Carpenter & Fredrickson, Academy of Management Journal, 44(3): 533-546
TOP MANAGEMENT TEAMS, GLOBAL STRATEGIC POSTURE,
AND THE MODERATING ROLE OF UNCERTAINTY
Mason A. Carpenter
Department of Management & Human ResourcesSchool of Business, University of Wisconsin-Madison
Grainger Hall, 975 University AvenueMadison, WI 53706-1323
(608) 262-9449, (608) 262-8773 [email protected]
James W. Fredrickson
Department of ManagementGraduate School of Business, CBA 4.202
University of Texas at AustinAustin, TX 78712-1174
(512) 471-5694, (512) 471-3937 FAX
_____________________________Special thanks to Syd Finkelstein, Don Hambrick, W. Gerry Sanders, and Jim Westphal for theirhelpful comments on earlier drafts of this manuscript
2
TOP MANAGEMENT TEAMS, GLOBAL STRATEGIC POSTURE,
AND THE MODERATING ROLE OF UNCERTAINTY
Abstract
The complexity surrounding globalization makes it particularly germane to the study of top
management team (TMT) demographic effects, and such complexity also provides a unique
context for assessing the moderating role of uncertainty. In a sample of U.S.-based firms, TMT
international experience, educational heterogeneity, and tenure heterogeneity were positively
related to firms' global strategic posture (GSP), while functional heterogeneity exhibited a
negative association. However, when uncertainty was introduced as a moderator, the
relationships between TMT tenure heterogeneity and functional heterogeneity and GSP were
negative in conditions of high uncertainty, but positive in the face of low uncertainty.
3
Contemporary firms are often encouraged to develop a commanding international
presence (Bartlett & Ghoshal, 1989; Zachary, 1996). However, while most large U.S. companies
are to some degree international, not all have been equally aggressive in responding to the
“global mandate” (Stopford, 1992). And despite assumptions to the contrary, the globalization of
U.S. firms significantly lags that of firms in other major industrialized countries (Makhija,
Kwangsoo & Williamson, 1997). Although research has emphasized a variety of factors as
determinants of globalization, with few exceptions (Lohrke & Bruton, 1997) it has typically
ignored the question of whether top management teams' (TMT) mix of backgrounds and
expertise are related to their firms' global strategic postures (GSP). Moreover, given the
increasing importance of globalization as a major and complex component of corporate strategy
(Prahalad, 1990), such a context may be particularly well-suited for observing TMT effects.
Therefore, the first purpose of this study was to examine the relationship between top
management team characteristics and GSP.
The influence of top executives on their organizations is often assumed to be a product of
their life experiences, as reflected in their demographic characteristics. But even though the
upper echelons research that considers relationships between TMT demographics and firm
outcomes is well developed (Finkelstein & Hambrick, 1996), an important condition for those
relationships has received little empirical attention. Specifically, TMT demographic effects are
likely to be strongest under conditions of high uncertainty (Finkelstein & Hambrick, 1996: 20;
Hambrick & Mason, 1984). However, upper echelons researchers have typically assumed that
uncertainty varies little across firms (i.e., it is typically "high"), and have thus ignored the
potential role of uncertainty as a critical moderator of TMT demographic effects. And just as the
complexity surrounding globalization makes it particularly germane to the study of TMT
4
demographic effects, so too does such complexity provide a unique context for assessing the
moderating role of uncertainty. Indeed, paying explicit attention to uncertainty may help to
explain the inconsistent findings of prior upper echelons studies. For instance, it is not known
whether TMT diversity has an independent effect on group conflict and cognitive heterogeneity,
or must be considered in concert with environmental uncertainty to understand diversity's true
impact on organizational outcomes. Therefore, the second purpose of this research was to test
the moderating role of uncertainty on the relationship between TMT characteristics and global
strategic posture.
THEORY DEVELOPMENT
Based on their assertion that organizations are "reflections of the values and cognitive
bases of powerful actors,” Hambrick and Mason proposed that “organizational outcomes –
strategic choices and performance levels – are partially predicted by managerial background
characteristics” (1984: 193). Subsequent upper echelons research has drawn on a variety of
theories, and indicates that the demographic characteristics of firms' top executives are indeed
related to a range of important organizational outcomes (Bantel & Jackson, 1989; Wiersema &
Bantel, 1992). For instance, it has been argued that background diversity or heterogeneity is
indicative of TMTs' sociocognitive diversity, their skill sets, and the breadth of their social and
professional ties (Finkelstein & Hambrick, 1996). This heterogeneity of cognitions, skills, and
ties is said to provide top management teams with diverse inputs and helps them be more
responsive to environmental complexity and change (Bantel & Jackson, 1989; Jackson, 1992).
Such diversity may similarly help TMTs overcome the information overload, complexity, and
domestic myopia that typically hamper globalization efforts (Ohmae, 1989; Sanders & Carpenter,
1998).
5
It should also be noted that some research has argued that demographic similarity or
homogeneity promotes behavioral integration among members of a team (O’Reilly, Snyder &
Boothe, 1993), while diversity is a likely source of conflict and a deterrent to productive group
functioning (O'Reilly et al, 1993; Wagner, Pfeffer & O'Reilly, 1984). However, others have
argued that the selection and socialization processes common among TMTs are likely to dampen
such conflict (Bantel & Jackson, 1989). Regardless of the valence of its impact, the upper
echelons perspective presumes that TMT heterogeneity will ultimately be reflected in firms'
actions. But it is unknown if such impact extends beyond the domestic operations of U.S. firms
to include decisions about global posture, or if top management effects are especially strong in
predicting this important organizational outcome. Moreover, a recent review of leading
management and international business journals concluded that top management teams were
"notably absent" from the literature on globalization (Lohrke & Bruton, 1997: 41). Therefore,
this study draws on the upper echelons perspective and attempts to extend that work by assessing
the relationship between top team diversity and GSP. As noted earlier, it also (re)introduces the
impact of environmental uncertainty into the upper echelons dialogue.
TMT Characteristics and Global Strategic Posture
Global strategic posture reflects the degree to which a firm depends on foreign markets
for customers and factors of production, along with the geographical dispersion of same. While
the literature on globalization typically stresses its potential benefits, it also highlights the
complexity inherent in going global. Despite such complexity, many management practitioners
and researchers maintain that companies' long-term success and survival are increasingly
dependent on their having a strong global presence (Bartlett & Ghoshal, 1989; Hitt, Hoskisson &
Kim, 1997; Maruca, 1994; Zachary, 1996). This is because an expansive GSP enables firms to
6
leverage R&D costs across countries and respond to foreign competitors in their domestic market
strongholds (Bartlett & Ghoshal, 1989; Kim & Mauborgne, 1991). At the same time, such global
activities are likely to increase the range of cultures (Gomez-Mejia & Palich, 1997; Hofstede,
1980), customers, competitors, and regulations (Porter, 1986) that a firm and its TMT face.
Therefore, the intricate web of activities and institutions that creates opportunities for global
firms also produces tremendous managerial complexity (Prahalad, 1990; Sanders & Carpenter,
1998).
Although few studies have linked TMTs to GSP (Lohrke & Bruton, 1997), some recent
upper echelons research is instructive. For example, Sanders and Carpenter (1998) found that
large top teams often lead firms that had expansive GSPs; they argued that firms with larger
TMTs can benefit substantially from a division of labor. Those same authors also suggested that
a TMT's size reflects the diversity of its members' backgrounds and information sources, and
ultimately their collective capacity to process the complex information arising from
globalization. However, team size alone is generally a poor gauge of TMT diversity (Finkelstein
& Hambrick, 1996). Indeed, the upper echelons perspective suggests that it is differences among
executives' experiences, as signaled by their demographic heterogeneity, which broadens TMTs'
networks and increases the diversity of their skill sets and world views (Hambrick & Mason,
1984).
Top management team demographic heterogeneity is relevant here since the diversity of
TMT network ties has been shown to be positively related to GSP (Athanassiou & Nigh, 1999;
Kim & Mauborgne, 1991). Similarly, the diversity in skill sets and world views that typically
accompany TMT demographic heterogeneity may also have implications for the expansiveness
of a firm's GSP. TMTs whose members have diverse skills and orientations can be more creative
7
and nimble in strategic problem solving (Dutton & Duncan, 1987), and perhaps more likely to
overcome their domestic myopia. Moreover, such diversity may build trust and perceptions of
procedural justice among a firms' product and geographic unit managers by signaling that the
TMT takes competing interests into account when allocating scarce resources around the globe
(Kim & Mauborgne, 1991). These heterogeneity effects can be related, in turn, to an expansive
GSP since such a strategy requires a firm to excel at interunit cooperation, to cross-subsidize
business and country units for overall corporate goals, and to swiftly coordinate actions globally
(Kim & Mauborgne, 1991; Gupta & Govindarajan, 1991). Finally, diversity in top teams has
also been argued to increase their level of sociocognitive complexity (Jackson, 1992; Wiersema
& Bantel, 1992). Therefore, TMTs that are sociocognitively complex may be better equipped to
make sense of changing international market opportunities, and "reconcile the conflicts and
paradoxes" presented by globalization (Murtha, Lenway, & Bagozzi, 1998: 112).
The arguments presented above suggest that firms are most likely to have expansive
global strategic postures when they are led by TMTs that have the diverse network ties, skills,
and world views that typically accompany demographic heterogeneity. Based on our interest in
(1) predicting the relationship between TMT demographic characteristics and GSP, and then (2)
considering the moderating effects of uncertainty on that relationship (discussed below), we
identified several TMT characteristics whose diversity seemed particularly pertinent to GSP, and
which represent variables of interest to upper echelons researchers (Bantel & Jackson, 1989;
Boeker, 1997; Hambrick & Mason, 1984; Roth, 1995; Wiersema & Bantel, 1992). Specifically,
this study investigates relationships between four TMT characteristics -- (a) breadth of
international work experience, (b) educational heterogeneity, (c) functional heterogeneity, and (d)
tenure heterogeneity -- and firms' global strategic posture.
8
International experience is an obvious choice for anyone considering demographic effects
and globalization. Moreover, it is considered one of the most broadening elements of executives'
backgrounds since it typically complements and expands on the role played by other experiences
(Reuber & Fischer, 1997; Roth, 1995; Sambharya, 1996). Similarly, top management team
educational heterogeneity is pertinent to GSP because it provides an indicator of the diversity of
skills and cognitive processes, as well as basic knowledge, embedded in the TMT (Bantel &
Jackson, 1989; Boeker, 1997; Wiersema & Bantel, 1992). The third characteristic, functional
diversity, captures the breadth of skill-sets and network information sources available to TMTs
(Ancona & Caldwell, 1992). The final characteristic, firm tenure heterogeneity, is relevant to
GSP because differences in tenure have been shown to influence the degree to which TMT
members use their networks to provide advice on international markets (Athanassiou & Nigh,
1999). In combination, these four characteristics have been shown to indicate variation in TMT
networks, skills, and orientations, and may result in executives considering of a wider range of
strategy alternatives. Therefore, we hypothesize that such breadth and diversity may prompt
some TMTs to overcome their domestic myopia, and will be related to global strategic posture.
H1: There will be a positive relationship between top management team (a)international work experience, (b) educational heterogeneity, (c) functionalheterogeneity, and (d) firm tenure heterogeneity and the expansiveness of GSP.
Contrary to the assumption underlying the above argument, that TMT selection and
socialization processes are likely to mitigate the potentially negative effects of demographic
heterogeneity (Wiersema & Bantel, 1992), several recent studies have shown the relationship of
team member heterogeneity with organizational outcomes to be curvilinear or simply negative.
For instance, Greening and Johnson (1997) reported curvilinear relationships between TMT
functional heterogeneity, tenure heterogeneity, and firms' responses to organizational crisis.
9
Also, O'Reilly and his colleagues (1993) found TMT tenure diversity to be negatively related to
strategic change in high technology firms. And, in the context of new high-technology product
teams and airline industry top teams, respectively, Ancona and Caldwell (1992) and Smith,
Smith, Olian, Sims, O'Bannon and Scully (1994) observed that aspects of demographic diversity
impeded team performance. Taken together, this research argues that while moderate levels of
team member heterogeneity may be beneficial because of the varied networks, skills, and world
views that diversity provides, excessive heterogeneity may lead to interpersonal conflict,
communication breakdowns, and thus detract from the TMT's overall level of behavioral
integration. Given the complexity characterizing high-GSP firms (Sanders & Carpenter, 1998),
and hence the need for high levels of teamwork in their TMTs (Kim & Mauborgne, 1991; Weick
& Van Orden, 1990), this evidence suggests that the relationship between TMT characteristics
and GSP might be curvilinear. Therefore, we offer the following hypothesis as an alternative to
the demographic relationships proposed by hypotheses H1a-d.
H2: There will be a curvilinear relationship between top management team (a)international work experience, (b) educational heterogeneity, (c) functionalheterogeneity, and (d) tenure heterogeneity and the expansiveness of GSP.
Environmental Uncertainty and TMT Demographic Effects
This section advances the proposition that if the relationship between TMT demographics
and GSP is contingent upon environmental uncertainty, then the level of such uncertainty will
moderate that relationship. Indeed, a notable aspect of the studies cited above that found a
negative or curvilinear relationship is that they were conducted in highly uncertain contexts (i.e.,
situations of crisis, or in industries otherwise characterized by high uncertainty). As discussed
earlier, the complexity surrounding globalization makes it particularly relevant to the study of
TMT demographic effects -- such complexity also provides a unique context for assessing the
10
moderating role of uncertainty. And while we recognize that there are other sources of
uncertainty (e.g., organizational), we emphasize environmental uncertainty here because of its
apparent centrality to the upper echelons perspective (Hambrick & Mason, 1984).
Uncertainty is a consequence of environmental factors that generally result in a lack of
information needed to assess means-ends relationships, make decisions, and confidently assign
probabilities to their outcomes (Duncan, 1972: 318). In developing their upper echelons
perspective, Hambrick and Mason (1984) made uncertainty a precondition for demographic
effects by anchoring it in the Carnegie School’s behavioral theory of the firm (e.g., March &
Simon, 1958). That theory, in turn, holds that decisions made under great uncertainty are more
likely to be “the outcome of behavioral factors rather than a mechanical quest for economic
optimization” (Hambrick & Mason, 1984: 194). The importance of uncertainty to the upper
echelons view is well articulated by the following observation:
"That top executives would act on the basis of their own predispositions is fully understandable.Senior managers are embedded in ambiguity, complexity, and information overload.….Thus, thetop executive faces the classic case of what the renowned psychologist Walter Mischel (1977)calls a 'weak situation,' that is, one in which the characteristics of the situation are not clear-cutenough to dictate a course of action. In such circumstances, the decision maker's personal frameof reference, not the objective characteristics of the situation, becomes the basis for action"(Finkelstein & Hambrick, 1996: 20).
Therefore, upper echelons predictions are contingent upon the view that top management
teams generally operate under highly uncertain conditions (i.e., conditions typically characterized
by ambiguity, complexity, and information overload). As a result, embedded in the upper
echelons perspective is an assumption that the more uncertain the decision making situation, the
more likely it is that TMT demographic characteristics will be manifest in organizational actions
and outcomes.
11
Such an assumption, however, is untested and ignores a vast literature which shows that
firms and their top management teams vary significantly in the amount of uncertainty that
surrounds them (Aldrich, 1979; Bergh & Lawless, 1998; Dess & Beard, 1984; Duncan, 1972;
Keats & Hitt, 1988). Therefore, to the extent that TMT members rely on their prior experiences,
the effects of functional background and other management characteristics are also likely to vary
depending on the uncertainty they face. And such reliance increases the likelihood of
demographic effects. Ironically, it may also increase the likelihood of interpersonal conflict as
demographic differences are made more salient by uncertainty, and TMT members are afforded
greater discretion (Finkelstein & Hambrick, 1990). This tension is compounded by the fact that
globalization is sometimes portrayed as a strategy explicitly designed to spread risk (Kim,
Hwang & Burgers, 1993), and that executives may seek to manage environmental uncertainty
and increase their discretion through an expansive GSP. Consequently, there is reason to expect
that such uncertainty may systematically effect the relationship between TMT demographics and
global strategic posture. In the context of this study, it suggests the following hypothesis:
H3: Environmental uncertainty will moderate the previously proposed linearrelationships (i.e., H1a-d) between TMT (a) international work experience, (b)educational heterogeneity, (c) functional heterogeneity, and (d) tenureheterogeneity, and the expansiveness of GSP.
RESEARCH METHOD
Sample
A random, stratified sample of 300 U.S. firms was developed from the Standard & Poors
Industrials (first stratum) and Standard & Poors Mid-cap indices (second stratum). These
sources were selected because (1) they are representative of major U.S. industrial corporations,
(2) detailed demographic data are available on their TMTs, and (3) there is substantial variance
12
in the expansiveness of their global strategic postures -- even when controlling for firm size and
industry. It is important to note that the sample was consciously restricted to U.S. firms for two
primary reasons: (1) they have been most often criticized for their domestic myopia (Hordes,
Clancy & Baddaley, 1995; Ohmae, 1989), and (2) reliable longitudinal data on the TMTs of
firms outside the U.S. are generally not available. As discussed in detail below, archival data
were collected on 207 firms from 1984 through 1996. A means test indicated that the excluded
firms were not statistically different from the sample firms in total assets, total sales, and
performance (return on assets, return on sales, return on equity, and stock market returns).
Analytical approach, variables, and measures
Analytical approach. Hypotheses were tested with two panels of lagged, cross-sectional
data; OLS regression with fixed-effects models was used to control for unobserved differences
between firms. While GLS random-effects regression has also been used in upper echelons
studies incorporating multiple panels of data (e.g., Finkelstein & Hambrick, 1990), the Hausman
specification test suggests that a fixed-effects model was more appropriate here than a random-
effects model (Greene, 1990). Moreover, the fixed-effects model is generally considered a more
conservative test, and generates an R2 that is interpretable, whereas that generated by the GLS
model is not (Greene, 1990).
Estimating a fixed-effects model is equivalent to adding a dummy variable for each firm
(Greene, 1990), and controls for all constant, unmeasured differences across firms that may
explain differences in the dependent variables. For instance, industry membership is considered
an essential control variable in organizational studies, and it clearly has implications for firm
globalization (Porter, 1986). However, since primary industry membership did not vary among
these firms during the time period studied, and because fixed-effects control for variance due to
13
time-invariant characteristics like industry membership, industry dummies are not necessary.
Hence, they are not included in the fixed-effects models.
Dependent variable. GSP was calculated using a variation of Sullivan's (1994)
composite measure of firm internationalization. The first dimension, foreign sales, was
calculated as a ratio of foreign sales to total sales. The second dimension, foreign production,
was gauged by foreign owned-assets taken as a percentage of total assets. The third dimension,
geographic diversity, was measured as the extent to which a firm had subsidiaries in any of the
ten cultural zones of the world identified by Ronen and Shenkar (1985); each zone was counted
as .10 (e.g., a firm's presence in three zones generated a score of .30). The first two dimensions,
sales and assets, address a firm’s dependence on foreign markets and foreign-placed resources
(Kim et al, 1993), respectively. The third dimension indicates the geographic and cultural variety
associated with the other two; Sullivan (1994) provides a detailed discussion of these three
dimensions.
Indicators of each of the dimensions discussed above (foreign sales, foreign production,
and geographic diversity) were summed to form a composite measure of GSP. The higher the
number, the more expansive the firm’s GSP. Summing is not a problem in this case because
each indicator has the same metric (i.e., gauged as a percentage), is normally distributed, and the
three are highly correlated. As reported later, analyses also showed that these three indicators
load on one factor with a high Eigenvalue and explained variance, and the alpha was acceptable
at .87 (Nunnally, 1978). Since each of the three components is a ratio variable ranging from 0.0
to 1.0, a firm's GSP can range from 0.0 (no GSP) to 3.0 (a very expansive GSP). In our sample,
global strategic posture averaged .81 (s.d. of .43), with a range of .02 to 2.05 (out of 3.0). To
reduce the effects of annual aberrations, GSP was averaged over 1990 and 1991, and 1995 and
14
1996. Data on global strategic posture were gathered from PC-Compustat and Dun’s Directory
of American Corporate Families and International Affiliates.
Independent variables. Select TMT demographic characteristics comprised the set of
independent variables used to test hypotheses H1(a-d) through H3(a-d). Following other upper
echelons studies, the top management team was defined as the top two tiers of the organization’s
management (e.g., CEO, Chairman, COO, CFO and the next highest management tier, as used in
Wiersema & Bantel, 1992). Such a definition was expected to capture the dominant coalition for
all firms in the sample (Demb & Neubauer, 1992), and has been applied in other research
concerned with strategic actions (e.g., Finkelstein & Hambrick, 1990; Greening & Johnson,
1997; Wiersema & Bantel, 1992). This definition yielded teams of approximately six members.
To develop the panels of TMT demographic characteristics, teams were assessed in two
years -- 1988 and 1993 -- which themselves preceded the GSP observations by two years each
(averaged observations in 1990/91 and 1995/96, respectively). The two-year lag period (versus
longer or shorter periods) between the TMT and GSP observations reduced the possibility of
other factors confounding the relationship between demographics and GSP, and also allowed
enough time for potential top team effects to become apparent. Consistent with this expectation,
sensitivity analyses indicated that the relationships reported below were weaker with shorter (i.e.,
1 year) or longer (i.e., 3 years) lags. Data for TMT characteristics were obtained from the
executives’ career histories reported in Dun's Reference Book of Corporate Managements.
As noted by Michel and Hambrick, the coding of most TMT characteristics is “clear-cut
and objective” (1992: 21). The TMTs' international work experience was calculated as the
percentage of the TMT’s total years of experience accrued in international assignments (i.e.,
designated as international years). Educational background and functional background are
15
categorical variables. Using Wiersema & Bantel's (1992) coding scheme, executives' educational
backgrounds were classified as being in the arts, sciences, engineering, business and economics,
or law. Absent any other indication of education specialty, TMT members with M.S. or B.S.
degrees were coded as science specialists (Wiersema & Bantel, 1992). Executives with Ph.D.
degrees in the sciences were also coded as science specialists. Following Bantel and Jackson
(1989), Michel and Hambrick (1992), and Wiersema and Bantel (1992), we coded team
member’s functional backgrounds as being predominantly marketing, distribution, sales, research
and development, production, engineering, finance and accounting, law, or general. The degree
of heterogeneity of each variable was calculated using Blau's (1977) index. This index is
calculated as 1-�(Pi)2, where Pi are the percentage of individuals in the ith category. Firm
tenure heterogeneity was calculated using the coefficient of variation of the TMT’s firm tenure
(standard deviation divided by the mean). Allison (1978) has noted that among inequality
measures, the coefficient of variation is preferable when interval-level data such as age or time
are used.
Moderator variable. Hypotheses H3(a-d) predict that uncertainty will moderate the
main-effect relationships hypothesized in H1(a-d). We gauged environmental uncertainty using
a variation of the environmental instability measure developed by Dess and Beard (1984). This
measure incorporates aspects of Aldrich's (1979) stability-instability and turbulence constructs by
capturing the volatility of a firm's net sales in each of its four-digit SIC industries (Aldrich, 1979;
Bergh & Lawless, 1998; Dess & Beard, 1984; Keats & Hitt, 1988). Following Bergh & Lawless
(1998: 91-92), uncertainty was calculated by regressing a variable for each year on a variable for
net industry sales (Keats & Hitt, 1988). Five years of data were used for each regression (e.g.,
net industry sales from 1984 through 1988 were used to predict volatility in 1989). Following
16
the equation, yt = b0 + b1t + at, where y = industry sales, t = year, and a = residual, volatility was
the standard error of the regression slope coefficient divided by average sales. Greater
environmental uncertainty is indicated by larger values of volatility, and values were averaged for
the two-year period between the TMT and GSP measurements. Firms' four digit industries were
identified through PC-Compustat and Compaq Disclosure, and data for calculating the
regressions were collected from Compustat.
Control variables. It is often assumed that larger firms have more extensive
international activities which, in turn, would be reflected in higher GSPs. In addition, firm size
has been argued to affect the relationship between executive characteristics and organizational
outcomes (Miller, 1991; Miller & Toulouse, 1986). Therefore, organizational size was included
as a control and was measured as the log of total employees. Additionally, firm performance,
degree of diversification, and R&D intensity have been suggested as being related to a firm’s
level of globalization (Hitt et al, 1997; Sanders & Carpenter, 1998). Therefore, performance was
controlled by using the two-year (between the TMT and GSP observations) average of return on
assets (ROA), current ratio (the ratio of current assets to current liabilities), and free cash flow.
Diversification level was controlled using Palepu’s (1985) entropy measure, where
diversificationa = Σ Pia ln(1/Pia), and Pia, is the proportion of firm a’s sales in business segment
i. R&D intensity was controlled using the ratio of research and development expenses to total
revenues. Because measures of TMT diversity are size-dependent, and because GSP has been
associated with team size in prior studies (Sanders & Carpenter, 1998), TMT size was controlled
by using the total number of executives on the top team. Period effects were controlled (1 = first
panel) using a dummy variable for the two panels of data.
17
Average tenure is a standard control variable when tenure heterogeneity is calculated
using the coefficient of variation. It is also a proxy for managerial power and discretion
(Finkelstein & Hambrick, 1990), which may otherwise influence GSP. Therefore, we also
controlled for TMT average tenure, using the average of team members' firm tenures. Finally,
because changes in the composition of the top team have the potential to change the team's
characteristics, TMT turnover was controlled for using the percentage of new executives on the
TMT for the intervening two-year period between the TMT and GSP measurements.
RESULTS
The means, standard deviations and bivariate correlations for all variables are presented
in Table 1. With regard to the demographic variables, the low intercorrelations among them
suggests that they can be included in one regression model (Finkelstein & Hambrick, 1990).
Moreover, variable inflation factor (VIF) scores for all the models were within acceptable
parameters; thus multicollinearity was not a problem (Chatterjee & Price, 1991).
* * * * * Insert Tables 1 and 2 About Here * * * * *
As mentioned above, the hypotheses were tested using fixed-effects models and OLS
regression. All regression results, for both main and moderated effects, are presented in Table 2.
Model 1 includes only the control variables while Model 2 adds the variables needed to test the
main effects. Hypotheses H1(a-d) predicted that select TMT characteristics would be related to
GSP; the control (Model 1) and full model (Model 2) were significant, as was the change in R2.
Moreover, as shown in model 2, the coefficients for international work experience, educational
heterogeneity, and firm tenure heterogeneity were significant and positive, and that for functional
heterogeneity was significant but negative (in supplemental analyses, effects similar to those for
firm tenure were found using industry tenure and tenure on the TMT). Thus, the linear main-
18
effect results provide general support for H1(a), H1(b) and H1(d), but contradict the direction of
the relationship predicted by H1(c) (i.e., functional heterogeneity).
Alternatively, hypotheses H2(a-d) predicted that the effects of TMT characteristics would
be curvilinear with respect to GSP. To test these hypotheses we added a squared-term for each
demographic characteristic to Model 2; however, as shown in Model 3, none of the squared-term
coefficients were significant and the main-effect results were unchanged from those reported in
Model 2. Similar results were obtained when we used the log value of the demographic variables
instead of using the linear and quadratic components (Wiersema & Bantel, 1992: 105). To
preserve space, these results are omitted from the table. Thus, contrary to some research but
similar to Wiersema & Bantel's findings, the curvilinear predictions did not hold.
The final hypotheses (H3a-d), suggested that the relationship between TMT
characteristics and GSP would be influenced by an interaction with environmental uncertainty.
Therefore, it was appropriate to analyze these hypotheses by estimating a regression equation
with a multiplicative interaction term (Roth, 1995), and including the moderator variable as a
control. The tests of the uncertainty interaction hypotheses are shown in Model 4. As noted
there, the interactions of uncertainty with tenure heterogeneity, educational heterogeneity, and
functional heterogeneity were significant, as was the full model’s increase in R2 over the main-
effects model; the regression coefficients for the interactions were unchanged when each variable
was run in separate models.
To better understand the nature of the moderated relationships, supplemental simple-
effects analyses were also conducted. To examine such simple-effects, we followed the
decomposition procedure outlined by Jaccard, Turrisi and Wan (1990: 27-28), and focused only
on those interaction relationships that were significant (i.e., education, functional, and tenure
19
heterogeneity). In the first of these, the slope of educational heterogeneity remained positive
over the entire range of environmental uncertainty. However, its slope was significantly more
positive (p < .05, adjusted Bonferoni procedure) at high uncertainty than at low levels of
uncertainty. Visual plots further confirmed the functional forms of these underlying
relationships, thus providing evidence that high uncertainty positively moderated TMT
demographic effects. In contrast, decomposition of functional and tenure heterogeneity revealed
their slopes to be non-monotonic with respect to environmental uncertainty. Specifically, while
the coefficients of both of these TMT characteristics were initially positive (at low moderator
levels), they became negative as the moderators passed their median levels (p < .05, adjusted
Bonferoni procedure).
DISCUSSION
The primary objective of this study was to bring top management teams firmly into the
literature on globalization. This was done by focusing on the relationship between the
characteristics of firms' TMTs and the expansiveness of their global strategic postures. In
support of our expectations, the results indicate that (1) TMT characteristics were related to the
degree to which firms were global and, and (2) such relationships were contingent upon the level
of environmental uncertainty confronting the top team. These findings have important
implications for work on both globalization and upper echelons. Moreover, as discussed below,
the results provide a clearer picture than was previously available of the conditions that affect the
complex relationship between executives' backgrounds and organizational outcomes.
Conclusions and implications from main-effects
The present study indicates that firms were most likely to be highly global when they had
diverse TMTs -- diverse in terms of the breadth of their international experience, and the
20
heterogeneity of their educational backgrounds and firm tenures. Moreover, the linear main
effects for these characteristics are consistent with the upper echelons view that the diverse
perspectives, skills, and information networks that are presumed to accompany international
experience and educational and tenure heterogeneity (Bantel & Jackson, 1989; Dutton & Duncan,
1987; Finkelstein & Hambrick, 1996: 125; Hambrick & Mason, 1984; Wiersema & Bantel,
1992), may equip firms to manage the complexity and high information processing demands
associated with expansive GSPs.
In contrast, and contrary to what might be predicted by an upper echelons view,
functional heterogeneity exhibited a negative linear relationship with GSP. However, such a
finding is consistent with arguments which suggest that demographic heterogeneity is likely to
detract from team cohesiveness, and promote disagreement about the organization and its goals
(Ancona & Caldwell, 1992; O'Reilly, Caldwell & Barnett, 1989; O'Reilly et al, 1993). Conflict
among top team members may, in turn, lead a firm to be more defensive and ethnocentric in its
strategic actions, and make it difficult to gain the commitment needed to expand beyond its
domestic position (Kogut, 1985; Ohmae, 1989). Team conflict may also decrease the TMT's
capacity to process information -- a likely outcome of Staw, Sandelands, and Dutton's (1981)
"threat-rigidity effect." Consequently, this result suggests that a firm's ability to realize an
expansive GSP may be stifled when its top team has executives with dramatically different
functional experiences.
Finally, while we found no support for curvilinear demographic effects (hypotheses H2a-
d), the negative relationship between TMT functional heterogeneity and GSP suggests a possible
reconciliation between the two competing perspectives on the valence of TMT demographic
heterogeneity. As noted earlier, studies reporting negative or curvilinear demographic effects
21
have been conducted in contexts characterized by high uncertainty such as environmental crises
(Greening & Johnson, 1997), teams in high technology firms (O'Reilly et al, 1993 and Ancona &
Caldwell, 1992), and the intensely competitive airline industry (Smith et al, 1994). Moreover,
since those studies held the level of uncertainty relatively constant (i.e., high uncertainty
contexts), they provided no way to disentangle the probable impact of environmental uncertainty
on TMT demographic effects. However, the present study does, and the implications of
uncertainty for upper echelons research are discussed below.
Conclusions and implications from the interactions
We argued that high uncertainty should moderate the relationship between TMT
demographics and GSP, and the results for hypotheses H3(b-d) support that view. Specifically,
the association between TMT educational heterogeneity and GSP was uniformly stronger in
highly uncertain environments. In contrast, while the relationships of TMT tenure and functional
heterogeneity with GSP were also consistent across the different contexts, they exhibited a
strikingly different functional form than those for education. In contexts characterized by low
uncertainty, TMT tenure and functional heterogeneity were positively related to GSP. Therefore,
the simple-effects underlying these uncertainty interactions are consistent with those of TMT
educational heterogeneity, and the broader body of upper echelons theory that argues for the
sociocognitive benefits of heterogeneity (Finkelstein & Hambrick, 1996; Jackson, 1992).
However, decomposition of the interactions revealed the simple-effects of TMT tenure
and functional heterogeneity on GSP to be negative in high uncertainty contexts. Such negative
simple-effects were exactly opposite those of our upper echelons predictions, and lend qualified
support (i.e., depending on the level of uncertainty) to the arguments of researchers who suggest
that the behavioral implications of top team heterogeneity may lead to conflict, and ultimately
22
hinder the team's ability to manage complexity (O’Reilly et al, 1993). As discussed earlier, some
past research has attempted to bridge these divergent positions by suggesting an inverted-U (i.e.,
curvilinear) relationship between demographic heterogeneity and organizational outcomes
(Jackson, 1992; Wiersema & Bantel, 1992). Nevertheless, we found no support for the
curvilinear predictions, and our results suggest instead a refinement to such theorizing.
Specifically, arguments about the non-monotonic effects of heterogeneity may hold true for some
demographic characteristics, but one must first take into account the level of uncertainty facing
the TMT, as well as the specific demographic characteristic being examined. Moreover, since
the main-effect relationship between functional heterogeneity and GSP was negative, its positive
simple-effect in the presence of low uncertainty would have been masked entirely had the
moderating role of environmental uncertainty been ignored; and absent our accounting for
uncertainty, the negative simple-effects of tenure heterogeneity would have been similarly
masked in high uncertainty environments. Therefore, these results emphasize that the complex
effects of some TMT characteristics may not be evident unless explicit attention is given to the
level of uncertainty confronting the top team.
As noted above, TMT educational heterogeneity demonstrated a persistently positive
relationship with GSP, while the effects of functional and tenure heterogeneity varied by level of
uncertainty. Such inconsistencies in our results across demographic characteristics may reflect
differences in the impact of work-related experiences on group processes (Pelled, 1996), versus
the effects of other experiences like education. For example, functional and tenure heterogeneity
are the outcome of workplace socialization processes -- in addition to the network ties, skill sets,
and orientations they represent. Consequently, it is in contexts with low levels of uncertainty that
heterogeneous top teams may be most likely to profit from their diversity of ties, skills, and
23
knowledge. Moreover, if uncertainty creates a basis for conflict among TMT members who are
diverse (Milliken & Martins, 1996; O'Reilly et al, 1993), then under low uncertainty they may
have time to resolve their work-related differences. In contrast, the time pressures accompanying
high uncertainty may prevent such resolution. And without it, TMTs' work-related differences
may become counter productive (Finkelstein & Hambrick, 1996), thus preventing a firm from
realizing an expansive GSP. Moreover, contrary to prior theorizing (Priem, 1990), our results
suggest that TMT member similarity can indeed be functional in the face of high uncertainty. To
the extent that executives' backgrounds reflect a complex amalgam of relationships, cognitions,
and socialization outcomes, the present results lend empirical support to arguments that each
demographic characteristic is likely to have unique implications for team members' behaviors
(Greening & Johnson, 1997; Milliken & Martins, 1996; Pelled, 1996). Therefore, TMT diversity
should not be viewed as an all-inclusive, generic concept.
Unanswered questions and future directions
Like all research, this study has left questions unanswered, which in turn suggest future
research opportunities. Four of these questions are particularly important. The first arises from
the fact that we did not actually measure top team behaviors, cognitions, or perceptions, but
instead inferred them from the characteristics of the top team. As a result, by following the
norms of upper echelons research, we have “black-boxed” important underlying processes and
causal mechanisms that may have been pertinent to our arguments. For example, studies of
cognitive complexity among executives (Calori, Johnson & Sarnin, 1994; Murtha et al, 1998), as
well as TMT "international advice networks" (Athanassiou & Nigh, 1999), have found that both
of these factors are related to firm globalization. What is missing, however, is research that
establishes a direct link between such factors and certain forms of top team heterogeneity.
24
Therefore, studies are needed that further illuminate the nature of the relationships between TMT
characteristics and the actual cognitions and behaviors of upper echelon executives (e.g., Boone,
de Brabander, & van Witteloostuijn, 1996; Pelled, Eisenhardt & Xin, 1999).
The second research direction is suggested by our sampling of only U.S. firms. As
mentioned earlier, this was done in part because their executives are most often criticized as
falling prey to domestic myopia (Makhija et al, 1997; Ohmae, 1989). Moreover, we quickly
discovered that reliable TMT demographic data, like those needed in this study and in other
upper echelons research, are not typically available for non-U.S. firms (even those in Western
Europe). Regardless, the question of whether our findings generalize to such firms has not been
addressed. Further, to the degree that executives from different countries vary in their tolerance
of uncertainty (e.g., Hofstede, 1980), there is reason to expect that demographic effects will also
vary. For example, Wiersema and Bird’s (1993) finding that TMT heterogeneity was more likely
to be related to turnover in Japanese TMTs (a low uncertainty-tolerance culture) than in U.S.
TMTs (a high uncertainty-tolerance culture), provides an indication that demographic effects do
indeed play out differently depending on attitudes toward uncertainty that are embedded in the
team's socio-cultural context. Obviously, tests with non-U.S. samples are needed.
The third question concerns causality. Indeed, while we used a data structure that allows
us to suggest causal relationships between TMT characteristics and GSP, we have not established
causality; nor have we been able to gauge firms' global intent. For example, it is possible that
global firms consciously recruit internationally-seasoned executives to manage and grow their
far-flung operations. If so, it could be that GSP is a predictor of TMT demographics rather than
the reverse. However, while no social science research can prove causality (Cook & Campbell,
1979), we have established that certain TMT characteristics preceded expansive GSPs, and have
25
identified and included those control variables most likely to provide alternative explanations if
omitted. However, it is still important to better understand whether some top management teams
(i.e., those with more international experience) are more likely to lead their firms to expansive
global strategic postures than are others.
Finally, the discussion of causality gives rise to a fourth question and research direction --
one that concerns the effects of TMTs on international strategy implementation and firm
performance. Specifically, global firms differ significantly in the degree to which their far-flung
operations are actually coordinated and integrated (Roth, 1995). Moreover, upper echelons
research suggests that the level of such interdependence influences the pattern of executive
characteristics that are ideal for top managers to contribute to firm performance (Michel &
Hambrick, 1992; Roth, 1995). In support of this contention, Roth (1995) found that CEOs'
backgrounds were predictive of differences in firms' global interdependence and integration. He
also noted that certain CEO characteristics helped performance in low interdependence contexts,
but hindered it in high interdependence contexts. Therefore, investigators should do the field
research that takes them inside firms to better understand the TMT’s role in global strategy
implementation and firm performance.
Concluding Remarks
This study attempts to contribute to the literatures on top management teams and firm
globalization by going beyond prior work on both. Specifically, it inserts TMTs into the
discussion of globalization, and suggests that top executives’ backgrounds play a significant role
in the degree to which U.S. firms go global. Moreover, we have tried to demonstrate how the
complexity surrounding globalization makes it particularly germane to the study of TMT
demographic effects, and that such complexity provides a unique context for assessing the
26
moderating role of uncertainty. In addition, the arguments presented here serve to remind
researchers that Hambrick and Mason’s (1984) assumption of environmental uncertainty is
fundamental to upper echelons research, and instrumental in studies that use TMT demographic
variables to predict organizational outcomes. Indeed, by exploring the notion that uncertainty
systematically moderates TMT demographic effects, we hope we have opened up a fruitful new
avenue for research on upper echelons. And while this work emphasizes that there are limits to
the impact and interpretation of executive demographic effects, it also suggests that such
demographics can continue to play an important role in research on top management teams.
27
REFERENCES
Aldrich, H. 1979. Organizations and environments. NJ: Prentice Hall.
Allison, P. 1978. Measures of Inequality. American Sociological Review. 43: 865-880.
Ancona, D. & D. Caldwell. 1992. Demography and design: Predictors of new product teamdevelopment. Organization Science. 3: 321-341.
Athanassiou, N. & D. Nigh. 1999. The impact of U.S. company internationalization on topmanagement team advice networks: A tacit knowledge perspective. Strategic ManagementJournal. 20: 83-92.
Bantel, K. & S. Jackson. 1989. Top Management and Innovations in Banking: Does theComposition of the Top Team Make a Difference? Strategic Management Journal. 10:107-124.
Bartlett, C. & S. Ghoshal. 1989. Managing Across Borders. Harvard Business School Press.
Bergh, D. & M. Lawless. 1998. Portfolio restructuring limits to hierarchical governance: Effectsof environmental uncertainty and diversification strategy. Organization Science. 9: 87-102.
Blau, P. 1977. Inequality and Heterogeneity. Free Press.
Boeker, W. 1997. Executive migration and strategic change: The effect of top managermovement on product-market entry. Administrative Science Quarterly. 42: 213-237.
Boone, C., B. de Brabander, & A. van Witteloostuijn. 1996. CEO locus of control and small firmperformance: An integrative framework and empirical test. Journal of Management Studies.33:3: 667-699.
Calori, R., Johnson, G. & Sarnin, P. 1994. CEO’s cognitive maps and the scope of theorganization. Strategic Management Journal. 15: 437-457.
Chatterjee, S. & B. Price. 1991. Regression analysis by example (2nd Ed.). NY: Wiley & Sons.
Cook, T. & D. Campbell. 1979. Quasi-Experimentation. Chicago: Rand McNally.
Demb, A. & F. Neubauer. 1992. The corporate board. NY: Oxford University Press.
Dess, G. & D. Beard. 1984. Dimensions of organizational task environments. AdministrativeScience Quarterly. 29: 52-73.
Duncan, R. 1972. Characteristics of organizational environments and perceived environmentaluncertainty. Administrative Science Quarterly. 17: 313-327.
Dutton, J. & R. Duncan. 1987. The creation of momentum for change through the process ofstrategic issue diagnosis. Strategic Management Journal. 8: 279-296.
Finkelstein, S. & D. Hambrick. 1990. Top-Management-Team Tenure and OrganizationalOutcomes: The Moderating Role of Managerial Discretion. Administrative Science Quarterly.35: 484-503.
Finkelstein, S. & D. Hambrick. 1996. Strategic leadership: Top executives and their effects onorganizations. St. Paul: West Publishing Company.
28
Gomez-Mejia, L. & L. Palich. 1997. Cultural diversity and the performance of multinationalfirms. Journal of International Business Studies. 28: 309-336.
Greene, H. 1990. Econometric analysis. NY: Macmillan.
Greening, D. & R. Johnson. 1997. Managing industrial and organization crises. Business andsociety. 36: 334-361.
Gupta, A. & V. Govindarajan. 1991. Knowledge flows and the structure of control withinmultinational corporations. Academy of Management Review. 16: 768-792.
Hambrick, D. & P. Mason. 1984. Upper Echelons: The Organization as a Reflection of its TopManagers. Academy of Management Journal. 15: 3: 514-535.
Hitt, M., R. Hoskisson, & H. Kim. 1997. International diversification: Effects on innovation andfirm performance in product-diversified firms. Academy of Management Journal. 40: 767-798.
Hofstede, G. 1980. Culture's Consequences: International Differences in Work-Related Values.Beverly Hill, CA: Sage Publications.
Hordes, M., J. Clancy, & J. Baddaley. 1995. A primer for global start-ups. Academy ofManagement Executive. 9: 7-11.
Jaccard, J., R. Turrisi, & C. Wan. 1990. Interaction effects in multiple regression. London: SagePublications.
Jackson, S. 1992. Consequences of Group Composition for the Interpersonal Dynamics ofStrategic Issue Processing. Advances in Strategic Management. 8:345-382.
Keats, B. & M. Hitt. 1988. A causal model of linkages among environmental dimensions, macroorganizational characteristics, and performance. Academy of Management Journal. 3: 570-598.
Kim, W., P. Hwang, & W. Burgers. 1993. Multinationals' diversification and the risk-returntrade-off. Strategic Management Journal. 14: 275-286.
Kim, W. & R. Mauborgne. 1991. Implementing global strategies: The role of procedural justice.Strategic Management Journal. 12: 125-143.
Kogut, B. 1985. Designing global strategies: Profiting from operational flexibility. SloanManagement Review. 27: 1: 27-38.
Lohrke, F. & G. Bruton. 1997. Contributions and gaps in international strategic management.Journal of International Management. 3:1: 25-57.
Makhija, M., K. Kwongsoo, & S. Williamson. 1997. Measuring globalization of industries usinga national industry approach. Journal of International Business Studies. 28: 679-710.
March, J. & H. Simon. 1958. Organizations. New York: Wiley.
Maruca, L. 1994. The right way to go global: An interview with Whirlpool CEO DavidWhitwam. Harvard Business Review. March-April: 135-145.
Michel, J.G. & D. Hambrick. 1992. Diversification Posture and Top Management TeamCharacteristics. Academy of Management Journal. 35:1:9-37.
29
Miller, D. 1991. Stale in the Saddle: CEO Tenure and the Match Between Organization and theEnvironment. Management Science. 1: 34-54.
Miller, D. & J. Toulouse. Chief Executive Personality, Corporate Strategy and Structure in SmallFirms. Management Science. 1389-1409.
Milliken, F. & L. Martins. 1996. Searching for common threads: Understanding the multipleeffects of diversity in organizational groups. Academy of Management Review. 21: 402-423.
Mischel, W. 1977. The interaction of person and situation. In D. Magnusson & N. Endler (eds.).Personality at the crossroads: Current issues in interactional psychology. NY: Erlbaum.
Murtha, T., S. Lenway, & R. Bagozzi. 1998. Global mind-sets and cognitive shifts in a complexmultinational corporation. Strategic Management Journal. 19: 97-114.
Nunnally, J. 1978 (second edition). Psychometric theory. NY: McGraw-Hill.
Ohmae, K. 1989. Managing in a borderless world. Harvard Business Review. May-June: 152-161.
O'Reilly, C., D. Caldwell., & W. Barnett. 1989. Work group demography, social integration, andturnover. Administrative Science Quarterly. 34: 21-37.
O'Reilly, C., R. Snyder, & J. Boothe. 1993. Effects of executive team demography onorganizational change. In G. Huber & W. Glick (eds.) Organizational change and redesign.147-175. New York: Oxford University Press.
Palepu, K. 1985. Diversification strategy, profit performance, and the entropy measure. StrategicManagement Journal, 6: 239-255.
Pelled, L. 1996. Demographic diversity, conflict, and work group outcomes: An interveningprocess theory. Organization Science. 7: 615-631.
Pelled, L., K. Eisenhardt, & K. Xin. 1999. Exploring the black box: An analysis of work groupdiversity, conflict, and performance. Administrative Science Quarterly. 44: 1-28.
Porter, M. 1986. Competition in Global Industries. Boston: Harvard Business School Press
Prahalad, C. 1990. Globalization -- The intellectual and managerial challenges. Human ResourceManagement. 29: 27-38.
Priem, R. 1990. Top management group factors, consensus, and firm performance. StrategicManagement Journal. 11: 469-479.
Reuber, A. & E. Fischer. 1997. The influence of top management team's international experienceon international behaviors of SMES. Journal of International Business Studies. 28: 807-825.
Ronen, S. & O. Shenkar. 1985. Clustering countries on attitudinal dimensions: A review andsyntesis. Academy of Management Review. 10: 435-454.
Roth, K. 1995. Managing international interdependence: CEO characteristics in a resource-basedframework. Academy of Management Journal. 38: 200-231.
Sambharya, R. 1996. Foreign experience of top management teams and internationaldiversification strategies of U.S. multinational corporations. Strategic Management Journal.17: 739-746.
30
Sanders, W. & M. Carpenter. 1998. Internationalization and firm governance: The roles of CEOcompensation, top team composition, and board structure. Academy of Management Journal.41: 158-178.
Smith, K., K. Smith, J. Olian, H. Sims, D. O'Bannon & J. Scully. 1994. Top management teamdemography and process: The role of social integration and communication. AdministrativeScience Quarterly. 39: 412-438.
Staw, B., L. Sandelands, & J. Dutton. 1981. Threat rigidity effects in organizational behavior: Amultilevel analysis. Administrative Science Quarterly. 26: 501-524.
Stopford, J. 1992. Directory of Multinationals. NY: Macmillan Publishers.
Stopford, J. & L. Wells. 1972. Managing the multinational enterprise. NY: Basic Books.
Sullivan, D. 1994. Measuring the degree of internationalization of a firm. Journal of InternationalBusiness Studies. 25: 325-342.
Wagner, G., J. Pfeffer, & C. O'Reilly. 1984. Organizational demography and turnover in topmanagement groups. Administrative Science Quarterly. 29: 74-92.
Weick, K. & P. Van Orden. 1990. Organizing on a global scale. In N. Tichy (ed.) HumanResource Management. 29: 49-62.
Wiersema, M. & K. Bantel. 1992. Top Management Team Demography and Corporate StrategicChange. Academy of Management Journal. 35:91-121.
Wiersema, M. & A. Bird. 1993. Organizational demography in Japanese firms: groupheterogeneity, individual dissimilarity, and top management team turnover. Academy ofManagement Journal: 36: 996-1026.
Zachary, G. 1996. Major U.S. companies expand efforts to sell to consumers abroad. The WallStreet Journal. June 13: A1.
31
Table 1Descriptive Statistics*
Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13 141 Organizational size 2.42 1.37
2 Return on assets .10 .09 .01
3 Current ratio .90 .51 .06 .06
4 Free cash flow 9.57 15.61 -.02 -.01 .01
5 Diversification level .26 .24 -.05 -.01 .03 .01
6 R&D Intensity .04 .04 .01 .30 .20 .01 -.02
7 TMT size 6.03 1.87 .05 .01 -.01 -.01 .11 .01
8 TMT average firm tenure 21.52 8.26 .01 -.01 -.08 .01 -.03 -.02 -.10
9 TMT turnover .09 .02 .08 -.01 .01 .01 -.08 .01 -.11 .04
10 TMT international experience .18 .33 -.02 .03 .01 -.03 -.08 -.01 .30 .05 .42
11 TMT educational heterogeneity .47 .21 -.01 -.03 -.05 -.04 .06 -.07 .36 -.06 -.04 -.16
12 TMT functional heterogeneity .61 .15 .07 -.01 .01 -.04 -.10 .03 .44 -.16 -.14 -.24 .29
13 TMT firm tenure heterogeneity .46 .26 -.03 .02 .02 -.03 .01 -.03 -.03 -.12 -.02 -.01 -.11 .01
14 Environmental Uncertainty 1.95 .25 -.09 .05 -.02 -.02 .11 .08 .08 -.08 -.05 .04 .06 .10 .07
15 Global strategic posture .81 .44 .01 -.02 -.16 -.02 -.13 .21 .11 .03 .08 .24 .05 .02 .05 .02
____ ____ __ __ __ __ __ __ __ __ __ __ __ __ __ __Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13 14
*n=414 (207 cases multiplied by two years), For bivariate correlations > .12, p < .05, two-tailed tests
32
Table 2Summary of OLS Fixed-Effects Regression Results with Standardized Coefficients:
Dependent Variable is GSP
Model 1 Model 2 Model 3 Model4(Controls) (Main Effects) (Quadratic) (Interactions)
Main EffectsTMT International Experience H1a .17 ** .17 ** .18 **TMT Educational Heterogeneity H1b .19 ** .18 ** .16 **TMT Functional Heterogeneity H1c -.10 * -.09 † -.07 †TMT Firm Tenure Heterogeneity H1d .13 ** .12 * .12 *
Curvilinear EffectsTMT International Exp. Squared H2a .01TMT Educational Het. Squared H2b -.03TMT Functional Het. Squared H2c -.05TMT Firm Tenure Het. Squared H2d -.05
Moderated EffectsEnvironmental Uncertainty X: TMT International Experience H3a .03 TMT Educational Heterogeneity H3b .29 * TMT Functional Heterogeneity H3c -.45 * TMT Firm Tenure Heterogeneity H3d -.19 *
ControlsOrganizational size .07 † .07 † .07 † .07 †Return on assets .06 .05 .05 .06Current ratio -.06 -.06 -.07 -.07Free cash flow .02 .03 .03 .03Diversification level -.05 -.09 † -.09 † -.10 †R&D Intensity .10 * .10 * .10 * .09 †TMT size .11 ** .14 ** .13 ** .11 **Period .09 † .09 † .08 † .05TMT average tenure .08 * .07 † .07 † .05TMT turnover .11 * .02 .02 .03Environmental uncertainty (H3) .11 †
Adjusted R2 .23 ** .29 ** .29 ** .33 **
Change in Adjusted R2 Over Prior Model .06 ** .00 .04 **n=414 (207 cases multiplied by two years), **p<.01, *p<.05, †p<.10