and Internationalization with Strategic Change
Parshotam Dass
UNIVERSITY OFARKANSASWhat are the relationships among firm size, initial product diversity, interna- filled with uncertainty and ambiguity, company executives must make their strategic decisions on the basis of personal
experi-tional diversification, and abrupt change in product diversity? Data analysis
using a sample of 555 firms in diverse industries from COMPUSTAT for ence, stereotypes, and intuition, but with little help from theo-ries of change.
the period 1978–90 characterizes these relationships in terms of fluidity
(versus rigidity) effects. Specifically, we find that initial product diversity, This state of research on strategic change is reflected when
international diversification, and firm size have significant linear, quad- Rajagopalan and Spreitzer (1997) observe that “organization
ratic, and cubic effects, respectively. Significant interactive relationships size has been found to have positive (e.g., Zajac and Kraatz,
among the variables reveal that increasing firm size at the extreme lower 1993) and negative (e.g., Fombrun and Ginsberg, 1990) effects
as well as higher levels enables internationally diversifying firms to change on the likelihood of strategic change. The theoretical quandary
more (or less) abruptly when they have higher (or lower) initial product of whether firm size is a source of inertia or a source of resources
diversity. These effects of firm size are less pronounced at the medium for strategic flexibility remains unanswered” (Rajagopalan and
levels of firm size or for firms with a low level of international diversifi- Spreitzer, 1997). Similarly, initial level of product diversity
cation. These strategic choices may be explained by firm capabilities and international diversification can provide resources and
developed through experience and learning as a result of high initial capabilities for abrupt change in product diversification. Alter-product diversification and subsequent international diversification and natively, they can increase bureaucratic costs, information-reinforced by varying levels of resources and inertia associated with firm processing requirements, and diseconomies resulting in strate-size. J BUSN RES2000. 48.135–146. 2000 Elsevier Science Inc. gic inertia (Hannan and Freeman, 1977). Hill and Hansen All rights reserved. (1991) reported negative relationships of initial diversification
with product diversification. Researchers have generally ne-glected the association of international diversification to prod-uct diversification (Hitt, Hoskisson, and Ireland, 1994; Hitt,
C
hange and diversification have been dominant themes Hoskisson, and Kim, 1997). This association warrants atten-in the busatten-iness world for decades and it is likely to be tion because with increased numbers of companies going so into the next century. Most firms in their lifetime global, it may be essential for them to become involved in engage in diversification, downscoping, or other changes. Such dramatic product diversification to be competitive at home changes may be essential for their very survival and success and abroad (McNamee, Magnusson, Fins, and King, 1990). because they have implications for a firm’s assumed level of In the field of organization and management theory, change risk, access to opportunities, resources and capabilities, and theorists have addressed the question: How do (and should) efficiency and effectiveness. However, understanding of how organizations change? This question has evoked two conflict-organizations change under different conditions remains at a ing responses. Some researchers have noted the benefits of primitive level because research findings on strategic change abrupt change (e.g., Miller and Friesen, 1982, 1984), whereas have not evolved into a coherent body of knowledge (Ginsberg, others have advocated the virtues of gradual change (e.g., 1988; Mintzberg and Westley, 1992; Rajagopalan and Spreitzer, Quinn, 1980). Lamont, Williams, and Hoffman (1994) found 1997). Research on diversification also has often beencharacter-equivocal evidence and reported that gradual transformations ized as mixed (Dess, Gupta, Hennart, and Hill, 1995;
Ramanu-result in greater performance declines during the transition jam and Varadarajan, 1989; Ramaswamy, 1992). In a world
but they recover more quickly than dramatic changes. The objective of this research is to test whether firm size, Address correspondence to Parshotam Dass, Sam M. Walton College of Busi- initial product diversity, and international diversification have ness Administration, University of Arkansas, Fayetteville, AR 72701. E-mail:
[email protected] enabling or disenabling effects on strategic change. I examine
Journal of Business Research 48, 135–146 (2000)
2000 Elsevier Science Inc. All rights reserved. ISSN 0148-2963/00/$–see front matter
them with respect to a frequently studied but still not clearly ronmental conditions described by the ecological (selection) as well as institutional theories of the organization (DiMaggio understood phenomenon: product diversification. In their
and Powell, 1983; Hannan and Freeman, 1984). Enduring widely cited review article, Ramanujam and Varadarajan (1989)
organizational myths, routines, ideologies, heuristics, political defined diversification as change in a firm’s diversity. Change in
coalitions’ vested interests in continuity, resistance to change, product diversity is distinguished from international geographic
and social construction of realities contribute to inertia in diversification in the sense that the former refers to resource
organizations. Pressures for change or adaptation occur in deployment across industries, whereas the latter deals with the
terms of strategic demands, stakeholder and institutional pres-same across countries (Grant, Jammine, and Thomas, 1988).
sures, and the cognitive stressors of net present value and declining performance that result from changing internal and
Theoretical Background
external environmental conditions (Friesen and Miller, 1986;Huff et al., 1992; Kossek, Dass and DeMarr, 1994; Zajac and Two contrasting models of strategic change currently
domi-Kraatz, 1993). These stressors may favor abrupt change due nate the literature. According to Meyer and colleagues (1990),
to the problems or excesses of inertia and the disruptions Almost everyone who spends much time thinking about caused by new realities or major decisions. They are likely change processes seems to conclude that the world changes to increase the mismatch of sub-systems internally and/or in two fundamentally different modes. . . . Continuous, or between the system and its environments externally, thereby first-order change, occurs within a stable system that itself reducing the effectiveness of current strategy.
remains unchanged. . . . Discontinuous, or second-order Strategic managers continually struggle to minimize costs change transforms fundamental properties or states of the of being mismatched within the organization and with the
system. external environment, and those of changing to avoid the
mismatch (Ginsberg, 1988; Huff et al., 1992; Miller and Frie-Miller and Friesen (1982, 1984) examine and find support
sen, 1984). Hence, internal and external forces indicate the for distinguishing between these two models of momentum
appropriateness of current strategy, the need for a new strat-(evolutionary) and revolutionary change in organizations. Meyer
egy, and the organizational capability to implement change. and colleagues (1990) extend this perspective by applying it
Whether organizations follow a gradual or abrupt model of to both firm and industry levels. Huff, Huff, and Thomas
change, thus, depends on the nature and balance of these (1992) continue its support, citing empirical as well as
theoret-forces which, in turn, hinges on firm size, level of product ical justification.
and international diversity, among other organizational char-Several scholars from Lindblom (1959) and Cyert and
acteristics. These insights and earlier research findings can be March (1963) to Starbuck, Greve, and Hedberg (1978) have
used in the study of product diversification by examining it favored the evolutionary or gradual mode of change. They
in large sample multivariate studies in order to test the associa-suggested that organizations should nurture small disruptions
tion of firm size, initial product diversity, and international and gradual re-orientations due to sequential task
interdepen-diversification to abrupt change in product diversity. dence as well as political, economic, cognitive, and other
advantages including risk reduction and learning. More
re-Hypotheses
cently, researchers such as Quinn (1980) and Johnson (1988)
have found a widespread role for gradual changes in organiza- The dependent variable in the study is the degree to which tions. Theorists in the organizational development tradition organizations diversify in a gradual or in an abrupt manner. (Beckhard and Harris, 1987; Beer, Eisenhardt, and Spector, Change in product diversity may come in small increments 1990) have also proposed carefully planned and orchestrated or in big leaps. A company that changes its product line in change as the preferred mode of transforming organizations. small increments every year will be engaging in gradual In contrast, the quantum change theorists view revolutionary change. A company that changes its products in fits and starts or abrupt change as a way to break organizational inertia (e.g., major overhauls every four or five years) will be engaging and allow organizations to quickly gain alignments among in abrupt change (Dass, 1992; Dass and Moch, 1997). In the different sub-systems (Miller and Friesen, 1982, 1984). following section, I relate firm size, initial product diversity, More recently, Ginsberg (1988), Huff and colleagues and international diversification to present the “fluidity” and (1992), and Zajac and Kraatz (1993) have proposed models, “rigidity” theses for each variable for testing in the study specifying conditions that propel organizations to adopt a (Haveman, 1993; Staw, Sandelands, and Dutton, 1981). particular strategy of change. They suggested a diametric
Firm Size
model of pressures for change and status quo. The diametric
envi-Figure 1. Firm size and strategic change.
tion take risks, withstand setbacks, and initiate changes (Kim- size and diversification to follow a curvilinear path. The newly berly and Evanisko, 1981). Increased size provides more mar- emerging inertial theories also suggest that inertial tendencies ket power to an organization to deal with its stakeholders in change in a complex non-monotonic relationship during the the technical as well as institutional environment (Pfeffer and life cycle of an organization (Baum, 1990; Dass and Moch, Salancik, 1978). These arguments favor a fluidity of size expla- 1998; Fichman and Levinthal, 1991). The view is consistent nation and are consistent with the adaptation theories of strate- with Haveman’s (1993) findings of fluidity as well as rigidity gic change (Baum, 1990; Haveman, 1993; Singh, House, and effects of size that were curvilinear in nature. Therefore, ac-Tucker, 1986; Tushman and Romanelli, 1985). cording to the adaptation perspective, positive curvilinear ef-On the other hand, increased size may exert an inhibiting fects of size will lead to an S-shaped curve between firm size influence on the organization’s ability to change. Much research and diversification. However, the selection perspective that has shown that larger organizations tend to be more formalized focuses on negative curvilinear effects of size will lead to an and bureaucratic (Blau and Schoenherr, 1971), and formal rules inverse-S-shaped (i.e., logistic) curve between firm size and and regulations can stifle innovation. Complicated structures diversification. Both these curves may have minimum efficient associated with organizational size can constrain change (Huff scale or the optimal level of diversification as the inflection et al., 1992). Organizations seek reliability and predictability point, as suggested in product and international diversification achieved by bureaucratic procedures or organizational routines by Grant and colleagues (1988) and Geringer, Beamish, and that increase inertia (Nelson and Winter, 1982; Singh et al., daCosta (1989), respectively. These two forms of the relation-1986). Large organizations may also have diverse groups of ship between size and change are portrayed in Figure 1. Similar entrenched stakeholders and find it more difficult to build con- to Haveman (1993), I will competitively test both these effects. sensus for change (Freeman, 1984). Buffering from external influ- Accordingly, it is hypothesized that:
ences may increase rigidity effects in large organizations as well.
H1a: Firm size and abrupt change in product diversity Size, therefore, can be a burden as well as a boon. If
will have a positive relationship consistent with the economies and diseconomies of scale, scope, and
specializa-fluidity of size thesis. tion, and bureaucratic costs are a linear function of size, the
H1b: Firm size and abrupt change in product diversity relationship between size and diversification will be linear. If
will have a negative relationship consistent with the the economies/diseconomies and costs increase and decrease
ment, and recoup their investments faster due to the
econo-Initial Product Diversity
mies of scale, scope, and learning resulting from global de-Unlike size, there is little research on the role of initial diversity
mand for products and services rendered by related and to subsequent diversification. However, it is plausible to
as-unrelated diversified firms (Kotabe, 1990). These advantages sume that similar to firm size, a high level of initial diversity
may provide further incentives to global companies by low-may have positive as well as negative influences on subsequent
ering their risk (Kim, Hwang, and Burgers, 1989). Lowering change in product diversity. On the positive side, high initial
risk may enhance the organization’s capabilities for innovation product diversity represents economies of scope, extensive
and new product development (Hitt et al., 1994, 1997). opportunities for diversification, diverse learning avenues, and
Global companies can use their existing competencies and access to varied dominant logics, and other resources and
build new ones for new product development by acquiring a capabilities for undertaking abrupt diversification. On the
neg-higher level of less imitable resources and capabilities to take ative side, there are diminishing returns to diversification.
advantage of the global product life cycle (Hamel, 1991). Higher diversification may spread a firm too thin in its markets
Other contributions of international diversification may come resulting in reduced incentives to diversify further.
Disecono-in stabilizDisecono-ing returns or Disecono-in helpDisecono-ing to take advantage of Disecono- infor-mies of scale, increased need for coordination, information
mation asymmetries (Buhner, 1987). Though firms may have overload, and increased information-processing requirements
limited funds and sometimes may have to choose between may increase bureaucratic costs that can wipe out positive
product and international diversification (Sambharya, 1995), effects and bring in inertial tendencies in organizations (Hill
international diversification is likely to reinforce the need and and Hansen, 1991).
capabilities for product diversification over time. All these It is expected that the effect of initial product diversity may
factors are likely to lead to a positive relationship between vary, depending on its own level. For example, at some levels
the extent of abrupt change in product and international diver-positive effects of initial diversity may outweigh negative
ef-sification. fects, whereas at other levels bureaucratic costs may outstrip
The opposite scenario may be less compelling but is plausi-learning effects. Thus, I will competitively test its positive and
ble. Firms have limited financial and other resources to deal negative curvilinear effects.
with product and market diversities in organizations. At any H2a: Initial product diversity will have a positive relation- time, projects in a firm compete for attention and resources. ship with subsequent abrupt change in product di- One type of diversification may be substituted with the other. versity consistent with the fluidity thesis. Chandler’s (1962) study suggested that historically firms have H2b: Initial product diversity will have a negative relation- become involved in product diversification before interna-ship with subsequent abrupt change in product di- tional diversification. However, currently, firms may under-versity consistent with the rigidity thesis. take international diversification to take advantage of their specialized distinctive competencies before embarking on
International Diversification
product diversification. This implies a negative relationship between international diversification and abrupt change in Numerous researchers have examined the role of internationalproduct diversity. diversification in companies (e.g., Buhner, 1987; Geringer et al.,
The most likely scenario is that international diversification 1989; Grant et al., 1988; Ramaswamy, 1992). However,
inter-and product diversification may compete at some levels, national diversification is mostly investigated in an isolated
whereas both may reinforce each other at other levels. It way in simplistic models, relating it to the dependent variable
implies that they may have a curvilinear relationship. Hence, of performance (Hoskisson and Hitt, 1990). Researchers have
I will competitively test for positive and negative curvilinear argued in favor of examining the interrelationship of
interna-effects, as follows: tional and product diversification (e.g., Hitt et al., 1994).
The relationship between international and product diver- H3a: International diversification will have a positive rela-sities can be positive or negative. Let us first consider the
tionship with subsequent abrupt change in product scenario for a positive relationship. Companies going to global
diversity consistent with the fluidity thesis. markets are likely to face a broader diversity of customer
H3b: International diversification will have a negative rela-needs and preferences. Customer expectations from global
tionship with subsequent abrupt change in product companies in the domestic and international markets are
antic-diversity consistent with the rigidity thesis. ipated to be higher in terms of new product development.
Companies in global markets need to invest in new products
Interactions
to be competitive at home and abroad (Prahalad, 1990). Global
companies through strategic alliances or direct investment can Researchers have come across nonsignificant effects of some of the antecedents involved in this study. For instance, Kelly take advantage of the human, as well as material, resources
develop-the models conceptualizing develop-the dynamics of forces resisting tive relationship between initial product diversity and abrupt change in product diversity.
and favoring change—stressors and inertia (Ginsberg, 1988; Huff et al., 1992; Zajac and Kraatz, 1993), the relationships
FIRM SIZE3INITIAL PRODUCT DIVERSITY3INTERNATIONAL
DI-of firm size, initial product diversity, and international
diversi-VERSIFICATION. Over and above the two-way interactions, it fication may not only depend on their own levels (the
curvilin-is plausible to assume that these three variables have their ear effects already hypothesized above), they may also depend
synergistic effects on the magnitude of change because they on the level of one another. Therefore, I hypothesize the
all can provide incentives as well as disincentives for abrupt following interactions.
change in product diversity. For example, a large highly diver-sified firm such as General Electric, NEC, or Mitsubishi with
FIRM SIZE3INITIAL PRODUCT DIVERSITY. If firm size and initial
high international diversification can make a subsequent product diversity each can have fluidity or rigidity effects,
abrupt change in product diversity because of its level of their combined effects are expected to be more than the sum
capital, expertise, and access to diverse products and markets of their parts because they work in a synergistic way to provide
in numerous countries of the world. Alternately, a large highly economies of scale and scope, resources and opportunities,
diversified company such as ITT with high international diver-and capabilities to take advantage of them. Alternately, each
sification can suffer from communication and coordination may reinforce the inertial tendencies of the other. Therefore,
problems that may hinder their subsequent abrupt product I will test their moderating effects, as follows:
diversification. Both effects are theoretically credible; hence, H4a: Firm size will moderate the positive relationship be- are best left to the empirical analysis. Thus,
tween initial product diversity and abrupt change in
H7a: There will be a positive three-way interaction among product diversity.
firm size, initial product diversity, and international H4b: Firm size will moderate the negative relationship be- diversification with respect to abrupt change in
prod-tween initial product diversity and abrupt change in uct diversity. product diversity.
H7b: There will be a negative three-way interaction among firm size, initial product diversity, and international
FIRM SIZE 3 INTERNATIONAL DIVERSIFICATION. Apart from
diversification with respect to abrupt change in prod-their main effects, it is expected that the effect of change in
uct diversity. international diversity will depend on firm size because of the
joint effects of the economies of firm size and the economies Figure 2 depicts the relationships hypothesized in the study. of international diversification. The reverse will be true to
the diseconomies of size and bureaucratic cost involved in
Methods
international diversification. Hence,
Sample
H5a: Firm size will moderate the positive relationship
be-tween international diversification and abrupt change Data used to assess the hypotheses were retrieved from the COMPUSTAT II database. The sample constituted 555 compa-in product diversity.
nies registered on U.S. stock exchanges and reporting com-H5b: Firm size will moderate the negative relationship
be-plete data on the variables of interest. Represented industries tween international diversification and abrupt change
range from mining to manufacturing to service. Their Standard in product diversity.
Industrial Classification (SIC) codes were from 10 to 87, with the inclusion of 99 (which represents the conglomerates). The
INITIAL PRODUCT DIVERSITY3INTERNATIONAL DIVERSIFICATION.
sample included a wide range of small, medium, and large The level of international diversification may moderate the
organizations. Size of the organizations, measured in terms effect of initial product diversity because a firm with diverse
of number of employees, varied from 11 to 130,254, with an products can have much better access to varied markets. A
average of 4839. The time span for variables in the study was firm with fewer products will have only limited markets to
from 1978–90, similar to Grant and colleagues (1988). The choose from. Alternately, a product diversified firm already
only exception was the international diversification variable, has the complexity and coordination requirements and may
which was measured from 1984–90 because of unavailability have to come up with administrative or communication
inno-of its data from 1978–84. vations before it can embark on international diversification.
Hence,
Measures
H6a: International diversification will moderate the
posi-DEPENDENT VARIABLE. Product diversity was measured using tive relationship between initial product diversity and entropy measures (Palepu, 1985). Entropy measures, devel-abrupt change in product diversity. oped in industrial organization economics (Jacquemin and
nega-Figure 2. Hypothesized relationships (Industry performance, firm risk, and slack were used as controls. H1to H3: main effects; H4to H6: two-way interactions; H7: three-way interaction.)
ment literature for measuring product diversity (Palepu, 1985; measure is used by Finkelstein and Hambrick (1990) in study-ing firm-level change. Our measure is strongly correlated (r5
Wiersema and Bantel, 1992, 1993). The construct validity of
0.73) with another measure of abrupt change in product diver-these measures has been established by recent studies
(Hoskis-sity at the firm level. The other measure is computed as son, Hitt, Johnson, and Moesel, 1993). The product diversity
standard error of the regression coefficient of product diversifi-of the companies in the sample averaged over the period
cation from 1978 to 1990 when diversification is regressed 1978–90 varied from 0 to 1.37, with a mean of 0.19. A
on time (Dass and Moch, 1997; Keats and Hitt, 1988). It is product diversity index of 0 represented an undiversified
com-the firm-level equivalent of Dess and Beard’s (1984) industry pany, that is, a company that has all of its sales in one industry
dynamism. segment.
The constructs of change in product diversity and change EXOGENOUS VARIABLES. Firm size was measured by calculat-in calculat-international diversity have been operationalized uscalculat-ing en- ing the log of the total number of employees in the company. tropy or Herfindahl measures previously (e.g., Grant et al., The logarithmic transformation was used as it helped reduce 1988; Hill and Hansen, 1991; Wiersema and Bantel, 1992, the degree of skewness in the measure of size (Blau and 1993). The degree of abrupt change in total product diversity Schoenherr, 1971; Kimberbly, 1976). I used the total number was measured by calculating the total variance of annual of employees instead of assets or sales, which were employed changes in product diversity over the time period of the study, for computing performance (adjusted return on assets) and following Miller and Friesen (1982) who used standard devia- entropy measures of diversity (that consider proportion of tion as a measure of gradual–abrupt change. Note that the sales in different industry segments) to avoid overlap among gradual–abrupt change is defined in terms of fluctuations in the exogenous variables and between the exogenous and de-annual changes of product and international diversity. There- pendent variables. Initial product diversity was the entropy fore, what is important is not the absolute amount of annual measure in 1978. International diversity was also computed change, but the fluctuations in annual changes. Higher fluctua- using entropy measures (Kim et al., 1989; Vachani, 1991; tions lead to higher standard deviation or variance, and lower Hitt et al., 1997). Its average for the period 1984–90 varied fluctuations lead to lower standard deviation or variance from 0 to 1.42, with a mean of 0.12. International diversifica-(Miller and Friesen, 1982, 1984). Mathematically, an organi- tion was measured as change of international diversity from zation making no changes or identical annual changes 1984–90 (Grant et al., 1988).
throughout the period would yield zero variance, representing
Controls
extremely gradual change. Another company that changes in
Table 1. Means, Standard Deviations, and Correlationsa
Variables Means SD 1 2 3 4 5 6
1. Change in product diversity 20.05 0.94
2. Industry performance 0.09 0.94 20.01
3. Risk 20.04 0.91 20.03 0.03
4. Slack 20.02 0.85 20.05 0.02 20.02
5. Firm size 20.02 0.97 0.07 20.06 20.20 20.10
6. Initial product diversity 20.23 0.87 0.31 20.01 20.07 20.07 0.31
7. International diversification 20.02 1.01 0.04 0.03 20.06 20.04 0.01 20.07
an
5555.
Correlations equal to or greater than 0.07 are significant at the 0.10 level; those5or.0.08 are significant at the 0.05 level; those5or.0.11 are significant at the 0.01 level, and those5or.0.14 are significant at the 0.001 level (two-tail tests).
potential confounding variables. Existing theory and research
Analysis
have shown that industry performance is likely to influence The dependent variable of the degree of abrupt change was strategic change in organizations (Dess and Beard, 1984; Gins- continuous and was subject to hierarchical regression analyses berg, 1988; Miller and Friesen, 1984). This finding is consis- using ordinary least squares. Since the sample contained com-tent with the industrial organization economics paradigm panies from different industries, the company-level indepen-(Christensen and Montgomery, 1981; Rumelt, 1974; Stimpert dent and dependent variables were standardized with respect and Duhaime, 1997) as well as with the resource dependence to their two-digit SIC code level industry means (computed theory, according to which an organization depends on its from all companies of the particular industry in the sample) environment for its resources (Pfeffer and Salancik, 1978). to render the measures comparable across industries. Industry income was computed as the average return on assets
in the industry, similar to the firm performance measure
men-Results
tioned above.
Organizational slack may be related to size and perfor- Table 1 presents the means, standard deviations, and correla-mance but there is support in the literature for their separate tion matrix. As can be noted from the means and standard influences (Hambrick and D’Aveni, 1988; Singh, 1986).
Simi-deviations, the independent variables were standardized with lar to size, slack can have two kinds of relationships with mean
5 0 and standard deviation 5 1. Inspection of the organizational changes (Bourgeois, 1981; Sharfman, Wolf, correlations among the variables reveals that most variables Chase, and Tansik, 1988). Although organizational slack have weak to moderate correlations among themselves. The makes it easier to carry out change, it lowers the motivation quadratic and cubic terms were computed after standardizing to undertake change (Ginsberg and Buchholtz, 1990). Organi- the linear term to help reduce the variance inflation to accept-zational slack and change may have a curvilinear relationship. able levels.
Therefore, linear as well as quadratic terms of slack were The results of the hierarchical regression analyses are pre-used to partial out any variance to improve confidence in the sented in Table 2. The first step was to enter the control relationships of the variables of main interest in the study. variables. Hypothesized linear, quadratic, and cubic terms of Slack was measured by the ratio of equity to debt that repre- the main effects were added to the control variables in steps sented unabsorbed slack (excess uncommitted resources) 2 to 5. Steps 6 to 10 involved addition of two-way and three-reflecting resource availability over the long term that is impor- way linear, quadratic, and cubic interaction terms in the usual tant for making changes in the strategy of the organization order. The first column depicts standardized regression
coeffi-(Bourgeois, 1981; Singh, 1986). cients at the point of entry in the hierarchical regression. The
Risk faced by firms also may influence the patterns of second column reveals the same in the full multiple regression change in diversity (Hill and Hansen, 1991; Lubatkin and model. Cohen and Cohen (1983) recommend measuring sig-Chatterjee, 1994; Singh, 1986). Since abrupt change may nificance at the point of entry and using coefficients in the involve higher risk-taking, the level of risk faced by the firms full model for plotting the curves.
was deemed appropriate as a control. Its measure was variance In the study, all regression equations of study variables of company returns in the past: this accounting-based measure (see Table 2) generated statistically significant F-ratios (p,
Table 2. Results of Regression Analysisa
bat
Exogenous Variables Entry bin Full Model R2 Adj. R2 F df DR2 DF
Step 1: Control variables
Industry performance 0.00 20.01
Firm risk 20.03 20.02
Firm slack 20.11† 20.08
Firm slack 0.08 0.05 0.01 0.00 0.82 4,550 0.01 0.82
Step 2
Firm size 20.03 20.23***
Initial product diversity 0.32*** 0.22***
International diversification 0.06 0.02 0.10 0.09 8.79*** 7,547 0.10 19.29***
Step 3
Firm size2 0.08†
20.02
International diversification2 0.11** 0.18*** 0.12 0.11 8.22*** 9,545 0.02 5.72**
Step 4
Firm size3 0.22*** 0.26** 0.14 0.12 8.79*** 10,544 0.02 12.33***
Step 5
Firm size3Initial product diversity 0.12* 20.15*
Firm size3International diversification 0.13** 20.02 0.16 0.14 8.61*** 12,542 0.02 6.78***
Step 6
Firm size23Initial product diversity 0.25** 0.06
Firm size23International diversification 0.13* 20.10 0.18 0.16 8.42*** 14,540 0.02 6.29**
Step 7
Firm size33Initial product diversity 0.33** 0.48***
Firm size33International diversification 0.40*** 0.09 0.21 0.19 8.95*** 15,539 0.03 10.55***
Step 8
Firm size3Initial product diversity3 0.15* 20.21† 0.22 0.19 8.81*** 16,538 0.01 5.49*
International diversification Step 9
Firm size2
3Initial product diversity3 0.54*** 0.09 0.24 0.21 9.20*** 17,537 0.02 12.48***
International diversification Step 10
Firm size33Initial product diversity3 0.51*** 0.51*** 0.25 0.22 9.14*** 18,536 0.01 6.42**
International diversification
an
5555.
Standardized regression coefficients are reported andp-values are for two-tail tests. Variables in the hierarchical regression were entered in the following sequence: Control variables, linear main effects, quadratic main effects, cubic main effects, two-way interactions (product of linear, quadratic, and cubic terms), and three-way interactions (product of linear, quadratic, and cubic terms), as recommended (Cohen and Cohen, 1983).
b†p
,0.10; *p,0.05; **p,0.01; ***p,0.001.
linear relationship. There was marginal support for a quadratic with initial product diversity moderating the effects of interna-tional diversification with the linear, quadratic, and cubic one, whereas the cubic relationship was strongly supported.
effects of firm size, as depicted in Figure 4. These curves The curve shows a largely positive effect (Hypothesis 1a;
Fig-reveal that increasing firm size inclines an internationally di-ure 3). Data supported linear positive effects of initial product
versified firm to make more abrupt change in product diversity diversity (Hypothesis 2a) but not its quadratic effects (not
when it has high initial product diversity. However, increasing shown in Table 2). Hypothesis 3 examined the relationship
firm size influences an internationally diversified firm to make between international diversification and abrupt change in
more gradual change in product diversity when it has low product diversity. There was no support for the linear
relation-initial product diversity. These effects are more pronounced ship. However, the quadratic relationship was supported. Its
at extreme levels of firm size than at the average firm size direction was positive, resulting in a U-shaped relationship.
in an industry. The effects of firm size in firms with low I hypothesized two-way and three-way interactions of firm
internationalization are far less pronounced. size, initial product diversity, and international diversification.
Interactive relationships of firm size with initial product
diver-sity and with international diversification were supported,
Discussion
with linear, quadratic, and cubic effects. I did not find anyFigure 3. Results of the relationship of firm size and abrupt change [——Equation Y 5 20.21 * size10.09 * size squared10.22 * size cubed (Multiple regression model at step 4)].
ables. Data on 555 companies retrieved from the industry, Initial product diversity and international diversification had positive and U-shaped relationships with abrupt change geographic, and annual data tapes of COMPUSTAT II from
1978–90 provided support for several main and interactive in product diversity, respectively. Firm size was found to have a significant cubic effect. The slope of the curve was positive relationships in the study.
at the lower and higher levels, whereas it was almost flat for the R2values are moderate. Another study of the effects of a similar set of variables including performance, size, and indus-the medium-sized firms. The findings of this study largely
try concentration reported an R2of 0.06 in their control model support the fluidity thesis of firm size, initial diversity, and
(Wiersema and Bantel, 1992). The R2 of other studies that international diversification favored by the adaptation
per-can be considered closest per-cannot be directly compared because spective in contrast with the rigidity thesis of the selection
they used logistic or generalized least squares regression and perspective.
did not provide R2(Ginsberg and Buchholtz, 1990; Haveman, I also found support for interactive effects of the study
1993; Hoskinsson and Galbraith, 1985; Miller and Friesen, variables. Significant interactive effects imply that the
relation-1982). Further research can improve upon these results by ships of these antecedents with change depend on the level of
considering other variables and additional measures of both one another. Therefore, the main effects need to be interpreted
exogenous and dependent variables. Considering multiple di-cautiously. The results reveal the complex nature of the
rela-mensions of performance, other variables and strategic change tionships of strategic change and its antecedents, as postulated
should improve the predictive utility of the sorts of models in the emerging inertial theories (Baum, 1990), pointing
to-specified herein. ward the utility of combining the adaptation and inertial
theo-ries (Haveman, 1993; Singh et al., 1986). Overall, I conclude that increasing firm size at the extreme lower and higher
Conclusions
levels enables internationally diversifying firms to change more
abruptly when they had higher initial product diversity, This study examined the relationships of firm size, initial whereas it motivates them to change more gradually when product diversity, and international diversification with strate-they had lower initial diversity. These effects of firm size are gic change. These antecedents were found to have significant less pronounced at the medium levels or for firms with low relationships with the degree of abrupt change in product levels of international diversification. diversity. This study makes a significant contribution to re-The results of this study should be interpreted in view of search on strategic change in several ways. First, the focus of its limitations. The sample included all the companies for the study was to examine the relationship of several well-which data on the study variables were available for the time known antecedents with a well-defined dimension of change period of the study. This represents a convenient sample, in product diversity. I found strong support for the complex although it is better than taking a random sample of the effects of firm size, initial product diversity, and international convenient sample, as is often done in studies based on archi- diversification, largely supporting the fluidity thesis instead val data. The results of the study should be considered tenta- of the rigidity thesis. Second, it answers a call by Dess and tive unless they have been replicated in other contexts. In colleagues (1995) to decompose the aggregated variable of addition, data on almost all the independent and the depen- diversification and provides an illustration of how researchers dent variables were from the same time period, and regression can examine a dimension of change in a bid to go beyond analysis revealed association rather than causal relationships earlier research on change or diversification. Third, the study among the variables. The use of the word “effect” or “influence” investigated and found significant association of the pattern is derived from theory rather than directly from the data. of change in product and international diversity, an area to The present study did not include all antecedents of their which Hitt and colleagues (1994, 1997) have drawn particular multiple dimensions. The focus was to examine relationships attention. Last, the study integrated ideas from the change of abrupt change in product diversification with a limited and diversification streams of literature informing each stream number of fairly well known antecedents. It also did not with research from the other area. The observed patterns of investigate the consequences of gradual–abrupt change. The relationships advance understanding of the phenomenon of study illustrates the benefit of analyzing a particular dimension change in organizations and should stimulate theoretical spec-of change with respect to its antecedents rather than a generic ulation and further empirical testing.
variable of change, which is often done in studies of strategic
change. This study needs to be complemented by research Thanks are due to Professor Michael Moch for his comments on previous drafts of the manuscript. I gratefully acknowledge the research support by the
examining dimensions such as related-unrelated change and
Sam M. Walton College of Business Administration, University of Arkansas.
convergent-divergent change to distinguish among multiple dimensions of change in order to specify different models of
change, their antecedents, and consequences. Further research
References
also needs to examine other antecedents of this dimension ofBaum, J.: Inertial and Adaptive Patterns in Organizational Change, change and their alternative operationalizations (Dass and inAcademy of Management Best Paper Proceedings, L. Jauch and
Moch, 1997, 1998). J. Wall, eds., Academy of Management, New York. 1990, pp.
Although they compare favorably with the range of 0.0004 165–169.
to 0.05 reported in a meta-analysis of comparable studies on Beckhard, R., and Harris, R.:Organizational Transitions: Managing
Beer, M., Eisenhardt, R., and Spector, B.: Why Change Programs Don’t Ginsberg, A.: Measuring and Modeling Changes in Strategy: Theoreti-cal Foundations and EmpiriTheoreti-cal Directions.Strategic Management Produce Change.Harvard Business Review68 (June 1990): 156–166.
Journal9 (November-December 1988): 559–575. Bettis, R., and Mahajan, V.: Risk/Return Performance of Diversified
Firms.Management Science31 (July 1985): 785–799. Ginsberg, A., and Buchholtz, A.: Converting to For-profit Status: Corporate Responsiveness to Radical Change.Academy of Manage-Blau, P., and Schoenherr, R.:The Structure of Organizations, Basic
ment Journal33 (September 1990): 445–477. Books, New York. 1971.
Geringer, J., Beamish, P., and daCosta, R.: Diversification Strategy and Bourgeois, L.J.: On the Management of Organizational Slack.Academy
Internationalization: Implications for MNE Performance.Strategic of Management Review6 (March 1981): 29–39.
Management Journal10 (March-April 1989): 109–119. Buhner, R.: Assessing International Diversification of West German
Grant, R. M., Jammine, A. P., and Thomas, H.: Diversity, Diversifica-Corporations.Strategic Management Journal 8 (January-February
tion, and Profitability Among British Manufacturing Companies, 1987): 25–37.
1972–1984.Academy of Management Journal31 (December 1988): Chandler, A. D.:Strategy and Structure: Chapters in the History of the 771–801.
Industrial Enterprise,MIT Press, Cambridge, MA. 1962.
Hambrick, D., and D’Aveni, R.: Large Corporate Failures as Downward Christensen, H. K., and Montgomery, C. A.: Corporate Economic Spirals.Administrative Science Quarterly33 (March 1988): 1–23.
Performance: Diversification Strategy Versus Market Structure.
Stra-Hamel, G.: Competition for Competence and Interpartner Learning tegic Management Journal2 (October-December 1981): 327–343.
Within International Strategic Alliances. Strategic Management Cohen, J., and Cohen, P.:Applied Multiple Regression/Correlation Analysis Journal12 (Special issue, Summer 1991): 83–103.
for the Behavioral Sciences,Lawrence Erlbaum, Hillsdale, NJ. 1983.
Hannan, M. T., and Freeman, J. H.: The Population Ecology of Organi-Cyert, R., and March, J.:A Behavioral Theory of the Firm, Prentice
zations.American Journal of Sociology82 (5) (1977): 929–963. Hall, Englewood Cliffs, NJ. 1963.
Hannan, M. T., and Freeman, J. H.: Structural Inertia and Organiza-Dass, P.: The Dynamics of Strategic Organizational Change:
Devel-tional Change.American Sociological Review49 (April 1984): 149– oping a Generic Model of Multidimensional Change, Paper Pre- 164.
sented at the Annual Meeting of the Academy of Management,
Haveman, H.: Organization Size and Change: Diversification in the Las Vegas, NV. 1992.
Savings and Loan Industry After Deregulation.Administrative Sci-Dass, P., and Moch, M. K.: Stress, Inertia and Asset Flexibility— ence Quarterly38 (March 1993): 20–50.
Interacting Determinants of the Abruptness of Strategic Change
Hill, C. W., and Hansen, G. S.: A Longitudinal Study of the Causes In Product Diversity, Working Paper, College of Business
Admin-and Consequences of Changes in Diversification in the U.S. Phar-istration, University of Arkansas, 1997.
maceutical Industry 1977–1986.Strategic Management Journal12 Dass, P., and Moch, M. K.: Global Diversification as a Function of
(March 1991): 187–200. Initial Diversity and Firm Size, Paper Presented at the Annual
Hitt, M., Hoskisson, R., and Ireland, R.: A Mid-Range Theory of the Meeting of the Academy of Management, San Diego, CA. 1998.
Interactive Effects of International and Product Diversification on Dess, G., and Beard, D.: Dimensions of Organizational Task
Environ-Innovation and Performance. Journal of Management 20 (2) ments.Adminstrative Science Quarterly29 (March 1984): 52–73.
(1994): 297–326. Dess, G. G., Gupta, A., Hennart, J. F., and Hill, C. W. L.: Conducting
Hitt, M., Hoskisson, R., and Kim, H.: International Diversification: and Integrating Strategy Research at the International, Corporate,
Effects on Innovation and Firm Performance in Product-Diversi-and Business Levels: Issues Product-Diversi-and Directions.Journal of Management
fied Firms. Academy of Management Journal 40 (August 1997): 21 (3) (1995): 357–393.
767–798. DiMaggio, P. J., and Powell, W. W.: The Iron Cage Revisited:
Institu-Hoskisson, R. E., and Galbraith, C. S.: The Effect of Quantum Versus tional Isomorphism and Collective Rationality in Organizational
Incremental M-Form Reorganization on Performance: A Time-Fields.American Sociological Review48 (April 1983): 147–160.
Series Exploration of Intervention Dynamics.Journal of Manage-Fichman, M., and Levinthal, D. A.: Honeymoons and the Liability ment11 (1) (1985): 55–70.
of Adolescence: A New Perspective on Duration Dependence in
Hoskisson, R. E., and Hitt, M. A.: Antecedents and Performance Social and Organizational Relationships.Academy of Management
Outcomes of Diversification: A Review and Critique of Theoretical Review16 (April 1991): 442–468.
Perspectives.Journal of Management16 (2) (1990): 461–509. Fiegenbaum, A., and Thomas, H.: Attitudes Toward Risk and The
Hoskisson, R. E., Hitt, M. A., Johnson, R. A., and Moesel, D. D.: Risk-Return Paradox: Prospect Theory Explanations.Academy of
Construct Validity of an Objective (Entropy) Categorical Measure Management Journal31 (March 1988): 85–106.
of Diversification Strategy. Strategic Management Journal 14 Finkelstein, S., and Hambrick, D.: Top-Management-Team Tenure (March 1993): 215–235.
and Organizational Outcomes: The Moderating Role of Managerial
Huff, J., Huff, A., and Thomas, H.: Strategic Renewal and the Interac-Discretion.Administrative Science Quarterly35 (September 1990):
tion of Cumulative Stress and Inertia.Strategic Management Journal 484–503.
13 (Special issue, Summer 1992): 55–75. Fombrun, C., and Ginsberg, A.: Shifting Gears: Enabling Change in
Jacquemin, A., and Berry, C.: Entropy Measure of Diversification Corporate Aggressiveness.Strategic Management Journal11
(May-and Corporate Growth. The Journal of Industrial Economics 27 June 1990): 297–308.
(June 1979): 359–369. Freeman, E. R.:Strategic Management: A Stakeholder Approach,Pitman,
Johnson, G.: Rethinking Incrementalism.Strategic Management Jour-Marshfield, MA. 1984.
nal9 (January-February 1988): 75–92. Friesen, P. H., and Miller, D.: A Mathematical Model of the Adaptive
Keats, B., and Hitt, M.: A Causal Model of Linkages among Environ-Behavior of Organizations.Journal of Management Studies23
Performance. Academy of Management Journal 31 (September Quinn, J. B.: Strategies for Change: Logical Incrementalism, Irwin, Homewood, IL. 1980.
1988): 570–598.
Rajagopalan, N., and Spreitzer, G. M.: Toward a Theory of Strategic Kelly, D., and Amburgey, T. L.: Organizational Inertia and
Momen-Change: A Multi-Lens Perspective and Integrative Framework. tum: A Dynamic Model of Strategic Change.Academy of
Manage-Academy of Management Review22 (January 1997): 48–79. ment Journal34 (September 1991): 591–612.
Ramanujam, V., and Varadarajan, P.: Research on Corporate Diversi-Kim, W. C., Hwang, P., and Burgers, W. P.: Global Diversification
fication: A Synthesis.Strategic Management Journal10 (November-Strategy and Corporate Profit Performance.Strategic Management
December 1989): 523–551. Journal10 (January-February 1989): 45–57.
Ramaswamy, K.: Multinationality and Performance: A Synthesis and Kimberly, J. R.: Organizational Size and the Structuralist Perspective:
Redirection, inAdvances in International Comparative Management, A Review, Critique, and Proposal,Administrative Science Quarterly
vol. 7, S. B. Prasad, ed., JAI Press, Greenwich, CT. 1992, pp. 21 (December 1976): 571–597.
241–267. Kimberly, J. R., and Evanisko, M. J.: Organizational Innovation: The
Robertson, P., Roberts, D., and Porras, J.: Dynamics of Planned Influence of Individual, Organizational, and Contextual Factors
Organizational Change: Assessing Empirical Support for a Theo-on Hospital AdoptiTheo-on of Technological and Administrative
Inno-retical Model.Academy of Management Journal36 (June 1993): vations.Academy of Management Journal 24 (December 1981):
619–634. 689–713.
Rumelt, R.:Strategy, Structure, and Economic Performance, Harvard Kotabe, M.: The Relationship Between Offshore Sourcing and
Innova-Business School Press, Boston, MA. 1974. tiveness of U.S. Multinational Firms: An Empirical Investigation.
Sambharya, M. B.: The Combined Effects of International Diversifica-Journal of International Business Studies21 (4) (1990): 623–638.
tion and Product Diversification Strategies on the Performance of Kossek, E. E., Dass, P., and DeMarr, B.: The Dominant Logic of
U.S. Based Multinational Corporations.Management International Employer-Sponsored Work and Family Initiatives: Human
Re-Review35 (March 1995): 197–218. source Managers’ Institutional Role. Human Relations 47 (9)
Sharfman, M. P., Wolf, G., Chase, R. B., and Tansik, D. A.: Anteced-(1994): 1121–1149.
ents of Organizational Slack.Academy of Management Review13 Lamont, B. T., Williams, R. J., and Hoffman, J. J.: Performance During
(October 1988): 601–614. “M-Form” Reorganization and Recovery Time: The Effects of Prior
Singh, J. V.: Performance, Slack, and Risk Taking in Organizational Strategy and Implementation Speed.Academy of Management
Jour-Decision Making.Academy of Management Journal29 (September nal37 (February 1994): 153–166.
1986): 562–585. Lindblom, C.: The Science of Muddling Through.Public
Administra-Singh, J. V., House, R. J., and Tucker, D. J.: Organizational Change tion Review19 (Spring 1959): 79–88.
and Organizational Mortality.Administrative Science Quarterly31 Lubatkin, M., and Chatterjee, S.: Extending Modern Portfolio Theory (December 1986): 587–611.
Into the Domain of Corporate Diversification: Does it Apply?
Starbuck, W., Greve, A., and Hedberg, B.: Responding to Crises. Academy of Management Journal37 (February 1994): 109–136.
Journal of Business Administration9 (2) (1978): 111–137. McNamee, M., Magnusson, P., Fins, A. N., and King, R.: Think
Glob-Staw, B. M., Sandelands, L. E., and Dutton, J. E.: Threat-Rigidity ally, Survive Locally.Business Week(November 26, 1990): 50–51.
Effects in Organizational Behavior: A Multilevel Analysis. Adminis-Meyer, A. D., Brooks, G. R., and Goes, J. B.: Environmental Jolts and trative Science Quarterly26 (December 1981): 501–524.
Industry Revolutions: Organizational Responses to Discontinuous Stimpert, J. L., and Duhaime, I. M.: Seeing the Big Picture: The Change.Strategic Management Journal11 (Special issue, Summer Influence of Industry, Diversification. Academy of Management
1990): 93–110. Journal40 (June 1997): 560–583.
Miller, D., and Friesen, P.: Structural Change and Performance: Tushman, M. L., and Romanelli, E.: Organizational Evolution: A Quantum Versus Piecemeal-Incremental Approaches.Academy of Metamorphosis Model of Convergence and Reorganization, in Management Journal25 (December 1982): 867–892. Research in Organizational Behavior,vol. 7, L. L. Cummings and B. M. Staw, eds., JAI Press, Greenwich, CT. 1985, pp. 171–222. Miller, D., and Friesen, P.:Organizations—A Quantum View,
Prentice-Hall, Englewood Cliffs, NJ. 1984. Vachani, S.: Distinguishing between Related and Unrelated Interna-tional Geographical Diversification: A Comprehensive Measure Mintzberg, H., and Westley, F.: Cycles of Organizational Change.
of Global Diversification.Journal of International Business Studies Strategic Management Journal13 (Special issue, Winter 1992):
22 (2) (1991): 307–322. 39–59.
Wiersema, M., and Bantel, K.: Top Management Team Demography Nelson, R., and Winter, S.:An Evolutionary Theory of Economic Change,
and Corporate Strategic Change.Academy of Management Journal Belknap Press, Cambridge, MA. 1982.
35 (March 1992): 91–121. Palepu, K.: Diversification Strategy, Profit Performance and the
En-Wiersema, M. F., and Bantel, K. A.: Top Management Team Turnover tropy Measure.Strategic Management Journal6 (July-September
as an Adaptation Mechanism: The Role of Environment.Strategic 1985): 239–255.
Management Journal14 (October 1993): 485–504.
Pfeffer, J., and Salacik, G.:The External Control of Organizations: A Zajac, E. J., and Kraatz, M. S.: A Diametric Model of Strategic Change: Resource Dependence Perspective,Harper and Row, New York. 1978. Assessing the Antecedents and Consequences of Restructuring in Prahalad, C. K.: The Changing Nature of Worldwide Competition. the Higher Education Industry.Strategic Management Journal14
![Figure 3. Results of the relationship of firmsize and abrupt change [——Equation Y ��0.21 * size � 0.09 * size squared � 0.22 * sizecubed (Multiple regression model at step 4)].](https://thumb-ap.123doks.com/thumbv2/123dok/3110268.1377544/9.612.230.555.479.735/figure-results-relationship-firmsize-equation-sizecubed-multiple-regression.webp)
