Professional Documents
Culture Documents
Managerial Discretion
Author(s): Sydney Finkelstein and Donald C. Hambrick
Source: Administrative Science Quarterly, Vol. 35, No. 3 (Sep., 1990), pp. 484-503
Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management, Cornell
University
Stable URL: http://www.jstor.org/stable/2393314 .
Accessed: 24/08/2013 04:47
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating
with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.
http://www.jstor.org
?
In an attempt to reconcile these competing views, recent
1990 by Cornell University. theorists have taken a contingency approach. In some cases,
0001 '8392/90/3503-0484/$1 .00.
environments determine organizational forms and fates; in
other cases, managers, through their choices, have a great
This project was supported by the Social role in affecting outcomes (Hrebiniakand Joyce, 1985). As a
Sciences and Humanities Research
Council of Canada, the Irwin Foundation, theoretical bridge between these extremes, Hambrick and
and Columbia University's Executive Finkelstein (1987) developed the concept of managerial dis-
Leadership Research Center. We would
like to thank Warren Boeker, Bert Can-
cretion to refer to the latitude of action available to top exec-
nella, Ming-Jer Chen, John Delaney, Jim utives. Discretion is a means of accounting for differing levels
Fredrickson, Sara Keck, Nandini Rajago- of constraint facing different top-management groups. Where
palan, Mike Tushman, and the anonymous
reviewers for their helpful comments on discretion is low, the role of the top-management team is
earlier drafts of this paper. limited, and upper-echelons theory will have weak explana-
484/AdministrativeScience Quarterly,35 (1990): 484- 503
Measures
Strategic persistence. This variable measures the extent to
which a firm's strategy remains fixed over time. Conceiving of
strategy as a pattern in a stream of important decisions
(Mintzberg, 1978), we used six strategic indicators to create
composite measures of persistence: (1) advertising intensity
(advertising/sales), (2) research and development intensity
(R&D/sales), (3) plant and equipment newness (net P&E/gross
P&E), (4) nonproduction overhead (SGA expenses/sales), (5)
inventory levels (inventories/sales), and (6) financial leverage
(debt/equity). These dimensions were chosen because (a)
they are potentially controllable by top managers; (b) they
may have an important effect on firm performance; (c) they
are complementary, each focusing on an important but spe-
cific aspect of the firm's strategic profile; and (d) they are
amenable to data collection and have relatively reliable com-
parabilityacross firms within an industry. All have been used
in previous strategy research; advertising intensity, R&D in-
tensity, and plant and equipment newness are basic resource
allocations, the SGA to sales ratio addresses a firm's ex-
pense structure, the inventory to sales ratio measures pro-
duction cycle time and working capital management, and the
debt to equity ratio is an accepted measure of financial le-
verage (Schoeffler, Buzzell, and Heany, 1974; Schendel and
Patton, 1978).
The composite persistence measures were calculated as
follows: Treating t as the focal year, the firm's five-year (for
t - 1 through t + 3) variance (E t, - 7)2/n - 1) for each
strategic dimension was computed.2 Next, variance scores
for each dimension were standardized by industry, using data
points from sample firms only (mean = 0, standard deviation
= 1), and multiplied by minus one to bring the measures in
line with the concept of persistence (i.e., absence of strategic
variance over time).
Finally, the six standardized indicators were summed to yield
an overall measure, Strategic Persistence 1 (SP1). Because
two of the indicators-advertising and R&D intensity-had
considerable missing data, particularlyfor the natural-gas dis-
2 tribution industry, an alternative measure, Strategic Persis-
We used this time period because we tence 2 (SP2), was also developed based only on the other
were interested in capturing changes from four indicators (items 3 through 6, above). The missing data
just before the focal year, as well as after.
Hence, this approach captures change by reduced the N for SP1 to 155.
top teams that break with the past. Re-
sults did not change when strategic per- The rationale for creating these composite measures (and
sistence was measured from t to t + 4. those that follow, for strategic and performanceconformity)
491/ASQ, September 1990
Data Analysis
The data includeboth cross-sectionaland time-series compo-
nents. Such data can be exploited by poolingthe cross sec-
tions together for availableyears, producingstatistics that
indicatethe average effect of the independentvariablesover
the full period. Because largersamples are possible, more
precise estimates can be obtained. However, poolingcross-
sectional data presents two problems. First,slope coeffi-
cients may varyover time, renderingpooled techniques
inappropriate(Maddala,1977). And second, ordinaryleast
squares (OLS)estimates may be biased because of enduring
individual-firmcharacteristicsthat are not considered in the
model, violatingassumptions on independence of observa-
tions (Hannanand Young, 1977).
The problemof model instabilityover time was assessed
using a proceduresuggested by Maddala(1977: 322-326).
F-valuesrangedfrom .33 to 1.08, all of which were insignifi-
cant, indicatingthat slope coefficients did not varyover time.
Hence, the data were pooled.
Because of potentialbias in OLScoefficients, we primarily
used modifiedgeneralizedleast squares (GLS)models to
493/ASQ, September 1990
RESULTS
Table 1 presents means, standard deviations, and correlations
among the variables. Although significance levels are noted,
they should be interpreted with caution because the data
have been pooled. There are several patterns worth noting.
First, with an average team tenure of about 22 years, our
sample seems to be weighted toward long-tenured top
teams. However, the standard deviation of tenure was about
12 years, providing significant variabilityfor statistical tests.
Second, team tenure was positively correlated with both stra-
tegic persistence measures and all three conformity mea-
sures, often with substantial magnitude. This preliminary
result is consistent with our hypotheses linking tenure to out-
comes.
Table 1
Pearson CorrelationCoefficients of All Variables in Study*
Variable Mean S.D. 1 2 3 4 5 6 7 8
1. SPi 0 3.8
2. SP2 0 2.6 .93-
3. SC1 -4.6 1.6 .23- .21-
4. SC2 -3.1 1.3 .20- .14- .85-
5. PC -1.7 1.5 .35- .23- .19- .21--
6. Tenure 21.6 11.7 .47- .34- .21- .30- .12-
7. Performance 26.0 12.0 .04 - .01 - .13 - .10- .02 - .12-
8. Size 19.1 40.3 .32- .12- .30- .05 .11- .19- .03
9. Slack 22.0 18.7 -.19- -.25-- .01 -.15- -.14- -.19- .16- -.01
Table2
Results of GLS Regressions on Team Tenure in Firm*
SPi SP2 SC1 SC2 PC
Variable (N = 155) (N = 500) (N = 155) (N = 500) (N = 500)
Third, both size and slack were correlated with strategic per-
sistence, positively for size and negatively for slack. This sug-
gests that firms with persistent strategies were larger and
had less slack than other firms. Size was also positively asso-
ciated with SC1 and PC. Given these correlations, size and
slack were included as controls in regressions on tenure.
Table 2 presents regression results for the full sample. Using
generalized least squares, we regressed each of the five de-
pendent variables on tenure, size, and slack. As noted above,
the five-year average return on equity (performance) was also
included as a control variable for strategic persistence.
Strong support was found for a relationship between team
tenure and measures of persistence and conformity. Four of
five regression coefficients were significant and in the hy-
pothesized direction.4 The association between team tenure
and strategic persistance was especially strong. Firms led by
long-tenured top-management teams were significantly more
likely to follow persistent strategies.5 One of two regression
coefficients for strategic conformity was also significant at p
< .01. Finally, tenure and performance conformity were mar-
ginally related (p < .10) in the hypothesized positive direction.
The control variables, size, slack, and performance, yielded
mixed results. Of note, however, is the positive association
between size and all five dependent variables, with the coef-
ficient in the SC1 model significant. While only marginal, this
4
result does seem to suggest that larger firms tended to make
Followinga suggestion of an ASQ re-
fewer strategic changes. Coefficients for slack were more
viewer,we tested the robustnessof these mixed, although slack and SP2 were negatively and signifi-
resultsfor firmsof varyingages. Thatis, cantly associated. Performance was positively, but not signifi-
high-performing olderfirmsmay see no
reasonto change a successful strategy.If cantly related to the dependent variables.
these firms have long-tenuredtop man-
agers as well, similarpatternsof persis- Table 3 reports regression results for each of the three indus-
tence and conformitymay result. tries. The computer industry represents the high-discretion
However,the inclusionof this age effect environment, where we expected the strongest associations
in regressionswas not significantand did
not change the results. between tenure and organizational outcomes. Conversely,
5
managerial discretion is much more constrained in the nat-
Alternatively,it is possible to arguethat
ural-gas distribution industry, and we expected team tenure
teams with low variancein tenures form to have a weak association with the outcome measures. Fi-
a cohortwhose memberstend to act very nally, the moderate levels of discretion prevalent in the
similarly,amplifyingstrategicpersistence,
but this hypothesiswas not supportedin chemical industry should result in tenure effects somewhere
our sample. between those in the other two industries.
495/ASQ, September 1990
Table4
Results of OLS Regressions on Team Tenure in Firm*
SPi SP2 SC1 SC2 PC
(N = 155) (N = 500) (N = 155) (N = 500) (N = 500)
*0
0~~~~~~~~~~
C
0~ ~~~~~~~~ 0
0-
in ~ ~ ~ TemTnr (nYas
in -2
0~
0~~~
%6.(
U) -
0
0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40
%6-)
c.
X I_ I l l
E
-2-
c; 0
mace-1friylrengtv
E(1
scoes ndcatlss onormty
499/ASQ,~
Setmer19
0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40
Team Tenure(in Years)
*ForStrategicConformityand Perfor-
Limitations
Several important limitations of the study warrant discussion.
First, we restricted our focus to top managers when, in fact,
lower-level employees may be influential in professional firms
or in those with emergent strategies. However, the highest
managers and the CEO retain considerable symbolic influence
that can convey their preferences for lower-level initiatives
(Hage and Dewar, 1973; Pfeffer, 1981). Additionally, agenda
setting at the top serves as an important guide for lower-level
managers to follow (Kotter, 1982). While influence may ema-
nate from numerous parts of the organization, in this study
we have at least expanded our scope beyond the predomi-
nant focus on the CEO alone.
Second, our sample was restricted to large firms, and our
findings therefore may not be generalizable to all firms in an
industry. Managerial effects may be more intense in smaller
firms because they are less constrained by organizational in-
ertia (Miller, Kets de Vries, and Toulouse, 1982).
Another important limitation of the study concerns causality.
While hypotheses were stated in associational terms, the
logic behind them implies that managerial tenure "causes"
organizational outcomes. Unfortunately, even with longitu-
dinal data it was not possible to completely disentangle
causal direction. While the temporal ordering of team tenure
and the strategic persistence measures was consistent with
theory, hypotheses on conformity were tested at one point in
time. Accordingly, it seems plausible that both managerial
tenures and persistence and conformity are mutually rein-
forcing, a view very much in line with Miles and Snow (1978).
Long-tenured teams follow stable strategies, and firms that
follow stable strategies do not make many changes in execu-
tive leadership.
500/ASQ, September 1990
CONCLUSIONS
Our results allow conclusions at two levels. First, it appears
that managerial team tenure has a profound influence on or-
ganizational outcomes, such as strategic persistence, stra-
tegic conformity, and performance conformity. The apparent
effect of tenure occurs over its full range of values and is not
just an artifact of the radical changes brought on by top man-
agers newly arrived from the outside.
At a second level, our results indicate that upper-echelons
theory should be extended to include a moderating role for
managerial discretion. As Pfeffer and Salancik (1978: 245)
have argued, "it is necessary to develop a model that sug-
gests how much constraint an administrator faces in formu-
lating action." This paper illustrates that such a position has
strong empirical support, at least in the context of managerial
tenure.
There is clearly opportunity for further research to investigate
both managerial discretion and upper-echelons theory. The
results reported here addressed managerial tenure; other im-
portant demographic characteristics such as age, functional
background, industry experience, and education may yield
significant findings as well. This paper also suggests that a
more careful examination of the strategic choice paradigm is
required. Such research will greatly improve our under-
standing of the role of executive leadership in organizations.
REFERENCES
Aldrich, Howard Bantel, Karen,and Susan Jackson Bourgeois, L.J.
1979 Organizationsand Environ- 1989 "Topmanagementand inno- 1981 "Onthe measurementof or-
ments. EnglewoodCliffs,NJ: vations in banking:Does the ganizationalslack."Academy
Prentice-Hall. compositionof the top team of ManagementReview, 6:
Arnold, Hugh J. make a difference?"Strategic 29-39.
1982 "Moderatorvariables:A clas- ManagementJournal,10: Brown, M. Craig
sificationof conceptual,ana- 107-124. 1982 "Administrative succession
lytic,and psychometric Barnard,Chester and organizationperformance:
issues." OrganizationalBe- 1938 Functionsof the Executive. The succession effect." Ad-
haviorand HumanPerfor- Cambridge,MA: HarvardUni- ministrativeScience Quarterly,
mance, 29: 143-174. versity Press. 27: 1-16.
Miller,Danny, and Peter H. Friesen Salancik, Gerald R., and Jeffrey Szilagyi, Andrew D., and David M.
1980 "Momentumand revolutionin Pfeffer Schweiger
organizational adaptation." 1977 "Constraintson administrator 1984 "Matching managers to strate-
Academyof Management discretion:The limitedinflu- gies: A review and suggested
Journal,23: 591-614. ence of mayorson city framework." Academy of
Miller,Danny, ManfredF. R. Kets budgets." UrbanAffairsQuar- Management Review, 9:
de Vries, and Jean-MarieToulouse terly, 12: 475-498. 626-637.
1982 "Topexecutive locus of control Schendel, Dan E., and G. Patton Thompson, James D.
and its relationshipto 1978 "A simultaneousequation 1967 Organizations in Action. New
strategy-making,structure,and model of corporatestrategy." York: McGraw-Hill.
environment."Academyof ManagementScience, 24: Tushman, Michael L., and Elaine
ManagementJournal,25: 1611-1621. Romanelli
237-253. Scherer, F. M. 1985 "Organization evolution: A
Miller,Danny, and Jean-Marie 1970 IndustrialMarketStructureand metamorphosis model of con-
Toulouse EconomicPerformance.Chi- vergence and reorientation."
1986 "Strategy,structure,CEOper- cago: RandMcNally. In L. L. Cummings and Barry
sonalityand performancein Schoeffler, Sydney, Robert Buzzell, M. Staw (eds.), Research in
smallfirms."AmericanJournal and D. Heany Organizational Behavior, 7:
of SmallBusiness, 10(3): 1974 "Theimpactof strategicplan- 171-222. Greenwich, CT: JAI
47-62. ningon profitperformance." Press.
Mintzberg,Henry HarvardBusiness Review, 52: Tushman, Michael L., Beverly
1973 The Natureof Managerial 137-145. Virany, and Elaine Romanelli
Work.New York:Harper& Schwenk, Charles, and Ming-Je 1985 "Executive succession,
Row. Tang strategy reorientation, and or-
1978 "Patternsin strategyforma- 1987 "Economicand psychological ganization evolution." Tech-
tion." ManagementScience, explanationsfor strategicper- nology in Society, 7: 297-314.
24: 934-949. sistence." Workingpaper, Vancil, Richard F.
Pfeffer,Jeffrey GraduateSchool of Business, 1987 Passing the Baton: Managing
1981 "Managementas symbolicac- IndianaUniversity. the Process of CEO Succes-
tion: The creationand mainte- Singh, Jitendra sion. Boston: HarvardBusi-
nance of organizational 1986 "Performance,slack, and risk ness School Press.
paradigms."In L. L. Cummings takingin organizationaldeci- Venkatraman, N., and Vasudevan
and BarryM. Staw (eds.), Re- sion making."Academyof Ramanujam
search in OrganizationalBe- ManagementJournal,29: 1986 "Measurement of performance
havior,3: 1-52. Greenwich, 562-585. in strategy research: A com-
CT:JAIPress. parison of approaches."
1983 "Organizational demography." Staw, BarryM.
Academy of Management Re-
In L. L. Cummingsand Barry 1976 "Knee-deepin the Big Muddy:
view, 11: 801-814.
M. Staw (eds.), Researchin A study of escalatingcommit-
Organizational Behavior,5: ment to a chosen course of Walsh, James P.
299-357. Greenwich,CT:JAI action."Organizational Be- 1988 "Selectivity and selective per-
Press. haviorand HumanPerfor- ception: An investigation of
mance, 16: 27-44. managers' belief structures
Pfeffer,Jeffrey, and Gerald R. and information processing."
Salancik Staw, BarryM., and FrederickV. Academy of Management
1978 The ExternalControlof Orga- Fox
Journal, 31: 873-896.
nizations:A Resource Depen- 1977 "Some determinantsof com-
dence Perspective.New York: mitmentto a previously Wanous, J. P.
Harper& Row. chosen course of action." 1980 Organizational Entry: Recruit-
HumanRelations,30: ment, Selection, and Socializa-
Porter, Michael E. 431 -450. tion of Newcomers. Reading,
1980 CompetitiveStrategy:Tech- MA: Addison-Wesley.
niques for AnalyzingIndustries Staw, BarryM., and Jerry Ross
and Competitors.Glencoe, IL: 1987 "Behaviorin escalatingsitua- Weick, Karl E.
Free Press. tions: Antecedents, proto- 1979 The Social Psychology of Or-
types, and solutions."In L. L. ganizing, 2d ed. Reading, MA:
Ranson, Stewart, Bob Hinings, and Cummingsand BarryM. Staw Addison-Wesley.
Royston Greenwood (eds.), Researchin Organiza-
1980 "Thestructuringof organiza- tionalBehavior,9: 39-78.
tionalstructures."Administra- Greenwich,CT:JAIPress.
tive Science Quarterly,25:
1-17. Stearns, Timothy M., Alan N.
Hoffman, and Jan B. Heide
Salancik, Gerald R. 1987 "Performanceof commercial
1977 "Commitmentand the control televisionstations as an out-
of organizational behaviorand come of interorganizational
belief." InBarryM. Staw and linkagesand environmental
GeraldR. Salancik(eds.), New conditions."Academyof Man-
Directionsin Organizational agement Journal,30: 71-90.
Behavior:1-54. Chicago:St.
ClairPress.