Description
Description explain about a study of corporate entrepreneurship and firm performance in the computer industry.
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
A Study Of Corporate Entrepreneurship
And Firm Performance In The Computer
Industry During Two
Technological Disruptions
Ronald M. Rivas, (E-mail: [email protected]), Canisius College
ABSTRACT
This study presents a model of firm performance for Corporate Entrepreneurship (CE) Entry-Growth–
Exit strategies categorized by focus of entrepreneurship, CE domain, and CE form. Over a longitudinal
sample, this paper examines U.S. computer firm’s CE strategies and firm performance during two
technological disruptions namely the introduction of personal computers and subsequently the internet.
This paper shows that computer companies had superior performance by implementing CE strategies of
sustained regeneration and organizational rejuvenation during growth. In addition, best performing
companies survived the internet technological disruption by implementing the CE strategy of domain
redefinition. This study concludes with recommendations for future research.
INTRODUCTION
hat is the impact of corporate entrepreneurship on firmperformance during technological disruptions?
Both scholars and practitioners are interested in understanding corporate entrepreneurship (Chandler &
Lyon, 2001; Dess, Ireland, Zahra, Floyd, Janney, & Lane, 2003; Gregoire, Noel, Dery, & Bechard,
2006). Corporate entrepreneurship (CE) is the process whereby a firmcreates a new organization or
instigates renewal or innovation within an existing organization (Sharma & Chrisman, 1999). Corporate Entrepreneurship
has long been identified as a key means for promoting and sustaining competitive advantage (Covin & Miles, 1999). CE is
a means of accumulating, converting, and leveraging resources for competitive purpose (Floyd & Wooldridge, 1999) to
rejuvenate and redefine the firmand its markets (Covin & Miles, 1999). CE is also a means to venture into new markets by
creating new organizations fromwithin the firm, or acquiring them, or engaging in complex alliances for the purpose of
market entry (Miles & Covin, 2002).
j
Although significant theorizing has been proposed to clarify the domain of CE , there is nevertheless a need to
examine what is the relationship between different CE forms and firmperformance (Dess, Ireland, Zahra, Floyd, Janney, &
Lane, 2003). An area that has not received much attention in CE research is the effectiveness of CE under technological
disruptions. There is substantial research on CE under dynamic markets (Martin & Eisenhardt, 2004) in particular in the
computer industry (Christensen, 1997; Eisenhardt & Tabrizi, 1995). However, the understanding of what type of CE, if
any, might yield superior performance under technological disruption is limited. A technological disruption (Christensen,
1997) is a period of rapid industry transformation characterized by the introduction and adoption of new technologies
resulting in the rise of new entrants and the obsolescence of incumbents (Danneels, 2004). Hence, this study contributes to
the understanding of the impact of CE on firmperformance during technological disruptions.
The main goal of this study is to propose a model of CE and firmperformance under technological disruption. A
specific objective of this study is to examine and test what types of CE yield superior performance during entry, growth and
exit on an industry under technological disruption such as the computer industry.
65
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
This study begins by briefly reviewing the CE literature and proposing the model of CE and firmperformance
under technological disruption. It follows a brief description of the successive technological disruptions of the computer
industry from1976 to 2007. Then it proceeds with an application of cross-section time-series regression analysis with
pooled data. Tests were designed to assess the impact of CE on firmperformance during entry, growth, and exit on the
computer industry. It concludes with a discussion of results.
THEORETICAL BACKGROUND
Danneels (2004) defined disruptive technologies as “those technologies that render established technologies
obsolete and therefore destroy the value of the investments that incumbents have made in those technologies,” (p.248).
Christensen (1997) claims that overshooting is the main self-defeating mechanismof incumbents during technological
disruption. On the face of disruptive technologies, incumbents tend to overshoot, that is, they offer improved versions of
their products to the point where their new features are beyond the quality expectations of customers. On the other hand,
new entrants will offer products often below customer quality expectations but with features that satisfy customer needs
better than old products.
Figure 1
Corporate Entrepreneurship and Firm Performance during Technological Disruptions
Focus of
Entrepreneurship
CE domain Entry Growth Exit
Internal
Corporate
Rejuvenation
Innovators dilemma:
Incumbents
overshooting yields
diminishing
performance
Corporate rejuvenation
yields higher
performance gains
Innovators dilemma:
Best performers have a
chance of surviving
using Domain
redefinition. Other CE
forms lead to
overshooting and
diminishing
performance
Corporate Venturing Innovators dilemma:
Tech Disruption favors
performance gains
through Internal
venturing of New
entrants
Uncertainty of Tech
disruption is resolved
and internal venturing
yields diminishing
performance
Innovators dilemma:
Internal venturing
leading to domain
redefinition could
yield performance
gains
External
Corporate
Rejuvenation
Innovators dilemma:
Incumbents emphasize
internal improvement
over domain
redefinition, however,
yielding diminishing
performance
Uncertainty of Tech
disruption is resolved
and strategic
positioning yield
increasing
performance gains
Innovators dilemma:
Incumbents exit
seeking to sell venture
to "rightful' owner
among new entrants to
overcome diminishing
performance
Corporate Venturing Innovators dilemma:
Incumbents enter
through acquisition to
overcome diminishing
performance of old
business model
Acquisitions increase
market share and
stabilize supply chain
yielding performance
gains
Innovators dilemma:
Incumbents exit
cutting losses by
divesting or liquidating
66
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
Covin and Miles (1999) proposed a useful typology to classify CE centered on the concept of firmlevel
innovation with the objective of rejuvenating or purposely redefining organizations, markets or industries to create
competitive advantage. According to these authors, CE could be labeled as sustained regeneration, organizational
rejuvenation, strategic renewal and domain redefinition. Sustained regeneration is related to continuous improvement.
Organizational rejuvenation focuses on the organizational changes necessary to improve firmprocesses. Strategic renewal
focuses on how the firmcompetes, and the necessary changes to adapt to new realities of competition and competitor
moves. Domain redefinition focuses on changes a firmneeds to implement to enter into future new markets.
Miles and Covin (2002) clarified the domain of corporate venturing and proposed that corporate venturing
differentiates by focus of entrepreneurship and presence of investment intermediation. According to these authors,
corporate venturing could be internal or external to the corporation. In addition, it could be brought by direct investment in
the venture through the corporations operating strategic budget or by indirect investment in the venture using financial
intermediaries.
This study builds on the above mentioned typologies of CE (Covin & Miles, 1999; Miles & Covin, 2002) and on
the call for re-examination of firmperformance during disruptive technologies (Danneels, 2004). Figure 1 (See Figure 1
below) shows this study’s proposed typology of CE and firmperformance during technological disruptions. The model
shows that firmperformance is affected by the innovator’s dilemma during entry and exit. An incumbent often overshoots
by emphasizing on CE of “sustained regeneration” rather than switch to CE of “domain redefinition” or to corporate
venturing. However, Christensen (1997) claims that companies that have become leaders through the practice of CE of
“sustained regeneration” cannot enact a CE of “domain redefinition.” This notion is challenged by Danneels (2004), who
calls for renewed empirical theorizing and testing on this matter.
The ultimate answer to whether firms can use (or learn to use) new CE forms is empirical. Covin and Miles
(1999) suggest that there is no causal relationship between one CE formand another. These authors state that “while
sustained regeneration and domain redefinition, for example, may be outcomes of similar entrepreneurial processes, firms
that regularly introduce new products or enter new markets (evidence of sustained regeneration) may never have an arena-
creating new product-market introduction (evidence of domain redefinition), and firms that achieve the latter may not be
frequent new product-market innovator” (Covin and Miles, 1999: p.55). Nevertheless, it is not clear whether firms could
effectively learn different CE processes. Disruptive technologies provide the definitive ground for an empirical test of CE
effectiveness. During technological disruptions, incumbents incapable of domain redefinition are rendered obsolete
resulting in declining firmperformance and the ultimate destruction of their investment in old technologies. Hence, other
things being equal -- such as business segment, location of main operations, and timing of entry -- incumbents that
implement CE of sustained regeneration/organizational rejuvenation during growth will have superior firmperformance,
and those that implement CE of domain redefinition will survive technological disruption. If there are learning across-CE
forms then the former leads to the latter. Three hypotheses emerge out of the preceding discussion.
Hypothesis 1: CE of sustained regeneration / organizational rejuvenation will yield superior performance during growth.
Hypothesis 2: Best performers during growth have a chance of surviving technological disruption by implementing CE of
domain redefinition.
Hypothesis 3: Other CE forms different from domain redefinition – e.g., that focuses on internal entrepreneurship --- will
lead to overshooting, diminishing performance, and ultimately the collapse of the firm.
The computer industry has been widely studied for its sequential waves of rapidly changing disruptive
technologies (Christensen, 1997; Eisenhardt, 1989; Eisenhardt & Tabrizi, 1995). Figure 2 (See Figure 2 below) shows a
stylized description of the sequence of disruptive technologies in the computer industry. The growth period between 1986
and 1995 is ideal to test CE effectiveness. First, it is necessary to avoid the period of turbulence that followed immediately
after the occurrence of a disruptive innovation. The effect of the entrepreneur dominates the period of turbulence
(Christensen, 1997), which obscures the other components of CE.
67
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
The introduction of the personal computer in 1976 by Apple Inc. initiated the personal computer technological
disruption. After a period of turbulence from1976 to 1985, large public companies took a dominant position in the
computer industry (Bergin, 2006). The period between 1985 and 1995 witnessed the rapid growth of the global computer
hardware industry main competitors. The year 1995 was marked by the explosive growth of the Internet (Bergin, 2006;
Newsweek, 1995). After 1996, most companies either initiated an aggressive turnaround or pressed forward on mergers
and acquisitions or divestments. The shake out period of 1997 to 2000 was exacerbated by the technology-bubble-bust
starting in March 2000 and lasting well into 2003 (Gregory, 2003). Technology companies were further shattered by ripple
effects of corporate scandals (Times, 2006). Thus, the ideal period in the computer industry to study growth due to CE is
between 1985 and 1995. This period excludes the turbulence added by entrance of new entrepreneurs to the personal
computer technology (1976-1984) and to the internet technology (1996-2000) and technology corporate scandals (2000-
2006).
Figure 2
Firm Performance Gains and Technological Disruptions in the Computer industry
METHODS
Data
Research Insight was the main data reference. Company names obtained fromResearch Insight were cross-
referenced with a variety of sources, including Marketguide, Hoovers, Bloomberg, and SEC filings. Ten years of data for
each firmwere computed to estimate a pooled cross-sectional time-series research design. Using techniques for analyzing
missing data (Little & Rubin, 1983), this study estimated R&D missing data points for five firms resulting in a final sample
of 330 firm-years of data. This study employs producer price indices for each segment fromthe Bureau of Labor Statistics
to adjust each firm’s variables to 1990 dollars. This adjustment enables cross-firmand cross-time comparisons.
Disruption I:
Personal
Computer
Introduction
1976-1984
Growth II:
1985-1995
Disruption III:
Computer/
Wireless
Telecom
2006/2007+
Disruption II:
Internet
Disruption
1996-2000
Tech Bubble
Bust
2000-2003
Time
Performance Gains
Computer
Introduction
And
Growth I:
1952-1976
ENTRY GROWTH EXIT
68
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
This study uses the method of structured content analysis to collect data to evaluate entry, growth, and exit
strategies (Jauch, Osborn, & Martin, 1980). This method has been applied before to analyses of the auto and airline
industries (Chen & Hambrick, 1995; Lieberman, Lau, Williams, & Williams, 1990). A three-pronged approach was used
to classify the strategies. First, a variety of sources were reviewed including company reports, firmproxy statements,
electronic SEC filings, Bloomberg merger and acquisitions reports and database, and press releases. Second, the main
author of this study categorized the information for each firmaccording to previously determined theory-driven categories.
A research assistant performed similar tasks independently. Third, we compared the two sets of categories. The conflicting
differences were resolved by mutually cross-referencing our categories with information fromOverview and History
Hoovers’ profiles. This method guarantees inter-rater reliability.
Sample
The sample comprises thirty-three firms fromthe computer hardware industry from1986 to 1995. These firms
compete in the U.S. personal computer, super computer, mini computer, and special computer segments of the industry.
The computer industry exemplified the quintessential technological and competitive environment that makes this a high-
velocity industry in which corporate entrepreneurship is critical (Bourgeois III & Eisenhardt, 1988; Eisenhardt & Tabrizi,
1995). The computer hardware industry is entrepreneurial per excellence, providing the ideal ground to test our model.
The composition of the sample was personal computers 15%, super computers 21%, mini computers 30%, special
computers 15% and conglomerates 18%. The sample was divided in two geographic regions, “West” and “Rest.” Two
sets of companies were not included in this analysis. The first group started before 1984 but was out of business before
1995 (61 firms). These firms did not survive the personal computer technological disruption. The second group of firms
started much later than 1984 and still had operations in 1995 (15 firms). These firms were in a much earlier stage of
growth than those started before 1984. The remaining sample of 33 firms was active between 1984 and 1995. Hence, the
sample comprises firms that faced both the first technological disruption (personal computers) and the second one
(internet).
Variables
The dependent variable is “performance gains,” which is the logarithmof yearly percent changes of operating
income per employee. Operating income is computed as sales at the end of the year minus the total costs-of-goods-sold,
and this value divided by the number of employees. These values were adjusted using price deflators for each business
segment with base US$1990 (Lieberman & Asaba, 1997; Lieberman & Demeester, 1999; Lieberman, Demeester, & Rivas,
2002).
The independent variables are “firm-fixed-effects,” “past performance,” “physical capital per employee gains,”
“R&D per employee gains,” and “cycle time reduction.” Fixed-effects are dummy variables for each firm(Greene, 1993;
Judge, 1985). The firmfixed-effects in a dynamic model are performance gains due to unique characteristics of the firm.
“Past Performance” is the natural logarithmof the value of two-year lagged operating income per employee. These values
were adjusted using US$ 1990 price deflators for each business segment. “Cycle time reduction” is the natural logarithm
of the percent rate of change between current value of “Cycle Time” and one-year lagged value. “Cycle Time” is measured
in weeks of work-in-process-inventory. These values were adjusted using US$ 1990 price deflators for each business
segment (Lieberman & Demeester, 1999; Lieberman, Demeester, & Rivas, 2002). Second, “R&D per employee gains” is
the natural logarithmof the percent rate of change between current values of “R&D per employee” and one-year lagged
value. “R&D per employee” is measured as the stock of R&D per employee. The stock of R&D is calculated using the
perpetual inventory method discounting the assets by 10% (Hulten & Wykoff, 1996). The resulting values were adjusted to
US$ 1990 using the GDP deflator.
“PP&E per employee gains” is the natural logarithm of the percent rate of change between current values
of “PP&E per employee” and one-year lagged value. “PP&E per employee” is the end of the year net property-plant-
and-equipment per employee. These values were adjusted using US$ 1990 price deflators for each business segment.
69
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
There are several categorical classification variables in this study. “Acquisitions per year” is a dummy
variable to measure acquisitions during growth. Firms were classified according to their mode of entry: 1) “Internal
Venture,” if entry is by developing a product in-house; 2) “Buy,” if entry is by fully acquiring another existing firm;
3) “Alliance,” if entry is established by strategic alliance or by investment in partial ownership of another existing
firm. Dummy variables were used to classify firms according to their business segment, “Super Computers,” “Mini
computers,” “Personal Computers,” “Special Purpose Computers.” These categories follow the practices of
Eisenhardt (1990), Eisenhardt and Tabrizi (1995). There is also included a dummy variable for Conglomerate
Growth, indicative of diversified firms. The exit strategies were 1) “Liquidate,” if firms filed for bankruptcy and sold
assets to pay debts; 2) “Divest,” if firms sold computer business-units to different firms and ceased to be active in the
computer industry; 3) “Sell,” if firm was fully acquired by another firm; 4) “Domain redefinition,” if firms divested
some assets but reinvested proceeds into new business-units that were related to the disruptive technology. Figure 3
(See Figure 3 below) shows the classification of CE forms used by computer firms during the personal computer and
the internet technology disruptions.
Figure 3
Corporate Entrepreneurship Forms Used by Computer Firms during Two Technological Disruptions
Focus of
Entrepreneurship
CE domain CE forms Entry Growth Exit
Sustained Regeneration Investment in
physical capital,
investment in R&D.
Past performance
Organizational
rejuvenation
Cycle Time
reduction
Strategic Renewal
Corporate
Rejuvenation
Domain Redefinition Domain redefinition
Direct-Internal Internal venture
Internal
Corporate
Venturing
Indirect-Internal
Sustained regeneration
Organizational
rejuvenation
Strategic
alliances
Strategic Renewal
Corporate
Rejuvenation
Domain Redefinition
Direct-External Entry through
acquisition,
Entry through
strategic alliance
Growth through
Acquisitions
Sell (Sell company
to another company),
Divest
(sell divisions to
different companies)
External
Corporate
Venturing
Indirect-External Liquidate
(intermediary
financial institutions
assisting bankruptcy)
The study includes control variables namely “Entry before Disruption I,” and location “West” and “Rest.” “Entry
before Disruption I” is a dummy variable for firms that went public before 1976 --Apple Computer Inc. had IPO in 1976.
The Location dummy variables classified the geographic region of main operations within the U.S.: namely “West” and
“Rest.” The dummy variable “West” indicates location of main operating activities in one of the western states of the U.S.
namely California, Washington, and Oregon, and includes Japanese companies. “Rest” measures location of main
operations in any of the continental states except the western states. Regression coefficient for regional CE effects were
obtained by multiplying location dummy variables by the CE variables (Judge, 1985).
70
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
Estimation Techniques
This study employed dynamic modeling with pooled cross-section and time-series data (Greene, 1993; Hsiao,
1986). Models were estimated with ordinary least squares OLS because the dependent variable is continuous. Regression
models were estimated with an overall intercept and one less dummy variable than the total number of firms. In doing so,
the dummy variable not considered in the new set of dummies becomes a benchmark. The intercept represents the
individual firm-effect of the benchmark company. Each dummy variable is an estimate of the difference between each
company and the benchmark company. Regression parameters’ standard errors were corrected by using the White-
consistent variance-covariance estimates to adjust significance levels (White, 1980). Multicollinearity was tested by using
the coefficient of Tolerance. The test of First and Second Moment Specification was used to test for generalized
heteroscedasticity, and the Durbin-Watson test checked for autocorrelation. No instrumental variables were used to
estimate the lagged variable “Past Performance.” This would entail a discussion of the effect of inertia in past performance,
which is beyond the scope of this paper.
RESULTS
Table 1 (See Table 1 below) shows the regression parameters for performance gains during Growth II (1985-
1995). The model corresponding to parameters for all firms (Nation) shows that the firmfixed-effects were significant
across the board (p
Description explain about a study of corporate entrepreneurship and firm performance in the computer industry.
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
A Study Of Corporate Entrepreneurship
And Firm Performance In The Computer
Industry During Two
Technological Disruptions
Ronald M. Rivas, (E-mail: [email protected]), Canisius College
ABSTRACT
This study presents a model of firm performance for Corporate Entrepreneurship (CE) Entry-Growth–
Exit strategies categorized by focus of entrepreneurship, CE domain, and CE form. Over a longitudinal
sample, this paper examines U.S. computer firm’s CE strategies and firm performance during two
technological disruptions namely the introduction of personal computers and subsequently the internet.
This paper shows that computer companies had superior performance by implementing CE strategies of
sustained regeneration and organizational rejuvenation during growth. In addition, best performing
companies survived the internet technological disruption by implementing the CE strategy of domain
redefinition. This study concludes with recommendations for future research.
INTRODUCTION
hat is the impact of corporate entrepreneurship on firmperformance during technological disruptions?
Both scholars and practitioners are interested in understanding corporate entrepreneurship (Chandler &
Lyon, 2001; Dess, Ireland, Zahra, Floyd, Janney, & Lane, 2003; Gregoire, Noel, Dery, & Bechard,
2006). Corporate entrepreneurship (CE) is the process whereby a firmcreates a new organization or
instigates renewal or innovation within an existing organization (Sharma & Chrisman, 1999). Corporate Entrepreneurship
has long been identified as a key means for promoting and sustaining competitive advantage (Covin & Miles, 1999). CE is
a means of accumulating, converting, and leveraging resources for competitive purpose (Floyd & Wooldridge, 1999) to
rejuvenate and redefine the firmand its markets (Covin & Miles, 1999). CE is also a means to venture into new markets by
creating new organizations fromwithin the firm, or acquiring them, or engaging in complex alliances for the purpose of
market entry (Miles & Covin, 2002).
j
Although significant theorizing has been proposed to clarify the domain of CE , there is nevertheless a need to
examine what is the relationship between different CE forms and firmperformance (Dess, Ireland, Zahra, Floyd, Janney, &
Lane, 2003). An area that has not received much attention in CE research is the effectiveness of CE under technological
disruptions. There is substantial research on CE under dynamic markets (Martin & Eisenhardt, 2004) in particular in the
computer industry (Christensen, 1997; Eisenhardt & Tabrizi, 1995). However, the understanding of what type of CE, if
any, might yield superior performance under technological disruption is limited. A technological disruption (Christensen,
1997) is a period of rapid industry transformation characterized by the introduction and adoption of new technologies
resulting in the rise of new entrants and the obsolescence of incumbents (Danneels, 2004). Hence, this study contributes to
the understanding of the impact of CE on firmperformance during technological disruptions.
The main goal of this study is to propose a model of CE and firmperformance under technological disruption. A
specific objective of this study is to examine and test what types of CE yield superior performance during entry, growth and
exit on an industry under technological disruption such as the computer industry.
65
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
This study begins by briefly reviewing the CE literature and proposing the model of CE and firmperformance
under technological disruption. It follows a brief description of the successive technological disruptions of the computer
industry from1976 to 2007. Then it proceeds with an application of cross-section time-series regression analysis with
pooled data. Tests were designed to assess the impact of CE on firmperformance during entry, growth, and exit on the
computer industry. It concludes with a discussion of results.
THEORETICAL BACKGROUND
Danneels (2004) defined disruptive technologies as “those technologies that render established technologies
obsolete and therefore destroy the value of the investments that incumbents have made in those technologies,” (p.248).
Christensen (1997) claims that overshooting is the main self-defeating mechanismof incumbents during technological
disruption. On the face of disruptive technologies, incumbents tend to overshoot, that is, they offer improved versions of
their products to the point where their new features are beyond the quality expectations of customers. On the other hand,
new entrants will offer products often below customer quality expectations but with features that satisfy customer needs
better than old products.
Figure 1
Corporate Entrepreneurship and Firm Performance during Technological Disruptions
Focus of
Entrepreneurship
CE domain Entry Growth Exit
Internal
Corporate
Rejuvenation
Innovators dilemma:
Incumbents
overshooting yields
diminishing
performance
Corporate rejuvenation
yields higher
performance gains
Innovators dilemma:
Best performers have a
chance of surviving
using Domain
redefinition. Other CE
forms lead to
overshooting and
diminishing
performance
Corporate Venturing Innovators dilemma:
Tech Disruption favors
performance gains
through Internal
venturing of New
entrants
Uncertainty of Tech
disruption is resolved
and internal venturing
yields diminishing
performance
Innovators dilemma:
Internal venturing
leading to domain
redefinition could
yield performance
gains
External
Corporate
Rejuvenation
Innovators dilemma:
Incumbents emphasize
internal improvement
over domain
redefinition, however,
yielding diminishing
performance
Uncertainty of Tech
disruption is resolved
and strategic
positioning yield
increasing
performance gains
Innovators dilemma:
Incumbents exit
seeking to sell venture
to "rightful' owner
among new entrants to
overcome diminishing
performance
Corporate Venturing Innovators dilemma:
Incumbents enter
through acquisition to
overcome diminishing
performance of old
business model
Acquisitions increase
market share and
stabilize supply chain
yielding performance
gains
Innovators dilemma:
Incumbents exit
cutting losses by
divesting or liquidating
66
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
Covin and Miles (1999) proposed a useful typology to classify CE centered on the concept of firmlevel
innovation with the objective of rejuvenating or purposely redefining organizations, markets or industries to create
competitive advantage. According to these authors, CE could be labeled as sustained regeneration, organizational
rejuvenation, strategic renewal and domain redefinition. Sustained regeneration is related to continuous improvement.
Organizational rejuvenation focuses on the organizational changes necessary to improve firmprocesses. Strategic renewal
focuses on how the firmcompetes, and the necessary changes to adapt to new realities of competition and competitor
moves. Domain redefinition focuses on changes a firmneeds to implement to enter into future new markets.
Miles and Covin (2002) clarified the domain of corporate venturing and proposed that corporate venturing
differentiates by focus of entrepreneurship and presence of investment intermediation. According to these authors,
corporate venturing could be internal or external to the corporation. In addition, it could be brought by direct investment in
the venture through the corporations operating strategic budget or by indirect investment in the venture using financial
intermediaries.
This study builds on the above mentioned typologies of CE (Covin & Miles, 1999; Miles & Covin, 2002) and on
the call for re-examination of firmperformance during disruptive technologies (Danneels, 2004). Figure 1 (See Figure 1
below) shows this study’s proposed typology of CE and firmperformance during technological disruptions. The model
shows that firmperformance is affected by the innovator’s dilemma during entry and exit. An incumbent often overshoots
by emphasizing on CE of “sustained regeneration” rather than switch to CE of “domain redefinition” or to corporate
venturing. However, Christensen (1997) claims that companies that have become leaders through the practice of CE of
“sustained regeneration” cannot enact a CE of “domain redefinition.” This notion is challenged by Danneels (2004), who
calls for renewed empirical theorizing and testing on this matter.
The ultimate answer to whether firms can use (or learn to use) new CE forms is empirical. Covin and Miles
(1999) suggest that there is no causal relationship between one CE formand another. These authors state that “while
sustained regeneration and domain redefinition, for example, may be outcomes of similar entrepreneurial processes, firms
that regularly introduce new products or enter new markets (evidence of sustained regeneration) may never have an arena-
creating new product-market introduction (evidence of domain redefinition), and firms that achieve the latter may not be
frequent new product-market innovator” (Covin and Miles, 1999: p.55). Nevertheless, it is not clear whether firms could
effectively learn different CE processes. Disruptive technologies provide the definitive ground for an empirical test of CE
effectiveness. During technological disruptions, incumbents incapable of domain redefinition are rendered obsolete
resulting in declining firmperformance and the ultimate destruction of their investment in old technologies. Hence, other
things being equal -- such as business segment, location of main operations, and timing of entry -- incumbents that
implement CE of sustained regeneration/organizational rejuvenation during growth will have superior firmperformance,
and those that implement CE of domain redefinition will survive technological disruption. If there are learning across-CE
forms then the former leads to the latter. Three hypotheses emerge out of the preceding discussion.
Hypothesis 1: CE of sustained regeneration / organizational rejuvenation will yield superior performance during growth.
Hypothesis 2: Best performers during growth have a chance of surviving technological disruption by implementing CE of
domain redefinition.
Hypothesis 3: Other CE forms different from domain redefinition – e.g., that focuses on internal entrepreneurship --- will
lead to overshooting, diminishing performance, and ultimately the collapse of the firm.
The computer industry has been widely studied for its sequential waves of rapidly changing disruptive
technologies (Christensen, 1997; Eisenhardt, 1989; Eisenhardt & Tabrizi, 1995). Figure 2 (See Figure 2 below) shows a
stylized description of the sequence of disruptive technologies in the computer industry. The growth period between 1986
and 1995 is ideal to test CE effectiveness. First, it is necessary to avoid the period of turbulence that followed immediately
after the occurrence of a disruptive innovation. The effect of the entrepreneur dominates the period of turbulence
(Christensen, 1997), which obscures the other components of CE.
67
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
The introduction of the personal computer in 1976 by Apple Inc. initiated the personal computer technological
disruption. After a period of turbulence from1976 to 1985, large public companies took a dominant position in the
computer industry (Bergin, 2006). The period between 1985 and 1995 witnessed the rapid growth of the global computer
hardware industry main competitors. The year 1995 was marked by the explosive growth of the Internet (Bergin, 2006;
Newsweek, 1995). After 1996, most companies either initiated an aggressive turnaround or pressed forward on mergers
and acquisitions or divestments. The shake out period of 1997 to 2000 was exacerbated by the technology-bubble-bust
starting in March 2000 and lasting well into 2003 (Gregory, 2003). Technology companies were further shattered by ripple
effects of corporate scandals (Times, 2006). Thus, the ideal period in the computer industry to study growth due to CE is
between 1985 and 1995. This period excludes the turbulence added by entrance of new entrepreneurs to the personal
computer technology (1976-1984) and to the internet technology (1996-2000) and technology corporate scandals (2000-
2006).
Figure 2
Firm Performance Gains and Technological Disruptions in the Computer industry
METHODS
Data
Research Insight was the main data reference. Company names obtained fromResearch Insight were cross-
referenced with a variety of sources, including Marketguide, Hoovers, Bloomberg, and SEC filings. Ten years of data for
each firmwere computed to estimate a pooled cross-sectional time-series research design. Using techniques for analyzing
missing data (Little & Rubin, 1983), this study estimated R&D missing data points for five firms resulting in a final sample
of 330 firm-years of data. This study employs producer price indices for each segment fromthe Bureau of Labor Statistics
to adjust each firm’s variables to 1990 dollars. This adjustment enables cross-firmand cross-time comparisons.
Disruption I:
Personal
Computer
Introduction
1976-1984
Growth II:
1985-1995
Disruption III:
Computer/
Wireless
Telecom
2006/2007+
Disruption II:
Internet
Disruption
1996-2000
Tech Bubble
Bust
2000-2003
Time
Performance Gains
Computer
Introduction
And
Growth I:
1952-1976
ENTRY GROWTH EXIT
68
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
This study uses the method of structured content analysis to collect data to evaluate entry, growth, and exit
strategies (Jauch, Osborn, & Martin, 1980). This method has been applied before to analyses of the auto and airline
industries (Chen & Hambrick, 1995; Lieberman, Lau, Williams, & Williams, 1990). A three-pronged approach was used
to classify the strategies. First, a variety of sources were reviewed including company reports, firmproxy statements,
electronic SEC filings, Bloomberg merger and acquisitions reports and database, and press releases. Second, the main
author of this study categorized the information for each firmaccording to previously determined theory-driven categories.
A research assistant performed similar tasks independently. Third, we compared the two sets of categories. The conflicting
differences were resolved by mutually cross-referencing our categories with information fromOverview and History
Hoovers’ profiles. This method guarantees inter-rater reliability.
Sample
The sample comprises thirty-three firms fromthe computer hardware industry from1986 to 1995. These firms
compete in the U.S. personal computer, super computer, mini computer, and special computer segments of the industry.
The computer industry exemplified the quintessential technological and competitive environment that makes this a high-
velocity industry in which corporate entrepreneurship is critical (Bourgeois III & Eisenhardt, 1988; Eisenhardt & Tabrizi,
1995). The computer hardware industry is entrepreneurial per excellence, providing the ideal ground to test our model.
The composition of the sample was personal computers 15%, super computers 21%, mini computers 30%, special
computers 15% and conglomerates 18%. The sample was divided in two geographic regions, “West” and “Rest.” Two
sets of companies were not included in this analysis. The first group started before 1984 but was out of business before
1995 (61 firms). These firms did not survive the personal computer technological disruption. The second group of firms
started much later than 1984 and still had operations in 1995 (15 firms). These firms were in a much earlier stage of
growth than those started before 1984. The remaining sample of 33 firms was active between 1984 and 1995. Hence, the
sample comprises firms that faced both the first technological disruption (personal computers) and the second one
(internet).
Variables
The dependent variable is “performance gains,” which is the logarithmof yearly percent changes of operating
income per employee. Operating income is computed as sales at the end of the year minus the total costs-of-goods-sold,
and this value divided by the number of employees. These values were adjusted using price deflators for each business
segment with base US$1990 (Lieberman & Asaba, 1997; Lieberman & Demeester, 1999; Lieberman, Demeester, & Rivas,
2002).
The independent variables are “firm-fixed-effects,” “past performance,” “physical capital per employee gains,”
“R&D per employee gains,” and “cycle time reduction.” Fixed-effects are dummy variables for each firm(Greene, 1993;
Judge, 1985). The firmfixed-effects in a dynamic model are performance gains due to unique characteristics of the firm.
“Past Performance” is the natural logarithmof the value of two-year lagged operating income per employee. These values
were adjusted using US$ 1990 price deflators for each business segment. “Cycle time reduction” is the natural logarithm
of the percent rate of change between current value of “Cycle Time” and one-year lagged value. “Cycle Time” is measured
in weeks of work-in-process-inventory. These values were adjusted using US$ 1990 price deflators for each business
segment (Lieberman & Demeester, 1999; Lieberman, Demeester, & Rivas, 2002). Second, “R&D per employee gains” is
the natural logarithmof the percent rate of change between current values of “R&D per employee” and one-year lagged
value. “R&D per employee” is measured as the stock of R&D per employee. The stock of R&D is calculated using the
perpetual inventory method discounting the assets by 10% (Hulten & Wykoff, 1996). The resulting values were adjusted to
US$ 1990 using the GDP deflator.
“PP&E per employee gains” is the natural logarithm of the percent rate of change between current values
of “PP&E per employee” and one-year lagged value. “PP&E per employee” is the end of the year net property-plant-
and-equipment per employee. These values were adjusted using US$ 1990 price deflators for each business segment.
69
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
There are several categorical classification variables in this study. “Acquisitions per year” is a dummy
variable to measure acquisitions during growth. Firms were classified according to their mode of entry: 1) “Internal
Venture,” if entry is by developing a product in-house; 2) “Buy,” if entry is by fully acquiring another existing firm;
3) “Alliance,” if entry is established by strategic alliance or by investment in partial ownership of another existing
firm. Dummy variables were used to classify firms according to their business segment, “Super Computers,” “Mini
computers,” “Personal Computers,” “Special Purpose Computers.” These categories follow the practices of
Eisenhardt (1990), Eisenhardt and Tabrizi (1995). There is also included a dummy variable for Conglomerate
Growth, indicative of diversified firms. The exit strategies were 1) “Liquidate,” if firms filed for bankruptcy and sold
assets to pay debts; 2) “Divest,” if firms sold computer business-units to different firms and ceased to be active in the
computer industry; 3) “Sell,” if firm was fully acquired by another firm; 4) “Domain redefinition,” if firms divested
some assets but reinvested proceeds into new business-units that were related to the disruptive technology. Figure 3
(See Figure 3 below) shows the classification of CE forms used by computer firms during the personal computer and
the internet technology disruptions.
Figure 3
Corporate Entrepreneurship Forms Used by Computer Firms during Two Technological Disruptions
Focus of
Entrepreneurship
CE domain CE forms Entry Growth Exit
Sustained Regeneration Investment in
physical capital,
investment in R&D.
Past performance
Organizational
rejuvenation
Cycle Time
reduction
Strategic Renewal
Corporate
Rejuvenation
Domain Redefinition Domain redefinition
Direct-Internal Internal venture
Internal
Corporate
Venturing
Indirect-Internal
Sustained regeneration
Organizational
rejuvenation
Strategic
alliances
Strategic Renewal
Corporate
Rejuvenation
Domain Redefinition
Direct-External Entry through
acquisition,
Entry through
strategic alliance
Growth through
Acquisitions
Sell (Sell company
to another company),
Divest
(sell divisions to
different companies)
External
Corporate
Venturing
Indirect-External Liquidate
(intermediary
financial institutions
assisting bankruptcy)
The study includes control variables namely “Entry before Disruption I,” and location “West” and “Rest.” “Entry
before Disruption I” is a dummy variable for firms that went public before 1976 --Apple Computer Inc. had IPO in 1976.
The Location dummy variables classified the geographic region of main operations within the U.S.: namely “West” and
“Rest.” The dummy variable “West” indicates location of main operating activities in one of the western states of the U.S.
namely California, Washington, and Oregon, and includes Japanese companies. “Rest” measures location of main
operations in any of the continental states except the western states. Regression coefficient for regional CE effects were
obtained by multiplying location dummy variables by the CE variables (Judge, 1985).
70
Review of Business Information Systems – Third Quarter 2007 Volume 11, Number 3
Estimation Techniques
This study employed dynamic modeling with pooled cross-section and time-series data (Greene, 1993; Hsiao,
1986). Models were estimated with ordinary least squares OLS because the dependent variable is continuous. Regression
models were estimated with an overall intercept and one less dummy variable than the total number of firms. In doing so,
the dummy variable not considered in the new set of dummies becomes a benchmark. The intercept represents the
individual firm-effect of the benchmark company. Each dummy variable is an estimate of the difference between each
company and the benchmark company. Regression parameters’ standard errors were corrected by using the White-
consistent variance-covariance estimates to adjust significance levels (White, 1980). Multicollinearity was tested by using
the coefficient of Tolerance. The test of First and Second Moment Specification was used to test for generalized
heteroscedasticity, and the Durbin-Watson test checked for autocorrelation. No instrumental variables were used to
estimate the lagged variable “Past Performance.” This would entail a discussion of the effect of inertia in past performance,
which is beyond the scope of this paper.
RESULTS
Table 1 (See Table 1 below) shows the regression parameters for performance gains during Growth II (1985-
1995). The model corresponding to parameters for all firms (Nation) shows that the firmfixed-effects were significant
across the board (p