Antecedents for defining management control system design : empirical evidence from Brazilian firms

This work analyzes the development profile of Brazilian firms regarding management accounting attributes. Much has been written in recent years about the design of management control systems and the use of management control artifacts, but there is a relevant antecedent that has been little explored, namely the profile of these attributes, involving aspects such as scope, timeliness, integration and aggregation (Chenhall & Morris, 1986; Silvola, 2008). These attributes define, limit and expand the power for management control to act (Berry, Broadbent & Otley, 2005) and they are interdependent, since each influences the others. They are the basis on which the organization’s management model is structured and developed. Within the universe of a firm’s management control, there are various artifacts that are implemented with the aim of providing support to managers’ decision-making process, besides coordination and Received on 11/29/2010; reviewed on 08/18/2011; accepted on 09/13/2011; available in 03/05/2012 Correspondência autores*: † Associate Professor at USP Address: Av.Prof. Luciano Gualberto, 908, FEA 3, sala 226, São Paulo SP – Brazil E-mail:frezatti@usp.br Telephone: (11) 30915820 – r 157  Ph.D. in Accounting from USP. Professor of UFES Address: Rua Constante Sodré, 869/1001 – Praia do Canto, Vitória – ES – Brazil E-mail: emanuel.junqueira@ gmail.com Telephone: (27 ) 4009-7700 ¥ Doctor of Business Administration from USP. Adjunct Professor at the University Mackenzie Address: Av. Santo Antônio, 472 – Vila Osasco – Osasco – SP – Brazil E-mail: diogenesbido@yahoo.co m.br Telephone: (11)36815780 ‡ PhD in Accounting and Accounting from USP. Assistant Professor, State University of Feira de Santana. Address: Avenida Transordestina, Novo Horizonte, Feira de Santana – BA – Brazil E-mail: artnascimento@gmail .com Telephone: (75) 31618052 Θ PhD from USP. Professor at USP Universidade de São Paulo. Address: Avenida Marechal Eurico Gaspar Dutra, 988 – Jardim São Paulo, São Paulo – SP – Brazil E-mail: taniarsr@usp.br Telephone: (11) 3091 5820, ramal 166 Nota do Editor: Esse artigo foi aceito por Antonio Lopo Martinez. Esta obra está licenciada sob a Licença Creative Commons – Atribuição-Uso não-comercial-Compartilhamento pela mesma licença 3.0 Unported License


INTRODUCTION
hy speak of antecedents?Because often some requirements that should be considered in examining a given theme undergo changes, including in importance of their configuration.
Relevant aspects for investigating a theme can cease to be important due to changes in the environment or technology or the emergence of substitutes.When reexamining antecedents, it is necessary to review the structural base and constructs, and in this process opportunities for organizational improvement arise.
In the present case, the antecedents are the structure of the management control system and its attributes (scope, timeliness, aggregation and integration).Although they are present in all organizations, the change that comes over time can cause certain aspects to be overlooked because their review is not valorized.In the final analysis, to change and progress it is often necessary to review and rethink pillars previously considered as solid and untouchable.
Management of organizations requires an information system that provides responses to the demands of users of accounting.This system is traditionally separated into two areas of study: financial accounting, which involves the information for external users, and management accounting, which deals with the information intended for internal use of the organization (Hansen & Mowen, 1997, p. 2).The study of management accounting is mainly concerned with: (i) supporting firms' decision-making process (Anthony & Govindarajan, 2002), and (ii) influencing and (iii) coordinating people (Hansen & Mouritsen, 2007).
In turn, within the management process, management control "is a process of guiding organizations into viable patterns of activity in a changing environment" (Berry, Broadbent & Otley, 2005, p. 18).That process cannot subsist without a management accounting system that is structured consistently with the demands, and should provide conditions to instigate, orient and clarify managers regarding the important elements for managing their organization.
For this to occur, the information system developed must be dynamic, regarding currency, flexibility, amplitude and diversification.However, the limit on the potential use is intrinsically related to the system's capacity and the manager's demand.
A firm's financial result depends on a series of management aspects, and the use of management accounting resources can at most make managers more prepared and more confident to perform their activities, besides more assertive due to the support of the process.Therefore, achieving better financial results depends on a complex combination of many internal and external factors that are very hard to identify and study.In this way, the existence of an information structure results from the belief that assertiveness improves the quality of management.
In this sense, one antecedent of the existence of artifacts (the management control techniques that will be employed) is the design of the management information system.
And in turn, a forerunner of the management information system's design is the presence of determined qualitative aspects, called attributes, of the management accounting system.
These regulate the scope and characteristics of the information provided.The attributes are the scope, timeliness, integration and aggregation of the system.
Weak and unambitious attributes in terms of scope imply limited designs to meet the demand for artifacts.To some extent the attributes can interact, influencing each other.
Alterations in the configuration of the attributes are certainly possible, but need to be consistent among the elements, without which the artifacts will not meet the demand or will do so only partly.
The importance of studying the attributes of the management control system rests in the fact that the choice of the characteristics of the information generated by the system (the system's attributes) are forerunners for development of the system and the choices of the artifacts that compose it.
Therefore, adequate knowledge of the antecedents for developing the management control system can enable artifacts to be developed and implemented in harmony with the needs of the particular firm.
Awareness of this condition allows avoiding wasted effort by the people involved and increases the efficiency and efficacy of the system, in the sense of contributing more decisively to the firm's competitiveness and profitability.In turn, these aspects potentially mean more employment and wealth for the country as a whole.As a consequence of these observations, our guiding research question can be stated as follows: Is there an association between the attributes of management control systems (scope, timeliness, BBR, Braz.Bus.Rev. (Engl.ed., Online) Vitória, v. 9, n. 1, Art. 6, p. 125-145, jan-mar. 2012 www.bbronline.com.braggregation and integration) and the profile of the artifacts of management information systems?

LITERATURE REVIEW
The basic elements to study the theme are: (i) evolution of the attributes, (ii) design of the management control system and (iii) management control artifacts.

Evolution of the attributes
Chenhall & Morris (1986), in a seminal article, present the attributes of management information systems: scope, timeliness, integration and aggregation.These attributes are essential to the design and use of management control systems and permit identifying their profile.
Regarding scope, which refers to focus, quantification and horizon (Chenhall & Morris, 1986), the authors indicate that different dimensions are possible: narrow, broad or intermediate, also called medium (Silvola, 2008).
The broad dimension should provide (i) internal and external information, (ii) financial and nonfinancial information, and (iii) information oriented to the past and the future.
Restricted scope means the information is more limited, basically involving internal historic and financial information.This information will have to be adjusted in relation to timeliness, aggregation and integration to allow balance in the management information system design.
According to Silvola (2008), there is an intermediate profile, which differs from the broad scope only by not including external information.
Scope should be the first attribute to be defined, and consequently the entire design of the management information system should depend of its amplitude.However, this may not happen in practice, because definition of the design occurs at a certain moment and in the future may not meet the information demands in a properly organized and holistic form, unless efforts are made to revise the system's attributes in relation to the current configuration and complexity of operations.
Timeliness is defined in terms of opportune response, that is, providing information when requested or in the frequency desired.It enhances the use of management information, providing rapid feedback to the decisions made.The timely availability of information is a relevant factor, and although the supply of historic information may not be so complicated, the supply of projected information about the future, for example detailed by business unit, is more complex and costly.Automation plays an important role in the generation and availability of information in standardized form, and if a firm's computer systems are not up to the task, timely use of information will not be possible, making for inefficient management.
Aggregation is the attribute that permits information to be offered at various levels of agglutination, from simple information considered basic, such as the amount of a raw material not yet processed, to information aggregated in various time periods or by cost centers or areas of interest, etc. Proper aggregation can bring relevant responses for management actions.To achieve this requires analysis and decisions in terms of relevance and priorities.Obviously, changes in the modus operandi and structural changes, such as mergers and acquisitions, demand comprehensive revisions.
Finally, integration entails the coordination of various segments within a firm or business unit/area.This may simply involve combining information on areas over time or it can involve internal transfer prices, acting to regulate the relationship between different units, revealing their interdependence.In other words, integration refers to the information about the operations and activities of the units of an organization and also information about the impact of one unit's decisions on the performance of the others (Bouwens & Abernethy, 2000) and of the organization as a whole.
Moores & Yuen (2001), revisiting approaches covered before (Merchant, 1981(Merchant, , 1997)), refer to attributes as being the formality of routines and procedures with intensive use of computers, the technical team and financial modeling, all of which must have certain characteristics to be selected and presented within the decision process.Therefore, the prior choice of these attributes, regarding the structuring or fit of the design of management control systems implies identification of the qualities that managers expect from the information provided by the system.

Management information system designs
Aguiar & Frezatti (2007)  These dimensions are: (i) system quality, with focus on the definitions of the desired characteristics of the information system; (ii) information quality, which defines the information to be provided by the information system and its quality; (iii) use, meaning the way the system is utilized; (iv) user satisfaction, which considers the attitudes of managers regarding the system; (iii) perceived utility, which is an indicator of how strongly managers believe that the use of a particular information system will have on their performance or that of the group or organization, by producing benefits; and (iv) net benefits, measuring the sum of all the past benefits and the expected future ones attributed to the use of an application of the system, as perceived by the shareholders.
The contribution of Moores & Yuen (2001)   Ferreira & Otley (2006;2009) make an important contribution by including cost as an attribute, as well as, based on the work of Simons (1990;1995), incorporating a use dimension that includes the way managers use information: (i) diagnostic use (monitoring performance to take corrective actions), (ii) interactive use (monitoring strategic uncertainties and permitting dialog between mangers and subordinates) and (iii) dysfunctional use (use for purposes other than those envisioned for diagnostic or interactive use).
Besides the characteristics of the information that will be provided, the design of the management control system depends on decisions about the artifacts that will integrate it (techniques utilized), which can be classified as traditional and contemporary (Chenhall & Langfield-Smith, 1998;Otley, 1999;Ferreira & Otley, 2006).
In this respect, besides the design, the proposal incorporates the use perspectives of Simons (1990;1995), so that management control systems can be understood in relation to two dimensions (Figure 1), one involving the design and the other the use (Ferreira & Otley, 2006;2009).Despite the interdependence of these two dimensions, our basic element in this study is the performance of the design of the management control system.Additionally, we treat this design solely in terms of the characteristics of the information produced by it (attributes of the management control system), rather than in terms of the practices employed (management control artifacts or techniques).
As can be seen from the above discussion, attributes are included in all the models, being defined as qualitative factors that affect the system as a whole.Therefore, study of the attributes, from the perspective of antecedents for development of the system, becomes the main backdrop, independent of its design and hence use.

Management control artifacts
Artifacts are the carriers of institutional values, products of human actions: hardware, software, technologies or ideas (Scott, 2001, p. 81).
From the standpoint of management control, artifacts correspond to all the instruments and techniques employed to support management of firms.An artifact can be a concept, a system, a method or a tool (Frezatti et al., 2009).The existence and use of artifacts should improve the chances of organizational success, and this requires consumption of resources (money, time, equipment, training, etc.) that will not be avaialable for alternative uses.However, not all these artifacts are considered here.Out criteria for selecting the artifacts were: (i) recognition as an instrument of some part of formal planning; (ii) greater relative security that the essence of the artifact was captured by the survey among the respondents; and (iii) elements that permit avoiding the perspectives of "is contained or contains" (e.g., EVA can exist by itself in the firm's analysis or it can be present in more than one artifact, such as strategic planning, budgeting or capital budgeting).
Therefore, from the list provided by Chenhall & Langfield-Smith (1998), we surveyed formal strategic planning, traditional annual budgeting, rolling forecasting, capital budgeting, product profitability analysis, customer and market analysis and real options to analyze investment projects.The stratification was based on sector and size (in terms of annual revenue).The distribution found in terms of turnover was relatively balanced.However, out of the 24 different sectors used by the magazine to classify the companies, we formed 19 groups (sectors), within a criterion of maximum homogeneity, due to the great dispersion in some sectors (Table 1).
The sample of firms was composed by using a probabilistic approach, based on random numbers.Once the sample size was determined, the firms were drawn within the stratification chosen.In the case of firms that did not answer the questionnaire, a new drawing was carried out to replace them.Of a total of 2,181 companies (population), the final sample was made up 125 companies.

Artifacts
All the attributes had a larger percentage of responses in the fourth option (artifact is known, made avaialable and utilized).Table 4 presents the distribution of the responses for each artifact.

Attributes as antecedents of the artifacts
In this step of the analysis we estimated a model in which the four attributes were predictors of the six artifacts.The PLS-PM enabled estimation of all the coefficients simultaneously.We estimated this model four times: complete sample (125 firms), sample with narrow scope (30 firms), intermediate scope (59 firms) and broad scope (36 firms).
The justification for this procedure is that if we had considered the latent variable scope to have a score that increased as it became broader, as Chenhall & Morris (1986) did, it would have been sufficient to analyze the full sample.However, we observed that the responses to the scope items did not follow this logic, but rather presented organizations that indicated extreme scope and information about the future, but did not indicate the other items.
The criterion used to classify the firms was the following: those with systems providing external information and information about the future (projections) were classified as having broad scope, while those only having external information were classified as having intermediate scope and the others were classified as having narrow scope.
In this respect, the sample size affected the study, because only the intermediate scope had a relevant number of respondents (practically half the sample).In any event, the explanatory power, in the classification of Cohen (1977), was medium in most of the analyses, which is very relevant for this type of study.For the social and behavioral sciences area, Cohen suggests R 2 = 2% for classification as small, R 2 = 13% as medium and R 2 = 26% as large.
To allow understanding the results of the survey, Table 5 presents the information of each artifact broken down into each of the types of scope, after which we comment on the significant effects (p < 0.05)a.
(v)  individual impact, which analyzes the perception of individual users, their understanding of the information received and the change this causes in decision behavior; and (vi) organizational impact, which considers the influence that the individual impact has on the organization and the impact of the individual decisions on the organization(Delon & McLean, 1992;Iivari, 2005).The management information system attributes are included in the information quality category.In the approach of DeLone & McLean (1992), the system's success depends on the balance, or at least the consistency and coherence, among all six dimensions.Seddon (1997) starts from the model of DeLone & McLean and also criticizes its qualitative potential, which can cause confusion.As a solution, he proposes the inclusion of new variables in the model: (i) expectations of decision-makers about the costs and benefits of future use of the information system; (ii) corporate (shareholders), organizational and individual consequences of the results attributed to the system's use; to management information system models was to include segmentation between different levels of formality regarding the nature of the selection and presentation of management accounting information over the life cycle of companies.They use a life cycle perspective to discriminate this model.In their framework, the attributes of management information systems are found in the presentation of the information.Henri (2006) presents a configuration that considers aspects regarding the diversity of the indicators: (i) subjectivity and objectivity, (ii) internal and external focus and (iii) financial and nonfinancial metrics; as well as regarding the nature of use: (i) monitoring, (ii) attention focusing and (iii) strategic decision-making legitimization.

Table 3 : Correlation between the attributes and evaluation of the validity and reliability
(Ringle, Wende & Will, 2005) probability of significance were estimated using SPSS 17.0 based on the factorial scores obtained in SmartPLS 2.0.M3(Ringle, Wende & Will, 2005)and the other results were obtained directly in SmartPLS.The values in boldface (in the diagonal) are the square root of the AEV.