- Jnt. Studies of Man.
-
76 Andre Laurent (France)
of the holy principle of the single chain of command and
themanagers' recognition that this was a strong element of theirown
belief system rather than a constant element in nature.
Every manager has his own management theory, his own setof
representations and preferences that in some way guide hispotential
behavior in organizations; and it is critical for man-agers,
management researchers, and educators to identify andunderstand
these implicit theories of management better. Thisconviction that
every manager has his own management gospelprovided the initial
ground for the research reported here, r e -search that was not
originally designed to be comparative.
In order to elicit managers' implicit theories of management,I
had developed a questionnaire proposing 56 different state-ments
about the management of organizations. A five-pointopinion scale
was attached to each statement to record the r e -spondent's degree
of agreement or disagreement with thosestatements. Owing to the
bilingual setting in which the surveywas to be.administered, the
questionnaire was developed inboth English and French
simultaneously rather than formallytranslated from one language
into the other. A great deal ofcare was taken to avoid words,
expressions, or sentence con-structions that could not communicate
fairly equivalent mean-ing in both languages.
The questionnaire was initially administered to a group of60
upper-middle-level managers from various companies at-tending an
executive development program at INSEAD, a Euro-pean management
institute located in France. This first groupwas composed of 40
French managers and 20 managers fromseveral-other European
countries.
The original idea was simply to use the recorded opinionsof the
managers as an input in the pedagogical process. Theobjective was
to explore and discuss the participants' implicit^theories of
management.
Since I was aware of the work conducted by my colleagueGeert
Hofstede, who was, at the time, analyzing his bank ofcomparative
sxirvey data from the Hermes multinational cor-poration (Hofstede,
1980a), I decided to compare the results
-
Diversity in Concepts of Management 77
of the group of 40 French managers with those of the 20
non-French managers. The differences in opinion between thesetwo ad
hoc cultural groups appeared to be so. great on so manyitems of the
questionnaire that they could not be ignored.
These results led to the mainworkinghypothesisof this
sub-sequent research study, namely, that the national origin
ofEuropean managers significantly affects their views of whatproper
management should be. National culture seems to actas a strong
determinant of managerial ideolog>'. The objectiveof the
research thus became to assess and identify some of thenational
differences in concepts of management.
Research Methodology
The 56-item Management Questionnaire was
systematicallyadministered to groups of upper-middle-level managers
at-tending the various INSEAD executive development programsbetween
1977 and 1979. The managers came from a large num-ber of different
enterprises and from a variety of Westerncountries. So as not to
influence the responses, the data werealways collected before
beginning a program. Most of thequestionnaires were administered in
their English version torespondents of all nationalities, all of
whom were fluent inEnglish. For a few French-speaking programs, the
Frenchversion was used.
The first part of this report summarizes the results ob-tained
with 817 respondents from 10 Western countries (9European countries
and the United States). National samplesize varies from a low of 32
for Italy to a high of 219 for France. Within each national sample
there is some variancein terms of function, educational background,
age, types ofcompanies, etc. The only common characteristic of the
re-spondents is their participation in management-education
pro-grams attracting upper-middle-level managers from a
large^number of business firms^ The data-collection strategy
wasdesigned to randomize, as much as possible, all variables
ex-cept nationality. Attempts at controlling other sources of
vari-
-
78 Andre Laurent (France)
ance in the data are presented in the second part of this
report.Statistical analysis of the data was performed by
computing
"ecological" correlations among country mean scores acrossthe 56
items. This approach of ecological factor analysis,which has been
used and advocated by Hofstede (1980a, Chap, 2),considers the group
national culture, in the present case as the unit of analysis.
Thus, correlations are run among coun-try scores, not individual
scores, in an attempt to identifygroups of questions in which the
distribution of scores for thevarious countries shows similar
patterns from low to highagreement across the clustered questions.
These groups ofquestions or indices suggest factors or dimensions
that maymeaningfully differentiate national cultures in terms of
theirmanagerial ideologies.
Four indices or dimensions emerged from the statisticalanalysis.
Three of them cluster three questions each; one ofthem clusters
four questions. The four dimensions have beenlabeled: organizations
as political systems (Table 1), organi-zations as authority systems
(Table 2), organizations as role-formalizatioh systems (Table 3),
and organizations as hier-archical-relationship systems (Table 4),
Ecological correla-tions among country mean scores computed across
the 56items and leading toward the selection of these four
dimen-sions are presented in Table 5,
The four dimensions or indices do not pretend to be exhaus-tive.
They have been selected strictly on the basis of the highlevel of
statistical association among countries, the number of items they
cluster, and the conceptual meaning of the clus-tered items.
The indices represent attempts to capture a structure
ofcollective managerial ideologies that meaningfully
differen-tiates national cultures. They do not account for
individual ways of thinking within a given culture. Indeed, whereas
cor-relations among country scores are very high across the
clus-tered items within a given index, correlations among
individualscores for a given country within the same index have
provedto be remarkably low, (1) Once again, the purpose here is
not
-
Diversity in Concepts of Management 79
to analyze the structure of individual opinions, but to
comparecountries.
Although statistical analysis was performed initially on
coun-try mean scores from five-point opinion scales, the resultsare
presented here in terms of percent average agreementscores
("strongly agree" and "tend to agree" responses) andpercent average
disagreement scores ("strongly disagree" and"tend to disagree"
responses) for ease of reading and interpre-tation. Each table thus
presents the percentage of managersagreeing or disagreeing with
each statement and their percentaverage score across the clustered
questions (last row of thetable) for each country. This last
measure indicates the coun-try score or position on the index. In
each table, countriesare ordered from left to right according to
their increasingdegree of identification with the measured
dimension.
Results
Organizations as Political Systems (Table 1)
The dimension "organizations as political systems" clustersthree
items dealing, respectively, with the political role playedby
managers in society (item No. 40), their perception of
powermotivation within the organization (item No. 49), and an
assess-ment of the degree to which organizational structures
areclearly defined in the minds of the individuals involved
(itemNo. 33). In countries such as France and Italy, where
manag-ers report a stronger perception of their political role in
so-ciety, they also emphasize the importance of power
motivationwithin the organization and report a fairly hazy notion
of or-ganizational structure. Danish and British managers, on
theother hand, express a significantly lower political
orientation,both within and outside the organization, and a clearer
notionof organizational structure.
These results may provide some insight into the extent towhich
managers from different countries tend to interprettheir
organizational experience in power terms. Clearly, the
-
80
AlVlI
3DIJVad
Qi'lVia^ZilHS
V s n
N3a3HS
Afmv!a39
SaNV1H3H13N
tJiviiaalV3a9
>/COL U
cc
szz:
1COCE.
1
CNI
cn
r o
oLA
oLA
CSJ
CSI
o -
CD
COl l J
CDO -
L U
UGH T
p i
CD
LU
CO
LUr
_ J
TICA
Q_
TANT
IMPO
R
1
L A
I O
CNIcr
CNI
UDCSI
CNIr o
LnCNJ
CO
sL U
CD
Ot2 :LUt o
" CXLU
T WI
TMA
NA
^ 1L'.' r^
s SOilCD cn
B ^
COL U
L U
CD
INGL U
CO
^ ^
LuHOd 9
mCD
WITH
1N3W
1 1 1
L U
ISA
. . .
i-HLO
roLA
OOr o
CDro
ro
LOro
CNJ
CNICNI
H -
CD
O
Oz:
L U
-a:
LU
- ^
COLU
1t o
r o
LU
CTUP
:::1_ j
OIJ
r^
ZL
. i< C
L U
CDCD
CMCD, ^
LOCO
CVJ
CDCOCDCOCMCOCDCM
"' '~-
h-LJ
LJLJ t-cr 2.(3) LJ
LU LJ
o> CO< Q
OO
-
Diversity in Concepts of Management 81
political orientation of Italian (index score 66) and French
man-agers (index score 62) appears much stronger than the
politicalorientation of Danish (index score 26) and British
managers(index score 32). Furthermore, these findings indicate an
in-teresting association at the country level between a lower
in-clination toward political behavior and a greater
perceivedclarity of structure (in Denmark and Great Britain) versus
agreater inclination toward political behavior and a greater
per-ceived haziness of structure (in Italy and France).
I should like to suggest at this point that it may not be by
ac-cident that the contrasting results demonstrated by the
Britishand the French managers seem to parallel to a
considerableextent the contrasting perspectives on organizations
taken adecade ago by mainstream schools of organizational
sociologyon each side of the Channel. While the British Aston
Schoolresearchers were conducting their rational analyses of
struc-tural characteristics of organizations, French
sociologistswere describing organizations as sets of games and
powerstrategies played by actors seeking to maintain some
uncer-tainty around their function so as to play even more
powergames. Organizations were certainly perceived and definedfar
more as political systems by the latter than by the former.Thus,
cultural differentiation may affect not only managers'implicit
concepts of organizations but also researchers' ex-plicit theories.
Organization and management theory may beas much culturally bounded
as the actual processes of organiz-ing and managing (Brossard and
Maurice, 1974; Derossi, 1978;Hofstede, 1980b, 1981).
In summary, this index suggests some important effects
onorganizational behavior of cultural differences in the
politicaloutlook of managers in neighboring Western countries.
Organizations as Authority Systems (Table 2)
"Organizations as authority systems" groups three ques-tions
dealing with a conception of hierarchical structure asbeing
designed to specify authority relationships (item No. 14),
-
82
3D;iid
AHVil
wni9i3a
SailVTd3Hi3N
Nitf irdaitf3a9
N3a3MS
)iawiN3a
AMWd39
afjvia3ZiiHs
V s n
817 MA
NAGE
RS FR
OM TE
N CO
UNTR
IES
cnC.J
LA-a-
CVJ.13-
190
CDLn
crLn
l O
C3Ln
SAMP
LE SI
ZE
Ln
otrv
uo
oo
oo
LO
Lnro
LnCSJ
oo11
% AG
REEM
ENT
WITH ;
14, TH
E MA
IN RE
ASON
FO
R HA
VING A
HIER
ARCH
ICAL
STRU
CTUR
E IS
SO
THAT
EV
ERYO
NE KN
OWS
WHO
HAS
AUTH
ORIT
Y OV
ER WH
OM.
LO
CTLO
. 3 -LO
CO
ha^T
OO
ODLO
^
oc
*(to
LOKO
LO
CMLn
cr
oLn
X AG
REEM
ENT
WITH :
43, TH
E MA
NAGE
R OF
TO
IIOR
ROW
WILL
BE
, IN
TH
EMA
IN, A
NEGO
TIAT
OR. .
LOCD
CD
CDO)
CO' ^
CDTi-CD
COCJCOoCO
% AV
ERAG
E AG
REEM
ENT
ooo :ooCD
J3ceo
-
Diversity in Concepts of Management 83
a perception of authority crisis in organizations (item No.
52),and an image of the manager as a negotiator (item No. 43).
Itdifferentiates three country clusters of managers' perceptionof
organizations as authority systems. Latin countries suchas Belgium,
Italy, and France, at the upper end of the con-tinuum (index score
61 to 65), present a sharp contrast to coun-tries such as the
United States, Switzerland, and Germany atthe lower end (index
score 30 to 34), and the remaining fourcountries fall in the middle
(index score 46 to 49).
The belief that "The main reason for having a
hierarchicalstructure is so that everyone knows who has authority
overwhom" is associated across countries with the perception
that"Today there seems to be an authority crisis in
organizations"'and the expectation that "The manager of tomorrow
will be, inthe main, a negotiator." However, national culture
stronglyaffects the popularity of such conceptions. For instance,
or-ganizations are seen significantly more frequently as
authoritysystems by French managers (65 percent average
agreementrate across items) than by American managers (30
percentaverage agreement rate).
French, Italian, and Belgian managers report a more per-sonal
and social concept of authority that regulates relation-ships among
individuals in organizations. American, Swiss,and German managers
seem to report a more rational and in-strumental view of authority
that regulates interaction amongtasks or functions. For the former,
authority appears to bemore a property of the individual; for the
latter, it appears tobe more an attribute of the role or
function.
This index reveals important national variations in man-agers'
views of authority in organizations that are likely toinfluence
their behavior.
Organizations as Role-Fornialization Systems (Table 3)
The three items clustered in "organizations as
role-formali-zation systems" all focus on the relative importance
of defin-ing and specifying the functions and roles of
organizational
-
84
( atlVlHHZilMS
AlVil
wnisiBa
jDiJvadNiviiaa
iV3a9
)iaV;lN30
sa;j\n~d3Hi3N
V s n
N3a3HS
817 MA
NAGE
RS FR
OM TE
N CO
UNTR
IES
N-\
LTV
cn11
190
crLO
CNJcr
CDLn
CDLO
SAMP
LE
SIZE
(cn
CT>oo
enscn
cnoo
oo
UDoo
oo
1
t o
Ln%
AGRE
EMEN
T WI
TH .
1, WH
EN TH
E RE
SPEC
TIVE RO
LES
OF TH
E ME
MBER
S OF A
DEPA
RTME
NT BE
COME
CO
MPLE
X, DE
TAIL
ED JO
B DE
SCRI
P-TI
ONS
ARE A
USEF
UL HA
Y OF
CL
ARIF
YING
.
cn
=Tcn
oo
N ^OO
ir>oo
moo
cr.
cn
LO
I AG
REEM
ENT
WITH ;
13, TH
E MO
RE CO
MPLE
X A
DEPA
RTME
NT'S
AC
TIVI
TIES
,
THE
MORE
IM
PORT
ANT
IT IS
FO
R EA
CH IN
DIVI
DUAL
'SFU
NCTI
ONS
TO BE
WE
LL-D
EFIN
ED,
11
r v .
cnLO
11
OOUD
L O
CSILn
^ '
oLn
t DI
SAGR
EEME
NT WI
TH ;
38,
MOST
I1
ANAG
ERS
WOUL
D AC
HIEV
E BE
TTER
RE
SULT
SIF TH
EIR
ROLE
S WE
RE LE
SS PR
ECIS
ELY
DEFI
NED,
inCOinCO
CDT
100oCD
oCO
CDCDCD
i n
% AV
ERAG
E AGREEMENT./
DISA
GREE
MENT
CXI
c:
I
-
Diversitj- in Concepts of Management 85members. 'They stress the
values of clarity and efficiencythat can be obtained by
implementing such organizational de-vices as detailed job
descriptions, well-defined functions, andprecisely defined roles.
Here the results seem essentially toindicate a relatively lower
insistence on the need for roleformalization in Sweden, the United
States, and the Netherlands(index score 57 to 67) than in the
remaining seven countries(index score 80 to 85).
Hence, there may be national variations in the degree
offormalization, often considered an important structural
char-acteristic of organizations, that is judged desirable.
Organizations as Hierarchical-Relationship Systems (Table 4)
The last index, "organizations as
hierarchical-relationshipsystems," which groups four questions,
shows sharp differ-ences in management attitudes toward
organizational relation-ships as one moves from Northern Europe and
the United Stateson the lower end of the continuum to the Latin
countries ofEurope on the higher end.
Across countries, the dream of eliminating conflict from
or-ganizations is associated with the belief that a manager
shoulddefinitely know more than his subordinates and that
organiza-tions should not be upset by such practices as bypassing
orhaving to report to two bosses.
As suggested elsewhere (Laurent, 1981), an index such asthis one
can provide some assessment of the feasibility of neworganizational
arrangements such as the matrix structure that deviate from more
classic hierarchical forms. Indeed,in matrix-type organizations, as
opposed to more classic hier-archies, potential conflicts of
interest about resources tend tosurface more, bosses can no longer
pretend to have answersto most of their subordinates' questions,
bypassing of authoritylines becomes more of a way of life, and,
obviously, some man-agers have to report to two or more bosses.
Thus, the contrasting results obtained from Swedish man-agers
(index score 25) and Italian managers (index score 66)
-
86
AlVi l
3DN\Ald
wni3i3a
ANVWy39
aN\na3Zi!HS
)iyvuN3aNiviiaa
LV3y9
SaNVia3Hi3N
V s n
N3a3HS
817 MA
NAGE
RS FR
OM TE
N CO
UNTR
IES
ro
cn11CSJ
LA.3-
CSJr^
r--\CO
Ln
ocn1(
es j
OLA
OLA
SAMP
LE SI
ZE
cr
.3"CNJ
r*vCO
to11
oo1(
cn1
r*^
CD
cr
% AG
REEM
ENT
WITH :
19, MO
ST OR
GANI
ZATI
ONS
WOUL
D BE BE
TTER
OF
F IF
CONF
LICT
CO
ULD
BE EL
IMIN
ATED
FO
REVE
R,
in
hn
^
CO^a-
oo
CSi
r^CSI
oor-H
O1(
X AG
REEM
ENT
WITH :
24,
IT IS
IM
PORT
ANT
FOR A
MANA
GER
TO HA
VE AT
HAND
PR
ECIS
E AN
SWER
S TO
MO
ST OF
TH
E QUE
STIONS
THAT
HI
S SU
BORD
INAT
ES MA
Y RA
ISE
ABOU
T TH
EIR
WORK
.
LAr^
Cs) ^ ^
CsJcr
CO ^
1t
N-V
firo
o^ro
CNJro
csjCSJ
1 % DI
SAGR
EEME
NT Wi
tH ;
2. IN
OR
DER
TO HA
VE EF
FICI
ENT
WORK
RE
LATI
ON-
SHIP
S, IT
IS
OF
TEN
NECE
SSAR
Y TO
BY
PASS
TH
EHI
ERAR
CHIC
AL LI
NE,
OO
rooo
s
r^
KO
cnuc
CDCO
^zrLn
LO
% AG
REEM
ENT
WITH ;
8, AN
OR
GANI
ZATI
ONAL
ST
RUCT
URE
IN WH
ICH
CER-
TAIN
SU
BORD
INAT
ES HA
VE TW
O DI
RECT
BO
SSES
SHOU
LD BE
AV
OIDE
D AT
AL
L CO
STS,
CDCDOinoin
" ^
CO
CO
CDCOCOCOCDCVinCNJ
% AV
ERAG
E AG
REEM
ENT/
DISAGREEMENT'
ooa..
cr.
.a:
-
Diversity in Concepts of Management 87
Table 5
Ecological Correlations*
Items4049
Index I(politics)
490.73
Index III
33-0.85-0.89
(formalization)Items
113
130.95
38-0.91-0.83
Index II(authority)
Items1452
Items1924
2
52 430.86 0.69
0.88
Index IV(hierarchy)
24 2 80.88 -0.90 0.82
-0.87 0.80-0.64
*Pearson r among country mean scores. Coun-try mean scores per
item were computed for eachof the ten national samples on the basis
of the man-agers' responses to the five-point opinion scale
at-tached to every questionnaire item.
suggest that matrix-type organizational arrangements mighthave
better prospects in Sweden than in Italy. National varia-tions in
conceiving organizations as hierarchical relationshipsystems may
affect the structuring of organizations in differentcountries and
have implications for the transfer of organiza-tional forms across
cultures.
Limitations of the Findings
The four dimensions reported above that seem to differen-tiate
national cultures in terms of their managerial ideologiesrepresent
one attempt at mapping some of the cultural differ-ences in
concepts of management and organization within theWestern world.
Looking at organizations as symbolic systems
-
88 Andre Laurent (France)
of social representation, this research elicits findings that
em-phasize the need to recognize and identify cultural
specificityas a critical element in the texture of organizations.
Further-more, the reported cultural diversity poses significant
chal-lenges for both management theory and practice and
seriouslyquestions claims of universality in both.
These findings have important limitations, however.'. First,
they reflect the limitations of the questionnaire itself.Another
researcher with different interests would have deviseddifferent
questions and would have obtained other dimensionsof
differentiation among cultures.
Second, these findings probably reflect the French
culturalidentity of the author through the questions he thought of
andincluded in the list in the first place. Indeed, it may not be
ac-cidental that whereas the positions of other countries tend
tovary from index to index, the position of the Latin
countries,including France, is on the high-score side on all four
indices.Thus, the research findings indicate that French
managers,more than managers from non-Latin countries, tend to
vieworganizations as political, authority, role-formalization, and
hierarchical-relationship systems.
From this perspective, the findings provide clearer indica-tions
about Latin countries such as France than about othersin the
sample. They reveal the widest gap in conceptions ofmanagement
between the Latin cluster (France, Belgium, Italy)and the Nordic
cluster of America and Europe (United Statesand Sweden). Would a
British questionnaire designer have -eli-cited dimensions
illustrating the same gap? Even though Hof-stede's work (1980a)
partly eliminates such questions by inter-correlating the results
of various independent comparativestudies, and in spite of attempts
by multicultural teams of r e -searchers to design research tools
jointly, the methodologicalchallenge persists.
A third limitation of the findings stems from the limitednumber
of countries represented in the sample. Similar datasubsequently
collected from Japan and Indonesia do indicatethe Western bias of
some of the dimensions. For instance.
-
Diversity in Concepts of Management 89
Japanese managers obtain both high and low scores on
ques;-tionnaire items regrouped in the same Western index,
thuschallenging the validity of the ecological factor. The
validityof the four dimensions is therefore restricted to the ten
West-ern countries represented in the initial sample.
Finally, one may seriously wonder whether the ad hoc sam-pling
method of surveying relatively small groups of managersattending
executive development programs at INSEAD is validfor making
inferences about national cultures. Since .legitimatequestions
could be raised concerning the representativenessof the population
surveyed in the initial study, several subse-quent studies were
conducted to assess the validity of the pre-liminary results.
Replications of the Study
A partial replication of the original survey was initially
ob-tained by administering the same Management Questionnaireto a
much younger population of British and French MBA stu-dents at
INSEAD. The results, reported elsewhere (Laurent,1981), confirmed,
on a preliminary index, the differences ob-served between British
and French experienced managers.
Another partial replication was performed with a group of55
French MBA students from ISA, a French business school.Their index
scores are reported in Table 6, along with the in-dex scores of the
experienced French managers from the ini-tial survey. The scores of
these French MBA students runparallel to the scores of the
experienced French managers,with a tendency toward perceiving
organizations even more asrole-formalization and
hierarchical-relationship systems thanthe managers.
Fuller replication was then obtained in two, large, U.S.-based,
multinational companies. Since the initial results hadbeen obtained
from ad hoc samples of managers in executiveprograms, the question
arose of whether similar differencesin management and
organizational concepts would persist with-in the potentially
homogenizing, corporate culture of a single
-
90 Andre Laurent (France)
Table 6
Index Scores of French MBA Studentsand French Experienced
Managers
Subjects55 French
MBA students(ISA)
219 Frenchexperiencedmanagers(INSEAD)
Index I(politics)
61
62
Index II(authority)
64
65
Index III(formali-
zation)
90
81
Index IV(hierarchy)
66
50
multinational company, or whether the multinational culture
wouldbe sufficiently strong and pervasive to swamp national
differences.
National versus Multinational Cultures
The MNC-A Study
In order to test the above hypothesis, carefully matched
na-tional samples of managers were selected from a large,
U.S.-based, multinational, chemical firm (MNC-A) with
subsidiariesin France, Germany, and Great Britain. Every attempt
wasmade through sampling to control for all conceivable sourcesof
variance other than nationality. A description of
samplecharacteristics is presented in Table 7.
Index scores are reported for the three national groups inTable
8, along with the index scores obtained by the ad hocINSEAD
national samples from the original study. Althougha few variations
perhaps due to corporate culture did ap-pear, the results clearly
indicated the consistent and pervasive
-
Diversity in Concepts of Management 91
Table 7
MNC-A Study Survey Sample Characteristics
Characteristics France488042
Germany399143
GreatBritain
569740
Sample sizePercentage of response rateAverage ageAverage number
of years of
work experience 17 18 18Average number of years of
service with company 15 15 14Average number of years
lived abroad 2.5 1 1.5Percentage of respondents work-
ing in prunarily_line_osition^ 70 65 59Percentage of respondents
work-
ing in prijnarUy_staff_^oition 30 35 41Percentage of respondents
with
higher education at universitylevel 91 81 85
Percentage of respondents withmajor professional experiencein
technical fields (production,engineering, maintenance, etc.) 60 58
59
Percentage of respondents withmajor professional experiencein
marketing 18 25 23
Percentage of respondents withmajor professional experiencein
administration 22 17 18
effects of national cultures for the three countries
involved.The average distance among the three countries' scores
acrossthe four indices remained essentially the same (INSEAD
sam-ple = 13.3; MNC-A sample = 13.0), as did the average range
-
92
S o
rt
CO
tnutDCi
a
QW
-aci
O
4
Oo M tn
C O
rt
rt
tD
> ^(M CO
tn
I CDQ