Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Customer satisfaction, customer bargaining power, and financial performance
(USC Thesis Other)
Customer satisfaction, customer bargaining power, and financial performance
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
CUSTOMER SATISFACTION, CUSTOMER BARGAINING POWER, AND
FINANCIAL PERFORMANCE
Copyright 2006
by
Xiaoling Chen
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)
August 2006
Xiaoling Chen
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3236488
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3236488
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements
I thank my dissertation committee: Mark Young (co-chair), Sarah Bonner (co-chair),
Ken Merchant, Tatiana Sandino, Wim Van der Stede, and Cheng Hsiao (outside member)
for their encouragement, guidance, and support. They have been great mentors to me
throughout my PhD program and I feel truly blessed. I am especially indebted to Ken
Merchant for providing access to the research site and to the managers and employees at the
research site for sharing their data and insights.
I also thank Nerissa Brown, James Gong, Mingyi Hung, Sam Lee, Hai Lu, Michal
Matejka, and workshop participants at University of British Columbia, University of Illinois
at Urbana-Champaign, University of Toronto, University of Utah, and University of
Washington for their valuable comments.
I am grateful to the Leventhal School of Accounting and Marshall School of
Business at the University of Southern California for financial support. I also thank the
Deloitte Foundation for a doctoral fellowship in my dissertation stage.
The completion of this doctoral dissertation would not have been possible without
the constant support and encouragement of my family. I thank my husband, Zhenzhong Lu,
my parents, Zhibiao Chen and Pingxiang Ruan, and my parents-in-law, Sheng’en Lu and
Huiqin Miao for their endless love and unwavering belief in me.
Last but not least, I thank my lovely son, Kevin Jiashu Lu, for giving new
significance to my life and for tolerating the many hours that I spent away from him to work
on my dissertation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table o f Contents
Acknowledgements.............................................................................................................. ii
List of Tables...................................................................................................................... iv
List of Figures...................................................................................................................... v
Abstract...............................................................................................................................vi
Chapter 1 . Introduction.........................................................................................................1
Chapter 2. Literature Review.............................................................................................. 1 1
Chapter 3. Evidence from a Large-Sample Study................................................................ 19
3.1. Hypotheses...............................................................................................................19
3.2. Sample and Measures...............................................................................................25
3.3. Empirical Analyses and Results............................................................................... 30
3.4. Discussion............................................................................................................... 49
Chapter 4. Evidence from the Field: The Case of a Health Insurance Company.................. 52
4.1. Research Site............................................................................................................52
4.2. Hypotheses.............................................................................................................. 55
4.3. Sample, Variables, and Research Design..................................................................61
4.4. Results..................................................................................................................... 70
4.5. Discussion............................................................................................................... 90
Chapter 5. Conclusion.........................................................................................................93
Bibliography...................................................................................................................... 96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IV
List of Tables
Table 1 : Sample Industry Composition and Average Herfmdahl-Hirschman
Index (HHI) by Industry
Table 2: Descriptive Statistics
Table 3: Pearson Correlations between Independent Variables and Control Variables
Table 4: The Effects of Customer Satisfaction, Switching Costs, and an Interaction
Term between Customer Satisfaction and Switching Costs on Future
Financial Performance
Table 5: The Effects of Industry-Relative Customer Satisfaction, Switching Costs,
and an Interaction Term between Industry-Relative Customer Satisfaction
and Switching Costs on Future Financial Performance
Table 6: The Effects of Customer Satisfaction, Switching Costs, and an Interaction
Term between Customer Satisfaction and Switching Costs on the Market
Value of Equity
Table 7: Exploratory Factor Analysis of Satisfaction Measures
Table 8: Descriptive Statistics
Table 9: Correlation Matrix
Table 10: Measurement Model for Independent Latent Variables
Table 11: Structural Equation Modeling (SEM) Analysis of the Relation between
Satisfaction Measures and Future Revenue for the 20-Quarter Period
2000-2004
Table 12: Structural Equation Modeling (SEM) Analysis of the Relation between
Changes in Satisfaction Measures and Changes in Future Revenue for
the 20-Quarter Period 2000-2004
32
33
33
38
40
44
64
71
73
75
78
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
List of Figures
Figure 1: Conceptual Model of the Relations among Customer Satisfaction,
Switching Costs, and Future Financial Performance
Figure 2: Graphical Presentation of the Hypotheses
Figure 3: The Interaction Effects of Customer Satisfaction and Switching
Costs on One-Year Ahead Sales Revenue
Figure 4: The Interaction Effects of Customer Satisfaction and Switching
Costs on One-Year Ahead Income before Extraordinary Items
Figure 5: The Interaction Effects of Industry-Relative Customer Satisfaction
and Switching Costs on the Rate of Return on Assets (ROA)
Figure 6: The Interaction Effects of Industry-Relative Customer Satisfaction
and Switching Costs on the Market Value of Equity (MVE)
Figure 7: HBP’s Hypothesized Business Model
Figure 8: Structural Equation Model Results for the Overall Sample
(Structural Model)
Figure 9: Structural Equation Model Results for the High Percentage Voluntary
Patient Subsample
Figure 10: Structural Equation Model Results for the Low Percentage Voluntary
Patient Subsample
Figure 11: Structural Equation Model Results for the High Market Penetration
Subsample
Figure 12: Structural Equation Model Results for the Low Penetration Rate
Subsample
23
24
45
46
47
48
57
79
83
84
85
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
This study examines the following research question in two complementary research
settings: How does customer bargaining power affect the relation between customer
satisfaction and future financial performance? Motivated by the mixed evidence on the
effects of customer satisfaction on financial performance across industries, the first research
setting investigates the moderating effect of customer switching costs, an important
component of customer bargaining power, in a large-sample cross-sectional study. Using the
American Customer Satisfaction Index and Compustat data, I predict and find that high
customer switching costs, as measured by the industry-level Herfindahl-Hirschman Index,
weaken the association between customer satisfaction and future financial performance. In
the second research setting, I obtain a proprietary database from a leading health insurance
company that measures satisfaction levels of multiple customer groups, including: (a) clients
that purchase insurance plans for their employees, (b) patients who use the insurance plans,
and (c) doctors who provide medical services. I find that the extent to which a customer
group influences the purchasing decision increases the effect of customer satisfaction on
future revenues. I also find that customer bargaining power enhances the impact of customer
satisfaction on future revenues. This study contributes to our understanding of the
performance consequences of nonfinancial value drivers and has implications for resource
allocation, performance evaluation, and compensation practices within firms.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Chapter 1. Introduction
A growing body of literature suggests that nonfinancial performance
measures are leading indicators of financial performance. In particular, customer
satisfaction is one of the most important and most widely studied nonfinancial
measures (e.g., Banker, Potter, and Srinivasan, 2000; Behn and Riley, 1999; Foster
and Gupta, 1997; Ittner and Larcker, 1998a; Smith and Wright, 2004). For example,
a survey of senior executives from 148 financial services firms showed that customer
relations are ranked as the most important driver of firm’s long-term organizational
success. In comparison, short-term financial performance was only ranked the fifth
most important (Ittner and Larcker, 2001). Because of the perceived importance of
customer satisfaction measures, most companies allocate a substantial amount of
resources to measure customer satisfaction and managers rely on these measures to
make operational decisions. In addition, Ittner, Larcker, and Rajan (1997) find that
37% of firms using nonfinancial measures in their executive bonus contracts include
customer satisfaction measures. Given the resources devoted to customer satisfaction
and the economic consequences, it is critical to understand the relation between
customer satisfaction and future financial performance.
Despite the perceived importance of customer satisfaction, prior research has
provided mixed evidence on the relation between customer satisfaction and future
financial performance. Several studies have found a positive relation between
customer satisfaction and future financial performance (Ittner and Larcker, 1998a;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
Behn and Riley, 1999; Banker, Potter and Srinivasan, 2000; Bernhardt, Donshu and
Kennett, 2000; Smith and Weight, 2004). However, other studies have found that the
financial impact of customer satisfaction varies across industries and firms
(Anderson, Fomell and Rust, 1997; Ittner and Larcker, 1998a). Although researchers
have conjectured that contextual factors may contribute to the mixed findings (Ittner
and Larcker, 1998a; Lambert, 1998), little empirical evidence exists on whether and
how contextual factors affect the satisfaction-performance linkage.
In addition, the prior literature on customer satisfaction has focused
exclusively on settings with a single customer group. However, in reality, companies
often deal with multiple customer groups and therefore it is challenging to determine
which customer satisfaction measure generates the greatest benefits for financial
performance. The term “customer” includes consumers (i.e., end users of products
and services), clients (i.e., organizations that dictate or influence the choice of end
users), distributors, and any other party that an organization serves (Kohli and
Jaworski, 1990). For example, an insurance company interacts both with clients that
purchase the insurance plans for their employees and patients who use the insurance
plans. A distributor in a supply chain conducts business with redistributors, retailers,
and end consumers. Outsource service providers should not only establish good
relationships with clients, who typically negotiate outsourcing deals, but also
understand the needs and preferences of end users of the services (Feeny, Lacity, and
Willcocks, 2005). Prior literature provides little guidance as to which customer
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
satisfaction measure is most strongly associated with future financial performance in
such settings.
The economics literature suggests that customer bargaining power is one
contextual factor that might account for cross-sectional differences in the relation
between customer satisfaction and future financial performance. Customer
bargaining power is the impact that customers have on a company (Porter, 1980).
Customers with greater bargaining power are more likely to change providers when
they are not satisfied. Therefore, the relation between customer satisfaction and
future financial performance should be stronger when customer bargaining power is
greater. In contrast, customers with little bargaining power are more likely to stay
with the current provider even if they are not perfectly satisfied. For example, Jones,
Mothersbaugh, and Beatty (2000) show that customer switching costs, an important
determinant of customer bargaining power, moderate the relationship between
customer satisfaction and repurchase intentions. This study employs two different
research settings to triangulate on the relevant issues and to examine the following
research question: How does customer bargaining power affect the relation between
customer satisfaction measures and future financial performance?
Motivated by the mixed evidence on the effects of customer satisfaction on
financial performance across industries, the first research setting (Chapter 3) of my
dissertation investigates one component of customer bargaining power, customer
switching costs, in a large-sample cross-sectional study. Specifically, I predict that
high switching costs result in a weaker association between customer satisfaction and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
future financial performance. This is because customers may not switch to another
service provider in the presence of high switching costs even if they are not satisfied
with their current service providers.1
I measure customer satisfaction with the American Customer Satisfaction
Index (ACSI), which is an independent measure of the level of customer satisfaction
in selected firms across different industries. Industry-level switching costs are
measured with the Herfindahl-Hirschman Index (HHI). Economic theory suggests
that higher switching costs often result in lower competition and higher profit margin
in an industry (e.g., Klemperer, 1985,1995; Viard, 2003). In line with economic
theory, I infer industry-level switching costs from a measure that is commonly used
to capture industry concentration and competition.
The sample of service industry firms used in this study is the intersection of
the ACSI and COMPUSTAT databases. To test the hypotheses, I use cross-sectional
and time-series data of customer satisfaction, switching costs, and financial
performance from the sample firms for the years 1995-2002 in a moderated
regression analysis. I find that the relationship between customer satisfaction (as
1 For example, before the “number portability rule” (which allowed cellular phone customers to
switch carriers without changing their phone numbers) came into effect in November 2003, cellular
phone customers, especially business customers, were often stuck with a dissatisfying carrier because
it was a major hassle to notify others of their new phone numbers, providing a disincentive to change
providers (Wall Street Journal, June 9, 2003). The portability rule greatly reduced switching costs for
customers and increased the number of defecting customers sharply (about a million a month were
switching). AT&T Wireless, which had the worst customer-complaint rate of any major U.S. cellular
provider, was so hard hit by the portability rule that the company actually reported a quarterly net loss
of customers in the first quarter of 2004 for the first time in its history and eventually merged with
Cingular {Wall Street Journal, May 24, 2004; October 26, 2004). This anecdotal evidence illustrates
the effect of switching costs on the relationship between customer satisfaction and financial
performance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
measured by raw ACSI scores and industry-relative scores) and future financial
performance (as measured by one-year-ahead sales revenue, income before
extraordinary items, and ROA) is stronger for firms in service industries with low
switching costs (as measured by HHI). I also find that the relation between industry-
relative ACSI scores and the market value of equity is stronger for firms in industries
with low switching costs.
Chapter 3 of this study documents one aspect of customer bargaining power
(switching costs) that affects the relation between customer satisfaction measures
and future financial performance. However, firm-level analysis of cross-sectional
data unavoidably suffers from omitted variables and endogeneity problems. One way
to overcome the limitations of such analysis is to focus on a setting where different
measures are available for the same firm, which characterizes the second research
setting (Chapter 4) of my dissertation. This setting enables me to perform a cleaner
test of the effect of customer bargaining power on the relation between customer
satisfaction measures and future financial performance.
In Chapter 4 of this study, I obtain a unique database from a leading specialty
health benefit plan provider (hereafter “HBP”) that measures satisfaction levels of
multiple customer groups: (1) clients that purchase insurance plans for their
employees; (2) patients who use the insurance plans; and (3) doctors who provide
medical services. Due to significant competition in the health insurance sector,
enhancing customer satisfaction is a key element of HBP’s strategy. Consequently,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
the company builds its performance measurement and incentive system around a
business model that contains hypothesized links between financial performance and
various satisfaction measures.
As a non-profit organization, HBP’s primary objective is revenue growth
rather than profit maximization. It seeks to serve as many customers as possible in a
sustainable manner. Therefore, I measure relevance in this context as the strength of
the relation between various customer satisfaction measures and future revenues. I
hypothesize that the relevance of customer satisfaction measures in a setting with
multiple customer groups depends on the extent to which a customer group
influences the purchasing decision. This research setting is characterized by a two-
tier customer structure, where services are provided to one party (the patients), but
purchasing decisions are to a large extent made by another party (the clients). Since
clients tend to have greater influence on the purchasing decision and purchase in
greater volume than patients, I predict that client satisfaction, in general, is more
relevant than patient satisfaction. However, a more precise test would be to examine
the effect of the variation in the purchasing influence of clients vs. patients on the
relevance of satisfaction measures. The research site provides a unique measure to
proxy for such variation. HBP serves two different types of patients: voluntary and
non-voluntary patients. Non-voluntary patients do not have much purchasing
influence because their employers make the purchasing decision and payment for
them. In contrast, voluntary patients have much greater purchasing influence because
they pay a portion or all of the insurance premiums and have the option to choose
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
among different insurance products. I predict that the relevance of patient (client)
satisfaction increases (decreases) in the proportion of voluntary patients.
In addition, using HBP’s market penetration rate as a proxy for its market
power and a reverse measure of the bargaining power of its customers, I predict that
customer satisfaction measures have greater effects on future revenues when the
market penetration rate is low.
I test my hypotheses using structural equation modeling on the 51 markets of
HBP over a 20-quarter period from 2000-2004.1 find that, as predicted by HBP’s
business model, client satisfaction and doctor satisfaction are positively associated
with future revenues. The performance impact of customer satisfaction measures is
economically significant: 10% increase in client satisfaction leads to 18.2% increase
in future revenues, whereas 10% increase in doctor satisfaction leads to 15.9%
increase in future revenues. However, contrary to the business model, patient
satisfaction is negatively associated with future revenues. This result is consistent
with my hypothesis that client satisfaction is generally more relevant than patient
satisfaction. More importantly, this result, driven by a negative relation between
patient satisfaction with coverage and client satisfaction, suggests a misalignment of
interests between clients and patients. This is consistent with the health care
literature, which suggests that patients and clients often have conflicting objectives
in that patients prefer comprehensive coverage with the least out-of-pocket
disbursements, whereas clients give priority to cost containment (Mascarenhas,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
1993). Since clients tend to have greater purchasing influence than patients, the
preferences of clients dominate those of the patients.
As predicted, I find that patient satisfaction is more relevant in markets with a
higher percentage of voluntary patients, whereas client satisfaction is more relevant
in markets with a lower percentage of voluntary patients. Specifically, in the markets
with a higher percentage of voluntary patients, the relation between patient
satisfaction and future revenues becomes significantly positive: 10% increase in
patient satisfaction translates into 7.5% increase in future revenues. This result lends
further support to the argument that customer purchasing influence affects the
relevance of satisfaction measures.
Finally, consistent with my hypothesis, I find that client satisfaction, patient
satisfaction, and doctor satisfaction are more strongly associated with future
revenues in markets with lower market penetration rates, where customers tend to
have greater bargaining power.
My dissertation research has implications for research and practice. From a
research perspective, this study makes several contributions. First, the contingent
effects of customer bargaining power documented in this study contribute to our
understanding of the reasons for the mixed evidence on the relationship between
customer satisfaction and future financial performance. By showing that customer
bargaining power, an important aspect of competition, affects the relevance of
customer satisfaction, this study extends a stream of research that examines the effect
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
of competition on the usefulness of accounting information and management
controls (Hansen, 1998; Khandwalla, 1972; Krishnan, 2005).
Second, this is the first accounting study to examine the performance
consequences of satisfaction levels of different customer groups simultaneously.
While most studies in the prior literature ignore interactions among various
nonfinancial value drivers, the dataset allows me to explore these interactions and
potential tradeoffs among different value drivers.
Third, this study contributes to an emerging literature that empirically tests
business models that involve financial and non-financial performance measures
linked to firm-specific strategies (e.g., Campbell, Datar, Kulp, and Narayanan, 2004;
Malina and Selto, 2001, 2004; Rucci, Kim, and Quinn, 1998). The results in Chapter
4 challenge several assumptions in the managers’ hypothesized business model and
demonstrate variations of the business model in different markets, which highlights
the importance of validating hypothesized links in a firm’s performance
measurement system (Campbell et al., 2004; Ittner and Larcker, 1998b; Kaplan and
Norton, 1996, 2000).
From a practical perspective, a better understanding of the complex relations
between non-financial measures and future financial performance can help
companies improve their business models. For firms that have multiple customer
groups, the results indicate that managers should consider which customer group has
greater influence on the purchasing decision. By specifying the varying importance
of nonfinancial performance measures under different operating environments, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
results of this study enable a more refined interpretation of performance measures.
These insights allow managers to use nonfinancial measures more effectively in their
decisions concerning resource allocation, performance evaluation, and compensation
practices.
The remainder of the dissertation is organized as follows. Chapter 2 reviews
the literature on the relation between customer satisfaction and future financial
performance. Chapter 3 presents evidence on the effect of customer switching costs
on the relation between customer satisfaction and financial performance in a large-
sample study. Chapter 4 presents evidence on the impact of customer bargaining
power on the relevance of customer satisfaction measures in a field setting, where
the research site serves multiple customer groups. Chapter 5 concludes and suggests
directions for future research.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
Chapter 2. Literature Review
A growing body of literature on nonfinancial performance measures shows
that some nonfinancial performance measures are leading indicators of future
financial performance (e.g., Amir and Lev, 1996; Banker et al., 2000; Ittner and
Larcker, 1998a; Nagar and Rajan, 2001; Said, HassabElnaby, and Wier, 2003).
While financial performance measures are predominantly short-term and backward-
looking in nature, many nonfinancial performance measures such as quality and
customer satisfaction are future-oriented. They drive future financial performance
and, hence, help managers focus their attention on the long-term aspects of the
business (e.g., Banker et al., 2000; Hemmer, 1996; Ittner and Larcker, 1998a, 1998b,
2001, 2003; Kaplan and Norton 1992, 1996).
In particular, customer satisfaction, one of the most widely used nonfinancial
measures, has attracted significant attention from researchers and practitioners alike.
Customer satisfaction is defined as an output resulting from the customer's pre
purchase comparison of expected performance with perceived actual performance
and incurred cost (Churchill and Surprenant, 1982). The marketing literature
suggests that customer satisfaction operates in two different ways: transaction-
specific and general overall (Yi, 1991). The transaction-specific concept concerns
customer satisfaction as the assessment made after a specific purchase occasion.
Overall satisfaction refers to the customer's rating of the brand, based on all
encounters and experiences (Johnson and Fomell, 1991). Measures used in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
current study capture the overall satisfaction concept, which can be viewed as a
function of all previous transaction-specific satisfactions (Jones and Suh, 2000).
Prior studies have predominantly documented positive associations between
customer satisfaction and future financial performance (Behn and Riley, 1999; Ittner
and Larcker, 1998a; Banker et al., 2000; Bernhardt et al., 2000; Smith and Wright,
2004; Dikolli, Kinney, and Sedatole, 2005). For instance, Bernhardt et al. (2000)
finds a positive and significant relationship between customer satisfaction and one-
year-ahead profitability a fast-food restaurant. Similarly, Banker et al. (2000) detects
a six-month lead-lag relation between customer satisfaction and financial
performance in the hospitality industry and finds that incentive bonus contracts that
incorporate customer satisfaction measures positively affect future financial
performance. Ittner and Larcker (1998a) establish positive relationships between
customer satisfaction and future financial performance at the customer level (in a
telecommunications firm), the business division level (in the banking industry), and
the firm level (using the American Customer Satisfaction Index). They also show
that market value of public companies is positively associated with customer
satisfaction for the transportation, utilities and communication sectors. The above
evidence is consistent with the marketing literature, which suggests that higher
customer satisfaction increases customer retention (Fomell, 1992), reduces price
elasticities, lowers marketing costs (Anderson, Fomell and Lehmann, 1994;
Zeithaml, 2000), and increases repurchase and referral intentions (Anderson and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
Sullivan 1993; Cronin and Taylor 1992), all of which should improve financial
performance.
Prior studies, however, have also demonstrated variations in the relation
between customer satisfaction and financial performance. For example, Anderson et
al. (1997) find that the positive effects of customer satisfaction are more prominent
in manufacturing firms than in service firms. Ittner and Larcker (1998a) find
customer satisfaction to be positively associated with market values in the
manufacturing and financial service industries, but negatively associated with market
values in the retail industry.
The economics literature suggests that customer bargaining power is one
contingency factor that might account for cross-sectional differences in the relation
between customer satisfaction and future financial performance. Customer
bargaining power, one of the fives forces of competition, is the impact that customers
have on a company (Porter, 1980). Porter outlines the determinants of customer
bargaining power, including customer concentration, customer volume, customer
switching costs relative to firm switching costs, brand loyalty, customer price
sensitivity, threat of backward integration, and customer information. When
customers are not satisfied with their current service provider, they will take into
consideration the costs and benefits of switching to an alternative service provider.
When they have more bargaining power, the benefits are likely to increase due to
more favorable trade terms with alternative service providers and the costs are likely
to decrease due to lower customer switching costs and less brand loyalty. Therefore,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
customers with greater bargaining power are more likely to change providers when
they are not satisfied. As a result, the relation between customer satisfaction and
future financial performance should be stronger when customer bargaining power is
greater. In contrast, customers with little bargaining power are more likely to stay
with the current provider even if they are not perfectly satisfied.
As Ittner and Larcker (2001) point out, another limitation of studies on non
financial performance measures is that they tend to examine only one of many
potential non-financial measures and ignore interactions with other potential non
financial measures. They caution that these limitations can result in misleading
inferences if non-financial measures are highly correlated, or if different non
financial measures are complements or substitutes. Along the same line, advocates of
a “business model” approach to performance measurement propose formulating
performance measurement systems around a diverse set of financial and non
financial performance measures that are linked to firm-specific strategies (Eccles,
1991). Kaplan and Norton (1996) argue that a balanced scorecard should not just be
a collection of financial and nonfinancial measures in various categories, but rather
an integrated set of measures developed from a business model that articulates the
cause-and-effect relationships between the selected performance measures and
outcomes. However, survey evidence suggests that only 23% of companies
consistently built and tested the strength and validity of the hypothesized business
models (Ittner and Larcker, 2003). Corresponding to the lack of extensive causal
modeling and validation among practitioners, the academic literature provides
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
surprisingly little evidence on how companies develop explicit business models,
whether these models can be validated, and how these models vary depending on
contextual factors such as strategy, competitive environment, and organizational
design (Ittner and Larcker, 2001).2
A well-known exception is the Sears model illustrated in Rucci et al. (1998).
More than 100 top-level executives at Sears, Roebuck and Co. spent three years
rebuilding the company around its customers. The Sears model hypothesizes a chain
of cause and effect running from employee behavior to customer behavior to profits.
This model shows that a 5-point improvement in employee attitudes drives a 1.3-
point improvement in customer satisfaction, which in turn drives a 0.5%
improvement in revenue growth. Rucci et al. also shows that the business model
works for Sears: customer satisfaction increased 4% after incorporating the results
from the model into the choice of quality/customer initiatives and the design of their
long-term performance plan. The increase in customer satisfaction led to an
estimated $200 million increase in revenues, and ultimately an estimated $250
million increase in market capitalization.
Two recent studies also provide empirical evidence on whether firm-specific
business models can be validated and how they vary with contingency factors.
Campbell et al. (2004) analyze a diverse set of financial metrics and non-financial
measures of strategy implementation and employee capabilities from the
2 Exceptions include Malina and Selto (2004) and Abemethy et al. (2005), which provide field-based
evidence on how companies develop explicit business models.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
performance measurement system of a convenience store chain that implemented,
and subsequently abandoned, a new operating strategy. They investigate the extent to
which a diverse set of performance measures, and the links between these measures,
reveal information about the quality and implementation of a firm’s strategy. They
find that the impact of an increase in the firm’s nonfinancial measure of strategy
implementation on financial performance depends on employee skill levels. These
findings underscore the importance of considering interactions among nonfinancial
measures of employee capabilities and strategy implementation when implementing
business model based performance measurement systems. Similarly, Nagar and
Raj an (2005) adopts a models view that considers customer relationships as a
multidimensional process that involves both financial and nonfinancial measures that
are causally linked to each other. They employ a unique and comprehensive cross-
sectional database on retail banks to demonstrate empirically how managers can use
a set of customer relationships measures to identify cause-and-effect
interrelationships among various customer relationship activities, including price
metrics, service metrics, customer satisfaction, customer usage and volume metrics.
They find that the metrics do not individually predict future earnings, but predict
future earnings when combined. They also demonstrate that the forward-looking
nature of nonfinancial measures depend on broader environmental features such as
strategy.
A business model approach is especially important for the analysis of
customer satisfaction because companies often serve multiple customer groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
Though enhancing customer satisfaction is the cornerstone of many companies’
competitive strategies, defining customers may not be straightforward. Historically,
in the 1920s and 1930s, the term “customer’ primarily referred to distributors who
purchased goods and made payments (McKitterick, 1957). Starting about the 1950s,
the focus shifted from distributors to end consumers and their needs and preferences.
Today the term “customer” includes end users, distributors, and any other party that
an organization serves (Kohli and Jaworski, 1990). Many firms have consumers (i.e.,
end users of products and services) as well as clients (i.e., organizations that may
dictate or influence the choice of end users). For examples, executives of several
packaged goods distributors indicated that it is critical for their organizations to
understand the needs and preferences not just of end customers but also of retailers
through whom their products are sold (Kohli and Jaworski, 1990). Keeping retailers
satisfied is important to ensure that they carry and promote the distributors’ products,
which in turn enables the distributors to cater to the needs of their end customers
(Kohli and Jaworski, 1990). Niraj, Gupta and Narasimhan (2001) make a similar
point and build a detailed activity-based cost model to understand the behavior of
costs and profits associated with different types of customers in a supply chain. One
limitation of Niraj et al. (2001) is that it does not consider externalities among
different types of customers. In today’s business environment with ever-increasing
outsourcing activities across many industries and functions, dealing with multiple
customer groups becomes even more prevalent.3 Outsource service providers should
3 Although most companies have outsourced selected projects and functions for many years,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
not only establish good relationships with clients, who typically negotiate
outsourcing deals, but also understand the needs and preferences of end users of the
services. Conflicts may arise when the interests of clients and end users are not
perfectly aligned.
outsourcing has grown tremendously since the 1990s as companies have sought to lower labor costs,
improve quality of service, and focus on core competencies (Langfield-Smith and Smith, 2003).
Recent studies indicate that 85% of all companies outsource at least one function (Elmuti, Kathwala,
and Monippallil, 1995). In a survey of more than 300 executives, 58% of the respondents said that
outsourcing is absolutely essential to their companies’ maintaining a competitive advantage. Thirty-
eight percent of the respondents said at least 11% of their total operations are outsourced, and 17%
outsource more than 20% of their total operations (Williams, 1997). Outsourcing activities cover a
wide range, including, but not limited to, the following functions: logistics, human resource, customer
service, manufacturing, engineering, research and development, finance and accounting, software
development, maintenance, and support, facilities management, telemarketing, and training and
development.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
Chapter 3. Evidence from a Large-Sample Study
3.1. Hypotheses
As discussed in Chapter 2, prior studies have demonstrated variations in the
relation between customer satisfaction and financial performance across industries.
Contingency theory suggests that an organization must be aligned with its
environment to achieve optimal performance (Chenhall, 2004). Consistent with this,
researchers argue that the relevance of financial or non-financial measures should be
affected by contextual factors (e.g., Shevlin, 1996). Nagar and Rajan (2005)’s study
of retail banks, for example, found that a nonfmancial measure, cross-sell ratio, is
more critical for banks that emphasize the sales of new products as their strategy.
Along the same lines, the mixed results on the satisfaction-performance linkage
imply that customer satisfaction is effective in improving future financial
performance of service firms only under certain situations. Though prior studies have
documented variations in the relation between customer satisfaction and financial
performance, little empirical evidence exists on contextual factors that moderate the
positive effects of customer satisfaction on future financial performance.
Studies in the economic and marketing literature suggest that one of the most
important factors that could potentially moderate the relation between customer
satisfaction and financial performance is “switching costs”, a component of customer
bargaining power (Porter, 1980). Switching costs take away power from customers
by: (1) creating an exit barrier to customers in a low satisfaction relationship; (2)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
decreasing the number of competitors in the market; and (3) increasing price levels
in the market. Switching costs are defined as the perceived or actual economic and
psychological costs associated with changing from one supplier to another (Jones et
al, 2000; Kim, Kliger, and Vale, 2001).4
Porter (1980) suggests that switching costs act as an exit barrier in a low
satisfaction relationship. In other words, even though customer satisfaction with a
relationship may be low, the customer may stay due to the high psychological and
economic costs of switching. Jones et al. (2000)'s study of customer switching costs
empirically tests Porter’s prediction. The authors propose that the effect of customer
satisfaction on repurchase intentions depends on the perceived magnitude of
switching costs in the service context. Using survey data collected from bank and
hairstyling/barber customers, the authors examine the moderating effects of various
switching costs, including switching costs, interpersonal relations, and attractiveness
of alternatives. The results support the hypothesis that the effect of customer
satisfaction on repurchase intentions decreases when customers perceive high
switching costs. The authors also find that the positive effect of switching costs on
4
Jones et al. (2002) clarify this construct by identifying the six dimensions of switching costs: 1) Lost
performance costs, 2) Uncertainty costs, 3) Pre-switching search and evaluation costs, 4) Post
switching behavioral and cognitive costs, 5) Setup costs, and 6) Sunk costs. Similarly, the economic
literature has identified economic as well as psychological origins of switching costs. For example,
economic origins of switching costs include learning costs, search costs, intertemporal product and
service compatibility, and informational investment in business relationships. Addiction and cognitive
dissonance are examples of psychological switching costs (Kim et al., 2001). Klemperer (1995)
summarizes consumer switching costs into six categories: a) Need for compatibility with existing
equipment; b) Transaction costs of switching suppliers; c) Costs of learning to use new brands; d)
Uncertainty about the quality of untested brands; e) Discount coupons and similar devices; and f)
Psychological costs of switching, or non-economic “brand-loyalty.”
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
repurchase intentions becomes more evident among companies where customer
satisfaction is low.
Economics-based analytical work has mainly examined the effect of
switching costs on prices and market power. For example, Klemperer (1987)
examines a two-period differentiated-product duopoly in which customers are
partially “locked-in” by switching costs that they face in a second period, which
results in higher prices in both periods compared to the non-switching cost case.
Klemperer (1995) generalizes this finding to a multi-period model. In addition,
Klemperer (1995) also shows analytically that switching costs discourage new entry
and reduce customers’ switching between firms, thus reducing the market’s
competitiveness. Empirical studies have generally supported these theoretical
predictions. For example, Ausubel (1991) and Stango (1998) study the credit card
market and find that switching costs are an important influence on high interest rates
in that market. Dahlby and West (1986) find that the high search costs in the liability
insurance industry results in high price dispersion. Similar results are found for auto
insurance (Schlesinger and von der Schulenburg, 1993). Kim, Kliger, and Vale
(2001) estimate the magnitude and significance of switching costs in the market for
bank loans and find that the average switching cost is about one third of the market
average interest rate on loans. They also find that more than a quarter of the
customer’s added value is attributed to the lock-in phenomenon generated by these
switching costs. Knittel (1997) shows that switching costs provide long distance
telephone carriers with market power. Similarly, Viard (2003) looks at 800-number
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
pricing before and after such numbers became portable and finds that lower
switching costs lead to lower prices and more competition.
The above findings from the economic literature suggest that customer “lock-
in” due to switching costs is commonly found in many service industries. As a result,
the effect of customer satisfaction on future economic benefits is weakened because
“locked-in” customers are not likely to switch to another service provider even if
they are not satisfied with their current service providers. In fact, due to the increased
importance of “new economy” information-based industries and the popularity of
“lock-in”, Hax and Wilde (2001) and Kaplan and Norton (2004) have articulated the
“lock-in strategy” as an alternative to the traditional “cost leadership” and
“differentiation” strategies. Under the “lock-in strategy”, companies seek to generate
long-term sustainable value by creating high switching costs. The most prominent
example of companies adopting this strategy is Microsoft. Even if customers were
not happy with the operating system of Microsoft, they would find it difficult to
switch to a different system because they would lose access to many application
programs that could only run on a Windows operating system.
The preceding discussion highlights the need to incorporate switching costs
as a moderator in the investigation of the relation between customer satisfaction and
financial performance. Figure 1 summarizes the linkages among customer
satisfaction, customer switching costs, and future financial performance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
Figure 1. Conceptual Model of the Relations among Customer Satisfaction,
Switching Costs, and Future Financial Performance
Switching
Costs
Customer
Satisfaction
Future
Financial
Performance
Based on prior literature and the above reasoning, I propose that high
switching costs attenuate the positive effects of customer satisfaction on the future
financial performance of firms in that industry. Specifically, I posit the following
hypothesis:
Hla: The positive association between customer satisfaction and future
financial performance is decreasing in switching costs.
Figure 2 illustrates H la .5
5 Dikolli et al. (2003) proposed a competing hypothesis, which predicts that the positive association
between customer satisfaction and future economic benefits is increasing in customer switching costs.
However, this hypothesis is reasonable only when switching costs are so low that no customer “lock-
in” will occur. The authors test their hypothesis on a sample of online retailers. However, both the
retail industry and Internet companies are characterized by very low switching costs (e.g. Bakos,
1991; Bakos, 1997). Therefore, although they find results consistent with their hypothesis, the
generalizability of their results to other industries is seriously weakened. In any given industry, it
remains an empirical question whether switching costs are large enough to create the “lock-in”
phenomenon and therefore weaken the association between customer satisfaction and future economic
benefits.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
Figure 2. Graphical Presentation of the Hypotheses
Future
Financial
Performance
Low Switching Costs
High Switching Costs
Customer Satisfaction
Prior literature suggests that the length of lag between customer satisfaction
and future accounting performance varies across firms. For example, Bernhardt et al.
(20 0 0)’s longitudinal study of a fast-food restaurant finds a one-year lead-lag
relationship between customer satisfaction and financial performance, whereas
Banker et al. (2000)'s longitudinal study detects a six-month lead-lag relation
between customer satisfaction and financial performance in the hospitality industry.
Because of this, there is a possibility that the effect of customer satisfaction on future
accounting performance may not reveal itself at the same time interval. As an
alternative way of assessing the strength of relationship between customer
satisfaction and future financial performance, I examine the association between
customer satisfaction measures and the market value of equity, after controlling for
information contained in contemporaneous accounting book values. Since the market
value of equity represents investors’ aggregate expectation of a firm’s future cash
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
flows, this test may alleviate the problem of differential time lags in the relationship
between customer satisfaction and future accounting performance under the
assumption of semi-strong market efficiency. In line with Hypothesis la, I posit the
following hypothesis:
Hlb: The positive association between customer satisfaction and the
contemporaneous market value o f equity is decreasing in customer
switching costs.
3.2. Sample and Measures
My sample consists of the intersection of service firms covered by the
American Customer Satisfaction Index (ACSI) and COMPUSTAT. ACSI is the only
national cross-industry indicator of customer satisfaction with the quality of goods
and services. Maintained jointly by the University of Michigan Business School and
the American Society for Quality, ACSI measures customer satisfaction of seven
economic sectors, 35 industries, 190 firms, and federal or local government agencies.
Although ACSI tends to cover major players in each industry,6 the measured
companies, industries and sectors in ACSI are broadly representative. Moreover, the
ACSI scores represent independent evaluation of customer satisfaction by a third-
party organization, which enhances the objectivity of the measures.
6 Within the selected two-digit SIC code industry groups, companies at the four-digit SIC code are
selected based on their total sales. As a result, companies covered by ACSI represent a major
proportion of sales of the selected two-digit SIC code industry groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
The service industries that ACSI covers are divided into five broad economic
sectors: finance/insurance,7 retail,8 transportation/communications/utilities,9
services,10 and E-commerce.1 1 Due to data requirements, the sample is reduced to 83
firms across 25 industries (4-digit SIC), and 504 firm-year observations.
I test the following equation:
Financial Performance = f (Customer Satisfaction, Switching Costs, Customer
Satisfaction x Switching Costs, Controls)
Customer Satisfaction
Customer satisfaction is operationalized by the ACSI score for each firm. The
ACSI data are available from the ACSI Website: http://www.theacsi.org.
The ACSI data are collected at the individual customer level, with indices for
a company's customers aggregated to produce company-level indices. Industry
indices consist of company indices, weighted by the sales of each company. The
ACSI scores come from telephone surveys of random samples of the customers of
1 9
rated companies. Since the first release of ACSI scores in December 11,1995, nine
7 The finance/insurance sector includes the following industries: banks, life insurance, and
property/casualty insurance.
8 The retail sector includes the following industries: department and discount stores, supermarkets,
fast food and gas stations.
9 The transportation, communications and utilities sector includes the following industries: parcel
delivery, US Postal Service, airlines, telecommunications, broadcasting/TV, publishing/newspapers,
and utilities.
10 The services sector includes the following industries: hotels, hospitals, and motion pictures.
11 The E-commerce sector includes the following industries: portals, retail, auction/reverse auction,
and brokerages.
1 2 ACSI scores are based on 15 questions on ten-point scales. The questions are formed into four latent
variables: perceived quality, customer expectations, perceived value, and customer satisfaction. The
first three variables are conceptualized as the antecedents of customer satisfaction. These latent
variables are linked in a causal model to two latent variables for customer satisfaction consequences:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
annual indices at the company, industry and sector level have been released. As
customer satisfaction may have a lead-lag relationship with financial performance, I
use the ACSI data from 1994 to 2002 to obtain customer satisfaction measures for
longer time periods.
Financial Performance
Financial performance is operationalized by four measures: sales revenue
(SALES), income before extraordinary items (PROFIT), return on asset (ROA), and
market value of equity (MVE). These data are available from COMPUSTAT. Since
different industrial sectors are measured during different calendar quarters, 131 use
the financial performance measures for the fiscal year-end closest to the month the
ACSI scores are released, as Ittner and Larcker (1998). Natural logarithm is taken for
each of these variables to ensure normality.
Switching Costs
Industry-level switching cost is measured with the Herfindahl-Hirschman
Index (HHI). Analytical work in economics has shown that consumer switching
costs lead to high prices and create deadweight losses of the usual kind in a closed
oligopoly, and may also discourage new entry and so further reduce the market’s
competitiveness (Klemperer, 1987; 1995). Ample empirical evidence supports these
theoretical arguments. Researchers have found that high switching costs reduce
customer complaints and customer loyalty. The scores for individual customers are re-scaled to range
from 0 (least satisfied) to 100 (most satisfied) (Fomell et al., 1996; Ittner and Larcker, 1998).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
competition in many industries, including credit card, auto and liability insurance,
banking, and telephony (e.g., Dahlby and West, 1986; Kim et al., 2001; Viard,
2003). Based on the economics literature, I infer switching costs from HHI, a
commonly accepted measure of competition and market concentration (Rhoades,
1993; DeFond and Park, 1999; Baker, 2001). It is calculated by squaring the market
share of each firm competing in the market and then summing the resulting numbers.
The HHI takes into account the relative size and distribution of the firms in an
industry and approaches zero when an industry consists of a large number of firms of
relatively equal size. The HHI increases both as the number of firms in the market
decreases and as the disparity in size between those firms increases. High HHI
values indicate high levels of industry concentration and hence high levels of
switching costs. To minimize time-series variations in HHI, the HHI value for each
firm is calculated as its industry (2-digit SIC) average HHI over the eight years of
this study (1995-2002).
Control Variables
The size of a firm has been shown to affect the purchase behavior of
customers because a high market share elicits a higher level of trust from customers
(Brynjolfsson and Smith, 2000). It is thus speculated that a high market share can be
self-perpetuating to some extent. In addition, Fama and French (1995) document that
13 Specifically, ACSI scores for the transportation, communications and utilities sector and the
services sector are released in the firs quarter every year, whereas ACSI scores for the
finance/insurance sector and the retail sector are released in the fourth quarter every year.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
larger firms have higher operating performance. Therefore, size is included as a
control variable. Size is measured as the natural logarithm of total assets.
As done by Ittner and Larcker (1998), past performance is included to control
for time-series trends as well as to examine whether the ACSI scores provide
incremental information on future performance. Past performance is measured with a
one-year lag. Though the choice of a one-year lag is admittedly arbitrary, it is
reasonable given the relatively short repurchase cycle in the service industries. For
example, Banker et al. (2000) documented a six-month lead-lag relation between
customer satisfaction and financial performance in the hospitality industry and
Bernhardt et al. (2000) documented a one-year lag in the fast-food restaurant
industry.
Firms adopting a “differentiator” strategy may put more emphasis on non-
financial performance measures such as customer satisfaction (Ittner et al., 1997) and
therefore achieve higher customer satisfaction. On the other hand, prior research
examining the financial performance of the “differentiator” vs. “cost leader”
strategies indicates that Differentiators may achieve superior financial performance.
For example, a number of studies have found Differentiators outperforming Cost
Leaders in terms of return on investment (De Castro and Chrisman, 1995; Wright et
al. 1991), sales growth (White, 1986), and new product and new market performance
(Porter, 1980; Valos et al., 2004). Therefore, strategy could potentially be a
correlated omitted variable. Market-to-book ratio is included to proxy for strategy.
Market-to-book ratio has been used in prior literature as one of the indicators of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
“Differentiator” strategy (Ittner et al. 1997; Said et al. 2003) since Differentiators are
expected to have greater growth potential than Cost Leaders.
Advertising intensity may be positively associated with customer satisfaction.
For example, prior research finds that intensive advertising reinforces feelings of
satisfaction for brands already being used and sometimes, through subsequent
reinforcement, helps to facilitate the development of a repeat buying habit (e.g.
Enhrenberg, 2000). Advertising intensity is also found to be positively associated
with future financial performance. For example, advertising expenditures have been
shown to affect sales and market share positively (Assmus, Farley, and Lehmann
1984; Tellis 1988). Hence, advertising intensity can potentially be a correlated
omitted variable. Adverting intensity is measured by advertising-to-sales ratio.
3.3. Empirical Analyses and Results
Table 1 presents sample industry composition and average HHI by Industry.
Sixteen 2-digit SIC service industries (25 4-digit SIC industries) are represented in
the sample, with utilities, general merchandise stores, and transportation by air
accounting for the largest percentages. There is a large amount of variation in HHI,
ranging from 0.107 to 0.429, with the average being 0.218 (median=0.225, standard
deviation=0.081). Table 2 summarizes descriptive statistics. The average ACSI
score is 72.91 (median=74.00) and the standard deviation is 5.58. The average
sample firm has $44,724 million (median=$ 15,966 million) in assets, $19,216
million (median=$ 13,540 million) in sales revenue, $934 million (median=$591
million) in profit, and $38,662 million (median=2,512 million) in market value of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
equity. As documented by prior literature (Ittner and Larcker, 1998), the sample
firms are larger and have more market share than the general population. This limits
the generalizability of research findings using the ACSI database.
Table 3 presents Pearson correlations between independent and control
variables. The customer satisfaction measure is negatively correlated with size (r=-
0.214, p<0.01) and contemporaneous profit (r=-0.110, p<0.05), but positively
correlated with contemporaneous ROA (r=0.165, p<0.01). The switching costs
measure is negatively correlated with size (r=-0.167, p<0.01) and profit (r=-0.093,
p<0.05), and positively correlated with sales revenue (r=0.163, p<0.01), ROA
(r=0.189, p<0.01), and market-to-book ratio (r=0.107, p<0.05). Advertising-to-sales
ratio is negatively correlated with size (r=-0.198, p<0.01), contemporaneous sales
(r=-0.477, p<0.01), contemporaneous profit (r=-0.339, p<0.01), and
contemporaneous ROA (-0.198, p<0.01).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
Table 1. Sample Industry Composition and Average Herfindahl-Hirschman
Index (HHI) by Industry
2 -digit
SIC
Industry description Average HHI
(square root)
Number of
observations
Percentage
42 Trucking and warehousing 0.370 8 1.587
45 Transportation by air 0.228 56 1 1 . 1 1 1
47 Transportation services 0.387 1 0.198
48 Communications 0.225 54 10.913
49 Electric, gas, and sanitary
services
0.107 115 22.817
51 Wholesale trade-nondurable
goods
0.198 3 0.595
52 Building materials and garden
supplies
0.336 4 0.794
53 General merchandise stores 0.297 77 15.278
54 Food stores 0.244 54 10.714
58 Eating and drinking places 0.267 2
0
4.365
59 Miscellaneous retail 0 . 2 2 0
L
1 0
1.984
60 Depository institutions 0.162 36 7.143
62 Security and commodity
brokers
0.429 6 1.190
63 Insurance carriers 0.199 32 6.349
70 Hotels and other lodging places 0.397
19
3.770
73 Business services 0.139 6 1.190
Total 0.218 504 1 0 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Table 2. Descriptive Statistics
ACSI
score HHI
Assets (in
millions)
Market value
of equity (in
millions)
Sales (in
millions)
Income before
extraordinary
items (in
millions)
N
504 504 504 504 504 504
Mean
72.91 0.218 44,724.00 38,662.00 19,216.48 934.49
Median 74.00 0.225 15,965.70 2,511.99 13,540.50 590.50
Std.
Deviation
25t h
5.58 0.081 86,326.27 121,659.51 23,466.06 1,613.78
percentile
70.00 0.162 8,498.35 623.95 7,125.80 219.95
75t h
percentile
76.00 0.267 35,943.00 12,354.55 24,250.00 1,183.00
Table 3. Pearson Correlations between Independent Variables and Control
Variables
SWITCH SIZE SALES PROFIT ROA MVE MB ADV
CS
.040 -.214** -.060 -.1 1 0 * .165** -.068 -.072 .019
SWITCH
-.167** .163** -.093* .189** .041 .107* .129
SIZE
.681** .728** -.540** .669** .477** -.198**
SALES
.678** -.053 .506** .337** -.477**
PROFIT
.185** .561** .376** -.339**
ROA
-.246** -.207**
_198**
MVE
.958** -.057
MB
.176*
Note: See Table 4 for variable definitions.
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
Hypothesis la predicts that customer satisfaction is more closely associated
with future accounting performance in low switching cost industries than in high
switching cost industries. I test the hypothesis with moderated regression analysis.
Following Hartmann and Moers (1999), both main effects and interaction effects are
included in the regression. As the main-effect terms and product terms may be highly
correlated, raising the concern of multicollinearity, which can make regression
coefficients unstable and difficult to interpret, main-effect terms and product terms
were centered to address these problems (Aiken & West, 1991). There has been
considerable debate on the appropriate functional form of the relation between
customer satisfaction and financial performance (Ittner and Larcker, 1998; Lambert,
1998). Banker et al. (2000) argues that because of the decaying effect of nonfinancial
measures over time, the permanent effect assumption of the changes model are
violated. The levels model is therefore more appropriate because residuals from the
levels model will have a lower autocorrelation than the changes model. Following
Banker et al. (2000), I use a levels model with a lagged dependent variable:
PERF i t+4 = q + Bi CS j,t + J 3 2 SWITCH, + fl3 CS i# t x SWITCH , + B 4 SIZE ijt+4 +
65 PERF i t +1^6 MB j t+4 + 67 ADV i t+4 + q j t+4
Three performance variables (PERF) will be examined: SALES, PROFIT, and ROA.
where:
t = the quarter that the ACSI score was released for a certain industry1 4
14 Each quarter, ACSI scores for one or two economic sectors of the U.S. economy are updated.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
SALES i, t+4 = Natural logarithm of sales revenue of firm i in quarter t+4
PROFIT j, t+4 = Natural logarithm of income before extraordinary items of firm i in
quarter t+4
ROA i, t +4 = Natural logarithm of return on assets for firm i in quarter t+4
CS i, t = Mean-centered ACSI score for firm i in quarter t (ACSI score - 72.91)
SWITCH j = Mean-centered Herfindahl-Hirschman Index (HHI) for the industry that
firm i belongs to (calculated as the mean of the sum of the squared market shares in
percentage of all firms in a 2-digit SIC industry, computed over the five years prior
to the event year) (HHI - 0.218)
SIZE i, t+4 — Natural logarithm of the total assets of firm i in quarter t+4
PAST PERF j, t = performance (REV, PROFIT, ROA, and ROE) of firm i in quarter t
MB i, t=4 = Market-to-book ratio of firm i in quarter t+4
ADV i, t+4 Advertising-to-sales ratio of firm i in quarter t+4
White’s test indicated that heteroscedasticity was not a problem for this
model. The Durbin-Watson statistic revealed no significant autocorrelation. Based on
prior literature, I predict the coefficient on customer satisfaction to be significantly
positive. I predict a significantly negative coefficient on the interaction term between
customer satisfaction and the measure of industry-level switching costs, which would
imply a significant moderating effect of industry-level switching costs in the relation
between customer satisfaction and future financial performance in the sample firms.
I do not make a prediction for the coefficient on switching costs. Finally, I predict
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
significantly positive signs for the coefficients on size, past performance, market-to-
book ratio, and advertising-to-sales ratio.
Table 4 presents the results of the test of Hypothesis la. Panel A summarizes
the results of the regression when sales revenue is the dependent variable. Panel B
summarizes the results when income before extraordinary items is the dependent
variable, and Panel C summarizes the results when ROA is the dependent variable.
The results indicate that, as predicted, the coefficient on customer satisfaction is
significantly positive when sales revenue (13=0.007, t=2.274, p=0.024) and income
before extraordinary items (13=0.024, t=2.186, p=0.031) are the dependent variables,
but the coefficient is insignificant when ROA is the dependent variable (13=0.020,
t=l .257, p=0.211). The coefficient on switching costs as measured by HHI is
insignificant when sales revenue is the dependent variable (13= -0.231, t=-0.604,
p=0.547) and significantly negative when income before extraordinary items (13= -
2.895, t=-2.373, p=0.019) and ROA (13= -4.727, t=-3.033, p=0.003) are the
dependent variables. The coefficient on the interaction term between customer
satisfaction and switching costs is significantly negative for all three performance
measures (13= -0.176, t= -1.941, p=0.054 for SALES regression, 1 3 = -0.623, t= -2.081,
p=0.039 for PROFIT regression, 1 3 = -0.657, t= -1.697, p=0.092 for ROA regression).
As expected, coefficients on past performance are significantly positive in all three
regressions. The coefficients on size are significantly positive when sales revenue
and income before extraordinary items are the dependent variables, and significantly
negative when ROA is the dependent variable. The coefficients on market-to-book
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
ratio are insignificant in all three regressions. Surprisingly, the coefficients on
advertising-to-sales ratio are significantly negative in all three regressions. The
results imply that for firms in service industries, customer satisfaction has
significantly greater association with future accounting performance in industries
with low switching costs than in industries with high switching costs. These findings
are consistent with Hypothesis la.
The above analysis is conducted with absolute levels of firms’ ACSI scores.
As a robustness check, I conducted the same analysis with firms’ ACSI scores
relative to industry benchmarks. The ACSI database provides customer satisfaction
scores for “All Others” for each industry. The “All Others” category represents the
remainder of the total industry market share, less the market shares of the ACSI-
measured companies. This score is an aggregate of a representative number of
customer interviews from each of potentially hundreds of smaller companies within
the industry. I reran the regression after replacing the absolute ACSI scores with the
differences between each firm’s ACSI scores and the “All Others” ACSI scores for
the relevant industry. Table 5 presents the results, which are qualitatively similar to
those using the absolute levels of ACSI scores. The coefficients on customer
satisfaction are significantly positive, whereas the coefficient on the interaction term
between customer satisfaction and switching costs are significantly negative in all
three regressions. The interaction effect is stronger for the ROA regression (13= -
0.295, t= -2.616, p=0.010) than in Table 4. Again, these results are consistent with
Hypothesis la.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
Table 4. The Effects of Customer Satisfaction, Switching Costs, and an
Interaction Term between Customer Satisfaction and Switching Costs on
Future Financial Performance
PERF; t+ 4 = 3. + B, C S; t + B 2 SWITCH; + B 3 CS M x SWITCH i + B 4 SIZE i t+ 4 +
6 5 PASTPERF i t + B g MB ; t +4 + B 7 A D V ; t+ 4 + q j t+ 4
Panel A: One-year ahead sales revenue (SALES) as the dependent variable
Independent Variables Expected
Sign
Unstandardized
Coefficients
Standardized
Coefficients
t Two-tail
p-values
Constant 0.164 1.344 0.181
CS + 0.007 0.036 2.274 0.024
SWITCH
?
-0.231 -0.008 -0.604 0.547
CS x SWITCH - -0.176 -0.033 -1.941 0.054
SIZE + 0.061 0.068 2.822 0.005
PASTPERM +
0.906 0.921 33.952 0 . 0 0 0
MB + 0.005 0 . 0 1 2 0.860 0.391
ADV + -0.062 -0.044 -2.782 0.006
N 504
Adjusted R2 0.979
Panel B: One-year ahead income before extraordinary items (PROFIT) as the dependent
variable
Independent Variables Expected
Sign
Unstandardized
Coefficients
Standardized
Coefficients
t Two-tail
p-values
Constant -0.834 -2.016 0.046
CS + 0.024 0.107 2.186 0.031
SWITCH
?
-2.895 -0.097 -2.373 0.019
CS x SWITCH - -0.623 -0 . 1 1 1 -2.081 0.039
SIZE + 0.377 0.387 5.811 0 . 0 0 0
PASTPERM + 0.524 0.537 7.935 0 . 0 0 0
MB + 0 . 0 1 2 0.029 0 . 6 6 8 0.506
ADV + -0.137 -0.090 -2.156 0.033
N 504
Adjusted R2 0.819
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
Table 4. (Continued)
Panel C: One-year ahead return on assets (ROA) as the dependent variable
Independent Variables Expected
Sign
Unstandardized
Coefficients
Standardized
Coefficients
t Two-tail
p-values
Constant -0.614 -0.814 0.417
CS + 0 . 0 2 0 0.146 1.257 0 . 2 1 1
SWITCH
?
-4.727 -0.260 -3.033 0.003
CS x SWITCH - -0.657 -0.196 -1.697 0.092
SIZE + -0.234 -0.376 -4.495 0 . 0 0 0
PASTPERM + 0.166 0.152 1.768 0.079
MB + 0.033 0.126 1.469 0.137
ADV + -0.286 -0.313 -3.653 0 . 0 0 0
N 504
Adjusted R2 0.283
Variable definitions:
t = the quarter that the ACSI score was released for a certain industry
PERM = SALES, PROFIT, or ROA
SALES t+ 4 = Natural logarithm of sales revenue of firm i in quarter t+4
PROFIT t+ 4 = Natural logarithm of income before extraordinary items of firm i in quarter t+4
ROA 1 1+4 = Natural logarithm of return on assets for firm i in quarter t+4
CS , = Mean-centered ACSI score for firm i in quarter t (ACSI score - 72.91)
SWITCH; = Mean-centered Herfindahl-Hirschman Index (HHI) for the industry that firm i belongs to
(calculated as the mean of the sum of the squared market shares in percentage of all firms in a 2 -digit
SIC industry, computed over the five years prior to the event year) (HHI - 0.218)
S IZ E t+ 4 = Natural logarithm of the total assets of firm i in quarter t+4
PAST PERF t = performance (REV, PROFIT, ROA, and ROE) of firm i in quarter t
MB +1_4 = Market-to-book ratio of firm i in quarter t+4
A D V t+ 4 = Advertising-to-sales ratio of firm i in quarter t+4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
Table 5. The Effects of Industry-Relative Customer Satisfaction, Switching
Costs, and an Interaction Term between Industry-Relative Customer
Satisfaction and Switching Costs on Future Financial Performance
PERF i,t+ 4 = 3 + Bi CS i.t + B 2 SWITCH (+ fi3 CS i t x SWITCH f + B 4 SIZE it+ 4 +
B 5 PASTPERF i t + B g MB j t+ 4 + B 7 ADV i t+ 4 + q j t+ 4
Panel A: One-year ahead sales revenue as the dependent variable
Independent Variables Expected
Sign
Unstandardized
Coefficients
Standardized
Coefficients
t Two-tail
p-values
Constant -0.261 -0.671 0.504
CS + 0.058 0.915 2.138 0.034
SWITCH
?
0.495 0.801 0.618 0.537
CS x SWITCH - -0.253 -0.917 -2.141 0.034
SIZE + 0.460 0.481 10.639 0 . 0 0 0
PASTPERM + 0.436 0.420 8.892 0 . 0 0 0
MB + 0 . 0 1 2 0.029 0.902 0.368
ADV + -0.300 -0 . 2 1 1 -6.370 0 . 0 0 0
N 504
Adjusted R2 0.877
Panel B: One-year ahead income before extraordinary items as the dependent variable
Independent Variables Expected
Sign
Unstandardized
Coefficients
Standardized
Coefficients
t Two-tail
p-values
Constant 0.124 0.207 0.836
CS + 0.082 1.218 1.905 0.059
SWITCH
?
-5.454 -0.199 -4.364 0 . 0 0 0
CS x SWITCH - -0.351 -1.203 -1.881 0.062
SIZE + 0.479 0.466 7.177 0 . 0 0 0
PASTPERM + 0.390 0.376 5.833 0 . 0 0 0
MB + 0.035 0.081 1.676 0.096
ADV + -0.188 -0.124 -2.701 0.008
N 504
Adjusted R2 0.819
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Table 5. (Continued)
Panel C: One-year ahead return on assets (ROA) as the dependent variable
Independent Variables Expected Unstandardized Standardized t Two-tail p-
Sign Coefficients Coefficients values
Constant -0.884 -1.243 0.216
CS + 0.106 2.319 2.822 0.005
SWITCH
?
-3.904 -0 . 2 1 1 -2.777 0.006
CS x SWITCH - -0.295 -1.610 -2.616 0 . 0 1 0
SIZE + -0.246 -0.354 -4.456 0 . 0 0 0
PASTPERM + 1.130 0.850 3.155 0 . 0 0 2
MB + 0.032 0.109 1.394 0.165
ADV + -0.290 -0.283 -3.945 0 . 0 0 0
N 504
Adjusted R2 0.300
Variable definitions are the same as in Table 4 except for the following:
CS i t = Mean centered industry-relative ACSI score for firm i in quarter t (ACSI score for firm i -
ACSI score in the “All others” category the industry that firm i belongs to in quarter t - 3.53)
Hypothesis lb predicts that customer satisfaction is more closely associated
with market value of equity in low switching cost industries than in high switching
cost industries after controlling for contemporaneous book values of assets and
liabilities. Following Ittner and Larcker (1998), I test the following model:
MVE i,t = ft + 1 3 ! CS i> t + 1 3 2 SWITCH; + 1 3 3 CS M x SWITCH j + 1 3 4 ASSET i t + 65
LIAB i t + Q i,t
Again, main-effect terms and product terms were both centered to alleviate
multicollinearity and allow easy interpretation (Aiken & West, 1991). Table 6
presents the results. Panel A summarizes the results when the raw ACSI scores are
used for each firm, and Panel B summarizes the results when the industry-relative
ACSI scores are used. As expected, coefficients on assets are significantly positive
and coefficients on liabilities are significantly negative. Coefficients on customer
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
satisfaction are significantly positive as predicted in both Panel A (13=0.026, t=4.518,
p=0.000) and Panel B (13=0.109, t=2.368, p=0.019). Coefficients on switching costs
as measured by HHI are significantly positive in both Panel A (13= 2.323, t=6.034,
p=0.000) and Panel B (13= 2.291, t=1.927, p=0.056). The coefficient on the
interaction term between customer satisfaction and switching costs is significantly
negative when industry-relative customer satisfaction scores are used (13= -0.458, t= -
2.295, p=0.023) but insignificant when raw ACSI scores are used (13= -0.023, t= -
0.259, p=0.796). Thus Hypothesis lb is partially supported. In other words, for firms
in service industries, customer satisfaction has a significantly greater association
with the market value of equity in low switching costs industries than in high
switching costs industries after controlling for contemporaneous book value of assets
and liabilities, but only when industry-relative customer satisfaction measures are
used.
As recommended by Aiken and West (1991), interactions were plotted by
deriving separate equations for the high and low switching costs conditions. The
regression equation is restructured to express the regression of financial performance
on customer satisfaction at different levels of switching costs:
PERF; t+ 4 = (q + 1 3 2 SWITCH 0 + (Bi + 1 3 3 x SWITCH i t) * CS j,t (Figures. 3-
5)
MVE i t = (q +132 SWITCH 0 + (1 3 i + 63 x SWITCH i t) * CSi t (Figure. 6)
The following values of switching costs are used for plotting the interactions:
Average switching costs (mean-centered) = 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Low switching costs (mean-centered) = -0.218 (one standard deviation below the
mean)
High switching costs (mean-centered) = 0.218 (one standard deviation above the
mean)
The horizontal axis represents ACSI scores, ranging from one standard deviation
below the mean-centered ACSI score (-5.58) to one standard deviation above the
mean-centered ACSI score (5.58). The vertical axis represents financial
performance. As shown in Figures 3-5, the link between customer satisfaction and
one-year ahead accounting performance is stronger for firms in service industries
with low switching costs. Similarly, Figure 6 shows that the link between customer
satisfaction and contemporaneous market value of equity is stronger for firms
in service industries with low switching costs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
Table 6. The Effects of Customer Satisfaction, Switching Costs, and an
Interaction Term between Customer Satisfaction and Switching Costs on the
Market Value of Equity
MVE i,t = 3 . + 8, CS i.t + B 2 SWITCHf + B 3 CS ijt x SWITCH, + 84 ASSET ,-t + 85 LIAB it + q f,
Panel A: Raw customer satisfaction score as the independent variable
Independent Variables Expected Unstandardized Standardized t Two-tail
Sign Coefficients Coefficients p-values
Constant - 1 . 0 1 2 -4.042 0.000
CS + 0.026 0.058 4.518 0.000
SWITCH
?
2.323 0.077 6.034 0.000
CS x SWITCH - -0.023 -0.003 -0.259 0.796
ASSET + 1.703 1.026 72.327 0.000
LIAB - -0.274 -0.089 -6.596 0.000
N 504
Adjusted R2 0.937
Panel B: Industry-relative customer satisfaction score as the independent variable
Independent Variables Expected Unstandardized Standardized t Two-tail
Sign Coefficients Coefficients p-values
Constant -1.912 -3.550 0 . 0 0 1
CS + 0.109 0.828 2.368 0.019
SWITCH
?
2.291 0.043 1.927 0.056
CS x SWITCH - -0.458 -0.801 -2.295 0.023
ASSET + 1.943 0.990 49.279 0.000
LIAB - -0.435 -0.093 -4.722 0.000
N 504
Adjusted R2 0.951
t = the quarter that the ACSI score was released for a certain industry
For Panel A, CS j t = Mean-centered ACSI score for firm i in quarter t (ACSI score - 72.91)
For Panel B, CS t = Mean centered industry-relative ACSI score for firm i in quarter t (ACSI score
for firm i - ACSI score in the “All others” category the industry that firm i belongs to in quarter t -
3.53)
SWITCH; = Mean-centered Herfindahl-Hirschman Index (HHI) for the industry that firm i belongs to
(calculated as the mean of the sum of the squared market shares in percentage of all firms in a 2 -digit
SIC industry, computed over the five years prior to the event year) (HHI - 0.218)
ASSET, = Natural logarithm of the total assets of firm i in quarter t
LIAB j = Natural logarithm of the total liabilities of firm i in quarter t
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Figure 3. The Interaction Effect of Customer Satisfaction and Switching Costs
on One-Year Ahead Sales Revenue
CS-Sales Relations
A High Switching Costs
^ Low Switching Costs
□ Avg Switching Costs
0.0
■ 6 6 -4 2 0 2 4
Note:
Average switching costs (mean-centered) = 0
Low High switching costs (mean-centered) = -0.218 (one standard deviation above the mean)
Low switching costs (mean-centered) = 0.218 (one standard deviation below the mean)
The horizontal axis represents ACSI scores, ranging from one standard deviation below the mean-
centered ACSI score (-5.58) to one standard deviation above the mean-centered ACSI score (5.58).
The vertical axis represents the natural logarithm of one-year ahead sales revenue.
The graph is based on the following models:
Average switching costs: Sales = 0.164 + 0.036 * Customer satisfaction
Low switching costs: Sales = 0.166 + 0.043 * Customer satisfaction
High switching costs: Sales = 0.162 + 0.029 * Customer satisfaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Figure 4. The Interaction Effect of Customer Satisfaction and Switching Costs
on One-Year Ahead Income before Extraordinary Items
o .o
-.2
-.4
-.6
-.8
- 1.0
CS-Income Relations - 1.2
A High Switching Costs
Low Switching CO sts
□ Avg Switching Costs
- 1.4
- 1.6
•6 ■ 4 2 0 2 4 6
Note:
Average switching costs (mean-centered) = 0
Low switching costs (mean-centered) = -0.218 (one standard deviation above the mean)
High switching costs (mean-centered) = 0.218 (one standard deviation below the mean)
The horizontal axis represents ACSI scores, ranging from one standard deviation below the mean-
centered ACSI score (-5.58) to one standard deviation above the mean-centered ACSI score (5.58).
The vertical axis represents the natural logarithm of one-year ahead income before extraordinary
items.
The graph is based on the following models:
Average switching costs:
Income before extraordinary items = -0.083 + 0.107 * Customer satisfaction
Low switching costs:
Income before extraordinary items = -0.813 + 0.131 * Customer satisfaction
High switching costs:
Income before extraordinary items = -0.855 + 0.083 * Customer satisfaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5. The Interaction Effect of Industry-Relative Customer Satisfaction and
Switching Costs on the Rate of Return on Assets (ROA)
10-
CS-ROA Relations -10-
A High Switching Costs
* Low Switching Costs
□ Avg Switching Costs -20
•6 2 0 2 6 -4 4
Note:
Average switching costs (mean-centered) = 0
Low switching costs (mean-centered) = -0.218 (one standard deviation above the mean)
High switching costs (mean-centered) = 0.218 (one standard deviation below the mean)
The horizontal axis represents ACSI scores, ranging from one standard deviation below the mean-
centered ACSI score (-5.58) to one standard deviation above the mean-centered ACSI score (5.58).
The vertical axis represents the natural logarithm of ROA.
The graph is based on the following models:
Average switching costs: ROA = -0.884 + 0.106 * Customer satisfaction
Low switching costs: ROA = -0.838 + 0.170 * Customer satisfaction
High switching costs: ROA = -0.930 + 0.042 * Customer satisfaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6. The Interaction Effect of Industry-Relative Customer Satisfaction and
Switching Costs on the Market Value of Equity (MVE)
- 2.0
- 2.5
- 3.0
- 3.5
- 4.0
- 4.5
CS-MVE Relations
A High Switching Costs
Low Switching Costs
□ Avg Switching Costs
- 5.0
- 5.5
■ 6 ■ 2 6 ■ 4 0 2 4
Note:
Average switching costs (mean-centered) = 0
Low switching costs (mean-centered) = -0.218 (one standard deviation above the mean)
High switching costs (mean-centered) = 0.218 (one standard deviation below the mean)
The horizontal axis represents ACSI scores, ranging from one standard deviation below the mean-
centered ACSI score (-5.58) to one standard deviation above the mean-centered ACSI score (5.58).
The vertical axis represents the natural logarithm of market value of equity.
The graph is based on the following models:
Average switching costs: MVE = -3.714 + 0.251 * Customer satisfaction
Low switching costs: MVE = -3.747 + 0.283 * Customer satisfaction
High switching costs: MVE = -3.681 + 0.219 * Customer satisfaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
3.4. Discussion
This chapter seeks to understand the effect of switching costs on the relation
between customer satisfaction and future financial performance in service industries.
I propose and test the hypothesis that the relationship between customer satisfaction
and future financial performance decreases as industry-level switching costs
increase. Using the intersection of ACSI and COMPUSTAT databases from 1994-
2 0 0 2 ,1 find that, for firms in service industries, the relationship between customer
satisfaction as measured by raw ACSI scores and industry-relative ACSI scores and
future financial performance as measured by one-year-ahead sales revenue, income
before extraordinary items, and ROA is stronger for firms in industries with low
switching costs (as measured by HHI). I also find that the relation between industry-
relative ACSI scores and the market value of equity is stronger for firms in industries
with low switching costs.
The results of this chapter should be interpreted with several limitations in
mind. First, as the sample firms are limited to the largest, surviving firms, the
generalizability of the findings is limited. Future studies can use field or survey
method to complement the archival evidence presented in this chapter.
Second, the Herfindahl-Hirschman Index (HHI) is only one proxy for
switching costs and may not capture the whole construct. Future research is needed
to identify more precise measures of switching costs. In addition, this study takes a
broad perspective and examines industry-level switching costs only. It is conceivable
that even within one industry, different firms may build different levels of switching
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
costs. For example, Dikolli et al. (2003) find variations in switching costs across
firms within the retail industry. This caveat aside, it is still meaningful to examine
the effect of industry-level switching costs due to its theoretical importance and
practical implications. For instance, if industry-level switching costs are found to
moderate the relation between customer satisfaction and future financial
performance, users of customer satisfaction information can adjust for industry
effects in predicting future financial performance of firms and in designing optimal
performance measurement systems.
Third, customer satisfaction is treated as an exogenous variable in this study.
However, prior researchers find that customer satisfaction can be a function of the
level of competition and differentiation in an industry (Fomell and Johnson, 1993;
Fomell et al., 1996). Anderson (1994) also finds that customer satisfaction tends to
be higher when competition, differentiation, involvement, or experience is high or
when switching costs, difficulty of standardization, or ease of evaluating quality is
low. In other words, companies may have already taken into account of the levels of
switching costs when choosing their “optimal” levels of customer satisfaction.
Though it is unlikely that all sample firms are optimizing at the same time, this
endogeneity issue will result in conservative tests of the interactive effects between
customer satisfaction and switching costs.
Finally, this chapter investigates only one moderator of the relation between
customer satisfaction and future financial performance. However, other variables
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
may also moderate the main effect of customer satisfaction on financial performance.
The next chapter explores other moderators of the relationship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
Chapter 4. Evidence from the Field: The Case o f a Health Insurance
Company
4.1. Research Site
This study’s research site, HBP, is a leading specialty health benefit plan
provider.1 5 It was established as a non-profit organization to provide high-quality,
cost-effective specialty health care benefits. As companies and government agencies
recognized the need to include this type of specialty health care in their employees’
benefits packages, HBP experienced rapid growth and strong partnerships with
doctors. HBP now provides a variety of flexible health insurance plans and operates
nationwide. It has approximately 30,000 clients, including about 250 Fortune 500
companies that purchase health insurance plans for their employees. HBP has a
network of over 25,000 doctors and over 3,000 employees. In 2004, the firm had
gross revenues of about $3.5 billion.
HBP offices in different markets are grouped into eight geographic divisions,
each with its own division manager. Managers are accountable for revenues in the
regions that they manage. Offices in different markets are homogeneous in many
aspects of their operations including organization structure, products and services
provided, business cycle, and performance measurement and incentive system, but
they vary in size, geographic location, market competition, and patient mix. To the
extent that large clients have employees in multiple markets, some interdependence
15 Under a confidentiality agreement, I have permission from the research site to use the company
information with disguise to ensure the anonymity of the company.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
exists between the different markets. HBP faces a number of national and regional
competitors for the services they provide and the market has become more
competitive and more vertically integrated since the 1990s. For example, a large
HMO purchased one of HBP’s competitors to expand into this specialty health care
area in order to supplement their medical plan offering. This is consistent with a
broad trend in the health care industry, which is characterized by increasing vertical
integration of health care delivery organizations with both upstream and downstream
activities (Abemethy et al., 2005).
As a non-profit organization, HBP’s primary objective has always been
growth. It seeks to serve as many customers as possible in a sustainable manner. In
an increasingly competitive and vertically integrated market, HBP believes two
factors are the most important drivers of growth. First, meeting or exceeding
satisfaction goals promotes growth both by generating more revenue from existing
customers and by maintaining HBP’s strong reputation in the marketplace. Second,
reducing administrative costs promotes growth by resulting in more competitive
pricing for prospective and renewing customers. To motivate employees to create
growth in the early 1990s, HBP implemented bonus plans that include customer
satisfaction as an important factor. Consistent with their dual strategy, HBP gives
satisfaction goals and cost goals equal weights in the bonus plans. The satisfaction
measures for each of the three customer groups served (doctors, patients, and clients)
also are given equal weights to convey the message that each customer group is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
equally important to HBP.1 6 Since HBP emphasizes the national scope of their
business and touts seamless administration across all regions, firm-level performance
is used in the bonus plan instead of using regional-level or state-level performance.
The weights on the different performance measures have remained the same since
the bonus plans came into effect, but the bonus targets have become more difficult to
achieve for both administrative costs and satisfaction measures. So far the
performance measurement system and incentive system seem to have successfully
supported HBP’s strategy. The firm has achieved high satisfaction scores among
clients, patients, and doctors, and has enjoyed spectacular revenue growth in the last
decade.
There are several reasons why this research site is appropriate for my study.
First, in contrast to a large amount of research on customer satisfaction involving a
single customer group, there is a lack of research in settings with multiple customer
groups. This research site offers a unique opportunity to fill the gap and extend
existing literature in fruitful ways. Second, since the firm has been pursuing a
strategy of providing high customer satisfaction, customer satisfaction measurement
and revenue implications are of particular importance to the firm. Third, similar
products, common technologies, and a common satisfaction measurement system
and incentive system minimize the heterogeneity problems prior research
encountered from pooling across multiple firms and industries (e.g., Ittner and
16 The bonus plan for the management team also includes an employee satisfaction measure.
However, employee satisfaction measures are not broken down by states, so they are excluded from
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
Larcker, 1998a). The cross-sectional variation in competitive environment, on the
other hand, provides the opportunity to isolate the effect of competition on the
satisfaction-revenue link.
4.2. Hypotheses
As discussed in Chapter 2, companies often serve multiple customer groups,
so the analysis of customer satisfaction should take this into account. In the health
care industry, where services are provided to one party (the patients), but purchasing
decisions are often made by another party (employers of the patients), organizations
unavoidably deal with multiple customer groups. In a comprehensive review of
accounting and control in a health care setting, Abemethy, Chua, Grafton, and
Mahama (2005) point out that a distinctive feature of the health care setting is
multiple and conflicting goals of different customer groups. In particular, they
mention the conflicts between administrators of health care organizations whose
priorities are dominated by efficiency concerns and medical professionals whose
goals are dominated by quality of care. However, empirical research on management
control systems in the health care sector is relatively sparse. In addition, among the
limited empirical evidence on the health care industry, most studies focus on
hospitals to the exclusion of other institutions such as pharmaceutical,
biotechnology, and insurance companies, which interact with hospitals and influence
the efficiency and quality of care of hospitals (Abemethy et al., 2005).
the analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
With a business model that involves multiple customer groups, the current
research site allows me to examine a broader set of satisfaction measures and their
interactions to overcome the limitations of prior literature. Like Rucci et al. (1998),
Campbell et al. (2004), and Nagar and Raj an (2005), I take a business model
approach to performance measurement and first examine the validity of a
hypothesized model unique to a firm’s strategy. I then examine the cross-sectional
differences in the business model depending on the different operating environments
in local markets. Unlike the above studies I focus my analysis on one type of
nonfinancial performance measure (customer satisfaction) and show the complexity
of customer satisfaction measures in a setting with multiple customer groups.
In the current research setting, HBP deals with three types of customers:
companies that purchase their plans (“clients”), members that use their plans
(“patients”), and doctors who provide services to the patients. As a non-profit
organization, HBP’s primary objective is revenue growth rather than profit
maximization. Therefore, I measure financial performance in this context with future
revenues.1 7 HBP builds its business model and performance measurement system on
the assumption that higher client, patient, and doctor satisfaction lead to greater
revenues in the future. My first set of hypotheses is therefore based on HBP’s
hypothesized business model (Figure 7). Specifically,
17 Previous studies document mixed evidence between customer satisfaction and profit and conjecture
that it could be driven by cost differences in achieving CS (Anderson et al., 1997; Ittner and Larcker
1998). However, my study suggests an alternative explanation: the differences could also be driven by
differences in revenue consequences.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Hla: Client satisfaction measures are positively associated with future
revenues.
Hlb: Patient satisfaction measures are positively associated with future
revenues.
HI c: Doctor satisfaction measures are positively associated with future
revenues.
Figure 7. HBP’s Hypothesized Business Model
Client Satisfaction
Hla: +
Hlb: +
Patient Satisfaction
Future Revenues
Hlc: +
Doctor Satisfaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Compared with a large stream of literature in financial accounting research
that investigates the value-relevance of financial metrics such as earnings,
management accounting research that examines the relevance of nonfinancial metrics
has been limited and sparse. Most prior studies focus on settings where constructing
a nonfinancial metric is relatively straightforward. For example, the customer
satisfaction literature has concentrated on settings where there is only one set of
customer responses. However, as mentioned above, constructing relevant customer
satisfaction measures may present a challenge because the term “customer” is
broadly defined. Prior research provides little guidance as to which satisfaction
measure is more closely associated with future financial performance in these
settings.
I propose that the performance consequences of various customer satisfaction
measures in a setting with multiple customer groups depend on the bargaining power
of each customer group. Although the purchasing decision is made jointly in any
two-tiered customer structure that characterizes the current research setting, there is
variation in the influence of the client vs. the end-consumer. Since clients tend to
purchase in large volume and have greater influence on the purchasing decision than
patients, I predict that client satisfaction, in general, is more relevant than patient
satisfaction. However, a cleaner test would be to examine the effect of the variation
in the bargaining power of clients vs. patients on the relevance of satisfaction
measures. The research site provides a unique measure to proxy for such variation.
HBP serves two different types of patients: voluntary and non-voluntary patients.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
While non-voluntary patients do not have much purchasing influence because their
employers make the purchasing decision and payment for them, voluntary patients
have much greater purchasing influence because they pay a portion or all of the
insurance premiums and have the option to choose among different products offered
in their benefits packages. Since patients have greater purchasing influence in
markets with a larger proportion of voluntary patients, patient satisfaction should be
more important in those markets than in the markets with a smaller percentage of
voluntary patients. Conversely, client satisfaction should be more important in the
markets with a smaller percentage of voluntary patients than in those markets with a
larger percentage of voluntary patients. Thus, I posit the following hypotheses:
H2a: Client satisfaction is more strongly associated with future revenues
than patient satisfaction.
H2b: The proportion o f voluntary patients increases the strength o f
relationship between patient satisfaction and future revenues.
H2c: The proportion o f voluntary patients decreases the strength o f
relationship between client satisfaction and future revenues.
In addition, I use HBP’s market penetration rate as a proxy of HBP’s market
power in each market and a reverse measure of customer bargaining power.1 8 In the
words of the VP of Marketing,
18 Bargaining power is the relative size of customers vs. providers. Due to data limitation, I couldn’t
measure the relative market power of customers, so I measured the relative market power of the
provider.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
In the markets where we don’ t have a big market share, our brand is not
well-known, the market is not mature, and employee benefits are not a big
concern fo r local employers. Since we haven’ t established relationships with
clients, clients will shop mainly on price.
Where market penetration is low, the bargaining power of clients is likely to increase
for two reasons. First, clients are price sensitive, increasing the potential sources of
supply and available substitutes. Since HBP competes on quality of service rather
than price, it is more difficult for HBP to retain price-sensitive clients if they are not
satisfied. Second, lack of brand loyalty and established relationships with clients
lower switching costs for clients. Similarly, the bargaining power of doctors is likely
to increase in low penetration markets because doctors tend to have a smaller
percentage of patients with HBP coverage in these markets, which decreases the
switching costs for them. In addition, lack of brand loyalty and established
relationships with HBP further lower switching costs for the doctors. Finally, low
market penetration means lower brand loyalty, so patients also are more price
sensitive and more likely to defect when dissatisfied, which increases the relevance
of patient satisfaction. Conversely, in those markets where HBP achieves a high
market penetration rate, it has greater market power, more established brand loyalty,
and more mature relationships with existing clients, patients, and doctors, which
significantly lower customer bargaining power.
Thus, I posit the following hypotheses:
H3a: HBP’ s market penetration rate decreases the strength o f the relation
between client satisfaction and future revenues.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
H3b: HBP’ s market penetration rate decreases the strength o f the relation
between doctor satisfaction and future revenues.
H3c: H BP’ s market penetration rate decreases the strength o f the relation
between patient satisfaction and future revenues.
4.3. Sample, Variables, and Research Design
4.3.1. Sample
The sample consists primarily of quarterly data obtained from HBP on
financial and satisfaction measures for 51 markets over 20 quarters, from the first
quarter of 2000 through the fourth quarter of 2004.1 supplement these data with
market penetration and patient mix information obtained from the HBP Marketing
department. In addition, I collect quarterly data from the Bureau of Economics
Analysis on economic conditions in the states during the same period. To gain
familiarity with the business environment of HBP, I spent two days at the firm’s
national headquarters discussing the project with the CFO and managers in the
Customer Service Department. I also had numerous phone and email discussions
with the CFO, VP Marketing, and Customer Service managers and employees during
the data collection period. Finally, I reviewed internal company documents about the
performance measurement and incentive system.
4.3.2. Variables
Client Satisfaction
The firm measures client satisfaction for a stratified random sample of clients
purchasing a health plan each quarter. The survey is administered by email for all
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
clients with an email address and via mail for the remainder of clients. The average
initial sample size is 2,500 each quarter and the average final sample size is 650,
resulting in an average response rate of 26%. The firm measures client satisfaction
based on a questionnaire. To maintain comparability across time, I retain 18
questions that are common throughout the 20 quarters in the sample period. Panel A
of Table 7 presents results of a principal component factor analysis using oblique
rotation, allowing the factors to be correlated.1 9 The factor loadings suggest two
dimensions of client satisfaction. I label the first factor “Client satisfaction with
customer service.” The second factor captures “Client satisfaction with value.”
Given that these two factors emerge consistently as separate dimensions across
divisions and over time, I construct weighted indexes for these two client satisfaction
dimensions using factor loadings and treat them as observed variables in the
structural equation modeling (SEM) analysis.
Patient Satisfaction
The firm measures patient satisfaction for a random sample of patients who
use their health care benefits within each quarter. This survey is administered via
mail. The initial sample size averages 3,500 each quarter and the final sample size
averages 630, yielding an average response rate of 18%. The firm measures customer
satisfaction based on a questionnaire. I retain 20 questions that are common
throughout the 20 quarters in the sample period for comparability purposes. Panel B
19 The Varimax rotation method, which assumes orthogonality between the factors, yields similar
results.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
of Table 7 presents the results of principal component factor analysis using oblique
rotation. Three dimensions of patient satisfaction emerge from the analysis. The first
factor is “Patient satisfaction with coverage”; the second factor is “Patient
satisfaction with quality of service”; and the third factor is “Patient satisfaction with
convenience.” I construct weighted indexes for the three dimensions of patient
satisfaction based on factor loadings and treat them as measures of patient
satisfaction in the subsequent SEM analysis.
Doctor Satisfaction
The firm measures doctor satisfaction for a random sample of doctors who
contract with HBP each quarter. The survey is administered via mail. The average
initial sample size is 2,400 and the average final sample size is 936, resulting in an
average response rate of 39%. The firm measures doctor satisfaction using a
questionnaire with 15 questions. Panel C of Table 7 presents the results of principal
component factor analysis with oblique rotation, which yields two dimensions of
doctor satisfaction. The first dimension assesses doctors’ absolute satisfaction level
with HBP’s service and support, and the second dimension assesses their relative
satisfaction level with HBP’s services and support when comparing HBP with other
plans that they participate in. I label these two dimensions as “Absolute doctor
satisfaction” and “Relative doctor satisfaction” respectively, construct factor-score-
weighted indexes for the two dimensions, and treat them as indicators of doctor
satisfaction in the SEM analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7. Exploratory Factor Analysis of Satisfaction Measures
Panel A. Client satisfaction
Survey Questionnaire Items Factor 1 Factor 2
The ease of doing business with HBP .571 .326
Your HBP account executive’s responsiveness to your questions .945 -.094
Thoroughness with which your issues are resolved by your HBP account executive .937 -.068
Your HBP account executive’s knowledge of HBP products and services .875 -.043
Your HBP account executive’s ability to advise you on your organization’s benefits strategy .815 -.035
Your HBP account executive’s frequency of contact to meet your needs .814 -.005
Your other HBP service contacts’ accessibility to answer your questions .914 -.036
Your other HBP service contacts’ responsiveness to your questions .937 -.052
Thoroughness with which your issues are resolved by your other service contacts .918 -.029
Courtesy shown to you by your other service contacts .878 -.043
Effectiveness of HBP’s communications with you .763 .113
Effectiveness of HBP’s communications with your employees/members .634 .225
The accuracy of your monthly bill .526 .238
HBP’s ability to meet your reporting needs .557 .274
HBP’s ability to provide a plan that meets the organization’s need .347 .561
The value received for the dollars spent on your HBP plan .167 .710
Selection of HBP doctors from which your employees can choose -.089 .909
Your employees/members satisfaction with HBP doctors -.003 .884
% of Variance Explained 64.731 6.439
Cronbach’s Alpha 0.872 0.854
Labels for Factors Customer service Value
Note: Panels A through C present results from principal component factor analysis with oblimin rotation.
o \
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7. (Continued)
Panel B. Patient satisfaction
Survey questionnaire items Factor 1 Factor 2 Factor 3
Selection of materials in your HBP doctor’s office that were fully covered by your plan .533 .218 .049
Your understanding of the plan .481 -.039 .372
Amount of coverage provided by your plan .826 -.072 .118
Amount of coverage in exam .887 .0003 -.043
Amount of coverage in medical materials .901 .005 -.047
Amount of coverage in specialty materials .890 -.008 -.099
Amount of coverage in optional materials .884 -.046 -.087
Value received for the dollars you spent on your care .740 .130 .060
Ease of using your plan .419 . 1 0 2 .390
Rate your HBP doctor in thoroughness of your exam -.004 .911 .029
Rate your HBP doctor in explanation of diagnosis and treatment -.032 .930 .007
Rate your HBP doctor in personal interest and courtesy -.046 . 8 8 6 .046
Length of time you waited to receive your service .266 .362 .197
Quality of service .395 .399 . 1 1 0
Rate your HBP doctor or staffs explanation of your overall coverage .354 .394 .141
Overall experience with your HBP doctor .079 .848 -.026
Selection of HBP doctors .237 - . 1 0 2 .693
Rate your HBP doctor in convenience of location -.109 .025 .893
Rate your HBP doctor in convenience of office hours -.051 .159 .780
Rate your HBP doctor in length of time between setting an appointment and your visit -.024 .249 .653
% of variance explained 52.260 10.296 4.266
Cronbach’s Alpha 0.933 0.910 0.835
Labels for Factors Coverage Quality of service Convenience
O s
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7. (Continued)
Panel C. Doctor satisfaction
Survey questionnaire items Factor 1 Factor 2
Ease of doing business with HBP .681 .205
Ease of administering HBP’s plans .658 .134
Ease of explaining HBP’s plans to patients .614 .117
Promptness of payment for claims submitted .849 -.043
Accuracy of payment for claims submitted .877 -.052
Ease of using your remittance advice .769 -.023
Resolution of payment issues . 8 8 8 -.049
Ease of obtaining eligibility/plan information .778 -.046
Usefulness of HBP doctor communications .711 .058
Responsiveness of HBP employee in resolving your issues .764 .037
Overall satisfaction with HBP compared to other plans you participate in .115 .752
Rate how HBP compares to other plans you participate in fairness of reimbursement fees for services provided .008 .791
Rate how HBP compares to other plans you participate in providing information/training on products and . 0 2 2 .767
Rate how HBP compares to other plans you participate in supporting private practice doctors in remaining -.026 .837
Rate how HBP compares to other plans you participate in patients’ understanding of what their plan covers -.028 .640
% of Variance Explained 52.664 9.843
Cronbach’s Alpha 0.817 0.876
Labels for Factors Absolute doctor Relative doctor
Satisfaction satisfaction
Os
Os
67
Financial Performance. I operationalize financial performance as the natural
logarithm of quarterly sales revenue for each market.
Market Penetration Rate. I measure market penetration rate for each market by the
percentage of the population that are HBP patients.
Percentage o f Voluntary Patients. I measure percentage of voluntary patients for
each market by dividing the number of voluntary patients by the total number of
HBP patients. I obtain this data for 16 quarters from 2001 through 2004.
Economic Conditions. General economic conditions have a direct impact on
financial performance of firms in the health care industry. Good economic conditions
enable employers to spend more money on health care benefits for their employees
and allow voluntary patients to purchase optional health care insurance. Therefore, I
control for economic conditions, as measured by quarterly state personal income.
This measure tracks the level of personal income received by people who live or
work in a state and is often used as a measure of state-level economic conditions. I
collect this data from the website of the Bureau of Economic Analysis
(www.bea.gov).
Past Performance. I include past performance to control for time-series trends as
well as to examine whether the satisfaction measures provide incremental
information on future performance (Ittner and Larcker, 1998a). I measure past
performance with a one-year lag to control for seasonality in the insurance business.
Specifically, I observe a big increase in revenues in the fourth quarter of each year
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 8
because clients tend to accelerate costs for tax purposes through prepayment of
medical benefits insurance.
4.3.3. Research Design
I analyze the data using structural equation modeling (SEM). Researchers in
management accounting have called for greater use of SEM in management
accounting research (Shields, 1997; Shields and Shields, 1998; Smith and Langfield-
Smith, 2004). SEM overcomes two limitations of OLS regressions: (1) the
assumption that all constructs are free of measurement error; and (2) the inability to
estimate multiple and interrelated dependent relations among variables. This study
uses SEM to test the validity of the HBP model in Figure 7. SEM allows me to
explicitly model the measurement error in latent variables (Bollen, 1989). This is
necessary in the current research setting because customer satisfaction, like many
nonfinancial performance measures, is a “soft” measure that potentially contains
considerable measurement error (Ittner and Larcker, 2003). SEM also permits
simultaneous estimation of the relations between satisfaction levels of different
customer groups and financial performance. Using SEM, I first test the business
model underlying HBP’s performance measurement system. I then examine whether
the model varies across different competitive environments.
Prior research suggests that the length of lag between customer satisfaction
and future financial performance varies across firms. For example, Banker et al.
(2000) document a six-month lead-lag relation between customer satisfaction and
financial performance in the hospitality industry, whereas Bernhardt et al. (2000)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
detect a one-year lag in the fast-food restaurant industry. In the current research
setting, the contracts between HBP and its clients, doctors, and patients provide some
information about the lags. I use contract length because satisfaction within the
contract period will increase the number of repeat customers and thus future
revenues.2 0 Most clients have two-year contracts with HBP and most patients have
one-year contracts with HBP. All doctors have evergreen contracts with HBP, which
allow either party to terminate the contract with ninety days of notice. Consistent
with this, I use 8-quarter lags for client satisfaction, 4-quarter lags for patient
satisfaction, and 1-quarter lags for doctor satisfaction.
Another design choice is the functional form. There has been considerable
debate on the appropriate functional form of the relation between customer
satisfaction and financial performance (Ittner and Larcker, 1998a; Lambert, 1998).
Banker et al. (2000) argues that, because of the decaying effect of nonfinancial
measures over time, the permanent effect assumption of the changes model are
violated. The levels model is therefore more appropriate because residuals from the
levels model will have a lower autocorrelation than the changes model. Following
Banker et al. (2000), I use a levels model with a lagged dependent variable.
2 0 However, I realize that this is still arbitrary, so I use alternative lags in the robustness checks.
Results are qualitatively similar.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
4.4. Results
4.4.1. Descriptive Statistics
Panel A of Table 8 provides descriptive statistics on the satisfaction
variables. The satisfaction measures are elicited with 5-point scales, where 1
indicates “Poor” and 5 indicates “Excellent”. Panel B of Table 8 provides
descriptive statistics on the other variables. The descriptive statistics indicate that
many variables are highly skewed. In order to obtain more normal distributions, I
apply a logarithmic transformation to each independent and dependent variable.
Table 9 presents Pearson correlations (after applying the logarithmic
transformation) among the variables. These results indicate that the correlations
generally are consistent with the hypotheses. The correlations between client
satisfaction measures and revenues are positive, and so are the correlations between
the doctor satisfaction measures and revenues. The only exception is patient
satisfaction, where the correlations between two measures of patient satisfaction
(satisfaction with quality of service and satisfaction with convenience) and revenues
are insignificant and the correlation between a third measure of patient satisfaction
(satisfaction with coverage) and revenues is significantly negative. I also find patient
satisfaction with coverage to be negatively correlated with client satisfaction with
value (r = -0.088, p < 0.05). These initial results suggest potential trade-offs between
patient satisfaction and client satisfaction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Table 8. Descriptive Statistics
Panel A: Customer Satisfaction Measures *
Sample
Size
Mean Standard Upper
Deviation Quartile
Median Lower
Quartile
Client Satisfaction with Customer Service 1 0 2 0 4.01 0.39 4.25 4.04 3.79
Client Satisfaction with Value 1 0 2 0 3.81 0.45 4.07 3.85 3.59
Patient Satisfaction with Coverage 1 0 2 0 3.53 0.43 3.76 3.52 3.29
Patient Satisfaction with Quality of Service 1 0 2 0 4.26 0.33 4.45 4.29 4.09
Patient Satisfaction with Convenience 1 0 2 0 4.06 0.37 4.27 4.08 3.87
Doctor Satisfaction - Absolute 1 0 2 0 3.49 0.33 3.7 3.51 3.28
Doctor Satisfaction- Relative 1 0 2 0 4.19 0.25 4.37 4.23 4.04
* This table presents descriptive statistics for satisfaction measures created by aggregating
questionnaire item scores weighted by factor loadings from factor analysis with oblimin rotation
(Table 1). Each observation represents a market-quarter. The satisfaction measures are elicited with 5-
point scales, where 1 indicates “Poor” and 5 indicates “Excellent”.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
Table 8. (Continued)
Panel B: Dependent Variables, Moderating Variables, and Control Variables *
Sample
Size
Mean Standard
Deviation
Upper
Quartile
Median Lower
Quartile
Gross Revenue
(millions)
1 0 2 0 $20.13 $54.07 $19.36 $7.88 $2.07
Net Revenue
(millions)
1 0 2 0 $3.16 $8.89 $2.79 $1.13 $0.34
Number of Clients 1 0 2 0 580 1,647 441 161 8 6
Number of Patients 1 0 2 0 608,350 1,357,399 645,926 295,776 69,082
State Personal
Income (millions)
1 0 2 0 $175,886 $207,558 $224,143 $107,164 $42,274
Market Penetration
Rate
1 0 2 0 11.46% 6 .8 6 % 15.21% 9.67% 6.38%
Percentage of
Voluntary Patients
816 44.76% 27.06% 68.82% 37.65% 22.81%
* Each observation represents a market-quarter.
Variable Definitions:
Net revenue = Gross revenue - claims payments to doctors
State personal income = Total income received by people who live or work in a state
Market penetration rate = Percentage of HBP patients out of the entire population in each market
Percentage of voluntary patients = Percentage of HBP patients who pay part or all of their HBP
insurance premiums
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 9. Correlation Matrix
1 2 3 4 5 6 7 8 9 1 0 1 1
1. Client Sat. with Customer Service 1 . 0 0 0
2. Client Sat. with Value .396** 1 . 0 0 0
3. Patient Sat. with Coverage .032 -.088* 1 . 0 0 0
4. Patient Sat. with Quality of Service .024 .035 .550** 1 . 0 0 0
5. Patient Sat. with Convenience .024 .096** .595** .648** 1 . 0 0 0
6 . Doctor Sat. - Absolute
2 2 4 * *
.183** .016 -.016 .066 1 . 0 0 0
7. Doctor Sat. - Relative . 0 2 0 .118** -.017 .005 .037 .664** 1 . 0 0 0
8 . Revenue .069*
2 4 2 * *
-.130** -.044 .018 .234** .333** 1 .0 0 0
9. State Personal Income .019 . 1 1 0 ** -.067 -.109** -.025 .247** .330*
7 4 4 **
1 .0 0 0
10. Market Penetration Rate -.019 . 1 2 0 ** -.062 .015
*
oo
p
.053 .079 .394**
1
©
0 0
1 .0 0 0
11. Percentage of Voluntary Patients .017 .071 -.015 - . 0 0 2 .006 .009 -.048 -.075* .2 1 1 **
_ 3 4 7 **
1 .0 0 0
Note: Natural logarithms are taken of each variable. See Tables 1 and 2 for variable definitions.
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
- j
74
4.4.2. SEM Analysis o f the HBP Model
I estimate the SEM models using AMOS 5.0. Before the parameter estimates
can be interpreted, it is important first to test the measurement model in order to
demonstrate that each proxy measure has a positive and statistically significant
relationship with its construct. In the measurement model, the relation between a
latent variable, £ ,i, and a survey item that comprises it, X j, is X j = A .jj ^ + 8jj, where A ,jj
is the loading, or degree of association between the latent variable and the manifest
variable, and is the measurement error associated with the survey item. To
identify the variance of the latent variables, I fix the loading of the item that I expect
a priori to best represent the construct at one. The error variance of each indicator
variable is the expected error (1-Cronbach’s alpha) of the measure multiplied by the
variance of the measure. Table 10 presents maximum likelihood coefficient
estimates of the measurement model. Maximum likelihood methods perform well
when data deviates from multivariate normality or are based on ordinal scales
(Boomsma and Hoogland, 2001; Distefano, 2001).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
Table 10. Measurement Model for Independent Latent Variables
This table reports the measurement model for the independent variables for a single representative
structural equation model (overall sample). The measurement model for these variables is stable
across different subsamples (i.e., subsamples with high vs. low percentage of voluntary patients and
subsamples with high vs. low market penetration rate).
Latent Variable
Indicators
Unstandardized
factor loading
(X)
Standard
error
z-statistic R Standardize
d factor
loadings
(te)
Client Satisfaction
Factor 1: Customer Service 1.000 0.541 0.735
Factor 2: Value 0.732 0.013 56.749 0.387 0.622
Patient Satisfaction
Factor 1: Coverage 1.000 0.541 0.735
Factor 2: Quality of Service 1.448 0.015 99.579 0.712 0.844
Factor 3: Convenience 1.744 0.054 32.053 0.601 0.758
Doctor Satisfaction
Factor 1: Absolute 1.000 0.541 0.735
Factor 2: Relative 0.953 0 . 0 1 2 79.344 0.516 0.719
Sample size 1 0 2 0
Degrees of freedom 2 1
Chi-square 97.86
RMSEA 0.063
CFI 0.923
Note: Chi-square/degrees of freedom (CMIN/DF) is the model chi-square fit index divided by degrees
of freedom in order to correct for sample size. Values of CMIN/DF less than 5 are considered
adequate fit. The comparative fit index (CFI) compares the existing model fit with a null model that
assumes the latent variables in the model are uncorrelated. By convention, CFI should be equal to or
greater than 0.90 to accept the model. The root mean square error of approximation (RMSEA)
corrects the Chi-square statistic for sample size as well as model complexity. By convention, there is
adequate model fit if RMSEA is less than or equal to 0.08 (Hu and Bentler, 1999; Fan et al., 1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
The standardized estimated loadings (Xs) of survey items on latent constructs
have the expected sign, are large enough to provide confidence that they measure
common latent constructs, and are statistically significant (p < 0.001, two-tailed).
Overall model fit is adequate, as indicated by three goodness-of-fit measures: the
Chi-square/degree of freedom (CMIN/DF) (= 4.66; < 5), the comparative fit index
(CFI) (= 0.923; > 0.90), and the root mean square error of approximation (RMSEA)
(= 0.063; < 0.08) (Hu and Bentler, 1999; Fan et al., 1999; Smith and Langfield-
Smith. 2004).2 1
Table 11 (Column 1) and Figure 8 present coefficient estimates of the
structural model for the relation between satisfaction levels of multiple customer
groups and future revenues. The model exhibits adequate model fit by the Chi-
square/degrees of freedom (CMIN/DF) (= 4.19; < 5), the comparative fit index (CFI)
(= 0.948; >0.90), and the root mean square error of approximation (RMSEA) (=
0.054; < 0.08). Overall, the evidence suggests that satisfaction levels of various
customer groups are related to future revenues after controlling for past financial
performance and general economic conditions.
The results indicate that, consistent with HI a, client satisfaction is positively
associated with future revenues (y = 0.546, z = 3.249). Also consistent with Hlc, the
21 Chi-square/degrees of freedom (CMIN/DF) is the model chi-square fit index divided by degrees of
freedom in order to correct for sample size. Values of CMIN/DF less than 5 are considered adequate
fit. The comparative fit index (CFI) compares the existing model fit with a null model that assumes
the latent variables in the model are uncorrelated. By convention, CFI should be equal to or greater
than 0.90 to accept the model. The root mean square error of approximation (RMSEA) corrects the
Chi-square statistic for sample size as well as model complexity. By convention, there is adequate
model fit if RMSEA is less than or equal to 0.08 (Hu and Bentler, 1999; Fan et al., 1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
coefficient on doctor satisfaction is significantly positive (y = 0.478, z = 3.703).
However, contrary to Hlb, which predicts a positive relationship between patient
satisfaction and future revenues, the coefficient on patient satisfaction is significantly
negative (y = -0.995, z = -2.665). I also find a negative correlation between client
satisfaction and patient satisfaction (r = -0.081 , P < 0. 10), contrary to the
expectations underlying the managers’ hypothesized business model. These results
imply that the interests of patients and clients are probably misaligned. This is
consistent with the health care literature, which suggests that patients and clients
often have conflicting objectives. Patients prefer comprehensive coverage with the
least out-of-pocket disbursements. Clients, on the other hand, give priority to cost
containment (Mascarenhas, 1993). Since clients tend to have greater purchasing
influence than patients, their preferences dominate those of the patients. Thus,
consistent with HI a, client satisfaction is more strongly associated with future
revenues than patient satisfaction. The correlation between doctor satisfaction and
patient satisfaction is positive but insignificant (r = 0.039). Finally, as expected, past
revenues and general economic condition are important predictors of current
revenues.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
Table 11. Structural Equation Modeling (SEM) Analysis of the Relation
between Satisfaction Measures and Future Revenue for the 20-Quarter Period
2000-2004
Overall High % Low % High Market Low Market
Sample Voluntary Voluntary Penetration Penetration
Patient Patient Subsample Subsample
Subsample Subsample
0 ) (2) (3) (4) (5)
4
0.496 -0.355 1.294 1.802 0.195
(1.386) (-0.540) (1.889)* (3.913)*** (0.388)
Client Satisfaction ,.8 0.546 0.501 0.716 0.486 0.679
(3.249)*** (1.581) (2.310)** (1.985)** (4.538)***
0.092 0.081 0.107 0.083 0.199
Patient Satisfaction,^ -0.995 0.226 -1.598 0.632 -1.385
(-2.665)* (1.653)* (-3.542)*** (1.451) (-3.291)***
-0.155 0.027 -0.254 0.076 -0.157
Doctor Satisfaction 0.478 0.598 0.380 0.353 0.529
(3.703)*** (2.228)** (5.168)*** (1.814)* (4.058)***
0.113 0.074 0.095 0.043 0.144
Revenue 0.769 0.742 0.586 0.846 0.797
(60.966)*** (33.364)*** (22.970)*** (34.984)*** (47.177)***
0.872 0.847 0.713 0.782 0.892
Economic Conditions, 0.275 0.357 0.470 0.296 0.264
(13.338)*** (9.233)*** (12.226)*** (16.493)*** (8.673)***
0.191 0.234 0.380 0.370 0.164
R2 0.799 0.773 0.661 0.753 0.823
Sample size 612 204 204 306 306
Degrees of freedom 40 40 40 40 40
Chi-square 167.609 103.704 123.915 97.429 84.576
RMSEA 0.054 0.061 0.080 0.058 0.046
CFI 0.948 0.936 0.924 0.927 0.954
Each cell reports the maximum likelihood coefficient estimate, the z-statistic (in parentheses), and the
standardized coefficient. ***, **, * indicate p-values of <0.01, 0.05, 0.10 in a two-tailed test.
Variable definitions:
Revenue (Dependent Variable) = Natural logarithm of total revenue in market i in quarter t
Client Satisfaction = Latent variable underlying the two indicators of client satisfaction (see Table 1,
Panel A and Table 4) in quarter t-8
Patient Satisfaction = Latent variable underlying the three indicators of patient satisfaction (see Table
1, Panel B and Table 4) in quarter t-4
Doctor Satisfaction = Latent variable underlying the two indicators of doctor satisfaction (see Table 1,
Panel C and Table 4) in quarter t-1
Economic Conditions = Natural logarithm of total income received by people who live or work in
state i in quarter t
High (Low) Market Penetration Subsample = Sample with the percentage of HBP patients out of the
entire population of each market greater (smaller) than 9.67% (median)
High (Low) Percentage Voluntary Patient Subsample = Sample with the percentage of HBP patients
who pay part or all of their HBP insurance premiums greater (smaller) than 37.65% (median)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Figure 8. Structural Equation Model Results for the Overall Sample (Structural
Model)
Client
Satisfaction
0.092***
0.872***
-0.081*
-0.155*
Patient
Satisfaction ,_ 4
0.113***
0.039
Doctor
Satisfaction
Economic
Conditions
Revenues M
Revenues t
Model fit statistics:
Ch-square = 167.609
RMSEA = 0.054
CFI = 0.948
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
4.4.3. SEM Multi-Group Analysis
H2b predicts that in those markets with a larger proportion of voluntary
patients, patients have greater purchasing influence and therefore patient satisfaction
should be more important than in the markets with a smaller proportion of voluntary
patients. Conversely, H2c predicts client satisfaction to be more important in the
markets with a smaller percentage of voluntary patients than in the markets with a
large percentage of voluntary patients. To test these hypotheses, I partition my
sample based on the percentage of voluntary patients. The subsample of 204 cases
with percentage of voluntary patients greater (lower) than 37.65% (median)
represents a setting where patients have more (less) purchasing influence. I test H2
by estimating a model that allows the structural model (SM) to differ between the
two sub-groups while assuming the same measurement model (MM). To establish
the appropriateness of a fixed MM, I first test whether differences exist in the factor
loadings between the two groups. Allowing the MM to vary freely between the two
groups does not significantly improve model fit (AX2 = 5.210, Adf = 4) (Bagozzi and
Yi, 1988). On the other hand, allowing the SM to vary between the two groups
significantly improves model fit compared to a model fixing the SM to be equal
(AX = 23.612, Adf = 5), suggesting that HBP’s overall business model differs
significantly across the two sub-samples.
Table 11 (Columns 2 and 3) and Figures 9 and 10 summarize the results of
this analysis. Consistent with H2b and H2c, I find that client satisfaction is more
relevant in the sub-sample with a lower percentage of voluntary patients, whereas
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
patient satisfaction is more relevant in the sub-sample with a higher percentage of
voluntary patients. The coefficient for client satisfaction is insignificant (y = 0.501, z
= 1.581) in the subsample with a high percentage of voluntary patients, but is
significantly positive (y = 0.716, z = 2.310) in the subsample with a low percentage
of voluntary patients. The Lagrange Multiplier (LM) test of the null hypothesis that
these two parameters are the same can be rejected (X = 2.729, df = 1, p < 0.10).
Similarly, the coefficient for patient satisfaction is significantly negative in the
subsample with a low percentage of voluntary patients (y = -1.598, z = -3.542), but
becomes positive and marginally significant (y = 0.226, z = 1.653) in the subsample
with a high percentage of voluntary patients. These two coefficients are statistically
different (X2 = 6.821, df = 1, p < 0.01). These results indicate that, consistent with
H2, customer purchasing influence affects the relevance of customer satisfaction
measures in a setting with multiple customer groups.
H3 propose that the associations between satisfaction measures and future
revenues are stronger in those markets where HBP has a low market penetration rate.
To test these hypotheses, I partition my sample at the median market penetration rate
(9.67%) and estimate the HBP model across the two sub-samples. I find that
allowing the MM to vary freely between the two sub-samples does not significantly
improve model fit (AX2 = 6.989, Adf = 4), which justifies using a fixed MM in the
multi-group analysis. However, allowing the SM to vary between the two sub
samples improves model fit compared to a fully constrained model (AX2 = 25.370,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
Adf = 5), which indicates that HBP’s business model differs significantly across the
two sub-samples.
Table 11 (Columns 4 and 5) and Figures 11 and 12 present results for the
two sub-samples. As predicted by H3a, the coefficients for client satisfaction are
positively related to revenues, and are significantly higher in the subsample with a
low penetration rate (y =0.679, z = 4.538) than in the subsample with a high
penetration rate (y =0.486, z = 1.985). These two coefficients are statistically
different (X2= 2.887, df = 1, p < 0.10). Similarly, the coefficients for doctor
satisfaction are positively related to revenue, and are significantly higher in the
subsample with a low penetration rate (y =0.529, z = 4.058) than in the subsample
with a high penetration rate (y =0.353, z = 1.814). These parameters are statistically
different (X2 = 3.906, df = 1, p < 0.05). Finally, the coefficients for patient
satisfaction are insignificant in the subsample with a high penetration rate (y =0.632,
z = 1.451), and are significant, albeit negative, in the subsample with a low
penetration rate (y =-1.385, z = -3.291). These parameters are statistically different
(X = 4.641, df = 1, p < 0 .05). I also observe that the structural equation squared
multiple correlations (i.e., the counterpart of R in OLS regressions) for the
subsample with low market penetration rate is 82.3%, compared to 75.3% for the
subsample with high market penetration rate. These results imply that, consistent
with H3, customer satisfaction is more relevant in environments where customers
have greater bargaining power.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
Figure 9. Structural Equation Model Results for the High Percentage Voluntary
Patient Subsample
Client
Satisfaction
0.081
0.847***
0.049
0.027*
Patient
Satisfaction t_ 4
0.074**
0.033 0.234***
Doctor
Satisfaction
Economic
Conditions
Revenues M
Revenues
Model fit statistics:
Ch-square = 103.704
RMSEA = 0.061
CFI = 0.936
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
Figure 10. Structural Equation Model Results for the Low Percentage
Voluntary Patient Subsample
Client
Satisfaction
Revenues
0.713***
0.254***
Patient
Satisfaction
Revenues
0.095***
0.380***
Doctor
Satisfaction
Economic
Conditions
Model fit statistics:
Ch-square = 123.915
RMSEA = 0.080
CFI = 0.924
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
Figure 11. Structural Equation Model Results for the High Market Penetration
Subsample
Client
Satisfaction ,_ 8
0.083**
-0.065
0.076
Patient
Satisfaction ,_ 4
0.043*
0.086 0.370***
Doctor
Satisfaction ,_ i
Revenues t ,4
Economic
Conditions,
Revenues
Model fit statistics:
Ch-square = 97.429
RMSEA = 0.058
CFI = 0.927
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
Figure 12. Structural Equation Model Results for the Low Penetration Rate
Subsample
Client
Satisfaction
Revenues
q 199* * *
0.892***
■0.157***
Patient
Satisfaction
Revenues
0 144* * *
Doctor
Satisfaction
Economic
Conditions
Model fit statistics:
Ch-square = 84.576
RMSEA = 0.046
CFI = 0.954
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
4.4.4. Robustness Checks
The above analysis does not consider the lack of independence of
observations in the panel data with multiple states across time. As a robustness
check, I include state dummies in the estimation. These dummies are jointly
insignificant and do not affect the inferences. Controlling for quarter dummies does
not changes the results either.
In addition, since the choice of time lags between various satisfaction
measures and future revenues is somewhat arbitrary, I test alternative lags (e.g., 4
instead of 8 quarters of time lag between client satisfaction and future revenues) and
obtain similar results.
Two alternative measures of performance are the number of clients and the
number of patients. I repeat the analyses using these measures as the dependent
variables. The results are robust to these alternative measures of performance.
Finally, I also estimate changes models instead of levels models, and the
results (summarized in Table 12) do not change substantively, except that patient
satisfaction becomes insignificant in the overall model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
Table 12. Structural Equation Modeling (SEM) Analysis of the Relation
between Changes in Satisfaction Measures and Changes in Future Revenue for
the 20-Quarter Period 2000-2004
Overall High % Low % High Market Low Market
Sample Voluntary Voluntary Penetration Penetration
Patient Patient Subsample Subsample
Subsample Subsample
(1) (2) (3) (4) (5)
4 .
0.013 0.014 0.016 0.012 0.018
(5.037)*** (2.959)*** (3.804)*** (3.522)*** (4.150)***
AClient Satisfaction ,.8 0.010 0.014 0.025 0.009 0.023
(2.186)** (1.494) (3.104)*** (0.923) (3.002)***
0.079 0.098 0.202 0.054 0.195
APatient Satisfaction ,_ 4 0.018 0.016 -0.022 0.013 0.014
(0.958) (2.215)** (-2.088)** (0.737) (1.494)
0.063 0.132 -0.135 0.048 0.098
ADoctor Satisfaction ,_ i 0.008 0.019 0.010 0.009 0.020
(1.669)* (3.019)*** (2.186)** (2.075)** (3.025)***
0.061 0.132 0.079 0.070 0.105
ARevenuet .i 0.281 0.440 0.480 0.295 0.266
(7.095)*** (6.560)*** (7.833)*** (5.555)*** (4.481)***
0.310 0.447 0.499 0.326 0.291
AEconomic Conditions, 4.581 16.377 19.131 4.064 8.401
(2.266)** (3.429)*** (5.223)*** (1.685)* (2.272)**
0.099 0.231 0.351 0.098 0.148
R2 0.211 0.329 0.118 0.134 0.248
Sample size 561 179 178 281 280
Degrees of freedom 40 40 40 40 40
Chi-square 168.712 179.518 100.130 185.649 98.112
RMSEA 0.060 0.058 0.043 0.069 0.037
CFI 0.933 0.931 0.958 0.922 0.967
Each cell reports the maximum likelihood coefficient estimate, the z-statistic (in parentheses), and the
standardized coefficient. ***, **, * indicate p-values of <0.01, 0.05, 0.10 in a two-tailed test.
See Table 5 for variable definitions.
ARevenue ((Dependent Variable)=(Revenuet- Revenue t _i) / Revenue t -i
The other changes variables are defined in the same way.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
4.5. Discussion
This chapter examines the following question in a setting with multiple
customer groups: How does customer bargaining power affect the relation between
customer satisfaction measures and future financial performance? Using structural
equation modeling, I find that, as predicted by the research site’s maintained
business model, future revenue is positively associated with client satisfaction and
doctor satisfaction. However, contrary to the expectations captured by the business
model, future revenue is negatively associated with patient satisfaction. The result
suggests conflicting interests of clients and patients. I also find that customer
purchasing influence determines which customer satisfaction measure is more
relevant in such settings. Specifically, I find that the proportion of voluntary patients
who have greater purchasing influence enhances the relevance of patient satisfaction
while decreasing the relevance of client satisfaction. Finally, I find that customer
bargaining power as measured (reversely) by the research site’s market penetration
rate increases the relevance of customer satisfaction, in that customer satisfaction
measures are more closely associated with future revenues when HBP’s market
penetration is low and customer bargaining power is high.
The results of this chapter should be interpreted with several caveats. First, an
inherent limitation of field-based research is the limited generalizability of the
empirical results. However, the results are likely to be generalizable to similar firms
in the health care insurance industry that have similar business models and face
similar challenges of balancing needs and preferences of multiple customer groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
More importantly, although the results per se may have limited generalizability, the
study is generalizable to the extent that the theory is generalizable. For instance, the
result that customer purchasing influence and customer bargaining power affect the
relevance of customer satisfaction measures should be generalizable.
Second, the satisfaction data of this study comes from third-party surveys
and, as a result, I have no control over the design and administration of the surveys.
Discussions with the HBP managers revealed that although they follow some
recommended procedures of survey design and administration such as random
sampling and follow-ups, they do not follow other procedures such as non-response
bias analysis (Van der Stede, Young, and Chen, 2005).
Third, consistent with HBP’s business model, the structural equation model
assumes linear relations between customer satisfaction and future financial
performance. The true relations, however, could be nonlinear (e.g., Ittner and
Larcker, 1998a). Nonetheless, the recent marketing literature finds that: (1) over a
reasonable range, the assumption of a linear relationship between satisfaction and
loyalty is acceptable (Yeung et al., 2002); and (2) models that assume non-linearity
between satisfaction and loyalty do not have superior explanatory power over linear
models (Streukens and Ruyter, 2004). These empirical results mitigate, but do not
completely eliminate, the concern with non-linearity.
Finally, there are other factors to consider when designing performance
measurement and incentive systems other than the informativeness or predictive
ability of measures (Banker and Datar, 1989; Feltham and Xie, 1994). Therefore,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
caution must be taken when applying the results of this study to the design of
performance measurement systems. For example, Datar, Kulp, and Lambert (2001)
provide insights into the trade-offs between congruity and sensitivity/precision in a
multi-action setting. They show that in a multi-action setting, the contract should
motivate both the intensity of the agent’s efforts and the allocation of the effort
across different dimensions. My research site also provides an example of the
different purposes that incentive contracts serve in an organization. HBP introduced
the incentive contracts both to motivate the allocation of employees’ efforts and to
communicate a customer-focused organizational strategy that seeks to balance the
needs of different customer groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
Chapter 5. Conclusion
This study employs two research settings to understand the effect of customer
bargaining power on the relation between customer satisfaction and future financial
performance. The findings of this study suggest that the strength of the linkages
between customer satisfaction and financial performance varies as a function of
customer bargaining power as measured by customer switching costs, customer
purchasing influence, and a company’s market power (reverse measure).
From a theoretical perspective, this study contributes to an emerging
literature that examines potential non-linearities in the relationships between non-
financial and financial performance measures. For example, Campbell et al. (2004)
conducted detailed analysis of the performance measurement system of a
convenience store and found that the impact of the firm’s nonfmancial measure of
strategy implementation on financial performance varies as a function of employee
skills levels. Like Campbell et al. (2004), the current study highlights the importance
of considering interactions between non-financial performance measures and
contextual factors in affecting future financial performance as well as the importance
of validating hypothesized links in a firm’s performance measurement systems
(Kaplan and Norton, 1996, 2000).
This study also has the following managerial implications. The most
important practical suggestion for managers is that they should take contextual
factors into consideration when formulating business models and designing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
performance measurement systems. One size simply does not fit all when it comes to
performance measurement systems. The results of the field-based study also
illustrate the importance of considering multiple customer groups at once when
trying to determine whether improving satisfaction will result in better financial
performance. For firms that have multiple customer groups, managers should pay
attention to potential spillover effects between the satisfaction levels of different
groups. In addition, by examining the circumstances under which customer
satisfaction has the greatest impact on future revenues, the results of this study can
help managers make more informed decisions related to resource allocation,
performance evaluation, and compensation. For example, the relative weights
assigned to customer satisfaction measures in a Balanced Scorecard performance
measurement system should probably be heavier in environments where customers
have greater bargaining power.
That said, these implications are not necessarily easy to implement. The
contingent effects of customer purchasing influence and customer bargaining power
must be carefully attended to by organizations before investing in customer
initiatives. However, market conditions are always evolving over time, requiring
managers to constantly monitor the performance implications of different customer
groups and revise their strategy and business model accordingly. The practical
implication for managers, therefore, is that they should consider the importance of
customer satisfaction given their current business environment while remaining
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
flexible to shift resources between customer initiatives and other operational
requirements as well as between different customer groups.
This study represents a step toward understanding the complex relations
between nonfinancial performance measures, contextual factors, and future financial
performance. The limitations of this study point to several directions for future
research. First, future studies may enlarge the research sample by including more
firms and industries to complement the evidence presented in this study. Second,
future research could examine other moderators of the relation between customer
satisfaction and future financial performance, e.g., repurchase cycle and strategy.
Finally, due to data limitations, this study focuses on a single non-financial
performance measure and a single aspect of customer relationship management.
However, customer relationships can be multifaceted, including customer
satisfaction, antecedents of customer satisfaction such as product and service
infrastructure, and consequences of customer satisfaction such as customer usage and
customer volume (Nagar and Raj an, 2005). Future research could explore other
dimensions of customer relationship management as well as more comprehensive
business models that include customer measures, employee measures, operational
measures, and financial performance measures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
Bibliography
Abemethy, M. A., M. Home, A. M. Lillis, M. A. Malina, and F. H. Selto. 2005. A
multi-method approach to building causal performance maps from expert
knowledge. Management Accounting Research 16: 135-155.
Aiken, L. S., and S. G. West. 1991. Multiple regression: Testing and interpreting
interactions. Newbury Park, CA: Sage Publications.
Aitchison, J., and D. C. Silvey. 1958. Maximum likelihood estimation of parameters
subject to restraints. Annuals o f Mathematical Statistics 29: 813-828.
Amir, E., and B. Lev. 1996. Value-relevance of nonfinancial information: The
wireless communications industry. Journal o f Accounting and Economics 22
(1-3): 3-30.
Anderson, E. W. 1994. Cross-category variation in customer satisfaction and
retention. Marketing Letters 5: 19-30.
Anderson, E. W., and C. Fomell. 2000. Foundations of the American Customer
Satisfaction Index. Total Quality Management 11 (7): 869-882.
Anderson, E. W., and M. W. Sullivan. 1993. The antecedents and consequences of
customer satisfaction for firms. Management Science 40 (3): 405-417.
Anderson, E. W., and V. Mittal. 2000. Strengthening the satisfaction-profit chain.
Journal o f Service Research 3: 107-120.
Anderson, E. W., C. Fomell, and D. R. Lehmann. 1994. Customer satisfaction,
market share, and profitability: Findings from Sweden. Journal o f Marketing
Research 58: 53-66.
Anderson, E. W., C. Fomell, and R. T. Rust. 1997. Customer satisfaction,
productivity, and profitability: Differences between goods and services.
Marketing Science 2: 129-45.
Assmus, G., J. V. Farley, and D. R. Lehmann. 1984. How advertising affects sales:
Meta-analysis of econometric results." Journal o f Marketing Research 21:
65-74.
Audubel, L. 1991. The failure of competition in the credit card market. The
American Economic Review 81: 50-81.
Bagozzi, R. P., and Y. Yi. 1998. On the evaluation of structural equation models.
Journal o f the Academy o f Marketing Science 16 (1): 74-94.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
Baker, L. 2001. Measuring competition in health care markets. Health Services
Research 36 (1): 223-251.
Bakos, Y. 1991. A strategic analysis of electronic marketplaces.” MIS Quarterly 15
(3): 295-310.
Bakos, Y. 1997. Reducing buyer search costs: Implications for electronic
marketplaces. Management Science 43 (12): 1676-1692.
Banker, R. D., and S. M. Datar. 1989. Sensitivity, precision, and linear aggregation
of signals for performance evaluation. Journal o f Accounting Research 27
(1): 21-39.
Banker, R. D., G. Potter, and D. Srinivasan. 2000. An empirical investigation of an
incentive plan that includes nonfinancial performance measures. The
Accounting Review 75 (1): 65-92.
Banker, R. D., S. Y. Lee, and G. Potter. 1996. A field study of the impact of a
performance-based incentive plan. Journal o f Accounting and Economics 21:
195-226.
Behn, B. K., and R. A. Riley. 1999. Using nonfinancial information to predict
financial performance: The case of the U.S. airline industry. Journal o f
Accounting, Auditing and Finance 14 (1): 29-56.
Bernhardt, K. L., N. Donshu, and P. A. Kennett. 2000. A longitudinal analysis of
satisfaction and profitability. Journal o f Business Research 47: 161-171.
Bitner, M. J. 1990. Evaluating service encounters: The effects of physical
surroundings and employee responses. Journal o f Marketing 54: 69-82.
Bollen, K. A. 1989. Structural equations with latent variables. New York: Wiley.
Boomsma, A., and J. J. Hoogland. 2001. The robustness of LISREL modeling
revisited. In Cudeck, R., S. de Toit, and D. Sorbom (eds.) Structural
Equation Models: Present and Future. Chicago: Scientific Software
International: 139-168.
Bouwens, J., and M. A. Abemethy. 2000. The consequences of customization on
management accounting system design. Accounting, Organizations and
Society 25: 221-241.
Brynjolfsson, E., and M. D. Smith. 2000. Frictionless commerce: A comparison of
Internet and conventional retailers. Management Science 46 (4): 563-585.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
Campbell, D., S. Datar, S. L. Kulp, and V. G. Narayanan. 2004. The strategic content
of non-financial performance measures. Working paper.
Chenhall, R. H. 2003. Management control systems design within its organizational
context: findings from contingency-based research and directions for the
future. Accounting, Organizations and Society 28: 127-168.
Chong, V. K., and Rundus, M. J. 2004. Total quality management, market
competition and organizational performance. British Accounting Review 36:
155-172.
Cook, D. P., C. H., Goh, and C. H., Chung. 1999. Service typologies: A state of the
art survey. Production and Operations Management 8 (3): 318-338.
Cronbach, L., and P. Meehl. 1955. Construct validity in psychological tests.
Psychological Bulletin 52: 281-302.
Cronin, J. J. Jr., M. K. Brady, and G. T. M. Hult. 2000. Assessing the effects of
quality, value, and customer satisfaction on consumer behavioral intentions in
service environments. Journal o f Retailing 76 (2): 193-218.
Cronin, J. J., and S. A. Taylor. 1992. Measuring service quality: A reexamination
and extension. Journal o f Marketing 56: 55-68.
Dahlby, B., and D. West. 1986. Price dispersion in an automobile insurance market.
Journal o f Political Economy 94: 418-438.
Datar, S., S. C. Kulp, and R. A. Lambert. 2001. Balancing performance measures.
Journal o f Accounting Research 39 (1): 75-92.
De Castro, J. O., and J. J. Chrisman. 1995. Order of market entry, competitive
strategy, and financial performance. Journal o f Business Research 33: 165-
177.
DeFond, M. L., and C. W. Park. 1999. The effect of competition on CEO turnover.
Journal o f Accounting and Economics 27: 35-56.
Dikolli, S. S., W. R. Kinney, Jr., and K. L. Sedatole. 2005. Measuring trading partner
relationship value: Inferences from customer switching data. Working paper.
Distefano, L. 2001. The impact of categorization with confirmatory factor analysis.
Structural Equation Modeling 9 (3): 327-346.
Drucker, J. 2004. Number portability’ rules to include more consumers. Wall Street
Journal, May 24: B5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
Drucker, J. 2003. Courts allows customers to keep cell phone numbers after switch.
Wall Street Journal, June 9: A8.
Drucker, J., and A. M. Sequeo. 2004. Cellular merger to alter service for millions;
AT&T Wireless-Cingular deal gains key regulatory approval; Impact on
coverage, pricing. Wall Street Journal, October 26: Dl.
Eccles, R. G. 1991. The performance measurement manifesto. Harvard Business
Review 69(1): 131-137.
Engel, R. 1985. Wald, likelihood ratio and Lagrange multiplier tests in econometrics.
In Griliches, Z. and M. D. Intriligator (eds.), The Handbook o f Econometrics,
North-Holland Publishing Co., Amsterdam: 776-826.
Enhrenberg, A. S. 2000. Repetitive advertising and the consumer. Journal o f
Advertising Research 40 (6): 39-48.
Fama, E. F., and K. R. French. 1995. Size and book-to-market factors in earnings
and returns. Journal o f Finance 50: 131-155.
Fan, X., B. Thompson, and L. Wang. 1999. Effects of sample size, estimation
methods, and model specification on structural equation modeling fit indexes.
Structural Equation Modeling 6: 56-83.
Feeny, D., M. Lacity, and L. P. Willcocks. 2005. Taking the measure of outsourcing
providers. MIT Sloan Management Review, 46 (3): 40-48.
Feltham, G., and J. Xie. 1994. Performance measures congruity and diversity in
multi-task principal/agent relations. The Accounting Review 69 (3): 429-453.
Fomell, C. 1992. A national customer satisfaction barometer: The Swedish
experience. Journal o f Marketing 56 (1): 6-21.
Fomell, C., and M. D. Johnson. 1993. Differentiation as a basis for explaining
customer satisfaction across industries. Journal o f Economic Psychology 14:
681-696.
Fomell, C., M. D. Johnson, E. W. Anderson, J. Cha, and B. E. Bryant. 1996. The
American Customer Satisfaction Index: Nature, purpose, and findings.
Journal o f Marketing 40 (4): 7-18.
Foster, G., and M. Gupta. 1997. The profitability implications of customer
satisfaction. Working paper.
Handheld, R. B., and Nichols, E. L. Jr. 1999. Introduction to Supply Chain
Management. Englewood Cliffs, NJ: Prentice-Hall.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
Hansen, S. 1998. Cost analysis, cost reduction and competition. Journal o f
Management Accounting Research 10: 181-204.
Hartmann, F. G. H., and F. Moers. 1999. Testing contingency hypotheses in
budgetary research: An evaluation of the use of moderated regression
analysis. Accounting, Organizations and Society 24: 291-315.
Hax, A,, and D. Wilde. 2001. The Delta project: Discovering new sources of
profitability in a networked economy. New York: Palgrave.
Hemmer, T. 1996. On the design and choice of “modem” management accounting
measures. Journal o f Management Accounting Research 8: 87-116.
Hoque, Z., L. Mia, and M. Alam. 2001. Market competition, computer-aided
manufacturing and use of multiple performance measures: An empirical
study. British Accounting Review 33 (1): 23-45.
Hu, L., and P. M. Bentler. 1999. Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling 6: 1-55.
Ittner, C. D., and D. F. Larcker. 1998a. Are nonfinancial measures leading indicators
of financial performance? An analysis of customer satisfaction. Journal o f
Accounting Research 36 (Supplement): 1-35.
Ittner, C. D., and D. F. Larcker. 1998b. Innovations in performance measurement:
Trends and research implications. Journal o f Management Accounting
Research 10: 205-238.
Ittner, C. D., and D. F. Larcker. 2001. Assessing empirical research in managerial
accounting: a value-based management perspective. Journal o f Accounting
and Economics 32: 349-410.
Ittner, C. D., and D. F. Larcker. 2003. Coming up short on nonfinancial performance
measurement. Harvard Business Review 81(11): 88-95.
Ittner, C. D., D. F. Larcker, and M. V. Raj an. 1997. The choice of performance
measures in annual bonus contracts. The Accounting Review 72 (2): 231-255.
Jones, M. A., D. L. Mothersbaugh, and S. E. Beatty. 2000. Switching costs and
repurchase intentions in services. Journal o f Retailing 76 (2): 259-274.
Jones, M. A., D. L. Mothersbaugh, and S. E. Beatty. 2002. Why customers stay:
Measuring the underlying dimensions of services switching costs and
managing their differential strategic outcomes. Journal o f Business Research
55: 441-450.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Kaplan, R., and D. Norton. 1992. The Balanced Scorecard: Measures that drive
performance. Harvard Business Review (Jan-Feb): 71-79.
Kaplan, R., and D. Norton. 1996. The Balanced Scorecard: Translating strategy into
action. Boston, MA: Harvard Business School Press.
Kaplan, R., and D. Norton. 2000. The Strategy Focused Organization. Boston, MA:
Harvard Business School Press.
Kaplan, R., and D. Norton. 2004. Strategy Maps. Boston, MA: Harvard Business
School Press.
Khandwalla, P. N. 1972. The effect of different types of competition on the use of
management controls. Journal o f Accounting Research 10 (2): 275-285.
Kim, M., D. Kliger, and B. Vale. 2001. Estimating switching costs and oligopolistic
behavior. Working Paper.
Klemperer, P. 1987. Markets with customer switching costs. Quarterly Journal o f
Economics 102: 375-394.
Klemperer, P. 1995. Competition when consumers have switching costs. Review o f
Economic Studies 62: 515-539.
Knittel, C. R. 1997. Interstate long distance rate: Search costs, switching costs, and
market power. Review o f Industrial Organization 12: 519-536.
Kohli, A. K., and B. J. Jaworski. 1990. Market orientation: The construct, research
propositions, and managerial implications. Journal o f Marketing 54 (2): 1-18.
Kreps, D. M. 1990. A Course in Microeconomic Theory. Princeton, NJ: Princeton
University Press.
Krishnan, R. 2005. The effect of changes in regulation and competition on firms’
demand for accounting information. The Accounting Review 80 (1): 269-287.
Krishnan, R., J. Luft, and M. D. Shields. 2002. Competition and cost accounting:
Adapting to changing markets. Contemporary Accounting Research 19: 271-
302.
Lambert, R. A. 1998. Customer satisfaction and future financial performance:
Discussion of “Are nonfinancial measures leading indicators of financial
performance? An analysis of customer satisfaction”. Journal o f Accounting
Research 36 (Supplement): 37-46.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
Lambert, R. A., and D. F. Larcker. 1995. The prospective payment system, hospital
efficiency, and compensation contracts for senior-level administrators.
Journal o f Accounting and Public Policy 14: 1-31.
Langfield-Smith, K. 1997. Management control systems and strategy: A critical
review. Accounting, Organizations and Society 22 (2): 207-232.
Malina, M., and F. Selto. 2001. Communicating and controlling strategy: An
empirical study of the effectiveness of the balanced scorecard. Journal o f
Management Accounting Research 13: 47-90.
Malina, M., and F. Selto. 2003. Causality in performance measurement models.
Working paper.
McKitterick, J. B. 1957. What is the marketing management concept? In Bass, F. M.
(ed.) The Frontiers o f Marketing Thought and Science. Chicago: American
Marketing Association: 71-92.
Mia, L., and B. Clarke. 1999. Market competition, management accounting systems
and business unit performance. Management Accounting Research 10: 137-
158.
Nagar, V., and M. V. Rajan. 2001. The revenue implications of financial and
operational measures of product quality. The Accounting Review 76 (4): 495-
513.
Nagar, V., and M. V. Rajan. 2005. Measuring customer relationships: The case of
the retail banking industry. Management Science 51 (6): 904-919.
Niraj, R., M. Gupta, and C. Narasimhan. 2001. Customer profitability in a supply
chain. Journal o f Marketing 65 (3): 1-16.
Porter, M. E. 1980. Competitive Strategy. New York: The Free Press.
Porter, M. E. 1985. Competitive Advantage. New York: The Free Press.
Rhoades, S. 1993. The Herfindahl-Hirschman Index. Federal Reserve Bulletin 79
(March): 188-189.
Rucci, A. J., S. P. Kim, and R. T. Quinn. 1998. The employee-customer-profit chain
at Sears. Harvard Business Review 76 (1): 82-97.
Said, A. A., H. R. HassabElnaby, and B. Wier. 2003. An empirical investigation of
the performance consequences of nonfinancial measures. Journal o f
Management Accounting Research 15: 193-223.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
Schlesinger, H., and J. M. G. Von der Schulenburg. 1993. Consumer information and
decisions to switch insurers. The Journal o f Risk and Insurance 60: 519-615.
Schlesinger, L. A., and J. Heskett. 1991. The service-driven company. Harvard
Business Review 69 (5): 71-81.
Sedatole, K. L. 2003. The effect of measurement alternatives on a nonfinancial
quality measure’s forward-looking properties. The Accounting Review 78 (2):
555-580.
Shevlin, T. 1996. The value-relevance of nonfinancial information: A discussion.
Journal o f Accounting and Economics 22 (1-3): 31-42.
Shields, J. F., and M. Shields. 1998. Antecedents of participative budgeting.
Accounting, Organizations, and Society 23: 49-76.
Shields, M. 1997. Research in management accounting by North Americans in the
1990’s. Journal o f Management Accounting Research 9: 1-61.
Smith, D., and K., Langfield Smith. 2004. Structural equation modeling in
management accounting research: critical analysis and opportunities. Journal
o f Accounting Literature 23: 49-86.
Smith, R. E., and W. F. Wright. 2004. Determinants of customer loyalty and
financial performance. Journal o f Management Accounting Research 16:
183-205.
Stango, V. 1998. Price dispersion and switching costs: Evidence from the credit card
market. Working Paper.
Streukens, S., and K. de Ruyter. 2004. Reconsidering nonlinearity and asymmetry in
customer satisfaction and loyalty models: An empirical study in three retail
service settings. Marketing Letters 15 (2-3): 99-111.
Tellis, G. J. 1988. The price elasticity of selective demand: A meta-analysis of
econometric models of sales. Journal o f Marketing Research 25: 331-341.
Valos, M. J., P. FitzRoy, and F. Mavondo. 2004. Marketing and financial
performance of the Porter strategy types. Working Paper.
Van der Stede, W. A., S. M. Young, and C. X. Chen. 2005. Assessing the quality of
evidence in empirical management accounting research: The case of survey
studies. Accounting, Organizations, and Society 30 (7/8): 655-684.
Viard, V. B. 2003. Do switching costs make markets more or less competitive? The
case of 800-number portability. Working Paper.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
Western, D., and R. Rosenthal. 2003. Quantifying construct validity: Two simple
measures. Journal o f Personality and Social Psychology 84 (3): 608-618.
White, R. R. 1986. Generic business strategies, organizational context and
performance: An empirical investigation. Strategic Management Journal 7
(3): 217-231.
Wright, P., M. Kroll, P., Chan, and K., Hamel. 1991. Strategic profiles and
performances: An empirical test of select key propositions. Journal o f the
Academy o f Marketing Science, 19 (3): 245-255.
Yeung, M. C. H., L. C. Ging, and C. T. Ennew. 2002. Customer satisfaction and
profitability: A reappraisal of the nature of the relationship. Journal o f
Targeting, Measurement and Analysis for Marketing 11 (1): 24-33.
Zeithaml, V. A. 2000. Service quality, profitability, and the economic worth of
customers: What we know and what we need to learn. Journal of the
Academy o f Marketing Science 28 (1): 67-85.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Does adding nonfinancial value drivers to a summary financial measure improve the learning and performance of managers?
PDF
Essays on technology innovation and new product development strategy
PDF
Information diffusion and the boundary of market efficiency
PDF
Effects of redundancy in media coverage on nonprofessional investors' earnings forecasts
PDF
An investigation of repeat visit behavior on the Internet: Models and applications
PDF
Androgyny and transformational leadership: Effects of gender and sex -role identity in the collectivistic context of Taiwan, R.O.C.
PDF
Essays on technological evolution and financial returns to innovation
PDF
Beyond skill development: The effects of training and development on the attitudes and retention of employees
PDF
Feeling good about control: Design considerations for accountability systems in schools
PDF
Influencing individual innovation through technology features that support cross -departmental understanding
PDF
Board involvement in monitoring and strategy making: Antecedents and consequences
PDF
Expression and functional studies of the Notch signaling pathway in feather development
PDF
Corporate ethical orientation and crisis management before and after September 11, 2001
PDF
Automatically and accurately conflating road vector data, street maps and orthoimagery
PDF
Design and performance of space -time codes
PDF
Demographic diversity, team process, and team performance: Assessing moderator effects of cognitive conflict management practices and task interdependence
PDF
Essays in supply chain management
PDF
Consumer choice under budget constraint: Why consumers overspend
PDF
Essays on new product preannouncements
PDF
Flow of fluids with yield stress in porous media
Asset Metadata
Creator
Chen, Xiaoling
(author)
Core Title
Customer satisfaction, customer bargaining power, and financial performance
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Business Administration
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
business administration, accounting,OAI-PMH Harvest
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Bonner, Sarah (
committee chair
), Young, Mark (
committee chair
), Hsiao, Cheng (
committee member
), Merchant, Kenneth A. (
committee member
), Sandino, Tatiana (
committee member
), Van der Stede, Wim (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-583928
Unique identifier
UC11336633
Identifier
3236488.pdf (filename),usctheses-c16-583928 (legacy record id)
Legacy Identifier
3236488.pdf
Dmrecord
583928
Document Type
Dissertation
Rights
Chen, Xiaoling
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
business administration, accounting