 
· 6k2 + 8M + 8kM −1 
where
f M ad  ac  k  
= kM k2 + 2ad  − M k2 M − 4ad 
+ 2M M − 2ad  − 2kM 2 2
M − 2ad 
2 additional value that a consumer will derive from
+ 4ac M k + M + kM 
g M ad  ac  k  
= M 2k ac + k + 4ad + k 2ad + M
To fully understand the impact of the DGM’s first-
period strategy on the firms’ profits, we need to look
at the effect of the first-period decisions on the second-
period quantities of both firms. The first-period deci-
sions that the DGM faces are: (1) total quantity to be
made available in the first period and (2) fraction of
this quantity to be leased.
Proposition 3.1. The second-period output of the
DGM and of the CPP are increasing and decreasing respec-
tively in the fraction  ∈ 0 1 of leasing employed in the
first period. Moreover, as complementarity k increases,
a given increase in  (more leasing) causes smaller in-
creases in the second-period output of the DGM and larger
reductions in the output of the CPP.
The DGM’s flexibility to control the second-period
market price of its product is directly related to the
Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
extent to which it employs leasing. By leasing its
products, the DGM retains the flexibility to increase
price by restricting the availability of the product in
period 2, thereby appropriating a larger share of the
surplus due to complementarity. However, because
the CPP will reduce its output in anticipation of this
opportunistic behavior, leasing can be self-defeating
for the DGM. Alternatively, by shifting to a greater
proportion of selling in the first period and effectively
relinquishing its flexibility to restrict the availability
of product and increase prices, the DGM can encour-
age a greater supply of the complementary product.
From (3.14) and (3.15), it is evident that the second-
period output quantities of both firms are influenced
by the amount of DG that is sold in the first period.
Assuming that both firms behave rationally and have
access to the same information, their second-period
profits can be expressed as functions of Q1 and . Let
us introduce the following definitions:
∗
2d Q1   = 2d Q1   Q2∗ Q1   y2∗ Q1  
∗
2c Q1   = 2c Q1   Q2∗ Q1   y2∗ Q1  
(3.14)
Let us now consider how the first-period prices are
affected by the DGM’s decisions about selling or leas-
ing the product. Because consumers value only the
service of the DG, the lease price in the first period
(3.15)
depends only on the total quantity of the DG and
the CP that are currently available. Thus, using (3.8)
and (3.9), the first-period lease price for the DG is
equal to pd Q1  y1 , while the first-period price for the
complement is equal to pc Q1  y1 .
To determine the price that the DGM can obtain
from any units that it sells, we must consider the
purchasing the product instead of leasing it. Under
the assumption that consumers can buy or sell the
DG through a perfect second-hand market, the addi-
tional value that a consumer derives from purchas-
ing instead of leasing is the discounted market price
for the product in period 2. This market price,
pd Q2∗ Q1   y2∗ Q1  , is independent of an individ-
ual consumer’s valuation and can be fully anticipated
by rational consumers. Therefore, in spite of having
different valuations for the DG, consumers will be
homogeneous with respect to the premium that they
will pay to buy instead of lease the product. That is,
all consumers whose valuations are sufficiently high
that they use the DG in period 1 will be indiffer-
ent between leasing and buying.1 Thus, the price for
1
Because the assumption of a perfect second-hand market ensures
that the Q2T units of the DG that are available in period 2 are
allocated to the consumers with the highest valuations and that
the market price will be pd Q2T  y2 , our model is indifferent with
respect to which of the individual first-period consumers lease ver-
sus buy the DG.
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS 1285

which the DGM can sell 1 − Q1 units in period 1,
given that it leases Q1 units and that consumers
anticipate profit maximizing behavior from both the
DGM and the CPP, is
s
p1d Q1   y1  = pd Q1  y1  +
pd Q2∗ Q1   y2∗ Q1  
(3.16)
where
is the discount factor applied to the antici-
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
at the bottom of this spectrum consume only the CP.
Thus, the appropriate interpretation of this result is
that as the number of consumers who consume only
the CP increases, the CPP’s pricing decision is less
heavily influenced by the availability of the DG. This
reduces the DGM’s incentive to use selling as a mech-
anism for committing to a greater availability of its
product.

pated second-period market price for the DG. Finally, to understand why the proportion of leas-
It follows that, at the beginning of period 1, the ing is increasing in , recall that represents the
profit functions for the two firms can be expressed as: strength of the relationship between a consumer’s
valuation for the DG and his or her marginal util-
s
d Q1   y1  = Q1 pd Q1 y1 + 1−Q1 p1d Q1 y1  ity for the CP. When = 0, all consumers have the
∗
+
2d Q1   (3.17) same marginal utility for the complement, so each
additional consumer who gains access to the DG will
∗
c Q1   y1  = y1 pc Q1  y1  +
2c Q1   (3.18) consume the same amount of the complement. In
contrast, when > 0, consumers with larger valua-
Proposition 3.2. (a) The DGM’s total profits are max- tions for the DG also have larger marginal utility for
imized when the CP. In this case, each additional consumer who
M 2k ac + ad + k + 4ad  + kM  gains access to the DG has less marginal utility for the
Q1∗ = complement. This tends to reduce the strategic incen-
6k2 + 8M + 8kM
tive for the DGM to commit to a greater availability
and the fraction of leasing that maximizes the DGM’s prof- of the product by selling.
its is The equilibrium prices and quantities for the DG
and the CP can be seen in Table 1. From the results
∗ = k4 + 8M k2  + M 2 + k3 + 2kM + k2 M 2
 in this table, we can establish the following:
· 4 k4 + 3k2 M + 2M 2  2 + 3k3 M Corollary 3.1. When the DGM follows her optimal
+ 4kM 2  + 2k2 M 2 2
−1  strategy (a) both the price and quantity of the complement
are higher in the second period than in the first, and (b) the
(b) 0 < ∗ ≤ 1, and ∗ = 1 if and only if k = 0. total availability of the DG is higher and implicit lease price
(c) ∗ is decreasing in k and increasing in M, , is lower in the second period than in the first.
and .
From the above proposition, pure leasing is opti- Table 1 Equilibrium Prices and Quantities
mal if and only if the there is no interaction between Decisions Equilibrium values
the two products k = 0. As the complementarity k DG
between the products increases, it becomes more AC
First-period lease price pd Q∗1  ∗ 
important for the DGM to commit to a larger avail- 2 2A + B
C
ability of product in period 2, and it does so by selling Second-period price pd Q∗2 Q∗1  ∗  y2∗ Q∗1  ∗ 
8

a greater portion of its output. First-period sale price s
p1d
C 4A
+
8 2A + B
However, the DGM’s incentive to sell decreases 
1 2 4

CM
First-period lease qty. ∗ Q∗1 − +
when there are increases in consumers’ sensitivity  A B A + 2B 8
2
k CM
to the price of the complement, the number M of First-period sale qty. 1 − ∗ Q∗1
8AB
consumers, or the tendency  for consumers who First-period total qty. Q∗1
MC
have high valuation for the DG to also have higher 22A + B
MC
marginal utilities for the CP. Higher sensitivity to the Second-period qty. Q∗2
8A

price of the complement tends to weaken the strength Optimal lease-sale ratio ∗ 1−
1 1
+
2
4 A B
of the complementary interactions, which causes the CPP
DGM to behave more like it would in the absence First-period qty. y1
MkD + 2AE
22A + B
of the complement, that is, by leasing more. The rea- Mk 3 C + kD + 2AE
Second-period qty. y2
son that an increase in the total number of consumers 8B2A + B
discourages selling is that, for a fixed value of ad , an First-period price pc Q∗1  y1∗ 
kD + 2AE
2 2A + B
increase in M would add consumers at the low end Ck 3 + 4BkD + 2AE
Second-period price pc QT2  y2∗ 
of the spectrum of valuations for the DG. Recall that 8 B2A + B

vd has a distribution that is uniform over the interval Note. A = k 2 + M + kM, B = k 2 + 2M + 2kM. C = 2kac + k +
ad − M ad  and that, at equilibrium, the consumers 4ad + k2ad + M. D = k 2 + 2ad + kM, E = 2ad  + 2ac − M.
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
1286 Management Science 51(8), pp. 1278–1290, © 2005 INFORMS

In the second period, time inconsistency pushes In this case, 0 < ∗ ≤ 1, and pure leasing, i.e., ∗ = 1 is
the DGM to increase the availability of the prod- optimal if and only if k = 0.
uct. In traditional models, this increased availability (b) If
serves only to drive down the implicit lease price 
of the DG. However, in the presence of a comple- 2ad − M 4a M 2 M
ac > k + d 3 + 5ad −
ment, the increased availability of the DG also stimu- M k 2
lates demand for the CP. The CPP responds to this by 8ad M 4ad  + M 2

increasing both its output and its price. Thus, the CPP + + 
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.

k2 k
is able to reap benefits of both higher margins and
larger output as a result of the strategic commitment then pure selling, i.e., ∗ = 0, is optimal and
that the DGM makes by selling some of its output in
ad
the first period. Q1∗ > 
2
3.2. Is Pure Selling Ever Optimal? Recall that when the CP is available in both peri-
In the previous section, we showed that, for our ods, pure selling is never optimal. However, the above
model, pure leasing is optimal when complementarity result shows that when the availability of the CP
vanishes, that is, k = 0. However, although the pro- is delayed, pure selling can be an optimal strategy
portion of leasing is decreasing in k and increasing in for the DGM. The intuition for why a delay in the
M, , and , it is always strictly positive. This leads to
availability of the CP tends to increase the propor-
the question of whether complementary effects could
tion of selling can be explained as follows: Regardless
ever be sufficiently strong that it would be optimal
of when the CP becomes available, selling stimulates
for a DGM to sell all of its output.
only the second-period output of the CP. When the
In seeking to answer this question, it can be ob-
availability of the CP is delayed, second-period profits
served that, in practice, CPPs often wait for some
represent a larger portion of the DGM’s total profits.
time after the introduction of a DG before introduc-
As a result, it becomes more important for the DGM
ing a product. For example, Electronic Arts and other
to provide a credible commitment to the quantity of
game producers waited three to four years after the
its product that will be available. Note that when the
introduction of Microsoft’s Xbox and Sony’s PlaySta-
marginal utility for the CP is very high (part b of
tion 2 before flooding the market with software re-
Proposition 3.3), pure selling alone does not provide a
leases (Business Week 2004). In addition, Gupta et al.
sufficiently strong level of commitment and the DGM
(1999) performed an empirical study that showed
increases its first-period output beyond the myopi-
how limited availability of programming restricted
cally optimal amount, that is, ad /2.
the demand of the HDTVs. To investigate how pro-
It is also worth noting that, although the optimal
duction lead times for complementary products affect
the DGM’s strategy, we consider a variation on our fraction of leasing in our original model, as shown in
original model and assume that the CP is not avail- Proposition 3.2, is independent of ad and ac , this is not
able until the second period, that is, y1 = 0 We charac- the case when the availability of the CP is delayed.
terize the optimal decisions of both firms in the same In this case, ∗ is increasing in ad and decreasing
way as in the previous sections. in ac . As consumers’ valuations for the DG increase,
it becomes increasingly important for the DGM to
Proposition 3.3. (a) If address its time inconsistency problem rather than to
 encourage greater output of the CP.
2ad − M 4a M 2 M
ac ≤ k + d 3 + 5ad −
M k 2 Corollary 3.2. The threshold level of complementar-
2 ity, kth , above which pure selling  = 0 becomes optimal,
8ad M 4ad  + M 
+ +  decreases with " = ac /ad , that is, #kth /#" < 0.
k2 k
the profits of the DGM are maximized with the following This is illustrated in Figure 2, where it can be seen
combined quantity of leasing and selling in period 1: that a larger value of " = ac /ad leads to a higher
threshold of complementarity k, above which pure
ad selling dominates.
Q1∗ = (3.19)
2
and the fraction of leasing is 3.3. Leasing with Option to Buy
In many durable goods markets, such as automobiles,
∗ = 2ad 2k4 + 4M 2  2 + 5k3 M + 8kM 2  musical instruments, etc., it is common for firms to
offer consumers leasing contracts that include options
+ 4k2 M  + M 2
 − k3 M 2 ac + k + M 
to purchase the product at the termination of the
· 4ad k2 + M + kM  k2 + 2M + 2kM −1  lease. GM’s Smart Buy, Chrysler’s Gold Key Plus, and
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS 1287

Figure 2 Optimal Fraction of Leasing When Availability of CP Is Delayed

0.4

0.3
Optimal fraction of leasing α
β = 1.0
0.2
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.

Hybrid selling
β = 1.5
and leasing
0.1
strategy

25 30 35 40 45 50
Pure selling
–0.1 Degree of complementarity k
M = 80
ρ = 1.0
–0.2 φ = 0.6

–0.3

Ford’s Red Carpet Option are examples of some such
contracts.
Because such leases are common, it is worth com-
menting on how such leasing arrangements relate to
our results. Recall that in our model, selling allows
the DGM to alter its second-period response function
and effectively commit itself to having more units of
the product in use than it would have if it relied
entirely on leasing. Consider now a DGM that leases
all of its output in period 1 but is able to precom-
mit to the price at which consumers can exercise an
option to buy at the end of the period. By specifying
an appropriate price at which consumers can exer-
cise that option, such a DGM can similarly alter its
second-period response function as a commitment to
have more units of the product in use. Under this type
of arrangement, it can be shown that at equilibrium
the exercise price specified by the DGM will be equal
to the market price for that product in period 2. It is
From a managerial standpoint, we can represent the
optimal strategy of the DGM as function of the ratio
" = ac /ad of the individual valuation parameters and
the level of complementarity k. This can be depicted
as shown in Figure 3. When a DGM anticipates a
large independent marginal utility for a strongly com-
plementary product, selling should represent a large
proportion of the total distribution of the product,
and if the availability of the complement is delayed,
even a pure selling strategy may be optimal. Con-
versely, when complementary effects are weak, due
to either a small individual marginal utility for the
CP or to a low level of interaction between the two
products, the firm should shift toward leasing. In all
Figure 3 Conceptual Leasing/Selling Strategy
Pu str
High

also possible to show that by leasing all of the output

re ate
Relative valuation of complementary market β

se gy

in period 1 and also specifying an appropriate option-

lli
If

ng
co

to-buy price, a DGM can induce the same availability

m
pl is d
em e

of the complement and achieve the same profitability

en laye
ta d

as is possible under a mixture of leasing and selling.

va
ila
bi
lit
y

4. Concluding Remarks
H
yb
rid str

In this paper we examine the effect that a comple-

di teg
st y

mentary product has on a DGM’s strategy to lease or

a
rib
ut

sell products. We show that in the case that the com-

io
n
g

plementary product is produced by another firm and

in
as

the extent of complementarity is sufficiently strong,

le
ss

the DGM’s preference for leasing will shift to selling.

Le

In the absence of complete contracts, selling can act

Low

as a commitment mechanism that can mitigate oppor-

tunistic behavior and encourage a larger supply of Low High

the CP. Degree of complementarity k

 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
1288
the cases we have considered, selling acts a Pareto-
improving mechanism by allowing higher profits for
both the CPP and the DGM. The firms are better off
because they are able to generate higher profits, and
the consumers are better off because greater quantities
of both products are available.
Although our analysis is based on the assumption
that a single firm produces the CP, most of our results
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.
extend qualitatively to the case in which the comple-
mentary market is open to entry. Moreover, in a sep-
arate analysis, we have confirmed that, as the cost
of entry to the complementary market decreases, the
DGM will tend to increase the proportion of selling
in its strategy.
One interesting situation in which the results of our
model could be applied is the one that arises when
demand for at least one of the two products would
not exist without the availability of the other. For
example, demand for video games that are compati-
ble with the Sony PlayStation would not exist with-
out the availability of the PlayStation and vice versa.
Such products that depend entirely on the availability
of another product have been described as contingent
(see Peterson and Mahajan 1978). To formally address
such products in our model would require setting
either ad = 0 or ac = 0, the latter of which is precluded
by Assumption (3.6). However, it is worth noting that
the purpose of this assumption is only to streamline
the mathematical analysis. Without Assumption (3.6),
it is possible that some consumers would consume
0 units of the CP, and yT pc  pd , as shown in (3.7),
would be quadratic in pc . Although this would make
the analysis much more messy, it would not affect the
qualitative insights.
In an attempt to validate our result that a DGM
should employ less leasing for products that have
strong complementary interactions, we contacted a
number of automobile dealerships that carry both
hybrid and traditional combustion engine vehicles.
We argue that hybrids have a moderately strong com-
plementary dependence on the availability of batter-
ies, which need to be replaced periodically. Note that
although both hybrid and traditional vehicles depend
on the availability of gasoline, this is a relatively weak
complementarity, since the output of a single manu-
facturer of a gasoline-consuming product has a neg-
ligible effect on the availability of gasoline. (In our
model, this would be represented by increasing the
parameter M to represent a situation in which the
number of consumers who have positive independent
valuation of a particular DG are only a small fraction
of the population that has some utility for the CP.) Out
of the five dealers that we contacted, all reported that
leasing represents between 50% and 67% less of their
business in hybrids than in traditional vehicles. This is
very consistent with our results, especially in view of
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS
the fact that the hybrid market is currently much less
competitive than the traditional vehicle market, which
would normally suggest that there would be more
leasing in the hybrid market (see Desai and Purohit
1999).
One can argue that the complementary dependence
between fuel cell vehicles and hydrogen fuel is even
stronger than that between hybrid vehicles and bat-
teries. Thus, our results would suggest that if early
entrants in the fuel cell vehicle market are dependent
on third parties to provide hydrogen fuel, they should
sell their products to encourage greater fuel availabil-
ity. Of course, an alternative strategy would be for
the manufacturer of such a vehicle to also provide the
fuel supply, eliminating the complementary hold-up
problem altogether.
It is also of interest to compare our results to the
recent developments in HDTV programming. While
leasing was initially the dominant model for dis-
tribution of HDTV set-tops used by cable service
providers, the last few years have seen an increasing
tendency for these firms to sell their HDTV set-tops
(Dow Jones Newswires 2003). Subsequently, there has
been a significant increase in the number of broad-
casters that are planning to air HDTV programming.
Since 2001, Fox, HBO, and ESPN have announced that
they would be entering this medium of communica-
tion in a big way (Business Week 2003, The New York
Times 2003). The results of our model are consistent
with these observations and may explain this inter-
play between the strategy for leasing or selling HDTV
set-tops and the decision of the broadcast firms to
enter the HDTV programming market.
This article makes three main contributions. First,
we identify the trade-offs faced by a DGM whose
product interacts with complements. Second, we have
developed a simple micromodel of consumer utility
for a durable good and a complementary product.
Finally, we have used this model to demonstrate how
a DGM can mitigate the adverse effects of its own
potential to behave opportunistically with respect to
both consumers and complementors.
One direction for future research would be to
consider other mechanisms that could induce greater
levels of investment from complementary firms. Com-
mitting to prices and quantities to mitigate hold-up
is prevalent when firms are related vertically. How-
ever, these mechanisms need explicit contracts; hence,
enforceability could be an issue when firms are not
directly related. Another direction in which our work
could be extended would be to relax the assumption
that both firms and consumers have perfect informa-
tion about cost structure and market characteristics.
This is certainly not true under all circumstances, and
it would be of interest to understand how a firm’s
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
Management Science 51(8), pp. 1278–1290, © 2005 INFORMS 1289

private information would affect its strategy for leas- (c) From part b it is clear that ∗ is decreasing in k. Dif-
ing and selling its products. An empirical analysis of ferentiating ∗ w.r.t. M, , and , we have that
the model of complementarity and durability as pre- #∗
sented in this paper would also serve to improve our = k2  + k  5k4 + 8M 2  2 + 12k3 M + 16kM 2 
#M
understanding of such markets. + 4k2 M 3 + 2M 2 

Acknowledgments · 4 k2 + M + kM 2 k2 + 2M + 2kM 2 −1 > 0


The authors are grateful to their colleagues, Vijay Mahajan #∗ k2 M 1
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.

and Robert Peterson, for their comments on a previous ver- =

# 4 k2 + M + kM 2
sion of this paper.
4
Appendix + >0
k2 + 2M + 2kM 2
Proof of Proposition 3.1. By differentiating (3.14) and 
(3.15) with respect to , it is easy to confirm that Q2 Q1   #∗ k3 M 1
=
is increasing in  and y Qa   is decreasing in . For the # 4 k2 + M + kM 2
second part of the proposition, note that 4
+ > 0

# 2 y  Q1  k2 + 2M + 2kM 2
= −Q1 3k4 + 4M 2  2 + 8kM k2 + M
#k#
+ k2 M 9 + 4M 2  Proof of Corollary 3.1. Please refer to Table 1 for
the equilibrium prices and quantities of both the products.
·  3k2 + 4M + 4kM 2 −1 < 0 Direct comparison gives us the above result.

# 2 Q2  Q1  2kMQ1 2 + k  Proof of Proposition 3.3. (a) As in the proof of Propo-
=− < 0 sition 3.2, we use the transformation of variables in which
#k# 3k2 + 4M + 4kM 2
Qs = 1 − Q1 and Ql = Q1 . Thus, when the complement
As the complementarity k increases, the impact of leas- is not available in period 1, the DGM’s profit function is
ing  on the second-period quantity of the DGM decreases dTr Qs  Ql  0, where dTr Qs  Ql  y1  is as shown in (A.1).
and the impact on CPP quantity increases.
As discussed in Proposition 3.2, the Hessian matrix for (A.1)
Proof of Proposition 3.2. (a) To identify the DGM’s is negative definite. The first-order conditions are satisfied
optimal leasing and selling strategy in period 1, it is help- at the following point:
ful to introduce a transformation of variables. Let Qs =
1 − Q1 be the number of units that the DGM sells and Ql∗ = 2ad 2k4 + 4M 2  2 + 5k3 M + 8kM 2 
Ql = Q1 be the number of units that it leases. Substituting
+ 4k2 M  + M 2
 − k3 M 2 ac + k + M 
into (3.17) and (3.18), we can derive the following trans-
formed profit functions of the DGM and CPP, which are as · 8 k2 + M + kM  k2 + 2M + 2kM −1 (A.3)
follows:
k2 M 2k ac + k + 4ad  + k 2ad + M
dTr Qs  Ql  y1  Qs∗ = 
8 k2 + M + kM  k2 + 2M + 2kM 

Ql
s
= Ql pd Qs + Ql  y1  + Qs p1d Qs + Ql  y First note that Qs∗ = 0 if and only if k = 0. For any k > 0,
Qs + Ql 1 both Qs∗ and Ql∗ are positive as long as

∗ Ql
+ 2d Qs + Ql  (A.1) 2ad 2k4 + 4M 2  2 + 5k3 M + 8kM 2  + 4k2 M  + M 2

Qs + Ql
3
> k M 2 ac + k + M 
cTr Qs  Ql  y1 
 which can be shown to be identical to the condition in the
∗ Ql
= y1 pc Qs + Ql  y1  + 2c Qs + Ql   (A.2) proposition. By the definition of ∗ = Ql∗ / Ql∗ + Qs∗ , it fol-
Qs + Ql
lows that 0 < ∗ < 1. Using the definition of Q1∗ = Ql∗ + Qs∗ ,
It can be verified that the Hessian matrix for (A.1) is nega- and some algebra, it can also be confirmed that Q1∗ = ad /2.
tive definite and the second derivative of (A.2) is negative,
(b) From (A.3), it can be seen that the requirement that
so first-order conditions are sufficient to determine a Nash
Ql ≥ 0 becomes binding (⇒Ql∗ = 0) when
equilibrium. Simultaneously solving FOCs w.r.t. to both the
DGM and CPP profit functions, we can find the optimal 2ad 2k4 + 4M 2  2 + 5k3 M + 8kM 2  + 4k2 M  + M 2

strategy of the DGM in terms of Ql and Qs . The result fol- 3
lows from transforming this solution back to the original ≤ k M 2 ac + k + M 
variables Q1 and .
which is identical to the condition in the proposition. So
(b) It is obvious that ∗ > 0. To see that ∗ ≤ 1, we need
when this condition is satisfied, pure selling is optimal.

only observe that ∗ is decreasing in k:
Proof of Corollary 3.2. From Proposition 3.3, we can
#∗ see that the threshold level of complementarity, kth , above
= −kM 2 + k  5k4 + 8M 2  2 + 12k3 M
#k which pure selling is optimal is given by the implicit
+ 16kM 2  + 4k2 M 3 + 2M 2
 function:
−1
· 4 k + M + kM  k + 2M + 2kM 2
2 2
<0 2ad 2k4 + 4M 2  2 + 5k3 M + 8kM 2 
and ∗ = 1, when k = 0. + 4k2 M  + M 2
 − k3 M 2 ac + k + M  = 0
 Bhaskaran and Gilbert: Selling and Leasing Strategies for Durable Goods
1290 Management Science 51(8), pp. 1278–1290, © 2005 INFORMS

By implicit differentiation, we can find the relationship Conlisk, J., E. Gerstner, J. Sobel. 1984. Cyclic pricing by a durable
between kth and " (relative size of the two markets) goods monopolist. Quart. J. Econom. 99 489–505.
Conner, K. R. 1995. Obtaining strategic advantage from being imi-
#kth tated: When can encouraging clones pay? Management Sci. 41(2)
= −2ad k3 M · k2 M 6"ad + 8k + 3M 
#" 209–225.
− 2ad 8k3 + 15k2 M + 8M 2  Desai, P., D. Purohit. 1998. Leasing and selling: Optimal market-
2 ing strategies for a durable goods firm. Management Sci. 44(11)
+ 8kM  + M −1 
S19–S34.
Desai, P., D. Purohit. 1999. Competition in durable goods markets:
Downloaded from informs.org by [139.80.123.44] on 26 July 2015, at 00:01 . For personal use only, all rights reserved.

It is easy to note that #kth /#" < 0 if the denominator is pos-

itive, which would occur if The strategic consequences of leasing and selling. Marketing
Sci. 18(1) 42–58.
ac
"= > 2ad 8k3 + 15k2 M + 8M 2  + 8kM  + M 2  Dhebar, A. 1994. Durable-goods monopolists, rational consumers,
ad and improving products. Marketing Sci. 13(1) 100–120.
− 8k + 3M k2 M · 6k2 Mad −1  (A.4) Farrell, J., G. Saloner. 1985. Standardization, compatibility and inno-
vation. RAND J. Econom. 16 70–83.
Since " = ac /ad , the above inequality is equivalent to
Farrell, J., G. Saloner. 1986. Installed base and compatibility: Inno-
ac > 2ad 8k3 + 15k2 M + 8M 2  + 8kM  + M 2
 vation, product preannouncements, and predation. The Amer.
Econom. Rev. 76(5) 940–955.
− 8k + 3M k2 M · 6k2 M−1  (A.5) Gupta, S., D. C. Jain, M. S. Sawhney. 1999. Modeling the evolution
of markets with indirect network externalities: An application
However, from Proposition 3.3, the threshold kth above to digital television. Marketing Sci. 18(3) 396–416.
which pure selling is optimal is the value of k that satisfies
Kahn, C. 1986. The durable goods monopolist and consistency with
 increasing costs. Econometrica 54(2) 275–294.
2ad − M 4a M 2 M
ac = k + d 3 + 5ad − Katz, M. L., C. Shapiro. 1985. Network externality, competition and
M k 2
compatibility. The Amer. Econom. Rev. 75(3) 424–440.
2
8ad M 4ad  + M  Katz, M. L., C. Shapiro. 1994. Systems competition and network
+ +  (A.6)
k2 k effects. J. Econom. Perspectives 8(2) 93–115.
It can be shown that the right-hand side of (A.5) is less than Kornish, L. J. 2001. Pricing of a durable goods monopolist
the right-hand side of (A.6) if under rapid sequential innovation. Management Sci. 47(11)
1552–1561.
k4 M + ad 12M 2  2 − 2k4 + 16kM 2  + 4k2 M  + M 2
 Kuhn, K.-U., A. J. Padilla. 1996. Product line decisions and the coase
> 0 conjecture. RAND J. Econom. 27(2) 391–414.
3k3 M
Levinthal, D. A., D. Purohit. 1989. Durable goods and product obso-
Since 2M > k2 according to Assumption (3.5), the above lescence. Marketing Sci. 8(1) 35–56.
condition is always true, and this implies that (A.4) is sat-
Lowry, T. 2003. HDTV: Coming into focus? Business Week (April 7)
isfied and that #kth /#" < 0.
58.
Parker, G. G., M. Van Alstyne. 2001a. Information complements,
substitutes and strategic product design. Mimeo, University of
References Michigan, Ann Arbor, MI.
Bagnoli, S., S. Salant, J. Swierzbinski. 1980. Durable goods monopoly Parker, G. G., M. Van Alstyne. 2001b. Unbundling in the presecnce
with discrete demand. J. Political Econom. 97 1459–1478. of network externalities. Mimeo: University of Michigan, Ann
Ball, J. 2003. Fuel-cell car concept raises a tricky question: Who will Arbor, MI.
pay for it? Wall Street Journal (March 7) A1. Peterson, R. A., V. Mahajan. 1978. Multi-product growth models.
Bond, E., L. Samuelson. 1984. Durable goods monopolies with ratio- Jagdish Sheth, ed. Research Marketing, Vol. 1. JAI Press, Green-
nal expectations and replacement sales. RAND J. Econom. 17(3) wich, CT, 201–232.
336–345.
Rochet, J. C., J. Tirole. 2001. Platform competition in two-sided mar-
Bucovetsky, S., J. Chilton. 1986. Concurrent renting and selling in kets. J. Eur. Econom. Assoc. 1(4) 990–1029.
a durable goods monopoly under threat of entry. RAND J.
Econom. 17 261–278. Sheng, E. 2003. Coming soon, cable companies to hawk services at
retail stores. Dow Jones Newswires.
Bulow, J. 1982. Durable goods monopolists. J. Political Econom. 90(2)
314–332. Stokey, N. 1981. Rational expectations and durable goods pricing.
Bulow, J. 1986. An economic theory of planned obsolescence. Quart. Bell J. Econom. 12 112–128.
J. Econom. 51 729–750. Subramaniam, B., K. Srinivasan. 1998. Modifying customer expec-
Chambers, E. 2004. Video gaming’s hardware hangup. Business tations of price decreases for a durable product. Management
Week Online (September 8) www.businessweek.com. Sci. 44(6) 776–786.
Choi, J. P. 1994. Network externality, compatibility choice and Taub, E. 2003. Fox television to add HDTV to primetime by Fall
planned obsolescence. J. Indust. Econom. 42(2) 167–182. 2004. New York Times (June 25) C5.
Coase, R. 1972. Durability and monopoly. J. Law Econom. 15 Waldman, M. 1993. A new perspective on planned obsolescence.
143–149. Quart. J. Econom. 108(1) 273–283.