.
ivdé
pfiW
. a. fl . ‘
avg
‘ . A:
.802
.anW.
. ...
i
«I

 

 

 

 

 

 

 

:3:
”mung”; ‘

.9)
”:90“!
w m

 

“huh.

5n v.5 In, IA
. , a... ......u,..:w Ears”. ,1 J fivhwwifififlfimfiflio ‘ APR," ,4. .. t 3 ‘

, .Eokrfil... .v mi .‘ ‘
’ ‘ .9 n

2...»...

. r

 

.1.
3006 LIBPARY
Michigan {mate
Univers ity

 

 

 

This is to certify that the
dissertation entitled

THE DYNAMICS OF MARKET POSITIONING: AN EMPIRICAL
ANALYSIS OF THE US. AUTOMOBILE MARKET

presented by
SENGU‘N YENIYURT

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Marketing

 

MW QQM’X‘I

 

Major Professor’ 3 SignatuIzé ‘
4’I 3 u I 3 2/0 0 5.

Date

MSU is an Affinnative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

JAN .3 l .2007

 

.1. J. ‘- gr '17"-

r.

 

“313:123330

 

 

 

 

 

 

 

 

 

 

 

 

2/05 cfilfiafitmjndd-DJS

 

THE DYNAMICS OF MARKET POSITIONING: AN EMPIRICAL ANALYSIS
OF THE U.S. AUTOMOBILE MARKET

Sengun Yeniyurt

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Marketing and Supply Chain Management

2005

ABSTRACT

THE DYNAMICS OF MARKET POSITIONING: AN EMPIRICAL ANALYSIS
OF THE U.S. AUTOMOBILE MARKET

By

Sengun Yeniyurt

This thesis explores the factors which influence the market positioning dynamics
and its short term and long term performance implications using a longitudinal approach
in the context of the U.S. automotive industry. The study derives on the co-evolutionary
theory, the organizational learning theory and the game theory to explore the external and
internal drivers of market positioning change. A special emphasis is given to the factors
that motivate and provide opportunity for change and the organizational capabilities
required for a market positioning dynamic. Market positioning dynamic is defined on
two dimensions, horizontal and vertical. Specific hypotheses regarding the effects of
several internal and external drivers on the propensity to engage in a market positioning
dynamic on each of the dimensions are developed. Further, the direction of the market
positioning dynamics and the short term and long term performance implications of
erratic and consistent market positioning dynamism are considered. Three separate
models are developed in order to test the hypotheses: a pooled hazard rate model of
market positioning dynamics, a competing hazards model of different types of market
positioning dynamics and a persistence model of the short term and long term

performance implications of market positioning dynamics.

Copyright by
SENGUN YENIYURT

2005

This dissertation is dedicated to my grandfather Izet Omer.

iv

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to those who provided their support
and assistance in completing this dissertation and receiving my Ph.D. I consider myself
lucky to have had on my side Zeyca Yeniyurt as Mom, tutor, and confidant; Ergun
Yeniyurt as brother and designated trouble solver; good friends from all around the
world; bright colleagues at Michigan State, among which Janell D. Townsend is the ideal
co-author and wonderful friend; and all the professors who I learned so much from during
my MBA. and Ph.D. studies, especially my Dissertation Committee, S. Tamer Cavusgil,
Roger J. Calantone, G. Tomas M. Hult, and Jonathan D. Bohlmann.

Thank you all.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................ viii

LIST OF FIGURES ................................................................................. x

CHAPTER ONE

INTRODUCTION .................................................................................... 1
RESEARCH QUESTIONS ................................................................ 2
EXPECTED THEORETICAL CONTRIBUTIONS .................................... 3
EXPECTED MANAGERIAL IMPLICATIONS ....................................... 4

CHAPTER Two ,

THEORETICAL FOUNDATIONS ............................................................... 5
MARKET POSITIONING DYNAMICS ................................................. 6
THE ROLE OF PRODUCT DIFFERENTIATION .................................... 9
INTERNAL DRIVERS OF MARKET POSITIONING DYNAMICS ............. 11
EXTERNAL DRIVERS OF MARKET POSITIONING DYNAMICS ............ 16
PERFORMANCE IMPLICATIONS OF MARKET
POSITIONING DYNAMISM .............................................................. 23

CHAPTER THREE

METHOD ............................................................................................ 27
DATA AND EMPIRICAL CONTEXT .................................................... 27
A HAZARD RATE MODEL OF MARKET
POSITIONING DYNAMICS ............................................................... 27
A COMPETING HAZARDS MODEL OF MARKET
POSITIONING DYNAMICS ............................................................... 35
A PERSISTENCE MODEL OF PERFROMANCE
IMPLICATIONS OF MARKET POSITIONING DYNAMISM .................... 40

CHAPTER FOUR

ANALYSIS AND FINDINGS .................................................................... 43
POOLED HAZARD RATE MODEL ESTIMATION...................................43
COMPETING HAZARD RATES MODEL ESTIMATION ........................ 56
PERSISTENCE MODEL ESTIMATION.................................................74

CHAPTER FIVE

DISCUSSION AND IMPLICATIONS ........................................................ 79
ANTECEDENTS OF MARKET POSITIONING DYNAMICS......................79
DIRECTION OF CHANGE ................................................................. 82
RETURN ON CONSISTENCY ......................................................... 83

Vi

CHAPTER SIX

LIMITATIONS AND FUTURE RESEARCH DIRECTIONS ........................... 85
APPENDICES ....................................................................................... 86

APPENDIX 1: VARIABLE DEFINITIONS ........................................... 87
REFERENCES ..................................................................................... 88

vii

Table 1:

Table 2:

Table 3:

Table 4:

Table 5:

Table 6:

Table 7:

Table 8:

Table 9:

Table 10:

Table 11:

Table 12:

Table 1 3:

Table 14:

Table 15:

LIST OF TABLES

Descriptive Statistics For The Covariates Employed In

The Pooled Model ................................................................. 33
Descriptive Statistics for Covariates Employed In

The Competing Risks Model .................................................... 38
Partial Likelihood Estimates of Covariate Effects on the

Propensity to Exhibit HMPD .................................................... 49
Partial Likelihood Estimates of Covariate Effects on the

Propensity to Exhibit VMPD .................................................... 50
Partial Likelihood Shared F railty Estimates of Covariate Effects

on the Propensity to Exhibit HMPD ............................................ 51
Partial Likelihood Shared Frailty Estimates of Covariate Effects

on the Propensity to Engage in VMPD ......................................... 52
Partial Likelihood Estimates of Covariate Effects on the

Propensity to Exhibit HMPD: Two Year Time Lag .......................... 53
Partial Likelihood Estimates of Covariate Effects on the

Propensity to Engage in VMPD: Two Year Time Lag ....................... 54
Summary of the Results for the Pooled Hazards Model ..................... 55

Partial Likelihood Estimates of Covariate Effects on the
Propensity to Exhibit HMPD Towards Market Center ....................... 57

Partial Likelihood Estimates of Covariate Effects on the
Propensity to Exhibit HMPD Away from Market Center ................... 58

Partial Likelihood Estimates of Covariate Effects on the
Propensity to Exhibit Downwards VMPD ..................................... 59

Partial Likelihood Estimates of Covariate Effects on the
Propensity to Exhibit Upwards VMPD ......................................... 6O

Partial Likelihood Shared Frailty Estimates of Covariate Effects
on the Propensity to Exhibit HMPD Towards Market Center ............... 61

Partial Likelihood Shared Frailty Estimates of Covariate Effects

viii

Table 16:

Table 17:

Table 18:

Table 19:

Table 20:

Table 21:

Table 22:

Table 23:

Table 24:

Table 25:

on the Propensity to Exhibit HMPD Away from Market Center ............ 62

Partial Likelihood Shared Frailty Estimates of Covariate Effects
on the Propensity to Exhibit Downwards VMPD ............................. 63

Partial Likelihood Shared Frailty Estimates of Covariate Effects
on the Propensity to Exhibit Upwards VMPD ................................. 64

Partial Likelihood Estimates of Covariate Effects on the Propensity
to Exhibit HMPD Towards Market Center: Two Year Time Lag .......... 65

Partial Likelihood Estimates of Covariate Effects on the Propensity
to Exhibit HMPD Away from Market Center: Two Year Time Lag ....... 66

Partial Likelihood Estimates of Covariate Effects on the Propensity

to Exhibit Downwards VMPD: Two Year Time Lag ........................ 67
Partial Likelihood Estimates of Covariate Effects on the Propensity

to Exhibit Upwards VMPD: Two Year Time Lag ............................ 68
Summary of the Results for the Competing Hazards Model ................ 69
The Brands Used In The Estimation Of The Persistence Model ............ 74
Impulse Response Function Summaries: ...................................... 77

The Comparison of the Short Term and Long Term Effects of
Consistent Versus Erratic MPD ................................................. 78

ix

Figure 1:

Figure 2:
Figure 3:

Figure 4:

LIST OF FIGURES

Theoretical Framework: Antecedents and Performance

Implications of Market Positioning Dynamics ................................. 8
Smoothed Hazard Rate Estimates for HMPD ................................. 47
Smoothed Hazard Rate Estimates for VMPD ................................. 48
The Impulse Response Functions for one of the Brands ..................... 76

CHAPTER ONE

INTRODUCTION

To improve is to change; to be perfect is to change often.

Winston Churchill

When you're finished changing, you're finished

Benjamin Franklin

The need for changing the marketing strategy as the environmental factors evolve
is widely acknowledged by the business literature. Facing environmental changes, the
organizations have to adapt the market strategies of the company through continuous
learning in order to remain competitive (Barr et al. 1992; Sinkula et al. 1997). Under
hostile environmental conditions companies are inclined to search for competitive
advantages by introducing new brands, or by repositioning their existing products in
more attractive market positions (Aaker 1997) to increase their profits (Carpenter 1989).
As such, market positioning dynamics is a result of complex managerial decision making
with respect to the market strategy the brand or company is following. Considering the
inherent complexities associated with strategic decision making under technological
turbulence and market dynamism, the internal characteristics and external conditions that
affect market positioning changes are of substantial interest.

Market strategy dynamism has been analyzed through a variety of frameworks,
including new market entry (e.g., Green et al. 1995; Han et al. 2001; Mahajan et al. 1993),

product line extensions (Green and Krieger 1985), and organizational learning (Sinkula et

al. 1997). Yet, research regarding the dynamic aspects of market positioning is relatively
scarce. Marketing strategy aims at developing a sustainable competitive advantage
through product differentiation and building a strong market position to the firm or the
brand. Hence, the short term and long term performance implications of such marketing
strategy changes are of great interest. The performance implications of several marketing
decisions have been explored before, yet the empirical research mainly focused on the
relative impact of different marketing actions, such as price promotions and advertising
attacks (e.g., Dekimpe et al. 1999a; Pauwels et al. 2004). Therefore, the performance

implications of market positioning dynamism remain to be explored.

RESEARCH QUESTIONS
This thesis explores the factors which influence the market positioning dynamics
and its short term and long term performance implications using a longitudinal approach
in the context of the U.S. automotive industry. Time series data between 1980 and 2004
regarding the automotive brands and the models they offer in the U.S. market are used to
test the research hypotheses. As such, this thesis employs the co-evolutionary theory, the
organizational learning theory and the game theory to explore the following research
questions:
0 What are the external and internal drivers of market positioning dynamics?
0 How is the market positioning expected to change, under which
conditions?
0 How do different types of market positioning dynamism relate to short

term and long term market performance?

The remainder of this thesis is organized in the following manner. First, the
extant literature related to market positioning dynamics is reviewed, theoretical
foundations are presented and research hypotheses are developed. The environmental
drivers and internal characteristics that are expected to impact market positioning
dynamics are presented and related research hypotheses are developed. Next, the short
term and long term effects of market positioning dynamism on performance are explored
and specific hypotheses are developed. To test these hypotheses, three empirical models
are proposed: a hazard rate model of the antecedents of market positioning dynamics, a
competing risks hazard rate model of different types of market positioning dynamics, and
a cointegration analysis of the short term and long term performance implications of

market positioning dynamics.

EXPECTED THEORETICAL CONTRIBUTIONS

Although extant research in market positioning is relatively rich, the drivers and
consequences of market positioning dynamism remain to be explored. This study
employs the co-evolutionary theory and derives on the organizational learning framework
to formulate specific hypotheses regarding the role of internal and external motivators,
opportunities for change, and brand level resources and capabilities in market positioning
dynamics. Established game theoretic rules are utilized in developing hypotheses
regarding the direction of the market positioning change. Further, the short term and long
term performance effects of different types of market positioning dynamics are explored.
Deriving on extant theoretical frameworks, three empirical models are developed to test

whether or not the hypothesized relationships hold. As such, this study contributes to the

extant marketing strategy literature in several ways. First, significant drivers of market
positioning dynamism are identified. Second, deriving on the extant theoretical
fi'ameworks, testable predictions regarding the direction of market positioning change on
the horizontal and vertical dimensions are formulated. Finally, the performance
implications of erratic and consistent market positioning dynamism are identified.
Overall, the study extends the co-evolutionary and organizational learning
theories by integrating game theoretic propositions and tests the predictions engendered

by the joint application of these theories empirically.

EXPECTED MANAGERIAL IMPLICATIONS

Increased competitive pressures and market dynamism are adding to the inherent
complexities of decision making regarding the optimal market positioning of the brand.
This study develops a framework that will help decision makers understand the dynamics
of market positioning. Of special interest for managers is the expected market
positioning change in different situations. Understanding the role different factors play in
market positioning dynamics will help them in formulating future marketing strategies
and foreseeing competitor strategic shifis. Further, knowing the short term and long
term market performance effects of different types of market positioning dynamism

managers would be able to better gauge the impact of their strategic decisions.

CHAPTER TWO

THEORETICAL FOUNDATIONS

Evolutionary economics posits a model of interaction among the environment, the
industry, and the firm through the dynamics of variation, selection, and retention (Nelson
and Winter 1982; Volberda and Lewin 2003). This framework incorporates multivariate
analysis of firm level behavior (managerial adaptation) and population ecology
(environmental selection) within a competitive environment (Volberda and Lewin 2003,
p.2113). Deliberate actions based on experience allow managers to develop anticipatory
models and systems of control in order to act in advance of blind environmental selection
(Volberda and Lewin 2003). Thus, managed selection drivers produce complex
microevolutionary processes, as indicated by adaptations that reflect critical time and
resource constraints on variations (Galunic and Eisenhardt 1996; Volberda and Lewin
2003).

In order for co-evolution to occur, firms must operate in a heterogeneous industry
where there is interaction and influence among competitors; in addition, the firms must
have a learning capability (V olberda and Lewin 2003). From a population perspective,
competition can be explained and predicted through consideration of relative niche
density; from a firm level perspective, organizational learning related to environmental
changes and organizational characteristics, helps predict future behavior.

Conceptually, this study draws on the tenets of organizational learning, or the
development of insights, knowledge, and associations based on past actions, the

effectiveness of those actions, and their connection with future actions (Fiol and Lyles

1985). From this perspective, experience is a pattern of recognition, a repetition of
activity undertaken previously, and future actions become a function of the accumulated
memory of the firm (Sinkula 1994; Slater and Narver 1995). Organizational memory is
the collective beliefs, behavioral routines, or physical artifacts that vary in their content,
level, dispersion, and accessibility (Moonnan and Miner 1997). Organizational routines,
procedures, and structures are vital components for controlling the behavior of the
organization and are accumulated over time, establishing conditions for subsequent firm
actions and activities (Cyert and March 1963; March and Simon 1958). Consequently,
organizational learning is a function of age and experience (Sinkula 1994). How a firm
applies experiential knowledge to its activities is a major source of capability (Grant
1996).

Organizations have to respond to the market based information generated and
disseminated though marketing actions (Kohli et al. 1993). Hence, under continuous
market dynamism and increased levels of competition, failing products, decreasing
market share and profit loss will require repositioning of existing products, dropping

product lines, and new product introductions. Change is required for survival.

MARKET POSITIONING DYNAMICS

There are two literature streams regarding the role of organizational actions in
organizational learning. Although some scholars believe organizational learning would
occur even if what is learned is not translated into changes in actions taken, many argue
that the learning is not completed unless the organizational learning is manifested through

change in actions (e.g., Argyris 1977; Argyris and Schon 1978; Fiol and Lyles 1985;

Garvin 1993; Sinkula et al. 1997). In this light, marketing strategy changes can be
regarded as both the expression of learning and a way to facilitate new learning through
trials and errors (Sinkula et al. 1997). Market oriented organizations possess strong
product innovation capabilities (Lukas and Ferrell 2000) and are more likely to have
successful new product introductions regardless of the environmental conditions (Slater
and Narver 1994),

Deriving on extant research (Sinkula et al. 1997), market positioning dynamic is
defined as the frequency with which the marketing positioning is modified (i.e.,
frequency of changes in product characteristics that effect the product differentiation and
market positioning of the brand). As such, market positioning dynamic can be achieved
through a variety of ways: repositioning existing products by changing their product
characteristics or price, launching new products, or dropping products from the product
line (Kaul and Rao 1995).

In this study, Miller and Chen (1994) ‘s framework of organizational change has
been adopted. According to this framework, three factors play a critical role in
organizational change. First factor consists of decision makers’ motivation to engage in
organizational change. Also, an opportunity for organizational change should exist, and
finally, the organization should possess the capability to change. As such, the
organizational and environmental characteristics that are related to these factors are

incorporated in the analysis. The theoretical framework can be seen in Figure 1.

 

 

 

compact—8m 839m o
beessa caspam .
Damiano 98 88.58%

 

 

 

 

EOE waoq
58,—. team

353:8qu mauve—.82

 

 

 

Beamom o

85 n

8558830 833.5 o
Gunman $885 maze—.82

 

 

 

 

 

wcmcoummom
3x82 80.55
.835 Dosage—om o
omega Com 832502

 

 

 

 

 

"8:80.“:on ~26on
commaamxm “83:2
b62095 2632380
Samantha “332 o
0958 we figs HEonm

 

3.82.3 mam-SEAS.— .3—32 mo 2:53am...— oouu—Eotom can fine—.33: 3.33083...— 183932;. 3 gang

 

 

 

THE ROLE OF PRODUCT DIFFERENTIATION

Product differentiation is defined as “offering a product that is perceived to differ
from the competing products on at least one element of the vector of physical and
nonphysical product characteristics” (Dickson and Ginter 1987: p. 6). Product
differentiation can be achieved through perceptual differences that are influenced by the
word of mouth, usage experience and promotion, or through actual differences created by
product characteristics, including price (Dickson and Ginter 1987). This study focuses on
quantifiable product differentiation, exploring the role of physical product characteristics
and price.

Product differentiation research can be dated back to Hotelling (l929)’s
horizontal differentiation model. According to his model, under equal pricing, competing
firms will choose to position at the market center. Yet, when price competition is
considered, maximum differentiation is optimal (d'Aspremont et al. 1979). Another
influential framework is Lancaster (1971)’s product space. Differentiation in the product
space is defined in terms of two dimensions: horizontal and vertical. Horizontal
differentiation is based on the premise that customer taste vary across the population and
product variety is required to satisfy this variation. Vertical differentiation refers to a
characteristic that is regarded as beneficial by all the customers, such as quality, yet their
willingness to pay for it varies. Earlier studies focused on only one of the differentiation
dimensions, whether horizontal (e.g., Hauser and Shugan 1983) or vertical (e.g.,
Gabszewicz and Thisse 1979; Shaked and Sutton 1982). The research stream evolved by
considering multiple differentiation dimensions simultaneously (e.g., Carpenter 1989;

Choi et al. 1990; Horsky and Nelson 1992; Kumar and Sudharshan 1988). Recent

research indicates that when two dimensions are considered, firms tend to maximize the
differentiation on one dimension while minimizing the differentiation in the other
(Economides 1989; Vandenbosch and Weinberg 1995). In vertical product
differentiation, price plays a crucial role by influencing the customer perceptions
regarding product attributes such as quality (Hauser and Simmie 1981). Therefore,
following on the extant research in product differentiation, the market position dynamic
is defined as having two dimensions: horizontal market positioning dynamic based on
physical product characteristics that provide variety for different customer preferences,
and vertical market positioning dynamic based on product characteristics that are
regarded as the more the better, such as quality, and captured by the average relative

price.

10

INTERNAL DRIVERS OF MARKET POSITIONING DYNAMICS

Past Market Positioning Dynamism

The marketing strategy a brand employs at a give time is dependent on previous
strategic positions the brand has held. Experience has a significant impact on firture
market strategy formulation (Huff 1982). For example, in the retailing context, past
pricing history is the most important determinant of the pricing strategy (Srinivasan et al.
2003)

Past marketing strategy changes can be regarded as an indicator of strategic
flexibility (Sanchez 1995). Organizations that possess strategic flexibility are able to
rapidly proliferate their product line in order to saturate a profitable market segment with
related product models (Sanchez 1995). Strategic flexibility can also be utilized in rapid
introduction of product upgrades to cannibalize successful products, in order to challenge
the competition (Conner 1988).

Organizations are more likely to change when the routines to make such changes,
including the development and acquisition of necessary resources are already developed
(Kelly and Amburgey 1991). As such, market strategy changes are likely to create an
internal momentum towards further marketing strategy changes, even in the absence of
environmental opportunities or threats (Greve 1998). The momentum created by past
strategic shifts is likely to decay over time, i.e. a marketing strategy change is expected to
have a strong positive effect on new marketing strategy changes shortly afier the event

has occurred (Amburgey et al. 1993).

ll

Hla. Products with recent horizontal market positioning changes are more
likely to exhibit changes in their horizontal market positioning.
Hlb. Products with recent vertical market positioning changes are more likely

to exhibit changes in their vertical market positioning.

Past Performance

There has been great interest in the academic literature regarding the effect of past
performance on different types of organizational change (Greve 1998; Miller and Chen
1994), risk taking (Bromiley 1991), innovation (Bolton 1993) and marketing actions
(Srinivasan et a1. 2003). This stream of research reveals that performance Shortcomings
are often motivators of change in organizations. As such, responsiveness and frequent
organizational change is not limited to only successful companies. Extant research
indicates that even less successful organizations are often active risk takers (Bowman
1982). The propensity to take risk and engage in organizational change is actually higher
when the performance levels are low (Bolton 1993; Bowman 1982). As such, it is
expected that poor market performance is a Significant internal motivator for horizontal
market positioning dynamic. On the other hand, successful brands have the tendency to
reposition vertically higher, and increase the price level in order to enjoy a more
beneficial market position (Aaker 1997). As such, it is expected that a success in the

market will engender an increase in the average price charged for their products.

H2a: Products with recent improvement in market performance are more likely

to exhibit change in their horizontal market position.

12

H2b: Products with recent decrease in market performance are more likely to
exhibit change in their horizontal market position.

H3a: Products with recent improvement in market performance are more likely to
exhibit change in their vertical market position.

H3b: Products with recent decrease in market performance are more likely to
exhibit change in their vertical market position.

H4: Products with recent improvement in market performance are more likely to

exhibit upwards vertical repositioning (i. e. increase the price).

Product Proliferation

Product line extensions are required to benefit from profitable market segments (Sanchez
1995). Yet, increased product proliferation may lead to increased complexity of
coordination and elevated inventory costs (Kekre and Srinivasan 1990) leading to
inflation in the total cost burden the brand has to sustain. As such, product line culling is
expected to lead to increased performance (Sorenson 2000). Therefore, a brand is
expected to aim towards and optimal product line breath through successive trials and

errors of adding and culling models from the product line.

H5. Product line proliferation has a non-monotonic ( fl shape) effect on the

propensity to exhibit horizontal market positioning dynamics.

13

Direction of horizontal market positioning dynamic

There are two strategic options that a company can choose in order to strengthen a
weak market position (Markides 1999): imitating the position of the dominant
competitors, or create a new market position through innovation. Therefore, decision
makers face a critical strategic dilemma. While repositioning closer to the market center
is expected to have a positive effect on performance through legitimation, positioning
away from the market center through differentiation is expected to decrease the
competitive pressure, also having a positive effect on performance (Deephouse 1999).
From a differentiation perspective, the bigger the distance of the focal brand position
versus the competition, the higher the benefit for the company (Carpenter 1989). Yet, in
order to be able to reposition in a new to world market position, organization have to
possess significant innovative capabilities (Han et al. 2001).

As such, it can be postulated that a company that possesses innovative capability,
i.e. has created innovative market positions away from the market center before, would
be more likely to create a new market position for itself, away from the market center.
On the other hand, organizations that lack such innovative capabilities, i.e. have not
repositioned themselves away from the market center before, would be more likely to
reposition themselves closer to the market center, imitating the dominant competitors.

It has been shown that through physical product characteristics differentiation in
the horizontal dimension a brand can reduce the competitive pressure and increase the
price charged for its products (Deephouse 1999). Hence, it is expected that brands that

follow a differentiation strategy will enjoy the benefits of higher margins, being able to

14

increase their average prices, while brands that aim towards the market center will be

forced to rely more on low price strategy to cope with the intense competitive pressure.

H6a:

H6b:

H7a:

H7b:

Products with recent horizontal repositioning away from the market
center are more likely to continue to be horizontally repositioned away
from the market center.

Products with recent horizontal repositioning closer to the market center
are more likely to continue to be horizontally repositioned closer to the
market center.

Products with recent horizontal repositioning closer to the market center
are more likely to exhibit downwards vertical repositioning.

Products with recent horizontal repositioning away from the market

center are more likely exhibit upwards vertical repositioning.

15

EXTERNAL DRIVERS OF MARKET POSITIONING DYNAMICS

The fast changing environment, including the dynamics of competitive and
technological shifts, has a continuous feedback relationship with the marketing strategy
the organization is able to implement (Sanchez 1995). The change in technological
standards and customer needs is increasing its velocity, creating vast opportunities for
innovative organizations (Markides 1998). Intensive industry competition and emerging
more profitable market segments are strong motivators of repositioning or new product
introduction and new brand creation (Aaker 1997). Therefore, the environmental
conditions provide significant opportunities and motivation for market positioning

dynamics.

Competitive Intensity

In organizational ecology, density dependence describes the consequences of
competition within a population; the size of a population at any time affects the rate of
birth and death from the population. The total density of the population thus becomes a
function of the carrying capacity of the niche, which represents and upper bound of the
aggregate activity that can be performed by an organizational form (Boone and
Witteloostuijn 1995). When few firms compete, they exert weak pressure for removing
firms from the environment (Harman and Freeman 1977).

In the absence of strong competition, a company may perform well even if does
not develop an adequate marketing strategy and differentiation and is not dynamic in

adapting this strategy to environmental changes because customers do not possess

16

alternatives. On the other hand, in industries with intense competition companies will be
forced to be extremely responsive (J avorski and Kohli 1993; Narver and Slater 1990).
One focus of ecological theory is the processes that lead to equilibria in the
environment. It is believed that as product populations compete for resources, they will
begin to specialize or differentiate. This can lead to changes and fragmentation of the
original population, what could be construed as endogenous population change
(Amburgey and Rao 1996). In theory, when two populations try to inhabit the same
niche, they cannot coexist in equilibrium; one will always try to outrnaneuver and thwart
the other in order to utilize the available niche resources, with the goal of competitive
exclusion (Geroski 2001). Hence, the competitive intensity, in terms of niche overlap
has a negative effect on long term survival of the focal organization (Dobrev et al. 2002).
Although fundamental, the evolutionary perspective has been limited in its ability
to explain and predict the competition for resources within a niche (Young 1998). While
parallels are often drawn between organizational ecology and the theory of industrial
organizations (Boone and Witteloostuijn 1995; Geroski 2001), it is interesting that an
integration of the theories provides means to investigate a multi-level phenomenon. As
an extension of the environment-strategy-performance perspective, strategic groups
theory, evolved as a means to explain how industry dynamics evolve in such a way that
firms, and in this case products, cluster together based on how competitors interact.
Essentially strategic groups are comprised of firms that compete for the same customers,
using similar resources (Hunt 1972). It is expected that certain groups within an industry
will pursue similar strategies, effectively impacting the conditions of competition. The

original argument was built around the concept of asymmetries within a market as criteria

17

for the identification of groups that shape the structure of an industry (Hunt 1972), and
focused heavily on mobility barriers created by these intra-industr'y groups (Caves and
Porter 1977). Further research has Shown that strategic groups act as reference points
for the group members when managers are developing competitive strategy (Fiegenbaum
and Thomas 1995). As such, the market positioning dynamism of the competition within
the same market segment with is expected to be a reference point for the focal brand
market positioning dynamics.

As differentiation increases, oligopolistic competition results; the more limited the
product differentiation, the market becomes closer to perfect competition (Hotelling
1929). As the likelihood of achieving a monopolistic position within a market increases,
so too does the opportunity for deriving economic rents. Average profits by all firms in
an industry fall when strategies become progressively more divergent, as well as when
there is an increase in rivalry (Cool and Dierickx 1993). Although empirical evidence
regarding the variation in performance within strategic groups has been limited, the
published findings indicate that there is a difference in performance between members of
an individual group (Cool and Schendel 1987).

One way to View the marketing segments in which the brands are competing is
based on the ecology theory and the niche concept. In the early studies, a niche was
defrned as all the combinations of resource levels at which the population can survive and
reproduce itself (Harman and Freeman 1977). The definition of niche evolved to include
the variability of customer demand (Freeman and Harman 1983) and the relative
differentiation on a resource space to assess the competitive intensity (Barnett and Carroll

1987; Baum and Haveman 1997; Baum and Mezias 1993). As such, “socially organized

18

market segments carry different information” (Sinchcombe 1990) and these differences
are reflected in the organizations that inhabit each market segment (Dobrev et al. 2002).
Although the definition of the niche is inherently multidimensional, the extant research
usually focuses only on one dimension of the niche in order to overcome the
operationalization constraints (e. g., Dobrev et al. 2002).

It can be expected that as competitive intensity within a market segment increases
there will be incentive to try to differentiate the product offering in order to gain a
competitive advantage. On the other hand, empirical evidence from industries with
significant environmental uncertainty indicates that the greater the product variety, the
less valuable the variety to a firm (Sorenson 2000). Therefore, it is expected that as the
product proliferation in the market segment increases, the market positioning changes are
less likely to occur.

Product offerings that are distinctively suited to meet the demand of a segment of
customers not addressed by rivals offers opportunity for marketplace advantage; as the
number of competitors increases, the chance to find the un-served segments decreases
(March 1991). A significant positive relationships has been found between the level of
new product success and measures of product differentiation including the presence of
unique features and high product quality (Song and Parry 1994). A such, new market
entries are expected to provide a strong motivation for adjusting some aspects of the
market positioning in order to defend the current market advantage (Hauser and Shugan
1983). Also, new market entries are likely to engender price decreases for the incumbent

brands due to a defensive reaction (Hauser and Shugan 1983).

19

H8: The higher the market competitive intensity, the higher the propensity of
the product to exhibit horizontal market positioning dynamic.

H9: The higher the market total product proliferation, the lower the propensity
of the product to exhibit horizontal market positioning dynamic.

HlOa: The higher the market competitive intensity, the higher the propensity of
the product to exhibit a decrease in its vertical market position.

HlOb: The higher the market competitive intensity, the lower the propensity of

the product to exhibit an increase in its vertical market position.

Market Dynamism:

Environmental dynamism has two dimensions: technological turbulence and
market dynamism. Technological turbulence is related to the extent to which
production/service technology in your principal market has changed over time (Narver
and Slater 1990). In industries that are subject to high technological turbulence,
technological changes provide big opportunities (Javorski and Kohli 1993). On the other
hand, market dynamism is related to the rate of change of the customer preferences,
market segments, and demand patterns (Javorski and Kohli 1993; Narver and Slater
1990). In such industries, the organizations have to adapt faster to the changing customer
demands.

One of the criticisms of strategic group theory is the lack of evidence regarding
the purported difference in performance results between groups (Cool and Schendel,
1987). The concept of strategic groups “captures the intuitive notion that within group

rivalry and between group rivalry differ” (Cool and Dierickx 1993). Strategic groups

20

within an industry exist, and provide means to understanding the variation in the level of
performance (Thomas and Pollock 1999). It is argued that group characteristics affect
performance that is independent of individual level or industry level effects (Dranove et
al. 1998). Thus, it is expected that the market dynamism varies across different segments.

The market strategy dynamism of the competition also signals the management of

the focal company regarding the relative attractiveness of different strategic alternatives.
Organizations overcome the uncertainty associated with new marketing positions by
mimicking competitors’ market positions (Greve 1998). Once proved successful, new
marketing strategies are swiftly adopted by the competition (Huff 1982). Recent
competitor market strategy dynamism acts as a facilitator of organizational learning of
the market structure and the future trends and opportunities in the customer needs (Greve
1998). Hence, market positioning dynamism has to be at least at the same rate with the
competition in order to have a positive effect on market performance (Sinkula et al. 1997).
Although the absolute market dynamism of the company is high, if it is below the market
segment average, long run performance would suffer.

H11: The higher the total market competitor horizontal market positioning
dynamism the higher the propensity of the product to exhibit horizontal
market positioning dynamic.

H12: The higher the total market competitor vertical market positioning
dynamism the higher the pr0pensity of the product to exhibit vertical

market positioning dynamic.

21

Market Expansion

Aggregate level demand has been the focus of several past studies (e.g., Hanssens
et al. 1990; Schultz and Wittink 1976). A recent study indicates that new competitive
entries in a market may create new demand, leading to an increase in the total market size
(i.e. market expansion) (Mahajan et al. 1993). As such, the new product introductions
and market repositionings can lead to market expansion, decreasing the incremental
competitive pressure created on the incumbent brands. In an expanding market,
marketing strategy changes are less likely to attack the customer base of the competition.
Hence, the strategic response of the competition is expected to be smaller in magnitude
(Dekimpe et al. 1999a). On the other hand, under the zero sum equilibria, which is the
case in no market expansion, strategic marketing attacks of one company are directed to
the customer base and market Share of the competition, inviting retaliation (Dekimpe et al.
1999a). Therefore, the focal brand is more likely to respond with marketing strategy
dynamism under stable market conditions.

H13: A brand is more likely to respond to stable or declining market total
demand conditions, than under expanding market conditions, by changing
its product’s horizontal market positioning.

H14a: A brand is more likely to respond to stable or declining market (segment)
total demand conditions, than under expanding market conditions, by
decreasing its product ’s vertical market positioning.

Hl4b: A brand is more likely to respond to expanding market conditions, than
under stable or declining market (segment) total demand conditions, by

increasing its product’s vertical market positioning.

22

PERFORMANCE IMPLICATIONS OF MARKET POSITIONING DYNAMISM

Extant research in strategic management literature reveals that there are several
distinct strategic orientations that have different performance implications. As such,
Miles and Snow (1978) proposed four strategy types: defenders, prospectors, analyzers,
and reactors. While the prospectors are characterized as aggressive and risky placing
high emphasis on new product introduction and market development seeking rapid
increase in market share (Miles and Cameron 1982), defenders have a narrow product-
market domain and focusing on efficiency, analyzers are operating in both stable and
changing market domains, searching for efficiency in the first and innovation in the later,
and reactors are not able to respond to changes effectively (Miles and Snow 1978).
Therefore, it is expected that organizations that adopt the aggressive and risky strategic
posture will enjoy higher sales and market shares in the short run while suffering in the
long run. On the other hand, analyzers and defenders, given their calculated behavior, are
more likely to envoy long term benefits and financial returns (Venkatraman 1989).

Early studies in performance response modeling of marketing actions focused
mainly on short term forecasting and optimization, assuming a stable environment (Wind
and Robertson 1983). Yet, the business environment is increasingly characterized by
high technological turbulence and market dynamism, and has to be incorporated in
contemporary performance response modeling of marketing actions (Dekimpe and
Hanssens 1995).

In line with contemporary research in performance response of marketing actions,
the short term effects of market positioning dynamism are defined as immediate effects,

while long term effects are defined as persistent effects which are derived from the

23

difference between the performance level before the market positioning change and
performance level afier the effects have stabilized (Pauwels et al. 2002).

Dynamism may occur due to organizational learning yet the short term market
performance may not improve because the improvement may not reach the thresholds
that would satisfy the customers, further changes being required in this case or because
the rate of learning is below the rate of the competitors (Sinkula et al. 1997). On the
other hand, promotional attacks and price discounts have significant positive effects on
Short term performance, but such marketing actions fail do provide benefit to the
organization in the long run (Dekimpe et al. 1999a; Pauwels et al. 2004). Game theoretic
models indicate that while firms are expected to achieve their short term goals via price
adjustments, the long run equilibrium is strongly influenced by product repositioning
strategies followed (Choi et al. 1992). Nevertheless, long term strategic orientation of the
brand may engender consistent vertical price adjustments to reposition in a performance
maximizing position. Although customers may react negatively to price adjustments,
they may adapt to higher prices over time engendering a short term drop in market
performance yet an increase in the long term performance (Dekimpe et al. 1999a).

In the long run, the market performance is expected to reflect the organizational
leaming capability of the firm, hence the market dynamism (Sinkula et al. 1997).
Although adverse risk taking behavior and intensive competitor orientation may result in
short run benefits such as increased market share, organization’s long term profitability
will suffer (Armstrong and Collopy 1996).

Pauwels et al. (2004) indicate that new-product introductions may have a long

term effect on revenues, while price discounts are known as having only a temporary

24

positive effect on performance (Nijs et al. 2001). Additionally, a new product
introduction can have a long term effect on performance because it transforms the
competitive capabilities of the brand (Geroski et al. 1993).

It is expected that brands that follow a consistent long term strategy are more
likely to enjoy stable long term returns. On the other hand, aggressive risk takers are
expected to be Short term oriented, being driven by short term performance results.
These brands are less likely to achieve a long term sustained market performance.
Finally, organizations that do not change their market positioning in line with the
changing customer requirements are not likely to improve their short term or long term

performance.

H15a: Erratic horizontal market positioning dynamism (changes in horizontal
market positioning that are not in the same direction with the previous
change) has a positive effect on short term market performance but no
significant effect on long term market performance.

H15b: Consistent horizontal market positioning dynamism (changes in horizontal
market positioning that are in the same direction with the previous
change) has a positive effect both on short term market performance and
on long term market performance.

H16a: Erratic vertical market positioning dynamism (changes in vertical market
positioning that are not in the same direction with the previous change)
has a positive effect on short term market performance but no significant

effect on long term market performance.

25

H16b: Consistent vertical market positioning dynamism (changes in vertical
market positioning that are in the same direction with the previous
change) has a positive effect on both short term and long term market

performance.

26

CHAPTER THREE

METHOD

DATA AND EMPIRICAL CONTEXT
The research hypotheses presented above are tested in the context of the U.S. automotive
market between years 1980 and 2003. The automotive industry is adequate to test our
hypotheses for several reasons. First, the automotive industry is one of the most
important industries in U.S. hosting one of every seven jobs in the economy (Tardiff
1998). Second, the industry is characterized by rapid change, and intense competition
with considerable variation in product specifications and market positioning across time.
Third, by focusing on the automotive industry we are contributing to the cumulative
knowledge gained through previous studies that utilized the same industry (e.g., Dobrev
et al. 2002; Dobrev et al. 2003; Harman 1998; Harman et al. 1998a; Harman et al. 1998b;
Pauwels et al. 2004).

The unit of analysis has been selected as the model. Product specifications, prices
and sales levels have been recorded for the major brands sold in U.S. during this period.
The data has been collected from a variety of secondary sources including Automotive

News, WARDS Automotive and the Auto Pacific Historical Battleground.

A HAZARD RATE MODEL OF MARKET POSITIONING DYNAMICS
A continuous time event history analysis with time varying covariates was
employed in order to estimate the effects of the independent variables on the probability

of a company changing its market positioning. This type of model provides the best tool

27

to analyze time based phenomena (Blossfeld and Rohwer 2002). The technique is similar
to multivariate regression, but the dependent variable is the unobserved probability of an
event to occur at a specific point in time. In our case, the dependent variable is the
propensity of a brand to change its market positioning in a given market segment at a
precise point in time. The significance of the estimates can be analyzed using
conventional hypothesis testing (Tuma and Harman 1984).

The determination of the probability distribution function of the transition rate
over time is a critical decision in event-history analysis. There are two basic approaches
to modeling time dependence: parametric (Tuma and Harman 1984) and semi-parametric
(Cox 1975). Parametric methods assume a specific distribution of hazard rate over time.
While the exponential model assumes constant hazard rates over time; Gompertz,
Makeham and Weibull models assume a monotonically increasing or decreasing rate over
time (Blossfeld and Rohwer 2002). The non-monotonic time dependence with a single
hump can also be modeled using parametric methods such as the log-logistic (Bruderl
and Schussler 1990) or lognorrnal distribution (Levinthal and Fichman 1988). On the
other hand, semi-parametric methods such as the Cox (1972) model have very few
assumptions regarding the distribution of the hazard rate over time by allowing for a
salient base rate function that is identical for all the members in the population, yet does
not follow any prescribed shape (Blossfeld and Rohwer 2002).

The semi-parametric Cox model does not assume an apriori shape of hazard rate
distribution over time and is thus appropriate for this analysis. The Cox model utilizes
partial likelihood maximization by sorting the data according to the event times. Given

that the data provides the event times up to the month of the strategic change, the results

28

 

of the partial likelihood estimation are reliable. Moreover, a relatively large portion of
our observations are right censored, which does not create any difficulty for the
estimation of the Cox model, such observations being included in the risk set at the event
time.

Two different hazard rates are defined: one for the propensity to change the
horizontal market positioning, and one for the propensity to change the vertical market
positioning. Hence, the propensity of changing the horizontal market positioning of a
product model at time t given that its previous positioning shifi has been at time t," n can

be defined as a continuous time hazard rate:

1im P(t,, S T < tn + At)
At —) 0 At

 

’10 I tn) =

Similarly, the propensity of changing the vertical market positioning of a product model

at time t can be defined as:

lim <
60""): P(t,, _T<t,,+At)
At —> 0 At

 

That is, the hazard rates are defined as the limit as At approaches zero of the probability
of having a market positioning change in the interval of time between the previous event
of a product model (tn), and tn + At. This approach of resetting the time axis after each
event is very Similar to the clock re-set model employed by Amburgey et a1 (1993).

The dataset has been split in monthly Spells. The variables will be updated for
each product model monthly or after each event. A product model is considered at risk
the day the it is first present in the passenger cars segment or the day after model’s last

market positioning change.

29

 

Market Positioning Dynamics

The market position at a given point in time has been measured through a variety
of methods. One such method resides on the niche definition of the company in terms of
the relative position of the mid point of the Spectrum covered in a technological space
(Dobrev et al. 2002; Podolny et al. 1996). For example, the amount of overlap of the
range of horse power of the cars produced was used as an indicator of niche overlap to
derive the intensity of competition, and the mid point of this range was utilized as a
measure of differentiation to indicate the relative market position (Dobrev et al. 2002).
As such, shifts in the midpoint of the product line offering on the horizontal dimension
could be used as an indicator of marketing positioning dynamics.

The size of the car has been identified as the most important product characteristic
in differentiation in the automotive industry (Arthur Andersen 1985). Although one
might argue that styling have a significant impact on differentiation, it has been shown
that styling changes do not pay off financially (Hoffer and Reilly 1984; Sherman and
Hoffer 1971). As such, wheelbase size has been used as an indicator of the horizontal
market positioning. Any changes in the mid wheelbase size of the product line (model)
offered are recorded as a change in the horizontal market position of the product. Hence,
in the first model the dependent variable consists of the propensity to change the mid
point of the product line span in terms of wheelbase through new product introductions,
product repositioning or product culling. The vertical market positioning dynamic is
captured via changes in the deflated manufacturer suggested retail base price. In order to

minimize noise in the data and capture only significant vertical repositioning, ten percent

30

 

was established as the arbitrary cut off value. AS such, changes bigger than ten percent

have been recorded as a VMPD event.

Covariates

Recent horizontal and vertical market positioning changes are captured as dummy
variables that are recorded as one if the product had a market positioning change in the
related dimension, within the last year, and zero otherwise. The recent change in market
performance is recorded as the percentage change in the unit sales of the product in the
last year, as compared to the previous year. Then, two different variables have been
recorded, one that is equal to the positive valences of the market performance change,
and one that is equal to the absolute value of negative valences of market performance
change. Product line proliferation is measured in terms of different wheelbase sizes a
product line has to offer at a given time.

The market segment competitive intensity is measured in terms of total number of
models competing at a time. Average market product proliferation is defined in terms of
the average number of wheelbase sizes the models competing at a given point in time
have to offer. Competitor horizontal market positioning dynamism is captured in terms
of total market dynamism of all the models in the market minus the market dynamism of
the focal model. The vertical market dynamism is measured similarly. Market expansion
is measured in terms of the percentage change in last year’s total sales in comparison
with the previous year. Finally, a set of exogenous factors are controlled for, including
the U.S. Gross Domestic Production per capita, Dollar/Y en exchange rate, oil barrel price,

and U.S. population. The GDPPC and the oil barrel price have been deflated and are

31

presented in 1980 Dollars. With the exception of recent HMPD and recent VMPD, all
covariates have been lagged by one year. The descriptive statistics of the covariates can

be found in Table 1.

32

Table 1: Descriptive Statistics For The Covariates Employed In The Pooled Model

 

 

 

 

 

 

 

 

 

 

 

Std. Correlations
Covariates Mean Dev. 1 2 3 4
1 Recent HMPD 0.148 0.355 1.000
2 Recent VMPD 0.124 0.332 .205 1.000
< .001
3 Performance increase 2.705 139.196 .010 -.007 1.000
.067 .195
4 Performance decrease 0.096 0.156 -.012 .074 -.012 1.000
.021 <.001 .018
5 Product proliferation 1.125 0.379 .204 .030 -.006 -.060

 

< .001 < .001 .236 < .001

 

 

6 Product proliferationsq 3 1.410 1.368 .205 .026 -.006 -.059
“f <.001 <.001 .283 <.001

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7 Competitive Intensity 135.896 28.727 .006 .033 -.008 .170
.282 < .001 .113 < .001

8 Average market 1.114 0.016 .027 -.004 -.009 .027
proliferation < .001 .451 .092 < .001

9 Recent competitive 18.652 7.361 .112 .041 -.010 .038
HMPD < .001 < .001 .055 < .001

10 Recent competitive 16.213 8.025 .038 .128 .027 .173
VMPD < .001 < .001 < .001 < .001

11 Market Expansion -0.004 0.064 .038 -.033 -.013 -.175
< .001 < .001 .014 < .001

12 GDPPC 17.512 1.668 -.008 .059 .031 -.080
‘ .123 < .001 < .001 < .001

13 Dollar/Yen 151.069 50.252 .006 .008 .025 -.094
.304 .149 < .001 < .001

14 Crude oil price 12.305 6.127 .018 .035 .050 -.076
< .001 < .001 < .001 < .001

15 U.S. population 254.532 18.644 -.026 —.091 -.025 .007

 

 

< .001 < .001 < .001 .154

 

 

 

 

 

 

 

 

33

 

Table 1 Continued

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Correlations
5 6 7 8 9 10 11 12 13 14
l
2
3
4
5 1.000
6 .977 1.000
<.001
7 .013 .008 1.000
.012 .135
8 .031 .033 .069 1.000
<.001 <.001 <.001
9 -.002 .000 .254 .174 1.000
.708 .939 <.001 <.001
10 .006 .008 .420 -.081 .302 1.000
.263 .162 <.001 <.001 <.001
11 .002 .008 -.124 .098 .165 -.381 1.000
.660 .119 <.001 <.001 <.001 <.001
12 -.006 .001 -.557 -.132 -.l73 .014 .120 1.000
.247 .884 <.001 <.001 <.001 .011 <.001
13 -.004 .006 -.709 -.025 -.140 -.l68 .273 .779 1.000
.455 .289 <.001 <.001 <.001 <.001 <.001 <.001 -
14 -.002 .005 -.710 -.120 -.075 -.009 .142 .705 .883 1.000
.671 .313 <.001 <.001 <.001 .084 <.001 <.001 <.001
15 -.003 -.012 .315 -.079 -.080 -.285 -.114 -.863 -.755 -.689
.634 .023 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001

 

 

 

 

 

 

 

 

 

 

 

34

 

 

A COMPETING HAZARDS MODEL OF MARKET POSITIONING DYNAMICS

In order to identify the significant differences in the effects of the covariates on
the probabilities of a change in the horizontal market positioning at a given point in time
towards the marketing segment center or away from the center, a semi-parametric Cox
model with competing risks was employed. A similar model was utilized to analyze the
difference in the covariate effects on the propensities to position vertically higher or
lower.

In order to estimate the competing risks model, the overall hazard rate of having a
horizontal market positioning change at time t is broken down into two different rates

(Hachen 1988):

 

lh(tlt)= lim P(t,,sT<t,,+At,H=h|T2t,,)
" At—+0 At

HEM

where, h = 1 indicates horizontal market positioning changes towards the market segment
center and h = 2 indicates horizontal market positioning changes away from the market
segment center.

Similarly, the overall hazard rate of having a vertical market positioning change at

time t is broken down into two different hazard rates:

35

 

lim < = >
5v(tltn)= P(t,,_T<t,,+At,V v|T_tn)
At—-)0 At

elm}

where, v = 1 indicates increases in the vertical market positioning and v = 2 indicates

 

decreases in vertical market positioning.
As such, the competing events are mutually exclusive and collectively exhaustive,

with:

Zi”(r|r..>=i(tltn)
h

25v“ I tn) = 5 (t I tn)

Similar to the pooled hazard rate model, the dataset will be split in monthly Spells.
Variables will be updated monthly or afier each event. The agent at risk is defined as the
brand-market segment combination. A product is considered at risk the day the focal
brand is first present in the market segment or the day after brand’s market positioning

change in the same segment.

Market Positioning Dynamics

Horizontal market positioning changes of the focal brand are evaluated in
comparison with the market segment center. The market segment center is defined in
terms of the average positioning in the product space on the horizontal dimension of the

four models with the largest market share in the market (Dobrev et al. 2002). Hence, the

36

horizontal market positioning dynamic is recorded as two types of events in terms of the
direction of the change of the average wheelbase of the focal model at a given point in
time: towards the market center, and away from the market center.

Vertical market positioning changes of the focal product at a given point in time
are recorded as two separate events also, in terms of whether the vertical market

positioning increases or decreases.

Covariates

The covariates of the competing hazard rates models include the variables
employed in the pooled models. Besides these variables, the direction of recent market
positioning changes will be captured via dummy variables. As such, recent HMPD are
recorded as two different variables: recent HMPD towards the market center, and recent
HMPD away from the market center. Similarly, dummy variables for recent vertical
market positioning are coded separately for market positioning increases and decreases.

The direction of change has been incorporated in the environmental drivers too.
The total market competitive HMPD has been recorded as two different variables: one
reflecting competition’s recent HMPD events that moved them towards the market center
and one reflecting competition’s recent HMPD event that were directed away from the
market center.

Further, in the competing risks models, recent HMPD events of the focal product
have been included as the covariates of VMPD competing hazards to test Hypotheses 7a
and 7b.

The descriptive statistics of the covariates employed in the competing risks

models are presented in Table 2.

37

 

Table 2: Descriptive Statistics for Covariates Employed In The Competing Risks

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Model
Std. Correlations
Covariates Mean Dev. 1 2 3 4 5
1 Recent HMPD, away 0.076 0.267 1.000
from MC
2 Recent HMPD, 0.072 0.261 -.090 1.000
towards MC <.001
3 Recent VMPD, 0.091 0.289 .172 .090 1.000
upwards <.001 <.001
4 Recent VMPD, 0.042 0.202 .005 .059 -.060 1.000
downwards .391 <.001 <.001 .
5 Performance 2.705 139.196 .018 -.005 -.007 -.004 1.000
increase <.001 .352 .258 .497
6 Performance 0.096 0.156 .009 -.027 .077 -.008 -.012
decrease .1 1 1 <.001 <.001 .173 .018
7 Product 1.125 0.379 .102 .174 .012 .023 -.006
proliferation <.001 <.001 .044 <.001 .236
8 Product 1.410 1.368 .099 .179 .007 .023 -.006
proliferation 3L <.001 <.001 .207 <.001 .283
9 Competitive 135.896 28.727 -.048 .055 -.016 .070 -.008
Intensity <.001 <.001 .005 <.001 .1 13
10 Average market 1.114 0.016 .008 .028 -.019 .016 -.009
proliferation .161 <.001 .001 .007 .092
11 Recent competitive 9.723 4.660 .047 .012 .061 -.034 .000
HMPD, towards MC <.001 <.001 <.001 <.001 .996
12 Recent competitive 9.760 5.147 -.017 .025 -.039 .041 -.021
HMPD, away from MC .002 <.001 <.001 <.001 <.001
13 Recent competitive 5.390 3.742 -.040 .028 -.026 .042 -.008
VMPD, downwards <.001 <.001 <.001 <.001 .178
14 Recent competitive 11.248 6.737 .038 -.007 .128 .008 .037
VMPD, upwards <.001 .236 <.001 .170 <.001
15 Market Expansion -0.004 0.064 .053 -.003 -.026 -.021 -.013
<.001 .567 <.001 <.001 .014
16 GDPPC 17.512 1.668 .048 -.058 .129 -.025 .031
<.001 <.001 <.001 <.001 <.001
17 Dollar/Yen 151.069 50.252 .066 -.058 .056 -.037 .025
<.001 <.001 <.001 <.001 <.001
18 Crude oil price 12.305 6.127 .063 -.039 .079 -.028 .050
<.001 <.001 <.001 <.001 <.001
l9 U.S. population 254.532 18.644 -.069 .035 -.133 -.005 -.025
<.001 <.001 <.001 .425 <.001

 

38

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

So.V So.V So.V So.V So.V So.V So.V So.V So.V So.V mmo. #3. v2.
ooo.- mm“..- mood- #26. wow: mag. 0mm. 23.. ohor 2m. NS... moor So. 3
So.V So.V So.V So.V So.V So.V So.V So.V So.V QM. So. So.V
ooo._ mww. mono N36 mmm. wmmr mhmr Sm. om? oz..- woo. woo... choc o—

So.V So.V So.V So.V So.V So.V So.V So.V owm. mmv. So.V
ooo._ ohho mumo mam ommr oomr com. mmor ooh: ooo. voor woof h—
So.v So.V So.V So.V So.V So.V So.V vww. new. So.V
ooo._ omfio com. vmmr moor Ev. mm? hmmr So. ooo.- ooo.- o—
So.V So.V So.V So.V So.V So.V o2. ooo. So.V
ooo._ owmr oomr ooo.- mom. woo. vmfi- woo. moo. mtr mu
So.V So.V So.V So.V So.V of. Sm. So.V
ooo._ moor mvor mum. $7 muo. woo. ooo. m2. 3
So. V So. V So. V ooo. one. wS. So. V
ooo._ mom. ommr 3o. Cm. moo... So. Nwo. Mm
So. V So. v So. V mmm. :m. So. V
oooA a _. SN. mom. So.. So... omo. N—
So. V So. V 3%. to. So. V
oooA omo. mom; voo. So... Nvor — —
So. V So. V So. V So. V
oooA ooo. mmo. So. umo. ou
m2. NS. So.V
ooo._ woo. mS. o2. o
So.V So.V
oooA who. omor o
So.V
ooo.— ooo.- b
w— 2 o— 2 3 m— Nm 3 on o a b o
gown—9.30

 

 

 

62.5230 N Saab

39

A PERSISTENCE MODEL OF PERFORMANCE IMPLICATIONS OF MARKET
POSITIONING DYNAMISM

Long term time series techniques provide the adequate means to identify the effects of
marketing actions on market performance (Dekimpe et al. 1999a). As such, cointegration
analysis is a frequently used method in persistence modeling to determine whether a
long-run equilibrium relationship exists among a set of variables (Dekimpe and Hanssens
2004)

The identification of whether a series is stationary or non-stationary is a crucial
step in the analysis. Stationary series revert to the same mean over time with only
temporary deviations. On the other hand, non-stationary series are evolving with no
tendency to revert to the original mean (Dekimpe and Hanssens 1995). Structural-break
unit root tests are employed to identify whether or not the non-stationary condition is met.
Once the revenues are determined as non-stationary, cointegration can be used to
establish the short term and long term effects of market positioning dynamism. Although
some performance measures such as market share and sales are generally stationary
(Dekimpe and Hanssens 1995; Dekimpe et al. 1999a), an analysis of moving windows of
time revealed significant evolution in brand performance over shorter periods (Pauwels
and Hanssens 2003).

In order to identify the effects of abrupt changes, which is the case for new
product introductions, repositioning and product culling, vector autoregressive (VAR) or
vector error-correction (VECM) models can be employed under the assumption that these
impulses do not change the structure of the data generating process (Dekimpe et al.

1999a). The utilization of VECM is contingent on the identification that a long run

40

equilibrium among the series exist via the cointegration test (Dekimpe and Hanssens
2004). If the cointegration test fails, a VAR model in differences has to be employed. A
VAR model in levels is adequate when the unit root test fails.

In such a model, the feedback loops between horizontal market positioning
dynamism and vertical positioning dynamism have to be accounted for. A successful
horizontal market repositioning may increase the differentiation and decrease the
competitive pressures enabling the brand to charge higher prices. Successful new
products introductions can create abnormal returns for the brand, broadening the resource

base, that can the utilized in further new product introductions (Pauwels et a1. 2004).

Variables

Horizontal and vertical market positioning dynamism are differentiated as erratic
and consistent. The variables are defined as set of dummy variables that is coded as one
for erratic change, and one for consistent change, while not change will constitute the
base case. For horizontal market positioning, erratic change is defined as change in
market positioning that is not consistent in direction with respect to the market center.
On the other hand, consistent horizontal positioning change is defined as a market
positioning change that is in the same direction as the last change, compared to the
market center. For example, if a brand’s horizontal positioning in a market center
changed towards the market center while the last change exercised by the same brand in
the same segment was away from the market center, the change is recorded as erratic.
Erratic and consistent vertical market positioning dynamism are defined in the similar

way, but the rule is adapted to increases and decreases in vertical market positioning.

41

 

Hence, a brand that increased its average manufacturer suggested retail price at a give
point in time by at least 5 percent while the last vertical positioning change of the same
brand in the same segment was decreasing its relative price is recorded as an erratic
vertical market positioning change.

Brand’s market performance is defined in terms of its total market revenue in the
given time period. Also, industry competitive intensity, total market dynamism, and total
market expansion are included in order to control for their effects on the horizontal and
vertical positioning dynamism and market performance of the focal brand. Additionally,
exogenous factors including the U.S. Gross Domestic Production per capita, and the Yen-
Dollar exchange rate are controlled for.

For each variable and brand, a time series is constructed encompassing a span of
23 years. Each time series consists of 276 monthly observations and is in line with extant
research in terms of data availability for estimation (e. g., Dekimpe et al. 1999b).

Because the variables are measured on a monthly bases the findings may be
subject to aggregation bias (Pesaran and Smith 1995). Yet, monthly aggregation is
common practice in extant research (e.g., Dekimpe et al. 1999b), and parallels the

information availability of the decision makers in the automotive industry.

42

 

CHAPTER FOUR

ANALYSIS AND FINDINGS

POOLED HAZARD RATE MODEL ESTIMATION

Wflgard rate estimationufliigures 2 and 3) indicate‘a non-monotonic
fiastienal- £9911. 9292.13m9f95 .,.b<>th...the- .hoazqmgi..m.-..venical-..mgrlsgtgggtioning
dynamics. Given the non-monotonic shape of the estimated hazard rates and in line with
similar studies (Carroll and Hannan 1989; Mitra and Golder 2002), we employ the Cox

model to identify the effects of different covariates on the hazard rates of horizontal and

vertical market positioning dynamics. Hence:

Mt Itn)=77/1(t'tn)expifl,1A(t)]a t>tn

5(t I tn)=776(t‘tn)exp1fl6 A(t)], t>tn
where 77,1 ( t — tn ) is the unspecified baseline hazard for horizontal market

positioning change, 170, ( t — tn ) is the unspecified baseline hazard for vertical market

positioning change, A ( t) is the vector of covariates evaluated for the focal product at
time t, and fl; and fl 5 are the coefficient vectors for the effects of the covariates on the

propensity to change the horizontal market position and on the propensity to change the

vertical market position respectively.

The Results of the Pooled Hazard Rate Models
Stata’s survival time module has been utilized to perform the analyses. The

effects of the covariates on the hazard rates will be estimated using the Efron (1977)

43

partial likelihood approximation method to deal with the event time ties; The standard
errors have been adjusted for clustering on each product line. The estimated parameters
can be seen in Tables 3 and 4.

The covariates have been included in the analysis on a hierarchical basis, starting

with a specification in which only the control variables have been employed and ending

with the full model spec1ficat1on The parameter estimates are relatively stable across

a ..-._....p..__._.,--- "»u M... -w
Hm -y-,...‘.... “ ——--.-o _,._fi., m“) p“..— W'

different specifications in terms of magmtude and significance, indicating that the

..A....-
7‘”. N51“. H!“- ~ _..

 

 

 

 

 

 

”ugh.-.“ J’M-w “M.
Mflm ”—.me " ->-—., o...

 

*UW

In the case of the horizontal market positioning dynamics, the results indicate that
the product lines with a recent change in the horizontal market positioning dynamic have
a significantly higher propensity to exhibit HPMD (B = 3.630, 2 = 3.53, p < .001),
supporting Hypothesis 1a. While the performance increase has a significant positive
effect on the hazard rate ([5 = .00028, 2 = 10.16, p < .001), the performance shortfalls do
not have a statistically significant effect ([3 = -.043, z = -.14, p > .10). Therefore, the
results strongly support Hypothesis 2a but fail to provide support for Hypothesis 2b. The

proliferation of the product line has the expected non-monotonic effect, with the linear
component having a statistically significant positive effect ([3 = 1.561, 2 = 6.58, p < .001)
and the quadratic component having a statistically significant negative effect on the
HMPD hazard rate (B = -.l71, z = -3.78, p < .001), indicating strong support for

Hypothesis 5.
Afier controlling for macro economic factors, the environmental drivers have

notable effects on the HMPD hazard rate. The competitive intensity has a statistically

44

significant and negative effect on the hazard rate (B = -.017, z = -2.02, p < .05), rejecting
Hypothesis 8. On the other hand, the average market product proliferation (B = 113.924,
2 = .27, p > .10) and its quadratic component (B = ~52.006, z = -.28, p > .10) do not have
statistically significant effects failing to support Hypothesis 9. Contrary to expectations,
recent competitive HMPD has a statistically significant and negative effect on the hazard
rate (13 = -.O24, z = -3.23, p < .001), rejecting Hypothesis 11. Finally, in contradictory to
the expectation of Hypothesis 13, market expansion has a statistically significant and
positive effect on HMPD (B = 2.166, 2 = 2.47, p < .05).

In the case of the VMPD, the results indicate that afier controlling for the macro
economic factors, the product lines that had a VMPD in the last year are more likely to
exhibit VMPD (B = 4.256, 2 = 4.13, p < .001). This result provides strong support for
Hypothesis lb. Yet, neither performance increases ([3 = -.044, z = -.76, p > .10), nor
performance decreases ([3 = .47, z = 1.48, p > .10) in the previous year have significant
effects on the propensity to exhibit VMPD, failing to provide support for Hypotheses 3a
and 3b.

The competitive intensity has an insignificant positive effect ([3 = .005, z = .60, p
> .10) on the propensity to exhibit VMPD, failing to provide support for Hypothesis 10.
Recent competitive VMPD has a negative and statistically significant effect ([3 = -.022, z
= -2.26, p < .05) on the VMPD hazard, contradicting the expectations of Hypothesis 12.
Finally, market expansion has a positive and statistically insignificant effect ([3 = 1.04, z
= 1.07, p > .10) on the VMPD hazard rate, failing to support Hypothesis 14.

Further, several shared frailty modelsmhayebeen estimated to test for unobserved

“WW-- ~" '- ‘W ”a... \.m a. "Wwp

 

 

effects at the brand and manufacturerflleve‘l.‘ As such, any omitted brand level and

fir ' “MM-'“w- . ‘mtr'h‘.‘hw..fi-
‘-

45

manufacturer level factors that have not been included in the estimation can be
statistically tested and controlled for. The results of these estimations can be seen in
Tableds 5 and 6 for the horizontal and vertical MPD respectively. Two different shared
frailty models have been estimated for each hazard rate: brand level, and manufacturer
level. The Wts (theta) are only marginally significant in all the
shared frailty models tested (p values ranging from .002 to .05), indicating that the brand
and manufacturer level effects not accounted for in the covariates are relatively limited.
Further, the coefficient estimates for the covariate effects are relatively stable in terms of
magnitude and significance between the ordinary specifications and shared frailty
specifications, indicating that the results are robust to unobserved brand levelnand

manufacturer level heterogeneity.

 

Further, in order to test the robustness of the results, the models where estimated

using a two_ year time lag between the covariates and the event. The estimated

" "”1,pr W’-

 

 

 

parameters for the horizontal and vertical market positioning dynamics are presented in

Table 7 and 8 respectively. The effects generally decay over time, and some coefficients
become insignificant. Yet, recent HMPD (B = 1.08, z = 5.26, p < .001), product
proliferation (linear component: B = 2.07, z = 6.30, p < .001; quadratic component: B = -
0.33, z = -4.21, p < .001) and competitive intensity (B = -0.019, 2 = -2.29, p < .05) still
have a significant effect on the HMPD hazard rate. In the case of VMPD, recent change
in base MSRP (B = 0.82, z = 3.71, p < .001) and recent competitor VMPD (B = -0.023, 2

= -2.3 8, p < .05) still have significant effects on the hazard rate.

A summary of the results along with the related hypotheses can be seen in Table 9.

46

A3395 2:: mam—«Em

com com one oo_. on o
L _ _

I

moo.

:03.

w '20.

omo.

I

mNo.

omo.

 

 

38 “:33,”

as: .8 83.5% 33— 2.3: .5585 a 95$...

47

ooN

Amazons was mmmbfia
om _. oo —. on o

woo.

oS.

 

NS.

88 23am

aux—E .8.“ 82::th 33— Eanal 358:5 “m 0.5me

ooo.

48

.\./

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.8%- 2%. 8.3g- Boéeaéuaa mod
53v 52v Sodv .5 Eu A
C o v ufi
mdom ménm mm.mm .om ED 255
o3 o3 ovv ”magma
3 ME ME “8935.
wmme wmme wmme “22320.30
Sodv onod- nwwood anmod- :od oovd- onood voN—od- oood owed- o—mood vmo—od- :owflzaom .m.D
ommd oomd nooSd _ommmo.o mnod own; omowod ovouod mmod oo—d mnde m—mmod 025:0 09:0
eood onod- vnNood monood- wmmd SNT Smood mmmood- ondd onmd- oomood wwmood- :BCBEQ
onod omud- wmoood ooSNd- Vmod omvd- oSood ooomod- o—od onm.~. nmomod nvom—d- GREGG
m—od omvd oonnod oonfid :ommcamxmuoxbfiz
Sodv o—nd. Snood nnMNod- DEE 03:59:00 :58
momd omod- ogmmd Nmemd- couwuogoaooxsfi owfio>
vvod oSdu Snood mvoSd- 36:85 0352580
Sodv ownd- o—mvod ononfio- Sodv onvd- ommvod nnnm—d- .om :ozfloEoE 8%on
Sodv oomd vooMNd nmooma Sodv ommd memmd omoomé scrunch—0E «2.68m
mood omdd- ovmdmd vivod- onod oddeanomd Snmod- 888% oogscotom
Sodv omod~ mooood wNoood Sodv ovmd— mooood nNoood 8885 oonanuotom
Sodv ommd Sonoé vmoNod Sodv omvd nomNoA o~o~md Dam—2E 880%

_~_Am N .._._m “—30 _N_Am a sim— uocU _N_Am N .....H H80 32:..ch

.3m .3m .3m
3.:ch .259: 2.59%

 

 

 

macuaofiuomm 82:82

 

 

 

9&2: «BE—NH 3 53:23.5 2: .5 magnum.“ 83.3.50 no 335th ecS—zS—E 15.....— um 035,—.

s

49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

33.- $2- $8. 82...»...582. mo.
.8. v .8. v .8. v em .5 A N.
o. N 4 u...
0.8 8.... 43.. .5 .6 2......
8m 8m 8m ”35....
So So 8... a.8....=m
am: an; .3: H2......E..§.o
2.... .33- 2...... 3.8...- 2...... .53- E8... 93.....- ......v 8...... 82...... .38...- 8.5.32. .m:
4.3... .08...- :08... 338...- MN... 8m..- .32.... wows..- .82 OS..- 36...... N25..- 8...:6 0.56
2...... $2- 28.... «$8...- 8.... 9e..- 28.... 38.....- 2...... .53- 22...... 38.....- =...>..~..on.
N8... 80.. 88.... ”at... 2...... .53 .23... «$2... 8m... ca... 338... was... 8.50
am... 2.... N38... N38. consaxmfias.
VS... .53- $8.... 3.8...- 22>
DiamaanO «Gooey:
an... OS... .38.... $8... b.m=o.=.o>....&eou
mm... 8.... Sam... $8.... $2 8... 82m... 8.3m... 3.3.8.U 85.883...
.3... 82.. N38... 82...... on... an..- 268... have..- assaooééotom
......v .5... man... mama... ......v 2...... 8.8.. 3.5.... 235.88
EA. N Em .afl .35 EA. N a... .5 .80 EA. N .5. ....m .80 8:550
3.53— umanom .253—
mneuaofioonm 82:82.

 

 

 

52> .33.... s 282...... 2.. .5 38.... 83.326 .9 835...... 3.5.2.... 3...... a. 2.39

50

Table 5: Partial Likelihood Shared Frailty Estimates of Covariate Effects on the
Propensity to Exhibit HMPD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IC Brand Level Manufacturer Level
ovariates Coef. Std. Err. z P>|z| Coef. Std. Err. z P>|z|
[Recent HMPD 3.61973 1.01563 3.5601<0.001 3.63590 1.01560 3.580 <0.001
E’erformance
'ncrease 0.00027 0.00013 2.020 0.043' 0.00026 0.00013 1.990 0.047'
Eerformance
ecrease -0.04910 0.31384 0160 0.876 -0.08277 0.31372 0260 0.792
Eroduct
roliferation 1.56796 0.35387 4.430 <0.001 1.57393 0.35404 4.450 <0.001
Eroduct
roliferation sq. -0.l7776 0.08824 -2.010' 0.044I-0.18212 0.08822 -2.060 0.039l
IlCompetitive
ntensity -0.01654 0.00792 -2.090| 0.037 -0.0l736 0.00793 -2.190 0.029I
Average market
roliferation -2.22403 3.67883 -0.600 0.545 -2.22566 3.67977 -0.600 0.545
lRecent
competitive
ll-IMPD -0.02355 0.00886 -2.660| 0.008l -0.02321 0.00885 -2.620 0.009l
arket
xpansion 2.13121 0.85485 2.490 0.0131 2.13217 0.85406 2.500 0.013]
GDPPC -0.19838 0.09727 -2.040 0.041 -0.l9634 0.09725 -2.020 0.0441
ollar/Yen -0.00795 0.00278 -2.860 0.004I-0.00789 0.00278 -2.840 0.00
Crude oil price 0.02407 0.02033 1.180 0.237 0.02317 0.02033 1.140 0.254
US. population -0.03537 0.00988 -3.580 <0.001-0.03621 0.00990 -3.660 <0.001
theta 0.01832 0.06509 0.03515 0.02409
Chibaqu (1) 2.720 0.050 8.370 0.002
Observations: 34538 34538
Groups: 40 29
Subjects: 738 738
ailures: 440 440
Wald Chi Sq. 148.1 146.8
d.f. 13 13
P > Chi Sq. <0.001 <0.001
[Log-likelihood -2401 .84 -23 99.02

 

 

 

 

51

Table 6: Partial Likelihood Shared Frailty Estimates of Covariate Effects on the
Propensity to Engage in VMPD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Brand Level Manufacturer Level
Std. Std.

Covariates Coef. Err. z P>|z| Coef. Err. z P>jz|
Recent VMPD 4.23164 1.01970 4.150I <.001 4.233601 1.01961 4.150I <.001
Performance
increase -0.03857 0.05683 -0.680 0.497 -0.04069 0.05789 -0.700 0.48
Performance
decrease 0.49337 0.31529 1.560 0.118 0.49789 0.31458 1.580 0.113
Competitive
Intensity 0.00351 0.00846 0.410 0.678 0.00338 0.00847 0.400 0.690
Recent
competitive
VMPD -0.02178 0.00982 -2.220| 0.027'-0.02152 0.00982 -2.l90| 0.0281
Market
Expansion 0.98399 0.99488 0.990 0.323 1.00661 0.99590 1.010 0.312
GDPPC 0.17792 0.09161 1.940 0.052 0.17623 0.09160 1.920 0.054
Dollar/Y en -0.00890 0.00339 -2.6301 0.009 -0.008801 0.00338 -2.600 0.0091
Crude oil price 0.00109 0.02275 0.050 0.962 0.00009 0.02273 0.000 0.997
U.S. population -0.03374 0.01136 -2.970 0.0031-0.03344 0.01134 -2.950 0.00
theta 0.07171 0.04309 0.06174 0.04524

Chibar Sq (1) 6.050 0.007 5.240 0.01 1
Observations: 34591 34591
Groups: 40 29
Subjects: 666 666
Failures: 366 366
Wald Chi Sq. 97.07 96.59

d.f. 10 10

P > Chi Sq. < .001 <.001
Log-likelihood -1981.10 -1982

 

 

52

Table 7: Partial Likelihood Estimates of Covariate Effects on the Propensity to
Exhibit HMPD: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Robust

kovariates Coef. Std. Err. z P>|z|
[Recent HMPD 1.07707 0.20476 5.260 <0.001
Erronnance increase -0.03383 0.05237 -0.650 0.518
[Performance decrease -0.04051 0.31168 -0.130 0.897
Ilfioduct proliferation 2.07343 0.32891 6.300 <0.001
[Product proliferation sq. -0.33390 0.07930 -4.210 <0.001
Competitive Intensity -0.01871 0.00818 -2.290 0.022
Average market proliferation 3.79723 3.8321 1 0.990 0.322
ecent competitive HMPD -0.00187 0.00885 -0.210 0.833
1Market Expansion 0.10793 0.82885 0.130 0.896
GDPPC -0.16516 0.10755 -1 .540 0.125
ollar/Y en 0.00242 0.0029] 0.830 0.405
Crude oil price -0.058741 0.02460 -2.390 0.01fl
US. population -0.03169 0.01064 -2.980 0.0031
Observations: 3091 3
Subjects: 687
ailures: 409
Wald Chi Sq. 304.57
d.f. 13
P > Chi Sq. <0.001
|Log pseudo-likelihood -2182.78

 

 

 

53

Table 8: Partial Likelihood Estimates of Covariate Effects on the Propensity to

Engage in VMPD: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

lC Robust
ovariates Coef. Std. Err. z P>|z|
[Recent VMPD 0.82373 0.22194 3.7101 <.001
[Performance increase -0.25330 0.18263 -1.390 0.165
firformance decrease 0.13616 0.40982 0.330 0.740
Competitive Intensity 0.00716 0.00963 0.740 0.457
[Recent competitive VMPD -0.02318 0.00973 -2.380| 0.017'
FMarker Expansion -O.85792 0.88402 -0970 0.332
GDPPC -0.13467 0.10427 -1.290 0.196
ollar/Y en 0.00700 0.00375 1.870| 0.062
Crude oil price -0.08430 0.02765 -3050] 0.002
.8. population -0.04466 0.01432 -3.120I 0.002
Observations: 3091 9
Subjects: 598
iFailures: 322
Wald Chi Sq. 81.31
d.f. 10
P > Chi Sq. <.001
[Log pseudo-likelihood -1699.55

 

 

 

54

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

votoamsm HoZ .uogoovm md - + - :ommcmmxm 6x32 3 .2
cocoon - + 22> .980 S
coco? - + 995 .980 :

votomqsm “oz md - “505 deg gov—.82 o
wotomasm SZ .voaooflom md + r + .3652: 3:33:50 3 .w
cocoaqsm C C 8:803on m
votommsm HoZ .wotomasm “oz .0: + .md + 030.8% “tom pm 5N
uncommnm “oz .wotommsm .m.: + + + 3385 “tom am am
cotommsm + + nag/S Eooum fl
wotommzm + + DEE Sound 3
DE); DAG/S 98:23 DE:

:0 Soto co Bebe co “auto :0 Soto

roamsm £85095 cofimfiumm wowoomxm onmEumm wouoaxm oBaE> £35095

 

 

.352 3.35“: 528.— 2: 58 3:53— 2: no bafifiam "a 039—.

55

COMPETING HAZARD RATES MODEL ESTIMATION

Similar to the pooled hazard rate estimation, the competing risks semi parametric

Cox models are specified as follows:

4”(t|t.)=n’i<t-tn)exp1/I’i A101, t>tn

5v(t|tn)=77:§(t'tn)expiflrs A(t)], t>tn

where 172 (t- tn) is the unspecified baseline hazard for horizontal market

positioning change, I]; (t-tn ) is the unspecified baseline hazard for vertical market
positioning change, A ( t) is the vector of covariates evaluated for the focal product at
time t, and ,6 Z and ,6 X are the coefficient vectors for the effects of the covariates on

the propensity to change the horizontal market position and on the propensity to change
the vertical market position respectively.

The Efron (1977) partial likelihood approximation method was employed to
estimate the effects of the covariates on each of the competing hazard rates independently
(Hachen 1988). Also, the standards errors were adjusted for clustering on each product
(i.e. car model). The partial likelihood estimates of the covariate effects on the hazard
rates of HMPD towards the market center (MC), HMPD away from the MC, downwards

VMPD, and upwards VMPD are presented in Tables 10, 11, 12 and 13 respectively.

56

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ease. _- 8S. _- eoeéofiwoeaoa m3

.8. v a... v lam Eu A a

m. .4. u...
5... am 90.43 .wm ED 285
Em Em ”33.8“.
”E woe ”mzoomem
m . mmm m . mmm ”30.62030
2.... v 3.....- m.m...... 33......- ...... v 3...? 22¢... 38.....- 5.52.2. .m.D
mow... own... ovao... vmmmo... mow... own... 33...... Nmmmo... 8E :0 ova-5
3...... :2.” own-2.... 25......- mmc... ... ..N- mung... b...........- cot-5:09
...... v 30m- mam... Some...- ...... v 30m- wamm.... amen}.- Oman—O
2...... ..~e.~ SEN. ammajr- w....... one.~ mac—NA van—d 56.896. .8132
2.... v cane“- ...w...... N33...- ...... v :36.- ..cw...... 33......- 0.2
ES. $26 .995 0352.88 E83.
mg... ..v~.- ~53... 39.....- 52.... ..N~.N- 3.2.... 3.2....- 0.2
$538 HES... 035388 2.83..
cum... co... .- ooo.-m6 933.c- onmd ow... .- .4936 oomoflo- 8:80.20... 8:38 umweo><
m8... ..m~.- $2.... 33.....- 3.... can.” 8mg... mung..- bficofi. 0352.800
5...... cg..- 3mm... macad- av..... Era..- ....wm.... venom..- .wm coups-EOE .26on
...... v ....~.m 33m... .53.”.— S... v Saar. 33m... 3mm... 22.80....838305
am . .o cm... 5.9.... onmmm... .3 .o one. am .3... 38m... 335% ooggoton.
2.... v and v........... mm......... 5... v ..mm.a v........... mm......... 0305:. ooggoton.
m 5... o. ..o x... mm... Some... 03. €838 HES... E83.
EA. N «Em— .Ew ”.30 EA u Em .3m ”.80 323.550

.253. .253.
33.358. m 83:82

 

.850 3:2 2:22:- 22: .323. 3 £2.23... 2.. no 38.... cateeeu he ceases. coca—3:.— Etam a... 2.3.

57

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.3. .- ...MN. .- 82:8...532. we.

...... v ...... v ...m ...o A ..

m. ... ....
mm... ....8 ...m ...o .25..
...N ...m H5......
...: ...: "38...;
m. 2. 2mm ”East-.30
8m... .3...- o.m...... 39.....- E... 9......- ..m...... a. .....- 8.3.2.2. .m...
3.... ..mg 8.2.... «.32.... .8... ....2 3.8.... 3.2.... 8.... ..o 0.56
2...... ......N- 3.3.... $2....- ......... ...-3- $25... 82.....- 5&539
8.... ca...- .§.... 82.....- mmm... ca..- 82.... $8....- 8.50
.3 . ... can. 83.... SN .... ...N... ....N. . . ... . .. Oman. soufimxm. Ea...
...... v ......m- 32.... 3......- ...... v ......m- $3.... 35.....- 0.).
EB. ..an HES... 03.89.80 833.
3... ....a... 28.... R .8... 8.... 8.... was... 88.... 0.). $339 ......2: 32.2.88 .58..
E... ..No... .....wa ... .00... gm... ..8... 3.9% ...-R; S..§...e.. .812. omega...
2...... ...... 5...... 22.... 3.... 2...... 32.... .88... £83... 03.3.88
8N... ...-....- aom... $8....- wmm... 8....- ES... 89.....- ...m 8.3.056... .88...
...... v ...... ...-.8... 35.. ...... v .53 .32... .88. 8.35:6... 652..
”mm... 93.... oo.o..... .89....- m..m... cg...- m.§.... 3......- assoc 8:28....
...... 8.....- 82.... N... ......- mcc... 9......- .. mg... 22.....- 0322.. 8525......
...N... .8... ”Sam... fig...- 02 so... .93 NE)... ...83.
EA. u ......— ....w .80 3A.. N ...-.m— ....m ...ooU 33.3.50

.253. .353—
mno..3u.oo. m 82:83.

 

 

 

.850 «3.3—... 8...... mafia GAS... ...—Eu.— 8 b.3099... o... E. 83.....— ouatago .... 82:53..— ..ooaaoxfi 35...... u: 2...;

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ema- 2.3m- 8053:5803 m3

Sod Sod .5 50 A m

S 2 .3
2.8 3% cm Eu 235
NS NS ”325
3m ea ”$8.35
ER. 32 “823530
33 $3- 883 3:3- SE 33. 0385 823- 8:238 .md
53 $3- $23 :83- Sad 83- SE3 383. 8% no 830
82 33- $83 $83- a Go 82- $83 883- autism
cad owed- 3&2 Ems? BS 83- 28% 523. 830
$3 8 H A $53 £33 ”Rd 2.2 SSE 8o :2 aoaamxm 332
82 23. SSE $83.. 22 8o. _- mg: 283. mega: .935 2:538 58%
GS 82- 823 Road- 83 83- 323 368.0. mosaaaou
dam); guacafig Euoom
m8... 8: as... was... 33... 35 5:3. «.38... £225 255%8
$3 83 332 $23 89o S 2 883 828 885% 85:83;
33 SM. H- 82 3 2 $3- 83 EN. 7 RS 3 maid- 0382: oofifiotom
32 ans. V893 SEQ- $3 2 2- 89:3 833.? 0.2 2:999 “9% 880m
83 $0. _- :aas 8:3- E; ES- mam}. £53...- 02 80¢ 3% .995 28%
as... 3.3- :22 3:3- musacaou .995 Begum
EA.— n .....m .3m ”.80 EAm N .....H .Em ..«ooU 32:..ch

~253— .253—
mnctauumuo. m 3:582

 

 

 

Gan—>5 mar—955:5 «Swim— 8 Emanoncum 95 .5 38b3— oaatago a: 32::th 62:53:1— EEaA "Nu 035—.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

h _ .03 _- ”29 7 goofing-883 we;

So.V So.V .5 Eu A m

S 2 .3
8.8 2w .5 Bo 235
EN EN ”825
0% 9% ”98.3%
ES 53 ”223530
Sc... .9...“- Ems... 2:3...- 83. 35.n- Ems... 2:3...- cougaa .md
Rec 33 $86 $33 $3 8% ”886 .323 8E =0 830
mm; 82- 233 $83- 83 $3- $85 883- autism
08d 22 R: 3 @385 $2 83 Re 3 SEE 9956
as... an: 32.3 SE: «S... 2.3 :32 £33 Sugaxm 332
was 82 $23 883 $2 83 E23 3.85 €339. .995 33298 68%
88 82- 8086 «$8.? £3 82. 883 ”885. $3398 .993 39298 58%
$3 8% t _ So 533 $3 82 a: :3 283 £825 26: 88
2 S 32 $de 222 mg... 83 ~23... 382. 0.320% uofiéotom
53 82- 529° 289°- 33 83- v8.86 :83- afiss 85.8898
32 Bed. 53$ 8%;- mowd ons- Sommd SEE- 02 $333 .999 58%
3:... if. Efig ”83...- 3...: 83. 2mg... name...- 02 50¢ 32. .999 308m
:5... 33- 9.8... mag;- %§§ .993 208m
EAm N ..Em ....m ...80 EAm u .....H .Em .uocU 32:32.0

unflacm «22‘6”—
:226592 m 82:82

 

 

 

995 2:22.: main a $22.2.— 2: .5 38mm 35.55 a. 8.339 25.535 3:3 "2 asap

6O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m _ .E _- SE _- Bofiez-mo
So.V So.V .am 302
E E .3
$8 3“? .5 Eu 2“?
EN EN ”85:“;
”2 won ”983%
am ow “mmzohO
2mm 2mm ”2523530
83 22 was 83 S cm 3&5
v8.86 @856 EA: 2686 see

8.7 8%. omes... £82.- §.v 2.3. ..33... :83.- Sagaawd
SE 88 8965 @286 was 8: @885 58.0 8:36 025
N36 83. 233 883- 95... £2- .58... 325.... autism
Sci 82. «$3... $33...- §.v Eam- N-maz... $33.- Page
33. 3.2 3d: 255.». 83. $2 mmmfig p.52."- sousmxmfisz
mg... 32. SE; 52...- 2:... Sn."- ES... 3%.;- 02 80¢ 3% .92: 32358283..
83 85- 3386 33.9 ~85 83- mwv-Sd 823.c- ozmuazedmzm 32358583:
Mag 83. woSS 82$- 82 cm:- EGG ”$2.0- Sfiézoafisa omen-E
as... 82- $23. $3;- 33 $2- 323. mans...- éaasoézaeou
63 EM:- maid 55o- vood on:- E38 N83,? .wmaoueezoaséé
:56 2:“...- 2%? 2mm? :56 En.»- vmsm... 3mm»; nose-..anséoa
N26 82 5:3 052 and 23 mass 883 amass 85:83;
as... 83 «.88... v.88... 25... :8.” 22.3. 988... 03205 8§§£§
82 83 ”ammo «£85 2% 83 $35 N385 uzmwaaedazmaosg
Ea N Em ....m ‘80 EA; N Em ....m .36 8.3::
_o>ow— hohaoauflfiwz _u>01— flaw—m U—

.S.-EU now—...“:

“Ego-F as: .523. 3 5.2235 2.. .8 385 33.380 ... 8355m— EE-a 3:25 8258.: Bra.— “3 ~33

61

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2am. .- 3mm. .- 8......e...-w3
...... v ...... v ...w ...o A ..
... ... ....
mm... a? .5 Eu as...
...N ...m Um........w.._
...: ....5 ”38.35
am ow .3520
2mm 2mm n2.03%....5
3.... ..m...~ .3... ..fi... ... cm .320
.88... 82.... 2%.... 3...... so...
8.... 8.....- $3.... 52......- mm... 2.....- 83...... 3.8...- 8.5.52. .3.
m8... .5." 83.... ”:8... 3.... ...-..N .mmmmc... ....S... 8.3.... 3.5
2...... 83- $3.... 32.....- 2...... ..NS. 22...... $2....- autism
8v... .5...- m......... $2...- .2... 3.....- 38... SR...- 0&8
8N... ... ... «8%.. 3.3... ..R... .... . .. 882 m 88.. 8.88%. 332
......v .53.- ..SS... .38...- ......v ....S- 33.... 2.3....- uz as. 32.9.2: 3.52.8868”
3m... ..m... 38.... 28.... am... ..m... $8.... 38.... 22 mass... HES... 2:838 ...-83..
a... ..No... 62% ......me Sm... ..a... ER? RES 8.3556... 3...“... 0932
NS... 8.... 82.... 82.... 2...... on... 82.... 82.... £22... 2.2.2.88
E... .5... owm... .38...- em... .5...- 8...... was..- ...mao.§o...e..§.§..
m....... ..m...~ 83v... «5.»... 2...... .3..." ......3... «2?... 8.3.05.0... .260...
SM... 8.....- 23...... 2:2.- mem... ca..- 83...... Ram...- amasu 85:53..
5.... ...N...- 5.8... 82.....- 5.... SN... 2%.... 82.....- 38.2.. 85.8.83;
an... 2....- ..mmmm... agno- mR... 8.... ...-mm... 2.5....- 02 as. $5,535.58
EA.— N ...—m .15 ..«ooU EAm N ...—n.— .Em ..«ooU magi-3.5
.954 sonata-«=52 .261— ..nahm U—
.S.—00 «...—.32

as. .22 9.2: .....im a £232.. 2.. .... 38.... 23.326 ... 8.55.. 2.2-.— .Eém €2.35 .55.. “m. 2...;

62

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.8...- 8.....- 8085.88.83
8.... 8.... ...m ...o A ..
m. m. ....
8.8 8.8 ...m ...o 2.3
8. 8. ”8.2....
8. 8. 5.8.35
mm am 89:80
58.. .8... 85.8880
8.... 88.. 8.... ..8... ... ..m 8.....0
88.... 88.... 88.... ~88... .8...
8m... 88...- 888... 88......- o..... 8...... ..88... 82.....- 5........o.. .m...
.8... 8.....- 2888 5......- ..8... 8...... 88.... 88.....- 8... .... 8.5
8.... ... m...- ....8..... 88.....- .8... .5...- 888... 88.....- .5583
8.... 88...- .E...8 88.....- 8...... 88.... .8..... 88.....- 9.8...
8.... ..8. 88.... N88. 8.... ...... 8.. .... .88.. 585...... 8......
8.... ..8..- N88... 88.....- .8... 88...- 888... 88.8...- 883.... d...>.> 8....258 .58..
8...... 88...- 8.8... 28....- .8... com..- ... .8... 88......- 853.38
.ngr 03.509200 «5.000%
..8... ..8. 8...... 888... 8.... ..8. 8...... ~88... 8.5.... 8...... .50
8.... ..8... 88m... ...; .m... .8... 8m... .88... 88...... 88.8.. 858.88..
88... 88.... mi... 8......- 88... 8.....- 8..... .88....- 888... 855.38..
8.... ..8. .- 88.... .88...- .8... ..8. .- 88.... 28.....- us. 88...... d8)... .58..
...-.... ......- 388 88......- ..8... ..8..- 88..... 8......- os. 50.. 53.. .98)... 888.
8...... .83. 888... 8.8... 8.... .83. 58.8... .88... 8535.... d...>.> .58..
.NF. N .... ....m .85 EA. N ...... ...... .80 88.580
.961— ..oufiuauauaz :35:— 6:95

 

 

 

 

52> ......EEA. .....a... ... 5.28%... 2.. .... 88.... 8.5.52.0 ... 82.5....— b...... 8.25 82...»... 1...... .... 0...-F

63

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

..m... . .- ...... . .- 82......7...
......v ......v ...m ...o A ..

m. m. ....
8.8 ......- ...m ...o 2.3
...N ...N H8......
.8 8m H38.3.5
mm mm 8.9.80
......m ...-..m H........>......o
2.... .8... 8.... ..m... J: ..m 5......
28.... 8.8... ..88... ....8... .8...

2...... ......N- «8...... .58....- ......... .53- 82... 83......- 5....=..o.. w...
8.... 0...... .88... 8. ...... ...... ....m... 8.8... 8...... ...... ... ......u
...... ......- ........... 58.....- 8.... ......- 8......... $8....- 5t....oo
Rm... ..8... .88.... .88... ...... ...... 3.8.... 8.8... 9.80
...... ...... 2.8.. .88.. 8.... ...... .83.. 8.8.. confirm-.88....
8.... ...N... 8. ...... 88.... .8... .8... 8m ...... 88.... 853.... .98}... .>..........8 .58.
...... ...... ..- .88... 88.....- 8... ......- 8..8.. v85...- 853§8d§> .>.......ao..58~.
.8... 8.... .88... 8...... ...... 8.... 3.2.... 82...... .55.... 3......58
8.... =8. 8.5.... 88.... 8.... ..8. 8.8... «8:... 88.8.. 858......
8.... 8...- $8.... 28.....- 8... cm...- 8.8... 88.....- 88.8.8558...
E... .8...- 88... .88...- .e... ....m...- .83... .88...- 02.2.5.8...255...
8.... ......- 883... 8......- .m.... ......- mmmz... .83...- us. ...... ..3..m....m.=.8m
......v ....~.m- 88m... 88...- ......v ....m.m- .83.... 8.....- 853.... .2235...
EA. . ...... ....m .80 EA. . ....— ....m ...... 88.526

.961— .oafiouusna—Z .96:— ...—«.5

 

 

52> ......E... ......E. .. 3.5%.... .... ... 38.... ......aso .. 8.3.8.. ....a... 8.25 88.58.... 3...... n: 2......

64

Table 18: Partial Likelihood Estimates of Covariate Effects on the Propensity to

Exhibit HMPD Towards Market Center: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Robust

Covariates Coef. Std. Err. z P>|z|
Recent HMPD, towards MC 0.47655 0.35951 1.330 0.185
Performance increase -0.091 36 0.09042 -1 .010 0.3 12
Performance decrease 0.03044 0.47577 0.060 0.949
Product proliferation 1.90684 0.45877 4.160 < .001
Product proliferation sgL -0.25417 0.10074 -2.520 0.012
Competitive Intensity -0.03429 0.01427 -2.400 0.016
Average market proliferation -6.351 16 5.96392 -1.060 0.287
Recent competitive HMPD, towards
MC 0.00669 0.02726 0.250 0.806
Recent competitive HMPD, away
from MC 0.08746 0.02646 3.310 < .001
Market Expansion -0.84525 1.05835 -0.800 0.424
GDPPC -0.39832 0.16076 -2.480 0.013
Dollar/Y en 0.00524 0.00417 1.260 0.209
Crude oil price -0.05471 0.03664 -1 .490 0.135
U.S. population -0.05425 0.01566 -3.470 < .001
Observations: 29850
Subjects: 664
Failures: 190
Wald Chi Sq. 174.93

d.f. 14

P > Chi Sq. < .001
Log pseudo-likelihood -1000.3

 

65

 

Table 19: Partial Likelihood Estimates of Covariate Effects on the Propensity to
Exhibit HMPD Away from Market Center: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Robust
Std.

Covariates Coef. Err. z P>|z|
Recent HMPD, away from MC 0.37889 0.30565 1.240 0.215
Performance increase -0.00695 0.03246 -0.210 0.830
Performance decrease -0.07960 0.45223 -0.180 0.860
Product proliferation 2.30345 0.45555 5.060 < .001
Productyroliferation sq. -0.38623 0.10748 -3.590 < .001
Competitive Intensity -0.01151 0.01434 -0.800 0.422
Average market proliferation 6.09313 5.62563 1.080 0.279
Recent competitive HMPD, towards MC -0.02493 0.02122 -1.170 0.240
Recent competitive HMPD, away from
MC 0.00845 0.02222 0.380 0.704
Market Expansion 1.03086 1.41 100 0.730 0.465
GDPPC -0.08187 0.16846 0490 0.627
Dollar/Y en 0.00435 0.00453 0.960 0.336
Crude oil price -0.l0096 0.05109 -1.980 0.048
U.S. population -0.02869 0.01714 -1.670 0.094
Observations: 29850
Subjects: 664
Failures: 197
Wald Chi Sq. 88.49

d.f. 14

P > Chi Sq. < .001
Log pseudo-likelihood -1034.44

 

66

 

Table 20: Partial Likelihood Estimates of Covariate Effects on the Propensity to
Exhibit Downwards VMPD: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Robust

Covariates Coef. Std. Err. z P>|z|
Recent VMPD, downwards -0.19285 0.52684 -0.370 0.714
Recent HMPD, away from MC -1.04768 0.55703 -l.880 0.060
Recent HMPD, towards MC -0.29779 0.41946 0710 0.478
Performance increase 008 l 39 0.09595 -0.850 0.396
Performance decrease 0.28888 0.65133 0.440 0.657
Competitive Intensity 0.00771 0.02013 0.380 0.702
Recent competitive VMPD,
downwards ~0.07176 0.03367 -2.130 0.033
Recent competitive VMPD, upwards -0.06252 0.02502 -2.500 0.012
Market Emansion -2.38354 2.55133 -0.930 0.350
GDPPC 0.02006 0.18382 0.110 0.913
Dollar/Y en -0.01331 0.00850 -1.570 0.117
Crude oil price -0.07440 0.07663 -0.970 0.332
U.S. population -0.05354 0.02828 -l.890 0.058
Observations: 27385
Subjects: 535
Failures: . 88
Wald Chi Sq. 49.9

d.f. 13

P > Chi Sq. < .001
Log pseudo-likelihood -449.82

 

 

 

 

 

67

 

Table 21: Partial Likelihood Estimates of Covariate Effects on the Propensity to
Exhibit Upwards VMPD: Two Year Time Lag

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Robust

Covariates Coef. Std. Err. z P>|z|
Recent VMPD, upwards 0.31233 0.42914 0.730 0.467
Recent HMPD, away from MC -0.75173 0.35538 -2.120 0.034
Recent HMPD, towards MC -0.24159 0.23426 -1.030 0.302
Performance increase -0.28936 0.30584 0950 0.344
Performance decrease 0.09476 0.54698 0.170 0.862
Competitive Intensity 0.00558 0.01357 0.410 0.681
Recent competitive VMPD, downwards -0.01080 0.02618 -0.410 0.680
Recent competitive VMPD, upwards -0.00966 0.01391 0690 0.487
Market Expansion -0.55567 1.30632 -0.430 0.671
GDPPC -0. 15866 0.15256 -1.040 0.298
Dollar/Y en 0.01488 0.00511 2.910 0.004
Crude oil price -0.1 1647 0.03581 -3.250 0.001
U.S. population -0.04526 0.01879 -2.410 0.016
Observations: 27385
Subjects: 535
Failures: 187
Wald Chi Sq. 71.23

d.f. 13

P > Chi Sq. <.001
Log pseudo-likelihood -926.27

 

 

68

 

Bugged I m 688.31: 335m I dd ”Enema—am I m ”cotommsm .A

32$ I .3 “83.83% “oz - .mz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.m ..mz + + .2 - comaamxm 2...va a: a:
.m.Z .md - .md - :oficamxm “of“; m—
.m. Z .m.: + .m.: + $839 .DmSE 0330950 “cocoa. Q
.m. Z .m.: + .md + $83558 HESS o>uuom88 “58% S

U
.m - + - + Boa .326 CASE 03:59:00 “cocoa 2
2.; .m.: + - + 02 $559 QASE 026.0988 Evenly? Z
.m.Z .md - md - .5on .coa 6x82 a
.m.z ..m .2 - + + £235 3:838 a: i:
..m .m.: + - + 56:85 gasoaoo w
.m.m + C C C consent—oi m
.mm ..m. Z + + .m.: + .m.: + .wd + 385% “to; pm .QN
.m.Z rm: .m.: + .md + .md + + + 03205 prom am .am
.M - + $529? HESS E83: 3
.m - + $33.58 HESS Becom— 3
.m - + 02 80¢ .33“ 9:2: Eooom— an
.m. Z .m.: + 02 $559 QASE EoooM— an
.mz ..mz .2 + u: see 32. 32: E83; .6 .3
.m.Z ..m.Z .m.: + 02 €839 nESE Encom— mo .2

< u L Huumxm cEEtun— Samar» . u r" Hoomxu . a r” Samar...—

tomanm €325: mEsBEEQ 33¢. $538.

2358? 32> ES: 03E§ 8%:

 

 

 

Bee—z mega: «i.e.—:80 a... ...: 3.58.: 2: he baa—Haw "NN 935,—.

69

In the case of horizontal market positioning dynamics, the results of the
competing risks model indicate that recent HMPD do not have a significant impact on the
competing hazards of HMPD towards (B = .036, z = .11, p > .10) and away (0 = -.374, z
= -1.25, p > .10) from the market center. These results fail to provide support for
Hypotheses 1a, 6a and 6b. Performance improvements have a significant effect of the
propensity to engage in HMPD towards the market center ([3 = .00035, 2 = 9.34, p < .001)
providing support for Hypothesis 2a. Yet, the same hypothesis is not supported for the
propensity to move away from the market center (0 = -.012, z = -.46, p > .10). Similarly,
performance decreases do not have a significant impact on the competing risks of HMPD,
failing to support Hypothesis 2b.

Product proliferation has a strong positive and statistically significant linear
impact on both of the HMPD competing risks (effect on HMPD towards market center: [3
= 1.86, z = 3.28, p < .001; effect on HMPD away from market center: [3 = 1.39, z = 4.11,
p < .001), while its quadratic component is negative and statistically significant for
HMPD towards the market center ([3 = -.27, z = -1.98, p < .05) and negative and
insignificant for the HMPD away from the market center ([3 = -.09, z = -1.27, p > .10).
These results provide strong support for Hypothesis 5 in the case of moving towards the
market center, and only partial support for the same hypothesis in the case of moving
away from the market center. In the later case, the effect is monotonic and positive.

Among the environmental drivers, the competitive intensity has a statistically
significant negative effect on the propensity to exhibit HMPD towards the market center
([3 = -.29, z = ~2.23, p < .05), and a positive but statistically insignificant effect on the
propensity to exhibit HMPD away from the market center (0 = .01, z = .76, p > .10),

70

rejecting Hypothesis 8 in the former and failing to support it in the later case. The market
proliferation has statistically insignificant effects on both HMPD competing hazards at
the .10 confidence level, failing to support Hypothesis 9. The HMPD of the competition
has a statistically significant and negative effect on the hazard of HMPD towards the
market center, regardless of the direction of the marketing dynamics exhibited by the
competition (competition’s HMPD towards market center: [3 = -.04, z = -2.24, p < .05;
competition’s HMPD towards market center: [3 = -.06, z = -3.29, p < .001). As such, in
the case of HMPD towards the market center, Hypothesis 11 is strongly rejected. On the
other hand, only the competitors’ HMPD away from the market center has a negative and
statistically significant effect on the hazard of exhibiting HMPD away from the market
center ([3 = -.07, z = -3.44, p < .001) while the competitive moves towards the market
center do not have a statistically significant effect ([3 = .02, z = .94, p > .10). As such,
Hypothesis 11 is rejected for the HMPD away from the market center too. Market
expansion has a positive and statistically significant effect on propensity to move toward
the market center (0 = 3.19, z = 2.62, p < .01), and statistically insignificant positive
effect on the hazard to move away from the market center ([3 = 1.41, z = 1.29, p > .10),
rejecting Hypothesis 13 for the former and failing to support it for the later.

In the case of vertical market positioning dynamics, recent vertical market
positioning dynamics of the same type with the competing risk, has a significant effect in
both of the models. As such, products with recent downwards VMPD are less likely to
engage in further downwards repositioning (B = -1.22, z = -2.33, p < .05). Similarly, a

product with recent upwards VMPD is less likely to reposition vertically higher (0 = -

71

1.09, z = -2.86, p < .01). These results are contrary the expectations associated with
Hypothesis 1b.

To test Hypothesis 7, recent HMPD towards and away from the market center
have been employed as covariates of the VMPD competing risks. While the recent
moves towards the market center have no significant effect on both of the competing
risks, recent moves away from the market center have a negative effects on the
propensity to exhibit downwards vertical market positioning change (0 statistically
significant at .10 in one of the specifications) and on the hazard of exhibiting an upwards
vertical market positioning dynamics (0 = -.47, z = -1.67, p < .10). According to these
results, while Hypothesis 7a is not supported, Hypothesis 7b is rejected.

Recent improvements in sales have no significant effect at the .10 confidence
level on either of the VMPD competing hazards, failing to provide support for
Hypotheses 3a and 4. Performance decreases have no significant effect on the propensity
to exhibit downwards VMPD, yet they have a significant positive effect on the hazard of
upward VMPD in one of the specifications (0 = .70, z = 1.80, p < .10). This result
partially supports Hypothesis 3b.

Competitive intensity has a significant positive effect on the propensity of the
product to exhibit downwards VMPD (B = .03, z = 1.72, p < .10), providing support for
Hypothesis 10a. On the other hand, Hypothesis 10b is not supported at the .10
confidence level in any of the alternate specifications of upwards VMPD. Recent
upwards and downwards VMPD of the competition have no significant effect on any of
the VMPD competing risks at the .10 confidence level, failing to provide support for

Hypothesis 12.

72

Market expansion has no significant effect on the hazard of exhibiting downwards
VMPD at the .10 confidence level, failing to support Hypothesis 14a. On the other hand,
its effect on exhibiting upwards VMPD is positive and statistically significant, supporting
Hypothesis 14b at the .10 confidence level ([3 = 2.03, z = 1.73, p < .10).

Similar to the pooled model, to test for the robustness of the results to unobserved
brand level and manufacturer level effects, two different shared frailty models have been
estimated for each hazard rate (see Tables 14 to 17). The shared frailty coefficients
(theta) are only marginally significant in all the shared frailty models tested (p values
ranging from .002 to .492), indicating that the brand and manufacturer level effects not
accounted for in the covariates are relatively limited. Further, the coefficient estimates
for the covariate effects are relatively stable in terms of magnitude and significance
between the ordinary specifications and shared frailty specifications, indicating that the
results are robust to unobserved brand level and manufacturer level heterogeneity.

Also, the competing risks models where estimated using a two year time lag
between the covariates and the event. The estimated parameters are presented in Table
18 to 21. Similar to the pooled model results, the effects are generally weaker for the two
year time lag, some coefficients becoming statistically insignificant.

A summary of the results of the competing risks models along with the related

hypotheses can be seen in Table 22.

73

PERSISTENCE MODEL ESTIMATION

In order to identify the differences in long term and short term effects of
consistent and erratic market positioning changes on the performance of the brand, 30
time series have been constructed for 30 different brands active in the U.S. passenger cars

segment between years 1980 and 2002. A list of these brands can be seen in Table 23.

Table 23: The brands used in the estimation of the persistence model

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Acura Lincoln
Audi Mazda
BMW Mercedes
Buick Mercury
Cadillac Mitsubishi
Chevrolet Nissan
Chrysler Oldsmobile
Dodge Pontiac
Ford Porsche
Honda Saab
Hyundai Subaru
Infiniti Suzuki
Isuzu Toyota
Jaguar Volvo
Lexus Volkswagen

 

For each brand, the Augmented Dickey-Fuller (ADF) test was utilized to test
whether or not the sales, consistent HMPD, erratic HMPD, consistent VMPD, and erratic
VMPD are stationary or evolving. The t statistic was larger in absolute value than the
5% critical ADF value for all brands and all the series of consistent and erratic market
Additionally, only 7 out of 30 brands included in the analysis

positioning dynamics.

have statistically insignificant ADF statistics for the monthly sales series, indicating that
74

an ample majority of the time series included in this analysis are stationary. As such, the

following Vector Autoregressive model in levels has been estimated:

 

 

 

 

 

 

 

 

 

"GNPPC,“
01L,
YEN,
M1,
Is 33’
CH: J ”21} ”2]; ”2}; ”214 ”215 CHI-j M4! "CHJ
:3 ..
’ ”4.1 ”4; ”4s ”4_4 ”4s "1 M6, CV"
~EV" fish ”512 ”53 ”514 ”515 ~EV’_jJ M7, yin/"J
M8,
M9,
M10,
_Mu.I

where J is the order of the model, III-1,; denotes the effect of the k-th variable lagged of the

j-th order (with the t-j subscript) on the i-th dependent variable, S stands for the brand
level monthly sales, CH is the consistent horizontal market positioning dynamics, EH
stands for erratic horizontal market positioning dynamics, CV stands for consistent
vertical market positioning dynamics, EV stands for erratic vertical market positioning
dynamics u stands for the error term associated with each dependent variable. B is the
coefficient matrix for the effects of the control variables, which consist of GNPPC (GNP

per capita), OIL (cruel oil barrel price), YEN (yen — dollar exchange rate) and a set of

75

dummy variables that denote the month in order to extract the seasonal effects (M1 to
M11).

The model has been estimated for each brand independently. The order of each
model (J) has been identified by the SBIC index and varies between 6 and 12. Once the
coefficients were estimated, the impulse response function (IRF) have been created to
visualize the short term and long term effects of each of the market positioning dynamics

variables on the sales. One such IRF can be seen in Figure 4.

Figure 4: The Impulse Response Functions for one of the Brands

 

Consistent Horizontal Consistent Vertical

3000 - MPD MPD

2000 ‘

1000 I - - A4 * .2, ‘
0 W'—
-1000 - u

Erratic Horizontal Erratic Vertical MPD
‘1 ‘ MPD ’

 

 

 

 

 

3000 ‘ '
2000 1
1000 ‘ " '

AA, z I.

0 V V‘" " I, j _ . "WM—v—
-1000-‘~~~ - 5. . ‘ “ ' ' ‘ '
0 20 40 60 0 20 40 60

 

 

 

 

 

 

In order to ease the interpretation of the results, the impulse response functions of

consistent and erratic market positioning variables have been summarized in Table 24.

76

Table 24: Impulse Response Function Summaries

 

 

 

 

 

 

 

 

 

 

HMPD VMPD
Consistent Erratic Consistent Erratic
Short Long Short Long Short Long Short Long
Term Term Term Term Term Term Term Term
Positive Effect 70% 60% 53% 33% 41% 31% 24% 20%
No Effect 27% 37% 30% 53% 41% 52% 56% 60%
Negative Effect 3% 3% 17% 13% 17% 17% 20% 20%

 

 

 

 

 

 

 

The results indicate that a relatively large ratio of the brands have positive short
term effects of the consistent and erratic HMPD (70% and 53% respectively). On the
other hand, while 60 percent of the brands have positive long term effects of consistent
HMPD on the sales, this ratio is only 33 percent for the erratic HMPD. While only 1 of
the brands has a negative short term and long term negative return on consistent HMPD,
17 percent of them have negative short term effect and 13 percent have negative long
term effects of erratic HMPD. While these results provide support for Hypothesis 15b,
their support for Hypothesis 15a is partial.

With respect to VMPD, consistent moves have positive short term effects on sales
for 41 percent of the brands, and long term positive effects for 31 percent of the brands.
In the case of erratic VMPD, only 24 percent of the brands have positive short term
effects and 20 percent have positive long term effects. On the other hand, while 17
percent of the brands have negative short and long term returns for consistent VMPD,
this ratio is 20 percent for erratic VMPD. These results provide partial support for

Hypotheses 16a and 16b.
77

Further, to better reflect the differences in effects of consistent and erratic moves,
their short term and long term effects on the monthly sales have been compared in terms
of magnitude of positive effect. The summary of these comparisons can be seen in Table

25.

Table 25: The Comparison of the Short Term and Long Term Effects of Consistent

 

 

 

 

 

Versus Erratic MPD.
Consistent HMPD Consistent VMPD
Short Long Short Long
Effect is: Term Term Term Term
Larger than Erratic 50% 47% 50% 42%
Same with Erratic 43% 43% 33% 38%
Lower than Erratic 7% 10% 17% 21%

 

 

 

 

 

 

 

According to these results, consistent HMPD has a larger short term positive
effect that erratic HMPD for fifiy percent of the brands. This ratio is 47 for the same
comparison with respect to the long term effect on sales. Only a few brands have less
positive short term (7%) and long term (10%) returns from the consistent HMPD than
from erratic HMPD.

For VMPD, in 50 percent of the cases consistent dynamics have a more positive
short term return and in 42 percent of the cases a more positive long term return than

erratic dynamics. Only 17 percent have more negative short term returns and 21 percent

have more negative long term returns on consistent VMPD.

78

CHAPTER FIVE

DISCUSSION AND IMPLICATIONS

ANTECEDENTS OF MARKET POSITIONING DYNAMICS

This study investigates the effects of several internal and external antecedents of
horizontal and vertical market positioning dynamics. Specific hypotheses are derived
from the tenets of evolutionary framework, organizational learning and game theory.
The analysis provides elucidating results that engender significant contributions to the
extant empirical market positioning literature.

First of all, the findings parallel and support the idea of organizational change
momentum. According to this framework, past market positioning changes are likely to
create an internal momentum towards further market positioning changes, even in the
absence of environmental stimuli (Amburgey et al. 1993; Greve 1998). As such, in the
light of our results, it can be concluded that the organizational change framework can be
extended to include the market positioning decisions. Market positioning dynamics
decisions can be regarded as a significant change that requires specific routines and
capabilities to be developed.

Although past research indicates that performance shortcomings are be a strong
internal motivator for change (Bolton 1993; Bowman 1982), the findings of this study do
not support this assertion in the context of market positioning. Yet, the findings indicate
that performance improvements are a significant driver of HMPD, implying that
companies tend to engage in significant changes in the product lines that have an up-trend

in market sales. On the other hand, past performance seems to not have a significant

79

impact on price setting, contrary the predictions of previous conceptual studies (Aaker
1997). One could speculate that companies may choose other means than vertical
repositioning to capitalize on accumulated brand equity, such as new product introduction
and brand extensions.

The hypotheses for the effects of product proliferation on HMPD are supported,
indicating an inverse U shape effect and that horizontal repositioning is a less viable
strategy when the product offerings are saturated. Again, one may assert that when the
product line is saturated in one differentiation dimension, other means of differentiation
such as image and style become a more attractive alternative.

Although past research indicates strong effects of competitive intensity on
responsiveness (Javorski and Kohli 1993; Narver and Slater 1990), the findings of this
study reject this hypothesis, indicating a negative effect of competitive intensity.
Similarly, the effect of average market product proliferation is insignificant. Yet, in the
context of this study, competitive intensity is captures the total number of products that
are on the market at a given time. As such, considering the intensive differentiation
efforts in the automotive industry, each product acts as a separate offering for the
customer. Extant empirical evidence from industries with significant environmental
uncertainty indicates that the greater the product variety, the less valuable the variety to a
firm (Sorenson 2000). Within the study context, increased competition can be equated
with increased product variety available for the customer, yielding in the saturation of
profitable market positions and decreasing the value and the opportunity of horizontal

repositioning.

80

The market dynamism resulting from competitors’ actions has a negative effect
on the propensity to exhibit both HMPD and VMPD. Although it has been postulated
that competitor dynamism acts as a facilitator of organizational learning of the market
structure and the future trends and opportunities in the customer needs (Greve 1998),
reject this assertion. It can be postulated that in the case of market positioning,
companies prefer to ‘wait and see’ the results of competitive actions before reacting.

Finally, according to extant game theoretic studies, the companies are inclined to
respond more aggressively to declining demand conditions than to stable or expanding
market size (Dekimpe et al. 1999a). Nevertheless, the results indicate that market
expansion has a positive effect on the HMPD indicating that companies are more willing
to take risks under expanding market conditions. These predictions are particularly
plausible when the costs associated with a HMPD are taken into account. As such, it can
be concluded that companies are more willing to bear the risk of a significant change in
the product line in expanding market conditions.

The results of this study also provide significant managerial implications. First of
all, the proposed framework and the empirical analysis help managers understand the
expected market positioning change of the competition in different situations. For
example, a new market entry or a market positioning change is more likely to be
encountered by a ‘wait and see’ strategy by the other competitors. A product line with a
significant HMP change is more likely to have a similar change within the next year.
Competitive products with increased market performance are more likely to be subject to
a significant market positioning change that the ones that show a decrease in sales.

Competitors that have a saturated product line are less likely to have HMPD that the

81

competitors that offer only limited product offerings. As such, understanding the role
different factors play in market positioning dynamics will help them in formulating future

marketing strategies and foreseeing competitor strategic shifts.

DIRECTION OF CHANGE

The results of the competing risks model shed more light on the complexities
associated with the market positioning dynamics. When the direction of change is taken
into consideration, the momentum of change (Amburgey et al. 1993; Greve 1998)
concept is not supported. Previous market positioning changes of the same type with
respect to the move direction have no significant effect on the hazard of exhibiting
HMPD towards the market center and HMPD away from the market center. Further, in
the case of VMPD, the results imply that companies do not prefer to engage in successive
significant price increases or decreases.

The findings also indicate that products with recent improvements in performance
are more likely to be repositioned towards the market center. Another interesting finding
is that products that decreasing sales may result in increased prices. This can be
attributed to the niche strategy. Companies may choose to compensate the revenues for
lower sales by increasing their price.

Contrary to the expectations, products with recent repositioning away from the
market center are less likely to exhibit upward repositioning. According to these findings,
increased horizontal differentiation from the market center does not engender price
increases. As such, it can be concluded that when mangers make their vertical

repositioning decisions are driven by other factors than vehicle size differentiation.

82

Important such factors include competitive intensity and market expansion. Under
increased competitive pressures products are more likely to exhibit downwards
repositioning, while under expanding market conditions are more likely to significantly
exhibit increased retail prices.

These finding also have significant managerial implications. They are unlikely to
face competitor products with successive price increases or decreases. Further, under
expanding market conditions they should expects significant price increases from their
competitors. Also, new market entries are likely to be responded to by price dumping.
As such, this study sheds additional light on competition actions, easing the managerial

strategic decision making process.

RETURN ON CONSISTENCY

This study contributes to the extant research regarding performance returns of
marketing strategy in several ways. First, we differentiate between short term and long
term effects of market positioning dynamics. Next, we differentiate between consistent
and erratic market positioning dynamics. Further, we empirically test whether consistent
MPD have a longer lasting and more positive effect than erratic MPD.

Overall, our results indicate that consistent horizontal MPD tend to have a longer
lasting positive effect on the sales than erratic horizontal MPD. Although the
hypothesized difference in effects is also supported for vertical MPD, the variation
between the effects of consistent and erratic dynamics is weaker in the case of vertical
market positioning. This findings are in line with the extant research that indicate that

price changes tend to have only short term impacts on sales (Dekimpe et al. 1999a;

83

Pauwels et al. 2004). Further, one could attribute this disparity between horizontal and
vertical market positioning dynamics to temporary adjustments in price that are not
perceived by the customers as erratic moves but just temporary discounts. Nevertheless,
future research should address this disparity between HMPD and VMPD. Also, further

investigation is required to identify the causes of negative returns of MPD.

84

CHAPTER SIX

LIMITATIONS AND FUTURE RESEARCH DIRECTIONS

The current study suffers from several limitations. First, only one dimension of
horizontal differentiation is considered. Size is the only horizontal differentiation
dimension included in this analysis. Yet, companies may choose to differentiate their
products on multiple dimensions such as style, image or horsepower. Although the
theoretical predictions would not change in any of the differentiation dimensions that
could be considered, the tradeoffs and interactions between different differentiation
dimensions is a promising future research avenue. As such, future research should
investigate the drivers and contingencies of the tradeoff or coupling decisions of different
differentiation criteria. The intangible product attributes are expected to have a
significant role in these decisions.

The context of this study is the U.S. automotive industry. Therefore, there are
some limits to the external generalization of the findings. For example, in a setting where
the product differentiation is not as prominent as in the U.S. automotive industry, as is the
case in a commodity industry setting, the effects of competitive intensity and competitive
market positioning dynamics may have a weaker or positive effect on the propensity to

change.

85

APPENDICES

86

 

EEO £3.53.
.283. .bEdU “meow EOEEA
65325 doEEU awn—EU nowa—

48E“
05 E 2ng “8:9: “womb: 05 HEB £358 Bow 2: mo 33633 owe?

82v
“2:00 HOV—SSH

 

 

 

 

 

 

o E E ... - E E .Boz 3032a 05 ES» 22289"
A6 + . 32 . $9 . 53 . :oEEQEoo E 83m :3 howsommaa ~89 mama» as E umEEo omficoohom «81m
.80» $2 05 E 835E Room 05 mo EEEmEAu BEBE BEES CASS
2t 35:: 35E 2: E mSEooE 2: =« no EEEEEV ESE :89? ESE 03:39:00
.53» “m2 2: E 83on Room 2.: mo BEBE? “BEE EEOESE 2t QASE
EEE woo—SE 2.: E £26on 05 =~ mo EEEEEV BEBE Baa—onto: :38. 02:89:00
Noon Echo 8 96: SEE 538032
E we; .33 .3 m _ A .33 E 2 E Eon 53w m 3 Eugen 2268 2: ENE owns—85> mo BEE: omfio>< 8x38 owfio><
Noon E .30 EEBE
3: £03 E of Swag E 2258 m: :oEm a E EoEwOm Emu “omqommaa E mEHomEOo 225E mo 898:: 30,—. 03:39:80
Nam 28 Woo ”mmEHobo 3330:? .oE» 53w a 5:802“:
m Us; 32 E EwBE Eso< 3 Echo E 830on a 5:3 E cove—5 830on a moEm 33323 €0.5th H2605

 

No3 E dem
.3 330820 8:; «$an «50¢.

032020 on .m EON on Him .39» £5329 05 9 389:8 mm .80» 3E
05 E 835K 2: mo 31% E5 05 E 338% 09350th .«O oEm> 8203<

383% ..tom

 

32 E
£3 .3 nomaoSE 33m EwBE «53.

.0382: o: a 83 8 Ham .8?» £5an 05 8 “02388
ma .30» SE 05 E “2603 05 90 3.8 E: 05 E omaoEoE owficoaom

838E “tom

 

mmwKw 98:8 owe E wvvda

 

 

 

 

 

 

$52 83 £3 .EwBE «33.. 8E owns :82 wowmowwsm SESMEENE Bacon 2: E mowEEU QASQ/

WW WmemEmowonomoWMWEMMHEMMMWM Second 0:: 630a 2: mo 36 8330:? EE 2: E momEEU EASE
«520$ .<,~O>OH .QMOm .20 853.255 mEoEcoa .AEEEoU HoqumEEWSH
83¢. .3on .Eom £20525 ..BEQ BEE. 8 .BRE. 05 3 BEN?“ 83. 2.8m
EwBE .mA 3on .3832 53:: ESE 0&6on a .3 cohobo $on 2&0on < 820on,
03::wa notice: 2.325»

 

 

 

 

SEE—Eon 033:3 "a EEGEFV

87

REFERENCES

Aaker, DA. (1997), "Should you take your brand to where the action is?," Harvard
Business Review, 75 (5), 135-43.

Amburgey, T.L., D. Kelly, and W.P. Barnett (1993), "Resetting the clock: The dynamics
of organizational change," Administrative Science Quarterly, 38 (1), 51-73.

Amburgey, TL. and H. Rao (1996), "Organizational Ecology: Past, Present, and Future
Directions," Academy of Management Journal, 39 (5), 1265-86.

Argyris, C. (1977), "Double Loop Learning in Organizations," Harverd Business Review,
55 (September-October), 115-25.

Argyris, C. and DA. Schon (1978), Organizational Learning.“ A Theory of Action
Perspective. Reading, MA: Addison-Wesley.

Armstrong, J .S. and F. Collopy (1996), "Competitor orientation: Effects of objectives and

information on managerial decisions and profitability," Journal of Marketing Research,
33 (2), 188-99.

Arthur Andersen, C. (1985), "Cars and Competition: Manufacturing Strategies." Chicago,
IL: Arthur Andersen & Co.

Barnett, W.P. and GR. Carroll (1987), "Competition and mutualism among early
telephone companies," Administrative Science Quarterly, 30, 400-21.

Barr, P.S., J.L. Stimpert, and AS. Huff (1992), "Cognitive change, strategic action, and
organizational renewal," Strategic Management Journal, 13 (8), 15-36.

Baum, J AC. and HA. Haveman (1997), "Love thy neighbor? Differentiation and
agglomeration in the Manhattan hotel industry, 1898-1990," Administrative Science
Quarterly, 42 (304-338).

Baum, J AC. and SJ. Mezias (1993), "Competition, institutional linkages, and
organizational growth," Social Science Research, 22, 131-64.

Blossfeld, H.-P. and G. Rohwer (2002), Techniques of Event History Modeling New
Approaches to Causal Analysis. New Jersey: Lawrence Erlbaum Associates.

Bolton, MK. (1993), "Organizational innovation and substandard performance: When is
the necessity the mother of innovation?," Organization Science, 4, 57-75.

Boone, C. and A.v. Witteloostuijn (1995), "Industrial Organization and Organizational
Ecology: The Potentials for Cross-Fertilization," Organization Studies, 16(2), 265-698.

88

Bowman, EH. (1982), "Risk seeking by troubled firms," Sloan Management Review, 23,
33-42.

Bromiley, P. (1991), "Testing a causal model of risk-taking and performance," Academy
of Management Journal (34), 37-59.

Bruderl, J. and R. Schussler (1990), "Organizational Mortality: The Liabilities of
Newness and Adolescence," Administrative Science Quarterly, 35 (3), 530-47.

Carpenter, GS. (1989), "Perceptual Position And Competitive Brand Strategy In A Two-
Dimensional Two Brand Market," Management Science, 35 (9), 1029-44.

Carroll, GR. and M.T. Hannan (1989), "Density Dependence in the Evolution of
Populations of Newspaper Organizations," American Sociological Review, 54 (4), 524-41.

Caves, RE. and ME. Porter (1977), "From Entry Barriers to Mobility Barriers,"
Quarterly Journal of Economics, 91, 241-46.

Choi, S.C., W.S. Desarbo, and PT. Harker (1990), "Product Positioning Under Price
Competition," Management Science, 36(2), 175 25 pages.

Choi, S.C., W.S. DeSarbo, and PT. Harker (1992), "A Numerical Approach to Deriving
Long-Run Equilibrium Soluti," Management Science, 38 (1), 75.

Conner, KR. (1988), "Strategies for Product Cannibalism," Strategic Management
Journal, 9 ((Special Issue)), 9-26.

Cool, K. and I. Dierickx (1993), "Rivalry, Strategic Groups, and F inn Profitability,"
Strategic Management Journal, 14 (1), 47-59.

Cool, K.O. and D. Schendel (1987), "- Unspecified Article Title," Management Science,
33 (9), 1102 23 pages.

Cox, DR. (1972), "Regression Models and Life Tables," Journal of the Royal Statistical
Society, B34, 187-220.

---- (1975), "Partial Likelihood," Biometrica, 62, 269-76.

Cyert, R.M. and J .6. March (1963), A Behavioral Theory of the Firm. New York:
Prentice-Hall.

d'Aspremont, C., J. Gabszewicz, and J .F. Thisse (1979), "On Hotelling’s 'Stability in
Competition'," Econometrica, 47, 1145-50.

Deephouse, D.L. (1999), "To be different, or to be the same? It's a question (and theory)
of strategic balance," Strategic Management Journal, 20 (2), 147-66.

89

Dekimpe, M.G. and D.M. Hanssens (1995), "Empirical Generalizations about Market
Evolution and Stationan'ty," Marketing Science, 14 (3,2), G109-G21.

---- (2004), "Persistence Modeling for Assessing Marketing Strategy Performance," in
Cool Tools in Marketing Strategy Research, D. Lehmann and C. Moorman, Eds.:
Marketing Science Institute.

Dekimpe, M.G., D.M. Hanssens, and J .M. Silva-Risso (1999a), "Long-run effects of
price promotions in scanner markets," Journal of Econometrics, 89, 269—91 .

---- (1999b), "Long-run effects of price promotions in scanner markets," Journal of
Econometrics, 89, 269-91.

Dickson, PR. and J .L. Ginter (1987), "Market Segmentation, Product Differentiation,
and Marketing Strategy," Journal of Marketing, 51 (2), 1-10.

Dobrev, S.D., T.-Y. Kim, and GR. Carroll (2002), "The Evolution of Organizational
Niches: U.S. Automobile Manufacturers, 1885-1981," Administrative Science Quarterly,
47, 233-64.

---- (2003), "Shifting gears, shifting niches: Organizational inertia and change in the
evolution of the U.S. automobile industry, 1885-1981," Organization Science, 14 (3), 264.

Dranove, D., M. Peteraf, and M. Shanley (1998), "Do Strategic Groups Exist? An
Economic Framework for Analysis," Strategic Management Journal, 19, 1029-44.

Economides, N. (1989), "Quality Variations and Maximal Variety Differentiations,"
Regional Science and Urban Economics, 19 (1), 21-29.

Efron, B. (1977), "The Efficiency of Cox's Likelihood Function for Censored Data,"
Journal of the American Statistical Association, 72 (359), 557-65.

Fiegenbaum, A. and H. Thomas (1995), "Strategic groups as reference groups: Theory,
modeling and empirical examination of industry and competitive strategy," Strategic
Management Journal, 16 (6), 461-77.

F iol, CM. and MA. Lyles (1985), "Organizational Learning," Academy of Management
Review, 10 (October), 803-13.

Freeman, J. and M.T. Hannan (1983), "Niche width and the dynamics of organizational
populations," American Journal of Sociology, 88, 1116-45.

Gabszewicz, J. and J .F. Thisse (1979), "Price Competition, Quality and Income
Disparities," Journal of Economic Theory, 20, 340-59.

90

Galunic, DC. and KM. Eisenhardt (1996), "The evolution of intracorporate domains:
Divisional charter losses in high-technology, multidivisional corporations," Organization
Science, 7 (3), 255-82.

Garvin, DA. (1993), "Building a Learning Organization," Harvard Business Review, 71
(July-August), 78-91.

Geroski, P., S. Machin, and J .V. Reenen (1993), "The Profitability of Innovating Firms,"
RAND Journal of Economics, 24 (2), 211.

Geroski, PA. (2001), "Exploring the niche overlaps between organizational ecology and
industrial economics," Industrial and Corporate Change, 10(2), 507.

' Grant, R.M. (1996), "Toward a knowledge-based theory of the firm," Strategic
Management Journal, 109-22.

Green, D.H., D.W. Barclay, and AB. Ryans (1995), "Entry strategy and long-term
performance: Conceptualization and empirical examination," Journal of Marketing, 59
(4), 1-16.

Green, PE. and A.M. Krieger (1985), "Models and heuristics for product line selection,"
Marketing Science (4), 1-19.

Greve, HR. (1998), "Performance, Aspirations, and Risky Organizational Change,"
Administrative Science Quarterly, 43 (1), 58-86.

Hachen, D.S.J. (1988), "The Competing Risks Model: A Method for Analyzing Processes
with Multiple Types of Events," Sociological Methods & Research, 17 (1), 21-54.

Han, J .K., N. Kim, and H.-B. Kim (2001), "Entry barriers: A dull-, one-, or two-edged
sword for incumbents? Unraveling the paradox from a contingency perspective," Journal
of Marketing, 65 (1), 1-14.

Hannan, M. T. (1998), "Rethinking Age Dependence in Organizational Mortality: Logical
Forrnalizations," American Journal of Soczology, 104, 85- 123.

Hannan, M.T., GR. Carroll, S.D. Dobrev, and J. Han (1998a), "Organizational Mortality
in European and American Automobile Industries, part I: Revisiting the Effects of Age
and Size," European Sociological Review, 14, 279-302.

Harman, M.T., GR. Carroll, S.D. Dobrev, J. Han, and J .C. Torres (1998b),
"Organizational Mortality in European and American Automobile Industries, part II:
Coupled Clocks," European Sociological Review, 14, 302-13.

Hannan, M.T. and J. Freeman (1977), "The population ecology of organizations,"
American Journal of Sociology, 82, 929-64.

91

Hanssens, D.M., L.J. Parsons, and KL. Schultz (1990), Market Response Models:
Econometric and Time Series Analysis. Boston: Kluwer Academic Publishers.

Hauser, JR. and SM. Shugan (1983), "Defensive Marketing Strategies," Marketing
Science, 2 (4), 319-60.

Hauser, JR. and P. Simmie (1981), "Profit maximizing perceptual positions: An
integrated theory for the selection of product features and price," Management Science,
27(1), 33-56.

Hoffer, GE. and R.J. Reilly (1984), "Automobile styling as a shift variable," Applied
Economics, 16, 291-97.

Horsky, D. and P. Nelson (1992), "New Brand Positioning and Price in an Oligopolistic
Market," Marketing Science, 11 (Spring), 133-53.

Hotelling, H. (1929), "Stability in Competition," Economic Journal, 39, 41-57.

Huff, AS. (1982), "Industry Influences on Strategy Reformulation," Strategic
Management Journal, 3 (2), 119-31.

Hunt, MS. (1972), Competition in the Major Home Appliance Industry: Harvard
University.

Javorski, B.J. and AK. Kohli (1993), "Market Orientation: Antecedents and
Consequences," Journal of Marketing, 57 (July), 53-70.

Kaul, A. and VR. Rao (1995), "Research for product positioning and design decisions:
An integrative review," International Journal of Research in Marketing, 12 (4), 293-320.

Kekre, S. and K. Srinivasan (1990), "Broader Product Line: A Necessity to Achieve
Success?," Management Science, 36 (10), 1216 26 pages.

Kelly, D. and TL. Amburgey (1991), "Organizational Inertia and Momentum: A
Dynamic Model of Strategic Change," Academy of Management Journal, 34 (3), 591-612.

Kohli, A.K., B.J. Jaworski, and A. Kumar (1993), "MARKOR: A Measure of Market
Orientation," Journal Of Marketing Research, 30 (4), 467-78.

Kumar, KR. and D. Sudharshan (1988), "Defensive Marketing Strategies: An
Equilibrium Analysis Based on Decoupled Response Function Models," Management
Science, 34 (7), 805.

Lancaster, K.J. (1971), Consumer Demand: A New Approach. New York: Columbia
University Press.

92

Levinthal, DA. and M. Fichman (1988), "Dynamics of International Attachments:
Auditor-Client Relationships," Administrative Science Quarterly, 33 (3), 345-69.

Lukas, BA. and QC. Ferrell (2000), "The effect of market orientation on product
innovation," Academy of Marketing Science. Journal, 28 (2), 239-47.

Mahajan, V., S. Sharma, and RD. Buzzell (1993), "Assessing the impact of competitive
entry on market expansion and incumbent sales," Journal of Marketing, 57 (3), 39-52.

March, J .G. (1991), "Exploration and Exploitation," Organization Science, 2, 71-87.

March, J .G. and HA. Simon (1958), Organizations. New York: Wiley.

Markides, C. (1998), "Strategic innovation in established companies," Sloan
Management Review, 39 (3), 31-42.

Markides, CC. (1999), "A dynamic view of strategy," Sloan Management Review, 40 (3),
55-63.

Miles, RE. and KS. Cameron (1982), Coflin Nails and Corporate Strategies. Englewood
Cliffs, NJ: Prentice-Hall.

Miles, RE. and CC. Snow (1978), Organizational Strategy, Structure and Processes.
New York: McGraw-Hill.

Miller, D. and M.-J. Chen (1994), "Sources and consequences of competitive inertia: A
study of the U.S. airline industry," Administrative Science Quarterly (39), 1-23.

Mitra, D. and RN. Golder (2002), "Whose Culture Matters? Near-Market Knowledge
adn Its Impact on Foreign Market Entry Timing," Journal of Marketing Research, 39 (3),
350-65.

Moorman, C. and AS. Miner (1997), "The impact of organizational memory on new
product performance and creativity," JMR, Journal of Marketing Research, 34 (1), 91 -
106.

Narver, J.C. and SF. Slater (1990), "The Effect of a Marketing Orientation on Business
Profitability," Journal of Marketing, 54 (October), 20-35.

Nelson, R. and S. Winter (1982), An Evolutionary Theory of Economic Change.
Cambridge, MA: Harvard University Press.

Nijs, V.R., M.G. Dekimpe, J .-B.E.M. Steenkamp, and DH. Hanssens (2001), "The
category-demand effects of price promotions," Marketing Science, 20 (1 ), l.

93

Pauwels, K. and D.M. Hanssens (2003), "Performance Regimes and Marketing Policy
Shifts in Mature Markets," in Working Paper. Tuck School of Business at Dartmouth
College.

Pauwels, K., D.M. Hanssens, and S. Siddarth (2002), "The long-term effects of price
promotions on category incidence, brand choice and purchase quantity," Journal of
Marketing Research, 34 (November), 421-39.

Pauwels, K., J. Silva-Risso, S. Srinivasan, and D.M. Hanssens (2004), "New Products,
Sales Promotions and Firm Value, With Application to the Automobile Industry,"
Journal of Marketing, Forthcoming.

Pesaran, M.H. and R. Smith (1995), "Estimating long-run relationships from dynamic
heterogeneous panels," Journal of Econometrics, 68, 79—113.

Podolny, J ., T. Stuart, and M.T. Hannan (1996), "Networks, knowledge and niches:
Competition in the world semiconductor industry 1984 - 1991," American Journal of
Sociology, 102, 659-89.

Sanchez, R. (1995), "Strategic flexibility in product competition," Strategic Management
Journal, 16((Special Issue)), 135-59.

Schultz, R.L. and DR. Wittink (1976), "The measurement of industry advertising
effects," Journal of Marketing Research, 13, 71-75.

Shaked, A. and J. Sutton (1982), "Relaxing Price Competition Through Product
Differentiation," Review of Economic Studies, 49, 3-13.

Sherman, R. and G. Hoffer (1971), "Does automobile style change pay off?," Applied
Economics, 3, 153-65.

Sinchcombe, A. (1990), Information and Organizations. Berkeley: University of
California Press.

Sinkula, J .M. (1994), "Market Information Processing and Organizational Learning,"
Journal of Marketing, 58 (January).

Sinkula, J .M., W.E. Baker, and T. Noordewier (1997), "A Framework for Market-Based
Organizational Learning: Linking Values, Knowledge and Behavior," Journal of
Academy of Marketing Science, 25 (4), 305-18.

Slater, SF. and J .C. Narver (1994), "Does competitive environment moderate the market
orientation-performance relationship?," Journal of Marketing, 58 (1 ), 46-55.

---- (1995), "Market Orientation and the Learning Organization," Journal of Marketing,
59 (3), 63-.

94

i; Hm“

 

Song, X.M. and ME. Parry (1994), "The dimensions of industrial new product success
and failure in state enterprises in the People's Republic of China," The Journal of Product
Innovation Management, 11 (2), 105.

Sorenson, O. (2000), "Letting the market work for you: An evolutionary perspective on
product strategy," Strategic Management Journal, 21 (5), 577.

Srinivasan, S., K. Pauwels, and V.R. Nijs (2003), "Disentangling the drivers of retail
price setting and their financial consequences," in Working Paper. Tuck School of
Business at Dartmouth.

Tardiff, J. (1998), "Motor Vehicles and Motor Vehicle Equipment," in US Industry
Profiles. New York: Gale Research.

Thomas, H. and T. Pollock (1999), "From I-O Economics' S-C-P Paradigm Through
Strategic Groups to Competence-Based Competition: Reflections on the Puzzle of
Competitive Strategy," British Journal of Management, 10, 127-40.

Tuma, NE. and M.T. Hannan (1984), Social Dynamics: Models and Methods. New
York: Academic Press.

Vandenbosch, MB. and CB. Weinberg (1995), "Product and Price Competition in a
Two-Dimensional vertical Differentiation Model," Marketing Science, 14(2), 224-49.

Venkatraman, N. (1989), "Strategic Orientation of Business Enterprises: The Construct,
Dirnensionality, and Measurement," Management Science, 35 (8), 942-62.

Volberda, H.W. and A.Y. Lewin (2003), "Co-evolutionary Dynamics Within and
Between Firms: From Evolution to Co-evolution," Journal of Management Studies, 40
(8), 2111-36.

Wind, Y. and TS Robertson (1983), "Marketing strategy: New directions for theory and
research," Journal of Marketing Research, 47, 12—25.

Young, RC. (1998), "Is Population Ecology a Useful Paradigm for the Study of
Organizaitons?," American Journal of Sociology, 94 (1), 1-24.

95