Вы находитесь на странице: 1из 19

The Contingency Theory of Managerial Accounting

Author(s): David C. Hayes


Source: The Accounting Review, Vol. 52, No. 1 (Jan., 1977), pp. 22-39
Published by: American Accounting Association
Stable URL: http://www.jstor.org/stable/246029 .
Accessed: 18/06/2014 20:09

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.

American Accounting Association is collaborating with JSTOR to digitize, preserve and extend access to The
Accounting Review.

http://www.jstor.org

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
THE ACCOUNTING REVIEW
Vol. LII, No. I
January 1977

The Contingency Theory of


Managerial Accounting

David C. Hayes

ABSTRACT: Working mainly from the literature of modern organization theory, it was
possible to develop a model of organizational performance for internal subunits. The model
hypothesizes three major contingencies which affect subunit performance: (1 ) factors
internal to the subunit (internal factors); (2) interrelationships with other subunits (inter-
dependency factors); and (3) interactions external to the firm (environmental factors).
The contingencies operate differentially across organizational subunits. Three propositions
derived from the model were examined empirically in a field study of large manufacturing
organizations. Strong support for two of the propositions was found, along with indica-
tions of support for the third, by using the analytical methodology of path analysis. The
final section of the paper explores the implications of the study approach and results for
managerial accounting theory and practice.

T HE role of managerial accounting in not provide a sufficiently rich description


business organizations generally is of performance. On the other hand, or-
perceived to be the provision of in- ganization theory provides a far more
formation for decision making by various complete approach to organizational
elements of the organization's hierarchy. assessment, in general, and subunit evalu-
The use of such information in the evalu- ation, in particular.
ation of performance is a critical aspect of The structure of the paper is as follows.
the decision-making function. Despite Contingency theory research on assess-
the Demski paper [1972], the impact of ment and a model derived from the liter-
analytical research in information eco- ature are outlined. Three propositions
nomics has tended to divert attention implied by the model are suggested, fol-
from performance assessment,1 one of lowed by a description of the empirical
the three managerial accounting func- research undertaken to test the model.
tions discussed by Simon, et al. [1954] in
their landmark study. 1 See, e.g., Feltham [1972], Horngren [1972] and
Demski and Feltham [1976].
The managerial accounting tools of
performance evaluation are restricted Many helpful comments on an earlier draft of this
paper were made by Philip Brown, Gerry Feltham and
largely to financially quantifiable data. members of the UBC Accounting Workshop. Tom Burns
Although in textbook discussions such and Jack Baumler provided a great deal of advice and
data are intended to serve a signal gener- assistance during the performance of the research.
ating function, implicit within its use is
the assumption that performance can be This paper was a 1976 manuscript con-
summarized and captured by a dollar test winner. David C. Hayes is Assistant
number. The rationale for the research Professor at the University of British
reported here is that this approach does Columbia.
22

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 23

The two dimensions concern the rela- associated with performance. This model
tive stability and homogeneity of the is presented in Figure 1.
environment faced by subunits. It is sug- The model is not intended to suggest
gested that the uncertainty associated that all aspects of each dimension are
with dealing with elements of the environ- appropriate for the evaluation of all sub-
ment is lowest in Cell I and greatest in units. Rather, only some aspects of each
Cell IV but that no ordering can be placed dimension would seem applicable for dif-
on the relationships of the other situa- ferent types of subunits. Some of these
tions in such a typology. are outlined in the propositions pre-
The nature of the environment faced sented.
by a subunit will determine the relative THE ASSESSMENT MODEL
difficulty such a unit will encounter in
dealing with environmental elements. It Two major contributions to a con-
would appear a priori that the more un- tingency theory of organizations are pro-
certain the environment, the more diffi- vided by Thompson [ 1967] and Lawrence
cult will be the tasks of the subunit and and Lorsch [1969]. Although conceptu-
the greater the extent of environmental ally abstract, Thompson deals more com-
influence on their functions. Thus, type pletely with the contingent aspects of
of environment is another contingent performance than Lawrence and Lorsch;
aspect of performance, leading to con- thus, the Thompson study serves as the
tingent methods of assessment. basis for much of the proposed model
The other major contingency aspects of development.
performance evaluation concern the in- The essence of any evaluation task can
ternal structure and functions of a sub- be captured by two dimensions per-
unit. The nature of the tasks performed, formance criteria and the ability to dis-
types of people, interpersonal relation- cern cause-effect relationships. Thomp-
ships and the ability to measure and son diagrams these via a 2 x 2 matrix as
quantify functions are all likely to vary presented in figure 2.
with subunit type. Situations posited by Cells I and II
A three-dimensional model of the as- theoretically provide little challenge to
sessment function can be constructed on the development of assessment devices
the basis of the three major contingencies and, consequently, encompass most of

FIGURE 1
THE ASSESSMENT MODEL
Productivity
______________________ Cost behavior
Internal Supportive relations
Manpower utilization
Work group cohesion

_ Reliability
Departmental
Effectiveness Interdependency < Cooperation
Flexibility
Planning ability
<_ Share of market
Environmental Dealer opinions
-\ Environmental stability
Environmental diversity

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
24 The Accounting Review, January 1977

FIGURE 2 that they belong to the same organiza-


POTENTIAL ASSESSMENT SITUATIONS tion; no exchanges occur between them,
(From Thompson [1967, p. 86])
and coordination involves the establish-
Beliefs about Cause-Effect Relationships ment of standardized procedures and
Complete Incomplete rules.
Sequential interdependence occurs in
cases where the outputs of one subunit are
Crystallized inputs to another, and thus a serial non-
symmetrical relationship exists between
Standards I II
of
them. Careful planning of the interde-
Desirability pendence is important because one unit
cannot act prior to the actions of another.
Ambiguous
However, some degree of flexibility and
III IV buffering can be built into such an ar-
rangement so that temporary failures by
one subunit will not jeopardize substan-
the currently employed managerial ac- tively the operations of the other.
counting assessment tools. However, A reciprocal interdependence occurs
Cell II does involve uncertainty, typified in cases where there is a symmetrical flow
by the allocation procedures employed between two subunits; i.e., an input to
and proposed by accountants. Cells III one is an output of the other. Coordina-
and IV of the matrix suggest difficulties tion in this complex situation may occur
not generally addressed by managerial via a process of mutual adjustment, with
accounting. If standards are ambiguous no fixed plans or rules. If a continual con-
(but not necessarily multiple and con- sultative relationship occurs, the poten-
flicting) the evaluation procedures em- tial exists for new information and ideas
ployed by accountants may fail. These to be introduced which may change the
cells suggest assessment methods con- intended nature of the activities from
tingent on the particular situation faced those originally specified. Thus, depend-
by the organizational subunit of interest; ing on the type of interdependency exist-
the matrix as a whole implies combina- ing between any two subunits, different
tions of contingent assessment proce- methods of coordination and, conse-
dures. quently, assessment can be proposed.
The contingency aspects explored The type of environment faced by an
within this study concern subunit inter- organization and its subunits can be
dependence, environmental relationships viewed similarly on several dimensions.
and factors internal to the particular sub- One typology suggested by Thompson is
unit of interest. presented in Figure 3.
Interdependence can be examined in
the typology suggested by Thompson Previous Research
[1967, Ch. 7] following March and Prior literature on contingent assess-
Simon [1958]. The interdependent rela- ment is limited. Most organizational
tionships and the procedures for their research in the area has been ad hoc in
coordination vary with organizational that the intent appears to have been data
complexity. Stable, pooled interdepen- gathering and examination rather than
dence occurs in those cases where the exploration in the context of well-struc-
only connection between two subunits is tured models and hypotheses. Despite

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 25

FIGURE 3 these reports there were indications that


ENVIRONMENTAL REPRESENTATION assessment methods differed across sub-
(From Thompson [1967, p. 72]) units.
Stable Dynamic
A study on production subunits of
varying size by Wahba and Shapiro
[1973] found results similar to the Ma-
H omogeneous
honey studies. The three dimensions they
posit with respect to effectiveness repre-
sent an expansion of the internal and
interdependency dimensions proposed in
this study.
Heterogeneous Other studies by Georgopoolos and
III IV
Tannenbaum [1957], Friedlander and
Pickle [1968], and Seashore and Yucht-
man [1967] highlighted major variables
this approach, the results are relatively in the assessment of organizational sub-
uniform and tend to provide empirical units which can be categorized easily as
support for the contingent model pro- corresponding to the three dimensions of
posed previously. the Figure 1 model. All of these studies
The studies by Mahoney [1967], Ma- were in a slightly different context than
honey and Weitzel [1969] and Mahoney both the others cited and the research
and Frost [1974] represent several dif- here.
ferent examinations of the same data. Accounting-oriented studies on effec-
Initially, the opinions of managers of tiveness are limited to those by Kinney
different organizational subunits were [1969] and Solomons [1965]. Other ap-
solicited on variables a priori related to proaches, such as the Hopwood [1974]
subunit effectiveness. The same managers and Klammer [1973] studies, are of only
also were requested to rate the overall marginal relevance to this research.
effectiveness of the subunits. This rating Kinney examined the effect on account-
was used as the dependent variable in the ing performance measures of factors out-
analyses of the latter two reports. In the side the control of the manager being
first study, the variables were factor- assessed. The approach to assessment
analyzed; the four highest loading fac- was narrow and only analyzed surrogate
tors were interpreted to mesh neatly into measures of management's ability to
the three dimensions of the model pro- cope with the task environment faced.
posed in Figure 1. Two of the factors, Although Solomons broadly addressed
planning and initiative, appear to be the assessment problem, its applicability
related to dealing with the task en- is limited due to the primarily verbal
vironment; the others, reliability and analysis performed.
productivity, appear to be related to the There is a limited amount of prior re-
interdependency and internal factors, search in the area of concern to this
respectively. study; it is apparent that the proposed
The second report utilized these fac- model can satisfactorily explain such
tors in a regression analysis, and the research. The following section outlines
third classified subunits on the basis of some propositions which may be drawn
the dominant technology of the unit prior from the model, followed by the research
to performing a path analysis. In both design and discussion of results.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
26 The Accounting Review, January 1977

Propositions the major evaluation factors employed.


Three propositions of interest can be Proposition 3. Environmental and in-
derived with respect to the model just terdependency variables provide approx-
outlined. imately equal contributions to the
explanation of the performance of
Proposition 1. Internal variables are marketing departments.
the major explanators of performance for A boundary-spanning unit such as
production departments, with a lesser marketing deals primarily with environ-
contribution made by interdependency mental constraints and contingencies
variables. operating on the firm. Thus, the nature of
Within production departments, it is the environment faced would be of criti-
expected that variables classified as in- cal importance in defining the role of
ternal will best explain their effectiveness. such a department. The role of a market-
Both cause-effect relations and standards ing unit is perceived to be two-fold,
or objectives are expected to be relatively thus defining the important assessment
well-established for these departments. variables. On the one hand, its relation-
Output and cost objectives of a produc- ships with production departments are
tion subunit can be evaluated readily by critical in that there is a sequential rela-
normal accounting techniques, such as tionship between them for products; the
standard cost variance analysis and rate output of production is an input to
of return. Thus, it is expected that inter- marketing. Evaluation of this interde-
nal variables will explain most of the pendence would seem relevant. On the
effective performance herein. However, other hand, its interrelationship with the
given the interrelationships with market- environment to which it transfers this
ing, on the product output side, and product represents its primary task for
R & D, on the technological input side, the firm. Thus, another element in the
interdependency factors are expected to overall marketing assessment is expected
have some explanatory power. to be the evaluation of this relationship.
Proposition 2. Interdependency vari-
ables are the major explanators of per- RESEARCH DESIGN AND METHODOLOGY
formance for research and development Manufacturing firms in Ohio fitting
departments. two major characteristics were chosen
Ambiguous desirability standards in for study. To overcome the problem of
research and development departments comparing the responses of individuals
suggest that the use of internal variables from very small with very large firms, an
in the evaluation of their effectiveness arbitrary size limit was selected as a cut-
will have limited impact in explaining off rule for firm inclusion. Although dif-
their performance. The extent to which ferent indicators of size can be used, such
they achieve budgeted cost levels in a as total assets or sales, number of em-
given period gives little insight into their ployees is a common measure in the
outputs or its value to the other subunits literature.2 Firms employing more than
of the firm which it serves. This implies 500 persons were selected for inclusion
that the nature of the interdependent in this study. The other major character-
relationship it possesses with these units, istic of the firms chosen was that each
and the degree to which it fulfills the ex- possessed research and development,
pectations of other units via the coopera-
tion achieved between them, would be 2 See, e.g., Pugh, et al. [1968].

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 27

TABLE I
ANALYSIS OF QUESTIONNAIRE RETURNS

Production Marketing R&D


Questionnaires Departments Departmients Departments Controllers Totals

Sent 107 109 105 109 430


Returned 55 65 70 84 274
Return percentages 51.4 59.6 66.7 77.1 63.7

marketing and production departments. for questionnaire studies.


Two questionnaires sent to different Two further comments concerning
individuals in each firm were prepared. Table 1 are necessary. First, it merely
The major questionnaire was sent to the examines the total number of returns
marketing, production and research and from each department group and does
development managers in each of the 109 not indicate the number of firms from
firms which met the required characteris- which more than one questionnaire was
tics. Two sets of responses were solicited returned. This information is provided
from each manager; one set dealt with in Table 2. Second, the response rates
their perception of internal and external differ markedly among departments,
factors that influence departmental per- most noticeably between research and
formance, and the other dealt with their development (66.7 percent) and produc-
perceptions of the effectiveness of inter- tion (51.4 percent). This may be due to
dependent departments. The second greater interest in general research by
questionnaire solicited from the con- R & D managers and their greater fa-
trollers of the respective organizations a miliarity with questionnaire surveys. The
judgment on the effectiveness of the large controller difference is probably
respective departments in the study. This due to the shortness of the questionnaire
judgment, independent from all of the and the inherent interest in an accounting
departments, was used as the dependent study.
variable in the path analysis.
TA13LE 2
Of the 430 questionnaires sent (4 to QUESTIONNAIRE RETURNS FOR 109 FIRMS IN SAMPLE
each of 103 firms, 3 to 6 others where
one department or another had indicated One
0o1
Complete
an unwillingness to participate), an over- Set of Three Two No
all response of 274, or 63.7 percent, was Questionnaire Retulrns Returns Retulrns Retulrns
achieved.3 This overall response rate is Number 24 30 29 26
broken down by departmental classifica-
tion in Table 1. As the questionnaire sent
to controllers was far shorter than that The nature of the research design
sent to the respective department man- necessitated returning an entire set of
agers, the same time commitment for questionnaires (four) from each firm so
completion was not required. Elimi- that a complete analysis could be per-
nating the responses of controllers from
the total (274 less 84), the overall re- 3 Following Oppenheim [1966] as far as practicable
sponse rate decreased to 59.2 percent given the nature of the questionnaire, examination for
nonresponse bias was performed. There were no readily
(190 replies received from 321 question- apparent differences between respondents and non-
naires sent). This still is a substantial rate respondents within the subunit categories.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
28 The Accounting Review, January 1977

formed. As Table 2 shows, of the firms obtained overall judgment on perfor-


that received questionnaires, 24 (23 per- mance by controllers (the dependent
cent) responded. Only the analyses re- variable) as well as summary overall per-
lating to the complete sets of returned formance judgments on each department
questionnaires are reported here.4 made by the respective managers and
A two-stage data analysis was under- their interdependent peers. The latter
taken on a departmental basis across were added to the set of independent
firms. The data initially was factor- variables and interpreted as reflecting all
analyzed; then it was inputted to a path aspects of departmental performance.
analysis as described below. Factor anal- The path-analytic methodology is es-
ysis was carried out for two reasons. The sentially similar to multiple regression,
first reason was that it provided a statis- however, its advantages are numerous.
tically objective means of variable selec- As discussed by Duncan [1966], the
tion for the path analysis. Given the method focuses heavily on variable inter-
number of firms analyzed, it was appar- pretation, and its diagrammatic ap-
ent that a set smaller than the total proach is a valuable tool in the search
number of variables (approximately 160) for alternative explanatory hypotheses.
would have to be utilized in the path However, its greatest strength lies in the
analysis. The highest loading variable on ability of the method to enable closure
each factor in the three separate factor to be forced on a system of interrelated
analyses was used as input to the path variables. This allows relative percent-
analyses to avoid the factor interpretation ages of explanation to be associated with
problem associated with using factor the independent variables by a partition-
scores as secondary analysis input. ing of R , the coefficient of determina-
Weighting and combining variables into tion, which cannot be done with ordinary
a renamed combination usually lead to multiple regression.
information loss and difficulty of under-
RESULTS
standing. It should be noted that the
variables selected for inclusion in the The model and derived propositions
path analyses by this approach concep- refer to contingent assessment of or-
tually represent the particular factors, ganizational subunits. Subsequent to the
not just themselves as separate entities. discussion of the results for each de-
The other reason for performing the partment in the study, some general
factor analysis was that it provided a implications of the research are outlined
first crude picture of the pattern of re- which examine the results as a whole.
sults. As can be seen from the representa-
tions of the factors in Tables 3-5, the Production Departments
factors are quite 'cleai' in that high The proposition on production de-
loading variables on each factor clearly partment assessment indicated that in-
relate to separate dimensions of the ternal variables would be the major
model in Figure 1, a first satisfactory explanators of effectiveness, with inter-
result of the analysis. dependency variables of secondary im-
Path analysis then was undertaken on portance. The buffering of production
the variables selected by the factor analy- departments from the environment was
sis. Stepwise regression was performed expected to result in little or no explana-
on this set of variables and a number of 4 See Hayes [1975] for more extensive data analyses
other variables e.g., the independently as well as for copies of the questionnaires utilized.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 29

TABLE 3
PRODUCTION DEPARTMENT FACTOR ANALYSIS: HIGHEST LOADING VARIABLES

Factor I Factor 2* Factor 3** Factor 4 Factor 5 Factor 6 Factor 7 Factor 8 Factor 9 Factor 10

Price-F Extent of Extent of Superior/ Product Importance Reliability of Budget Delivery Leadership
.75634 cooperation cooperation subordinate quality-D of budget personnel capturing of schedules-D ability
.82582 .72677 relations .86001 .88200 .81004 performance .71081 .66984
.87118 .82198

Customer Confidence Task Skill Product Performance Financial Delivery


service-D expressed cooperation utilization quality-F pressure data- schedules-F
.90141 .89126 .66858 .76202 .74707 .66846 capturing .78016
performance
.83109

Customer Task Productive Meeting of Company Capturing of


service-F cooperation conflict budget reputation-D performance
.88201 .92090 settlement .65886 .59351 .83109
.76062

Customer Productive Team


requirements-D conflict feeling
.91629 settlement .84754
.86747

Customer Team Creative


requirements-F feeling problem
.90247 .82021 solutions
.80435

Creative Information
problem exchange
solutions .84208
.93583

Information Flexibility
exchange .67791
.75786

Flexibility
.89668

* Interdependency variables; as perceived by marketing departments.


** Interdependency variables; as perceived by research and development departments.
NOTE: All variables followed by D or F represent the extent of influence of each on the performance of the department's daily or future planning
tasks, respectively.

tory power residing in the set of environ- where


mental variables.
Zi Py2i = direct effects
It is apparent from examination of
Figure 4 that this proposition is well-
supported by the analysis. Although each and
of the three performance dimensions are Z Py.iPy.jrij = indirect effects.5
present in this analysis, the relative sizes i~j
of the path and determination coeffi- In this case, suppressing the residual
cients associated with the internal vari- value,
ables indicate their explanatory impact.
The computation of the determination
5 The P,.i's refer to the individual path coefficients,
coefficients presented in the table ac- which can be shown to be equivalent to standardized
companying the figure is straightforward. regression beta weights. For an extensive discussion of
Complete determination is given by the merits of the analysis and its utility, see Land [1969].
Further discussion of the methodology as utilized in the
social sciences can be found in Aldrich [1972] and
ZPp2i + E p jPy.jrij 1 (1) Boudon [1968] and in its original application in genetics
i i~j in Wright [1921, 1934].

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
30 The Accounting Review, January 1977

TABLE 4
RESEARCH AND DEVELOPMENT DEPARTMENT FACTOR ANALYSIS: HIGHEST LOADING VARIABLES

Factor J* Factor 2** Factor 3 Factor 4 Factor 5 Factor 6 Factor 7 Factor 8 Factor 9 Factor 10

Extent of Extent of Innovation-D Reliability Skill Budget- Product Superior/ Leadership Delivery
cooperation cooperation -.58619 of personnel utilization capturing quality-D subordinate ability schedules-D
-.80188 .68762 .70805 -.82457 performance .75514 relations -.74834 .71828
.75287 .89703

Confidence Confidence Customer Price-F Financial Product Delivery


expressed expressed service-D .83197 data- quality-F schedules-F
-.88417 .57247 -.93558 capturing .84283 .63185
performance
.85949

Performance Performance Customer Company


pressure pressure service-F reputation-D
-.64060 .64712 -.78077 .81213

Task Task Customer Company


cooperation cooperation requirements-D reputation-F
-.95992 .93685 -.74815 .82944

Productive Productive Customer


conflict conflict requirements-F
settlement settlement -.74768
-.81206 .83407

Team feeling Team feeling


- .83261 .85014

Creative Creative
problem problem
solutions solutions
-.91046 .83295

Information Information
exchange exchange
- .88354 .91635

Flexibility Flexibility
-.89618 .81403

* Interdependency variables as perceived by marketing departments.


** Interdependency variables as perceived by production departments.
NOTE: All variables followed by D or F represent the extent of influence of each on the performance of the department's daily or future planning.

(.5471)2 + (.2415)2 + (.3039)2 + (- .2385)2 partmental performance, has both the


P2 p22 p2 p2 largest path coefficient and provides the
Y1 ly.2 y.3 y.4

? [(.5471)(.2415)(.2726) greatest explanation of variation in de-


Py 1 Py.2 rl2 partmental effectiveness. As indicated
? (.5471)(.3039)(- .2795) + earlier, this variable was interpreted as a
summary combination of all elements
Py.1 Pyp3 rl3
relating to performance in the internal
(- .2385)(.3039)(- .3135)] dimension of the model.
Py.4 Py 3 143 It is noteworthy that interdependence
=.2993 + .0583 + .0923 + .0569 with marketing as summarized by the
marketing manager judgment variable
+ [(- .0427) + (.0050) - (.0392)
provides the second largest direct ex-
-(.0251)] =.4048 = R2 planation after the internal variables, as
(with rounding errors). (2) hypothesized by the proposition. Al-
though there is a positive association with
The overall summary variable, pro- effective performance (Py-= .3039),
duction manager's own judgment on de- correlation with production manager

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 31

TABLE5
MARKETING DEPARTMENT FACTOR ANALYSIS: HIGHEST LOADING VARIABLES

Factor I* Factor 2** Factor 3 Factor 4 Factor 5 Factor 6 Factor 7 Factor 8 Factor 9 Factor 10 Factor 11

Extent of Confidence Within- Customer Innovation-D Budget- Product Product Price-D Extent of Performance
cooperation expressed group work service-D .72776 capturing price-F quality-D - .83286 cooperation pressure
.62786 .90771 relations .62682 perfor- -.94706 .66290 .74215 .82156
.85158 mance
-.74570

Confidence Task Superior, Customer Company Financial Delivery


expressed Coopera- subordinate service-F reputation-D data- schedules-D
.84995 tion relation- .84035 .83393 capturing .64272
.83195 ships perfor-
.86727 mance
- .79265

Performance Productive Leadership Customer Company


pressure conflict .83871 require- reputation-F
.74266 settlement ments-F .71002
.95920 .89487

Task Team
cooperation feeling
.91442 .89592

Productive Creative
conflict problem
settlement solutions
.88451 .86215

T:am feeling Information


.88160 exchange
.88373

Creative Flexibility
problem .80287
solutions
.88118

Information
exchange
.93492

Flexibility
.90476

* Interdependency variables, as perceived by research and development departments.


** Interdependency variables, as perceived by production departments.
NOTE: All variables followed by D or F represent the extent of influence of each on the performance of the departments daily or future planning.

judgment is negative (r13= - .2795). A the direct (.0569) and net effects (.0318)
joint interpretation of this result suggests computed for the coefficient of determi-
that while interdependence is an impor- nation support this proposition. Further,
tant element in explaining production assuming some consonance between fu-
department effectiveness (P. 3 > 0), in- ture production and delivery schedules,
terdependence and internal factors may the negative path coefficient (PY 4=
be working against each other (r13<0). -.2385) suggests an inverse relationship
This highlights potential conflict and with effectiveness. This implies that pro-
indicates the continual difficulty of suc- duction departments are being buffered
cessful differentiation and integration successfully from environmental influ-
[Lawrence and Lorsch, 1969]. ence by sufficient order lead times. Thus,
Limited association with effective per- the proposition concerning production
formance was expected from variables department effectiveness is supported
on the environmental dimension. Both strongly by the analysis.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
32 The Accounting Review, January 1977

FIGURE 4
PRODUCTION DEPARTMENTS*

/ 1. Own judgment on
/ performance
/ ~~~(internal)
Residual .5471
1 I / ~~~~ ~ ~ ~
~~~~~~~~.2726\\ ~

.7716 2. Leadership ability of


/ , ~~~~supervisors -.2795 \
/ .2415 / (internal) \/ (r, 3)

Y. Effectiveness 2(r23) (r474)

\ ~~~~~3.
Marketing managers' /.2698 /
\ on performance
~~~~~judgment /
\ ~~~(interdependence)/
(r34)~ ~ ~ ~(r24)~~~~~(r4

\- 2385 , /
\ ~~~4.
Importance of meeting /
future delivery
schedules
(environmental)

Coefficients of Determination

1 2 3 4

Direct effect: p2 .2993 .0583 .0923 .0569


Indirect effect: p).ipY.jrij -.0427 .0050 -.0392 -.0251

Net effect: .2565 .0633 .0531 .0318

* Straight lines refer to paths of influence, in direction of arrow; respective path coefficients are indicated on each.
Curved lines and associated numbers referto correlations between variables. Size of respective coefficients of determina-
tion included in accompanying table.

would lead to as ready a rejection of the


Research and Development Departments
proposition as acceptance in the case
The proposition advanced concerning of production departments. The size
research and development (R&D) de- of the path and determination coeffi-
partments stated that interdependency cients associated with the first variable,
variables would be the main explanators financial data, requires some explanation
of effectiveness. Internal and environ- prior to hasty rejection of the propo-
mental variables were expected to have sition.
little or no influence. A t-test comparison of answers to
Cursory examination of Figure 5 selected questions was performed on the

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 33

FIGURE 5
RESEARCH AND DEVELOPMENT DEPARTMENTS*

1. Ability of financial data


to measure performance
p q t
((internal)m
.6021/ \

.7106 /-.106 \

YCEffectiveness e 2. tasos 3m on joint


Cooperation
l m ~~~~~~~~~with
marketing .18
\ ~~~(interdependence) (,3

\ \ ~~~~~~~~-
.351 9 /
\-.2217 l32 3 \

\ 3. Importance of improving /
product quality
(environmental)

Coefficients of Determination

1 2 3

Direct effect: py. 1 .3625 .0910 .0492


Indirect effect: py ipjrij .0197 .0147 -.0421

Net effect: .3823 .1052 .0071

* Straight lines refer to paths of influence, in direction of arrow; respective path coefficients are indicated on each.
Curved lines and associated numbers referto correlations between variables. Size of respective coefficients of determina-
tion included in accompanying table.

respective responses of all departmental their departments. It is suggested that


managers. (See Table 1 for the numbers other judges of performance, having
in each group.) On the question of how little or no experience with nonfinancial
well they perceived financial data to measures and experiencing difficulty in
measure their department's performance, evaluating the tasks performed by R&D
the mean response of R&D managers departments, resort to financial data as
was 2.62 on a scale of from 1-5. The the means of evaluation. Note that the
R&D mean response was significantly independent judges of performance in
lower (P<.001) in both instances than this study were controllers.6
those of marketing managers (mean=
3.24) and production managers (mean= 6
Controllers were selected for pragmatic reasons. It
3.64) to the same question. Clearly, R&D was necessary to obtain responses from independent
judges with positions high enough in the organizational
managers perceive financial measures of hierarchy that they would have sufficient knowledge of
performance to be inappropriate for the departments involved in the study. As different

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
34 The Accounting Review, January 1977

To avoid forcing their responses in it is suggested that refutation is not


particular directions, the controllers in warranted at this stage.
the study were not questioned about the
bases upon which they made their evalu- Marketing Departments
ations. The explanation advanced is thus Approximately equal contributions to
essentially an hypothesis generated from the explanation of marketing department
further examination of the data. It effectiveness were expected from inter-
should be noted that its tenor is not such dependency and environmental variables.
as to deny that financial information on Limited explanation was expected from
R&D departments is important; such internal variables.
information may serve as a useful signal- Examination of the coefficients of de-
generating device for managers. The termination of the respective variables
hypothesis is an empirical question (Figure 6) shows that environmental
worthy of investigation, for empirical variables had by far the greatest impact.
support would aid in the search for alter- The sum of net effects of the two en-
native assessment methods. vironmental variables (4 and 5) is ap-
Both the interdependence and environ- proximately 35 percent; the direct effects
mental dimensions were represented in are even greater. The interdependent
the analysis of R&D performance. The relationship with production is evidenced
negative path coefficient (Py 2 =-.3017) by the inclusion of variables 2 and 3, the
on cooperation with marketing suggests summary judgment and pressure vari-
an inverse relationship with effective- ables. However, the values of the path
ness, implying some difficulty in co- and determination coefficients are quite
operative activities. Perhaps differences low. The total direct effect of the two
in perceptions as to the urgency of par- variables is less than 9 percent, lower than
ticular tasks, which lead to fruitless that of the summary internal variable
meetings, may explain this result. Law- (.094). It would appear that interde-
rence and Lorsch [1969] found a number pendence is less important than pro-
of differences in the personalities of posed. However, the major environ-
marketing and R&D personnel, among mental variable, concern with product
which was a difference in time horizons. quality, has some implications for the
Implicit in the development of the production/marketing relationship. Its
proposition on R&D departments was relative size and impact in explaining
the assumption that innovative pressures variation in effectiveness implies that
from the environment would be chan- marketing personnel may be able to
neled via marketing departments. The exercise some control over product qual-
environmental variable present concerns ity by the provision of feedback from the
innovation; to some extent it can be inter- environment. The correlations (r24=
preted as relating to interdependence. -.2843) with the summary judgment
Indications of the importance of the variable and the felt pressure variable
interdependence dimension of the model (r34=.2655) suggest this interpretation.
in assessing R&D effectiveness are pro-
vided by the results. The primary con- titles usually accompany similar executive levels in dif-
clusion, however, must be that further ferent firms, it appeared that only controllers and presi-
exploration is required on the proposi- dents could be reached in each firm to assure some con-
sistency across judges. It was felt that there would be a
tion advanced. Given the hypothetical greater chance of obtaining responses from controllers,
explanation of the financial data variable, so they were utilized.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 35

FIGURE 6
MARKETING DEPARTMENTS*

.. a Own judgment on
performance \
/ (internal)\ \

Y/ Efeu 2. Production managers'judgment


/ ~~~~on performance
Residual 5 ( r(interdependence)dpnec
}
/ ~~~~~~~.2432\/\
.7148 / /(r2 3) \ .1079 \ .0824
/ /' \ v
r3)
(r, 4)
~~~~~~~~~~~~~~~~(,
1,f teltpressure to perform
odu3. p e (
.1l(660 highly on joint taskse/nt
Y. Effectiveness -Cwith productionof Dtr
.5977 (interdepeence) (r)

\ \ ~~-.2655>\ ~
'~~~~~

Direct effect:.0407current product quality .


Ind\rec efet:.01 -047 -.(environmental) 2
\2245a
Net effect: .54 0360235 .280.665 .1885
1
\ ~~~~~(r4 5) (r3 52l

S g n fo s no5. \ Importance. of meeting d o


3.66e
\ future product require- / (1- id

ments of customers z
(environmental)

Coefficients of Determination

tion icddnao2 3 4 5

Direct effect: py2 i 0940 .0582 .0275 .3573 .0504


Indirect effect: py. py.jrij .0214 -.0547 -.0041 -.0733 .0122

Net effect: .11l54 .0036 .0235 .2840 .0626

*Straight lines refer to paths of influence, in direction of arrow; respective path coefficients are idctdon each.
Curved lines and associated numbers refer to correlations between variables. Size of respective coefficients of determina-
tion included in accompanying table.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
36 The Accounting Review, January 1977

The absolute values of the correlation accounting is that of budgeting and its
r45 (- .1565) and r14 (.0824) are sub- behavioral ramifications. The standard
stantially less than those with the inter- research approach has been the explora-
dependency variables, indicating that it tion of factors such as the extent of par-
has a relatively closer relationship to the ticipation in budget setting and the per-
interdependency variables than the formance effects of tight and loose
others. standards. The emphasis of this study
Another important relationship to note has been on more fundamental issues,
is that between the internal summary and these other aspects were intention-
variable (1) and concern with customers' ally excluded. What have been explored
future requirements (5); r15=.3622. One are the perceptions of managers with
aspect of this association is the depart- respect to the accuracy and importance
mental ability to perform its boundary- of budgets and other financial data
spanning role between the firm and the utilized in performance evaluation. The
environment. It suggests that internal results indicate a general dissatisfaction
task-performance is associated with satis- with such measures, tending to reinforce
fying environmental requirements. As the arguments from previous work that
such, it provides supportive argument they are unsatisfactory assessment tools
for the proposition and its rationale as [Argyris, 1953, Lowe and Shaw, 1968].
outlined in an earlier section of the paper. In fact in most instances, environmental
Examination of the variables and and interdependency factors were gen-
their interrelationships indicates substan- erally more important than financial
tive support for the proposition ad- data. The implication is that budgets,
vanced concerning marketing depart- which are essentially surrogates for the
ment effectiveness. However, environ- entire set of factors which influence per-
mental variables may contribute more to formance, do not perform the surroga-
the explanation than interdependencies. tion function well. By explicitly recog-
Further refinement and empirical work nizing controllable and uncontrollable
are required to clarify the various inter- influences [Horngren, 1972, pp. 162-163]
relationships. the link is formed to the concept of
Evaluating the results for all three de- responsibility accounting.
partments suggests strong support for the
propositions on production and market- Responsibility, Accounting
ing departments, and indications of such The basic tenet of responsibility ac-
support for research and development counting is that subunit managers should
departments. only be held accountable for aspects of
subunit performance over which they
IMPLICATIONS have control. As Caplan [1971, p. 19]
The research results reported here argued, accountants usually mean by
indicate that the traditional managerial this responsibility for costs. However,
accounting tools were relatively poor the notion really is broader than costs
explanators of effectiveness. The follow- for the extent of controllability of the
ing exploration of some implications of factors cause certain costs to occur,
this finding is performed within the and that really delimits the extent of re-
constraints on generalizability imposed sponsibility. From the theoretical de-
by the obvious limitations of the study. velopment and empirical results reported
One of the problem areas of managerial here, it is argued that these factors fall

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 37

into three classes. Controllability and outlined a number of strategies an organi-


responsibility are differentials across the zation may employ. For example, or-
classes. ganizations attempt to protect their
The personnel variables included in technical cores by buffering them from
the internal dimension are of intuitive environmental fluctuations. In cases
importance in determining the actual where this is not completely possible,
costs that a subunit will incur. Managers they either will try to smooth input and
regard the variables in this manner, as output transactions or try to anticipate
demonstrated by their uniformly high and adapt to environmental fluctuations.
responses to questions concerning the Depending on the strategy employed, the
importance of each variable to the ability of units that are interdependent
effective functioning of the subunit with production to decrease the influ-
[Hayes, 1975]. Cost data are not good ence of the environment on the produc-
surrogates, especially in marketing and tion department could be evaluated
R&D departments. The path analyses readily. The number of short lead-time
showed that environmental influences orders or unplanned product specifica-
had a substantial impact on the per- tion changes requested are examples of
formance of departmental tasks. In many such measures. The path analysis of
instances, the inferred ability of the de- production departments indicated that
partment to deal with these influences the protection offered them from en-
was of primary importance in the expla- vironmental influence was not optimal
nation of the results. Another aspect of in that some environmental effects on
causation behind actual cost figures is the performance were disclosed. As such
interdependent relationship among sub- factors behind cost figures are control-
units. To the extent that this relationship lable to some extent, they could be
is more complex, the influence on the assessed in overall subunit evaluation.
performance of a particular subunit and A number of environmental control
the costs incurred therein, too, would be strategies can be employed by the or-
greater. With each class of items which ganization as a whole [Thompson, 1967,
influence and lay behind the cost data, pp. 35-36]. Contracting entails obtain-
the question to be addressed (as far as re- ing agreement on the undertaking of
sponsibility accounting and budgeting future activities with task environment
are concerned) is the extent of controlla- elements. A more restrictive means of
bility. attempting to gain some environmental
Personnel factors are controllable to control is cooptation, whereby elements
a large degree by various levels of the of the task environment are absorbed into
management hierarchy. The abilities or the leadership of the organization; e.g.,
skills of the workforce may not be con- the inclusion of external directors on the
trollable by immediate supervisors, but board. Finally, joint ventures may be
they are controllable by those performing undertaken with other organizations; this
the personnel function in the firm. The is referred to as a coalescing strategy.
way in which such skills are utilized,
inter- and intra-group relationships, and 'This provides a slightly different rationale for
other aspects are under the influence of "human asset accounting" than that generally advoacted
subunit management.' [Flamholtz, 1972]; i.e., accountants should measure
personnel factors (not in financial terms) because the
With respect to control over the en- current cost data obtained are poor surrogates for this
vironment, Thompson [1967, pp. 20-22] causal influence behind costs.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
38 The Accounting Review, January 1977

Although such strategies refer to a charged with performing the integration,


broader spectrum than that addressed in responsibility could be assessed directly.
this study, the extent to which an organi- The overall implications of the study
zation undertakes them will directly in- for managerial accounting are severe,
fluence departmental functioning. Re- for they suggest a major reorientation
sponsibility for adoption of some of from the narrow examination of sur-
these strategies may rest with marketing rogates to the evaluation of the underly-
departments and could provide an as- ing causal variables. This implies not
sessment alternative to simply costs, only employing methods and techniques
both for marketing and production sub- which are not necessarily universally
units. applicable but also having a major re-
Direct evaluation of interdependence orientation in the approach to assess-
is an explicit part of such environmental ment. Indeed, the primary implication is
impact measures. However, there are that a contingency approach be taken to
other controllable aspects of interde- managerial accounting, the relevant as-
pendence, such as the performance of the sessment methods being determined by
integrating function as explored by Law- subunit type, subunit interrelationships
rence and Lorsch [1969], which imply and the extent of environmental influ-
further evaluation approaches. Depend- ence on subunit functioning.
ing on the hierarchical level of those

REFERENCES
Aldrich, Howard E., "Technology and Organization Structure: A re-examination of the Findings of the
Aston Group," Administrativte Science Quarterl/v Vol. 17, No. I (March 1972), pp. 26043.
Argyris, Chris, "Human Problems with Budgets," Harvard Business Reuiewt (January-February 1953),
pp. 97-110.
Boudon, Raymond, "A New Look at Correlation Analysis," in Hubert M. Blalock and Ann B. Blalock,
eds., Methodology in Social Research (McGraw-Hill, 1968), pp. 199-235.
Caplan, Edwin H., Managemnent A ccountinq and Behatioral Science (Addison-Wesley, 1971).
Demski, Joel S., "Optimal Performance Measurement," Journal ofA,4ccounting Research, Vol. 10, No. 2
(Spring 1972), pp. 243-258.
and Gerald A. Feltham, A Conceptual Approach to Cost Determination (Iowa State University
Press, 1976).
Duncan, Otis 1)., Path Analysis: Sociological Examples," American Journal of Sociology, Vol. 72, No. I
(July 1966), pp. 1-16.
Feltham, Gerald A., "Information Evaluation," Studies in Accountinq Research f5 (American Accounting
Association, 1972).
Flamholtz, Eric G., "Towards a Theory of Human Resource Value in Formal Organizations," THE
ACCOUNTING REVIEW(October 1972), pp. 666-678.
Friedlander, Frank and Hal Pickle, "Components of Effectiveness in Small Organizations," Administra-
tive Science Quarterly (September 1968), pp. 289-304.
Georgopoulos, Basil S. and Arnold S. Tannenbaum, "A Study of Organizational Effectiveness," American
Sociological Review, Vol. 22 (October 1957), pp. 534-540.
Hayes, David C., "The Contingency Theory of Managerial Accounting: An Empirical Test of an Assess-
ment Model" (Ph.D. diss., The Ohio State University, 1975).
Hopwood, Anthony G., "Leadership Climate and the Use of Accounting Data in Performance Evalua-
tion," THE ACCOUNTING REVIEW (July 1974), pp. 485-495.
Horngren, Charles T., Co~stAe ountinq:. A Managerial Emphais, 4th ed. (Prentice-Hall, 1972).
Kinney, William R., Jr., "An Environmental Model for Performance Measurement in Multi-Outlet Busi-
nesses," Journal of Accounting Research (Spring 1969), pp. 40-52.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions
Hayes 39

Klammer, Thomas, "The Association of Capital Budgeting Techniques with Firm Performance," THE
ACCOUNTING REVIEW (April 1973), pp. 353-364.
Land, Kenneth C., "Principles of Path Analysis," in E. F. Borgatta, ed., Sociological Methodology
(Jossey-Bass, 1969).
Lawrence, Paul R. and Jay W. Lorsch, Organization and Environment (Irwin, 1969).
Lowe, E. A. and R. W. Shaw, "An Analysis of Managerial Biasing: Evidence from a Company's Budget-
ing Process," The Journal of Management Studies (October 1968), pp. 304-315.
Mahoney, Thomas A., "Managerial Perceptions of Organizational Effectiveness," Management Science,
Vol. 14, No. 2 (October 1967), pp. B76-91.
and William Weitzel, "Managerial Models of Organizational Effectiveness," Administrative
Science Quarterly, 14(3) (1969), pp. 357-365.
and Peter J. Frost, "The Role of Technology in Models of Organizational Effectiveness," Orga-
nizational Behavior and Human Performance, Vol. I I (February 1974), pp. 122-138.
March, James J. and Herbert A. Simon, Organizations (Wiley, 1958).
Oppenheim, A. N., Questionnaire Design and Attitude Measurement (Basic Books, 1966).
Pugh, Derek S., David J. Hickson, C. R. Hinings and C. Turner, "Dimensions of Organization Structure,"
Administrative Science Quarterly, Vol. 13 (1968), pp. 65-105.
Seashore, Stanley E. and Ephraim Yuchtman, "Factorial Analysis of Organizational Performance,
Administrative Science Quarterly (December 1967), pp. 377-395.
Simon, Herbert A., Harold Guetzkow, George Kozmetsky and Gordon Tyndall, Centralization Vs.
Decentralization in Organizing the Controller's Department (Controllership Foundation, 1954).
Solomons, David, Divisional Performance. Measurement and Control (Irwin, 1965).
Thompson, James D., Organizations in Action (McGraw-Hill, 1967).
Torgerson, Warren S., Theory and Methods of Scaling (Wiley, 1958).
Wahba, Mahmoud A. and Harris J. Shapiro, "Managerial Assessment of Organizational Components,"
Academy of Management Journal, Vol. 16, No. 2 (June 1973), pp. 277-284.
Wright. Sewall, "Correlation and Causation," Journal of Agricultural Research, Vol. 20, No. 7 (Jan-
uary 1921), pp. 557-585.
,""The Method of Path Coefficients," Annals of Mathematical Statistics, Vol. 15 (September 1934),
pp. 161-215.

This content downloaded from 62.122.72.154 on Wed, 18 Jun 2014 20:09:27 PM


All use subject to JSTOR Terms and Conditions

Вам также может понравиться