Five theoretical formulations have been proposed in the literature. They are as follows:
(A) Efficient design of a management accounting system and a choice of control mechanisms that depend on the structure and context of an organization.3
The contextual variables that shape the organizational structure are assumed to be technology and environment. Technology is conceptualized as variable, from routine to nonroutine, based on the nature of raw materials and search processes. Environment is mapped on a continuum from highly predictable to highly un- predictable. The properties of the organizational structure that are shaped by technology and environment are the distribution of authority and power, the question of centralization versus decentralization, and the issue of procedure specification. In other words, the distribution of organizational authority and the extent to which procedures can be specified depend on technology and environ- ment. The type of organizational structure, in turn, is assumed to affect man- agement accounting processes such as planning, resource allocation, and measures of performance.
(B) Gordon and Miller proposed a contingency framework for the design of accounting information systems that takes into account the environment, organ- izational attributes, and managerial decision-making styles.4The environment is
characterized by three key dimensions: dynamism, heterogeneity, and hostility. The organizational attributes include decentralization, differentiation, integra- tion, bureaucratization, and resources. Finally, the decision-making style of ex- ecutives is characterized by the following six dimensions: analysis of decisions, decision time horizons, multiplexity of decision making, adaptiveness, proactiv- ity, and consciousness of strategies. These contextual factors and their key di- mensions are assumed to have an impact on such prerequisites of the accounting information system as information load, centralization of reporting, cost allo- cation methods, frequency of reporting, method of reporting, time element of information, performance evaluation, measurement of events, and valuation methods. While the number of permutations of these variables may suggest an unmanageable number of situations, Gordon and Miller suggested, in fact, that “it seems that environmental, organizational, and decision style traits are not distributed randomly but actually cluster together to form commonly accruing
configurations.”5 Three archetypes—the adaptive firm, the firm running blind,
and the stagnant bureaucracy—are presented as evidence of the need for a con- tingency approach in the design of an accounting information system.
(C) Macintosh and Daft investigated the relationship between one character- istic of the organization and the control system design.6 By interdependence
they meant the extent to which departments depend on each other and exchange information and resources to accomplish a task. It is also a variable relevant to control systems. Interdependence can be (1) pooled when the departments are relatively autonomous and little work flows between them, (2) sequential when the departments are linked in a serial fashion, with the output of one department used as the input of the next department, and (3) reciprocal when the depart- ments work jointly on a project and work flows back and forth between them.7
The management control system is viewed in terms of three control subsystems: operating budgets, statistical reports, and standard operating procedures and pol- icies. The hypothesized relationships and the use of management control systems are as follows:
1. In the case of pooled departmental interdependency, the preferred means of control is standardization and a greater reliance on standard operating procedures than on either operating budgets or statistical reports.
2. In the case of sequential departmental interdependency, the preferred means of control are planning and measurement, with more reliance on operating budgets and statistical reports than on standard operating procedures.
3. In the case of reciprocal departmental interdependency, the preferred means of control is mutual adjustment; less reliance is put on operating budgets, statistical reports, and standard operating procedures.
The results of Macintosh and Daft’s field study showed that when interde- pendence is low, control is focused on the use of standard operating procedures; when it is moderate, control rests on budgets and statistical reports; and when it is high, the role of the three control systems diminishes.
(D) Macintosh proposed a contextual model of information systems that em- braces both a macroorganizational concept—technology and a human informa- tion processing system construct—and personal decision style.8 Basically, the
model combines personal decision style, technology type, and organizational structure to derive an information system style. These variables are defined as follows:
1. Driver and Mock’s decision-style model is used to define the decision-style variables.9 The model assumes two dimensions of information processing:
amount of information used (from minimum to maximum) and degree of focus in the use of data (from one solution to multiple solutions). These two dimen- sions are combined to derive four distinctive styles: decisive, flexible, hierarchic, and integrative.
different meanings at different times. Decisive individuals look for efficiency, speed, and consistency in the information to be used. They prefer brief com- munications and summary reports focusing on one solution, results, and action. They like to be in hierarchic organizations with a short, clear span of control and clear rules.
The flexible style assumes the use of a minimum amount of data to generate different meanings at different times. Flexible individuals look for speed, adapt- ability, and intuition rather than developing and operating in accordance to a plan. They prefer brief communications that focus on a variety of solutions. They favor loose and fluid organizational patterns.
The hierarchic style assumes the use of masses of data to generate one firm opinion. Hierarchic individuals look for thoroughness, precision, and perfec- tionism. They prefer long, formal, thorough reports that present problems, meth- ods, and data and generate one best solution. They like to be in a classic organization with a broad span and control as well as elaborate procedures.
The integrative style assumes the use of masses of data to generate a multitude of possible solutions. Integrative individuals look for the creative use of infor- mation in experiments, simulations, and games. They prefer complex and fluid communication that emphasizes discussion rather than reports. They like to work in nonautocratic teams and in nonhierarchic organizations of the matrix type.10
2. Perrow’s categories of technology are used to define the technology vari- able.11The model assumes two dimensions of technology: task knowledge (from
analyzable to unanalyzable) and task variety (from low to high). These two dimensions derive from distinctive categories of knowledge: (a) craft technology (analyzable task knowledge and low craft technology task variety); (b) routine technology (analyzable task knowledge and low task variety); (c) research tech- nology (unanalyzable task knowledge and high task variety); and (d) technical- professional technology (analyzable task knowledge and high task variety). Each of these categories of knowledge is assumed to be best served by a distinctive organizational structure that fits the special needs of the task.
3. Finally, four information styles are differentiated in terms of two dimen- sions: amount and ambiguity. Macintosh defined them in the following manner:
The concise information system. Small to moderate amounts of information
that are precise and unambiguous, and are used in a quick and decisive way.
The elaborate information system. Large amounts of information, frequently
in the form of databases or simulation models, which tend to be detailed and precise; recipients normally use such information in a slow and deliberate man- ner.
The cursory information system. Small amounts of information, neither pre-
cise nor detailed and frequently superficial, that are used in a causal yet decisive way.
The diffuse information system. Moderate to large amounts of information,
covering a wide range of material, frequently ill-defined and imprecise, that typically are used in a slow, deliberate manner.12
(E) Ewusi-Mensah investigated the impact of the external organizational en- vironment on management information systems.13The organizational environ-
ment was classified as either static or dynamic, and as controllable, partially controllable, or uncontrollable. Variations in organizational environments are assumed to require different decision processes and, consequently, different in- formation characteristics, including information quality, information availability, information value, impact on decision making, organizational interaction, or- ganizational search, response time, time horizon, information source, and infor- mation type.
EMPIRICAL STUDIES IN CONTINGENCY THEORY