没有合适的资源?快使用搜索试试~ 我知道了~
201206Dynamics of Consumer Demand for New Durable Goods.pdf
1.该资源内容由用户上传,如若侵权请联系客服进行举报
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
版权申诉
0 下载量 127 浏览量
2022-06-26
19:04:45
上传
评论
收藏 1.04MB PDF 举报
温馨提示
试读
48页
201206Dynamics of Consumer Demand for New Durable Goods.pdf
资源推荐
资源详情
资源评论
Dynamics of Consumer Demand for New Durable Goods
Author(s): Gautam Gowrisankaran and Marc Rysman
Source:
Journal of Political Economy
, Vol. 120, No. 6 (December 2012), pp. 1173-1219
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/10.1086/669540
Accessed: 27-06-2016 12:41 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
The University of Chicago Press
is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Political Economy
This content downloaded from 132.239.1.231 on Mon, 27 Jun 2016 12:41:16 UTC
All use subject to http://about.jstor.org/terms
Dynamics of Consumer Demand for New
Durable Goods
Gautam Gowrisankaran
University of Arizona, HEC Montre
´
al, and National Bureau of Economic Research
Marc Rysman
Boston University
Most new consumer durable goods experience rapid prices declines
and quality improvements, suggesting the importance of modeling
dynamics. This paper specifies a dynamic model of consumer pref-
erences for new durable goods with persistently heterogeneous con-
sumer tastes, rational expectations, and repeat purchases over time.
We estimate the model on the digital camcorder industry using panel
data on prices, sales, and characteristics. We find that the 1-year elas-
ticity in response to a transitory industrywide price shock is about
25 percent less than the 1-month elasticity. Standard cost-of-living in-
dices overstate welfare gain in later periods due to a changing com-
position of buyers.
If you don’t need a new set, don’t rush to buy one. Prices
will no doubt continue to drop over time, ½and you’ll have
more sets to choose from. ðConsumer Reports on 3D high-
definition televisions: http://www.consumerreports.org/cro
/electronics-computers/tvs-s ervices/tvs/tv-buying-advice
/tv-3d/tv-3d.htmÞ
We thank Dan Ackerberg, Victor Aguirregabiria, Ana Aizcorbe, Rabah Amir, Lanier
Benkard, Steve Berry, Sofronis Clerides, Tim Erickson, Simon Gilchrist, Avi Goldfarb, Igal
[ Journal of Political Economy, 2012, vol. 120, no. 6]
© 2012 by The University of Chicago. All rights reserved. 0022-3808/2012/12006-0005$10.00
1173
This content downloaded from 132.239.1.231 on Mon, 27 Jun 2016 12:41:16 UTC
All use subject to http://about.jstor.org/terms
I. Introduction
In many durable goods settings, the choice of when to buy is as impor-
tant to consumers as what to buy. Particularly for consumer electronics,
the quote from Consumer Reports conveys the conventional wisdom: prices
will fall and new choices, often higher-quality ones, will arrive. As a result,
many consumers purposely delay their purchase of these goods. When
they do purchase, consumers often have in mind that they will rep lace the
model with a superior model in the foreseeable future. Thus, dynamic
behavior is an important element of demand for consumer durable
goods. This paper specifies a structural dynamic model of consumer
preferences for new durable goods, estimates the model using data on
digital camcorders that are primarily at the level of the camcorder model,
and uses the model to evaluate elasticities and cost-of-living indices for
this market.
The digital camcorder industry is an important sector, with about
11 million units sold in the United States from 2000 to 2006. During
this time, the sector also experienced a huge evolution. Average digital
camcorder prices dropped from $930 to $380 while average pixel counts
rose from 580,000 to 1.08 million. The number of models available grew
from fewer than 30 to almost 100. Annual sales grew by 2.6 times in
4 years. The rapidly evolving nature of the characteristics and sales—
together with the fact that consumers are advised to consider the dy-
namic implications of their decisions—suggests that modeling dynamics
is empirically very important for estimating consumer preferences in
this industry. These issues have broad applicability: rapidly falling prices
and improving features have been among the most visible phenomena
in a large number of other new consumer durable goods markets, in-
cluding computers, digital video disc players, and high-definition televi-
sions.
In our model, dynamically optimizing consumers may choo se among
the set of available camcorders or wait. Camcorders are durable, so pur-
chase provides flow utility into the future. The available prices, quality,
and variety may improve over time, so waiting is valuable. In addition,
while consumers can hold only one camcorder at a time, they may sub-
stitute a new camcorder for an old one, so consumers continue to eval-
uate the market even after purchase. Our model allows for product
differentiation, endogeneity of prices, a changing number of models,
Hendel, Kei Hirano, Firat Inceoglu, Sam Kortum, John Krainer, Aviv Nevo, Ariel Pakes,
Minsoo Park, Rob Porter, Jeff Prince, John Rust, Pasquale Schiraldi, Andy Skrzypacz, Mo
Xiao, and seminar participants at several institutions for helpful comments; Mingli Chen,
Haizhen Lin, Ryan Murphy, Kathleen Nosal, David Rapson, Alex Shcherbakov, and Shirley
Wong for research assistance; and the NPD Group and ICR-CENTRIS for providing data.
The comments of the editor and anonymous referees substantially improved the paper. We
acknowledge funding from the National Science Foundation. All errors are our own.
1174 journal of political economy
This content downloaded from 132.239.1.231 on Mon, 27 Jun 2016 12:41:16 UTC
All use subject to http://about.jstor.org/terms
persistent consumer heterogeneity, and endogenous repeat purchases
over time. As our model is dynamic, we need to specify consumer per-
ceptions over future states of the world. To make the problem tractable,
we focus on a major simplifying assumption: that consumers expect
that the evolution of the value of purchase will follow a simple one-
dimensional Markov process. In this sense, consumers use a reduced-
form approximation of the supply-side evolution to make predictions
about the value of future purchases. We also examine a number of
alternative specifications for perceptions, including multidimensional
processes and perfect foresight.
The dynamics of our model build on the traditional vintage capital
models ðsee Solow et al. 1966Þ in that consumers in our model hold one
product at a time and endogenously reallocate to new models as the
technology for capital ði.e., camcordersÞ improves. Our framework dif-
fers from this literature in that we model a sunk cost of acquiring new
technology—namely the purchase price—which makes the consumer
purchase decision dynamic. In this way, it is similar to the Rust ð1987Þ
model of bus engines, in which the agent must decide when to replace
the engine.
While the Rust model concerns an industry with one homogeneous
product, the camcorder industry has almost 100 models at some time
periods, each with different prices and characteristics. To understand
purchase decisions in this industry ðand other consumer durable goods
industriesÞ, we need to model both the “when” to buy of the dynamic
literature and the “what” to buy. A different literature started by Bres-
nahan ð1981 Þ and Berry, Levinsohn, and Pakes ð1995Þ has modeled static
consumer decisions for differentiated products systems with many het-
erogeneous products. This literature emphasizes that incorporating
consumer heterogeneity into differentiated product demand systems is
important in obtaining realistic predictions. Our paper nests a Berr y et al.
style demand system within the dynamic replacement framework. By al-
lowing for persistent consumer heterogeneity, we relax the assumption
that choices are conditionally independent given the observed state, an
assumption that is typically required for dynamic estimation. The cost is
computational complexity. We develop a new estimation procedure that
draws on the techniques of Berry et al. for modeling consumer hetero-
geneity in a discrete-choice model and on Rust ð 1987 Þ for modeling
optimal stopping decisions. Our primary methodological advance is in
developing a feasible specification that allows us to combine these two
separate methods.
Over the last 15 years, a substantial literature has used static Berry
et al. style models to investigate questions of policy interest. This lit-
erature has analyzed questions that include ðbut are by no means lim-
ited toÞ horizontal merger policy ðsee Nevo 2000aÞ, trade policy ðsee
consumer demand for new durable goods 1175
This content downloaded from 132.239.1.231 on Mon, 27 Jun 2016 12:41:16 UTC
All use subject to http://about.jstor.org/terms
Berry, Levinsohn, and Pakes 1999Þ, and the value of new goods ðsee
Petrin 2002Þ. Many of these papers investigate industries, such as au-
tomobiles, for which goods are durable. To the extent that dynamics
are important for many industries, our paper may be useful in deriving
better estimates for these and related questions. Indeed, recent work is
using and extending our methods to examine the importance of soft-
ware in the video game industry ðLee 2012Þ, scrapping subsidies for
automobiles ðsuch as cash-for-clunkers programs; see Schiraldi 2011Þ,
markups for digi tal cameras ðZhao 2008Þ, and switching costs in con-
sumer banking, Medicare managed care, and subscription television
ðsee, respectively, Shcherbakov 2009; Ho 2011; Nosal 2012Þ, among other
research questions.
We estimate very different price and characteristic elasticities between
our dynamic model and similar static models. Moreover, the dynamic
results are more intuitive, for example, with negative price elasticities
and positive elasticities on important characteristics such as LCD screen
size. Thus, it is useful to consider how static estimates of durable goods
models might be biased. By modeling dynamics, we have essentially
transformed a consumption problem into a capital investment problem.
Firms will invest in a piece of capi tal if the flow of services from the
capital is greater than the rental cost of the capital. The rental cost of
capital is essentially the difference in the present-value price of capital
between this period and next period.
1
Thus, for the camco rder industry,
static models would predict a steady increase in sales as prices drop,
while dynamic models would predict that sales would increase the most
only when prices stop dropping. By incorrectly using price instead of
the difference in price, a static estimation applied to a durable good
purchase decision with falling prices will then result in mismeasure-
ment that may tend to bias the price coefficient toward zero. Our model
adds complexity in that we have heterogeneous agents. In particular, the
sales increase from price declines will be moderated over time since de-
mand endogenously falls as high-value consumers accumulate the good.
Not accounting for heterogeneity will further cause the static model
to understate the importance of price and quality, since the population
response to product improvement over time is smal ler than the aver-
age individual response, because of the changing population of avail-
able consumers. Finally, note that we observe both dynamic and cross-
sectional variation. It is possible for the dynamic model to generate
substitution patterns within time periods and across time periods that
1
Gandal, Kende, and Rob ð2000Þ show formally that in a simple dynamic model with
one model, sales are a linear, decreasing function of the forward difference in price net of
discounted future price ðp
t
2 bp
t11
Þ.
1176 journal of political economy
This content downloaded from 132.239.1.231 on Mon, 27 Jun 2016 12:41:16 UTC
All use subject to http://about.jstor.org/terms
剩余47页未读,继续阅读
资源评论
samLi0620
- 粉丝: 754
- 资源: 1万+
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功