The effect of formal control tightness on performance in Professional Service Firms : what is the impact of customer involvement?

(1)

Amsterdam Business School

The effect

of formal control tightness on performance in

Professional Service Firms:

What is the impact of customer involvement?

Name: Nick Schavemaker Student number: 10867872

Thesis supervisor: dr ir S.P. van Triest Document: Thesis

Date: June 18, 2016 Word count: 17.440

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam

(2)

Statement of Originality

This document is written by student Nick Schavemaker who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

Organizations should design their management control system in such a way that it ensures that employees act in accordance with the organizational objectives and strategies (Merchant & Van der Stede, 2012). This study examines the effect of formal control tightness in professional service firms. Professional services firms are of interest because they are considered as special type of firms which need distinctive theories of management and research on management control systems in these firms is lacking (Hall, 1968; Raelin, 1989; Von Nordenflycht, 2010). Professionals have a demand for autonomy and they want to be evaluated on the results of their activities rather than on the conformity of their actions to the rules and regulations of the enterprise (Raelin, 1985; Maister, 2003). Therefore, in hypotheses 1 and 2, it is expected that action control tightness has a negative effect and that tight results control has a positive effect on individual and business unit performance in professional service firms. Furthermore, it is expected that customer involvement affects the relation between formal control tightness and individual performance negatively and the relationship between formal control tightness and business unit performance positively. The hypotheses are tested using a multiple regression analysis on survey data of 275 professionals. The null hypothesis with respect to the first two hypotheses could not be rejected. Contrary to expectations, this study finds that tight action control tightness is positively related to business unit performance. It appears to be that the effect of action control tightness on both individual and business unit performance is more positive when the degree of customer involvement is high. The results indicate that there is no relation between results control tightness and individual or business unit performance, not even for a higher or lower degree of customer involvement.

(4)

1 Introduction

For decades, researchers have investigated the design of firms’ management control systems (MCSs). The purpose of MCSs is to ensure that employees’ actions are in accordance with the organizational objectives and strategies (Lowe, 1971; Merchant, 1985). Therefore, MCSs can be used as a solution to the agency problem, which is the existence of a conflict of interest between the agent and the principle (Eisenhardt, 1989). The contingency theory tells us that the design and use of the MCS is dependent upon the organizational setting and variables (Fisher, 1998, p. 48). An organization must design the MCS so that it matches the organization’s context because a better matched MCS increases the organizational and individual performance (Abernethy & Stoelwinder, 1991; Fisher, 1998). Organizations can design their MCS by choosing which type of controls to use or by deciding the tightness of the controls.

Merchant and Van der Stede (2012) distinguish the following three types of controls: results controls, action controls and soft controls. Results and action controls are seen as formal controls in which the former measures and rewards individual performance and the latter holds employees responsible for certain actions/behavior that must be performed (Ouchi, 1979; Merchant & Van der Stede, 2012). Soft controls, also referred to as informal controls, try to ensure that employees are able to do a good job, increase the engagement into self-monitoring and are designed to create an organizational culture which encourages mutual-monitoring by employees (Merchant & Van der Stede, 2012). Control tightness does not have a consistent definition and operationalization in the literature (Van der Stede, 2001; Fisher, 1998; Simons, 2013). In this research, the tightness of control refers to the degree of flexibility in the MCS. This depends on the extent or scope of the MCS and the level of tolerance for deviations within the MCS. Tight results control means the use and monitoring of performance targets more frequent, objective and precise. Tight action control means the use and monitoring of many specific standardized processes, policies and procedures, job descriptions and rules which employees must follow.

In this study, professional service firms (PSFs) are investigated because they are considered as special type of firms which need distinctive theories of management and research on MCSs in these firms is lacking (Hall, 1968; Engel, 1970; Sorensen & Sorensen, 1974; Raelin, 1989; Von Nordenflycht, 2010). PSFs are characterized by ‘knowledge intensity’, ‘low capital intensity’, ‘professionalized workforce’, ‘task complexity’ and

(6)

‘customer involvement’ which distincts them from bureaucratic organizations (Raelin, 1985; Abernethy & Stoelwinder, 1995; Løwendahl, 2005; Maister, 2003; Alvesson, 2004; Von Nordenflycht, 2010). Furthermore, the differentiation within PSFs has implications for the effectiveness of control mechanisms and the organizational design (Von Nordenflycht, 2010). For example, the degree of customization or standardization has been noted as a common point of differentiation with important strategic and organizational implications (Maister, 1993; Von Nordenflycht).

In this study, the research question is: “What is the effect of formal control tightness on individual and business unit performance in Professional Service Firms, and does this differ for the degree of customer involvement?”. To examine this research question, first, the effect of formal control tightness on individual and business unit performance will be investigated. Subsequently, the moderating effect of customer involvement on the relationship between formal control tightness and performance will be examined. With this research question, this paper aims to get a better understanding of the effects from the use of formal controls in the professional services context. This paper contributes to current research by empirically examining the relationship between formal control tightness and performance of individuals and business units in professional service contexts. This study investigates the implications of the level of customer involvement on the relation between formal control tightness and performance of individuals and business units in PSFs.

This paper is structured as follows. The next chapter discusses literature and presents the research hypotheses. This chapter provides a theoretical framework which results in presumptions about some specific relations. Chapter three illustrates the research methodology and data used for this study. Chapter four discusses the results of the analysis and shows if the proposed hypotheses can be confirmed or not. Finally, the paper provides a conclusion as well as a review of the implications and limitations of this study.

(7)

2 Theoretical framework and hypotheses development

In this chapter, relevant literature is explored in order to understand the topic, its concepts and to develop hypotheses. The first section explains the agency theory and clarifies the need and relevance for MCSs. The next section introduces MCSs, the different types of control and the concept of control tightness. The following sections explain the PSFs and customer involvement. Finally, based on this theoretical framework, the hypotheses are provided.

2.1 Agency theory

Jensen and Meckling (1976) defined the agency relationship as follows: “a contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent” (p. 310). The agency theory is the existence of a conflict of interest (the agency problem) between the agent and the principle (Eisenhardt, 1989). Agency problems arise because contracts are costly written and enforced (Fama & Jensen, 1983, p. 327). The agency theory is concerned with resolving two problems occurring in an agency relationship (Eisenhardt, 1989). The first problem is the agency problem that occurs when the principle and the agent have different goals and it is difficult or expensive for the principle to determine if the actual behavior of agent is in line with the firm (Eisenhardt, 1989). The second problem is that the principal and the agent have different attitudes towards risk and therefore prefer to take different actions (Eisenhardt, 1989). In both cases, the agent will not always act in the best interests of the principal (Jensen & Meckling, 1976).

The principal can monitor and reward the agent’s work to ensure that the agent does what is in his own best interest (Jensen & Meckling, 1976). Due to monitoring, the agent is more likely to act on behalf of the principal, so the principal controls the agent’s behavior (Eisenhardt, 1989, p. 59). Another way to align the interests of the principle and the agent is to bind the agent by a contract (Jensen & Meckling, 1976). In this way, the agent is motivated by rewards to take actions which are desired by the organization. The costs of monitoring and bonding form together with a residual loss (which is the loss due to the divergence between the agent’s decisions and the decisions which would maximize the welfare of the principles) the total of the agency costs (Jensen & Meckling, 1976).

(8)

2.2 Management Control Systems

In the literature, there is not one constant definition of MCSs (Simons, 1990). The definition of an MCS has evolved over the years from one focusing on more formal, financial information to assist managerial decision making to one that embraces a much broader scope of information and strategy formulation (Simons, 1990; Otley, 1994). First, this section discusses the need for MCSs and some definitions of MCSs. Hereafter, the definitions of an MCS and the different types of MCSs according to Ouchi and Merchant and Van der Stede are discussed. This section ends with an explanation of the concept of control tightness.

In recent years, the MCS literature had problems with the definition of its concepts (Otley, 1994; Chenhall, 2003; Tessier & Otley, 2012). Terms as management accounting systems, control systems, management accounting and control systems, organizational control, control mechanisms, management control and MCS are often used by each other (Ouchi, 1979; Chenhall, 2003; Malmi & Brown, 2008; Tessier & Otley, 2012). This makes it difficult to compare the results of different studies (Tessier & Otley, 2012). The purpose of MCSs is to ensure that employees’ actions are in accordance with the organizational objectives and strategies (Lowe, 1971; Merchant, 1985). If all employees act in the organization’s best interest, MCSs would not be necessary (Merchant & Van der Stede, 2012). But because employees do not always act in the best interest of the firm (agency problem), managers must take actions to prevent them of undesirable behavior and encourage desirable behavior (Merchant & Van der Stede, 2012). MCSs can therefore be a solution to the agency problem.

Anthony is one of the first researchers who gave a definition to management control (Simons, 1990; Otley, 1994; Hofstede, 1981). He defined management control as “the process by which management assure that resources are obtained and used effectively and efficiently in the accomplishment of the organization’s objectives” (Anthony, 1965, p. 17). In his theoretical study, Anthony (1965) distinguished management control from both strategic planning and operational control as three separate processes of organizational planning and control. This definition of management control focuses on strategy execution and implementation, the strategies are taken as given.

Lowe (1971) describes management control as a process of ensuring that what in some senses 'ought' to be done is done and of detecting when it is (p. 3). He defines an MCS as a system of organizational information seeking and gathering, accountability, and feedback

(9)

designed to ensure that the enterprise adapts to changes in its substantial environment and that the work behavior of its employees is measured by reference to a set of operational sub-goals (which conform with overall objectives) so that the discrepancy between the two can be reconciled and corrected for (Lowe, 1971, p. 5).

Simons (1990) has a slightly different view on management control than the authors mentioned before. While Anthony and Lowe focus, in their definition of management control, mainly on assuring that organizational objectives are achieved (strategy execution and implementation), and thereby, they separate this from strategy formulation (Anthony, 1965; Lowe, 1971). In Simons’ (1990) opinion, there is a strong relationship between management control and strategy formulation. Management control does not only concerns monitoring of employees but also includes motivating the firm to anticipate on the current and expected strategic uncertainties (Simons, 1990). In his definition, strategy formulation plays an important role in the MCS. He defines MCSs as the formal, information-based routines and procedures managers use to maintain or alter patterns in organizational activities as the formalized procedures and systems that use information to maintain or alter patterns in organizational activity (Simons, 1990, p. 128).

Merchant and Van der Stede (2012) define management control as all the devices or systems managers use to ensure that the behavior and decisions of their employees are consistent with the organization’s objectives and strategies (p. 6). In their opinion, the function of management control is to influence behaviors in desirable ways, and therefore, management control has a strong behavioral orientation (Merchant & Van der Stede, 2012). Management control can help an organization to achieve their objectives (Merchant & Van der Stede, 2012). Merchant and Van der Stede (2012) explain that management control is, together with objective setting and strategy formulation, one of the processes of the management function. Management control is the back-end of the management process and has the same meaning as strategy execution and implementation. Merchant and Van der Stede (2012) tell that an MCS is the collection of the management controls that are used in an organization.

2.2.1 Type of MCSs

The contingency theory tells us that the design and use of the MCS is dependent upon the organizational setting and variables (Fisher, 1998, p. 48). An organization must design the MCS so that it matches the organization’s context, a better matched MCS increases the

(10)

organizational and individual performance (Abernethy & Stoelwinder, 1991; Fisher, 1998). An organization can design its MCS by the choice of using certain type of control mechanisms. Merchant and Van der Stede (2012) distinguish three types of controls: results controls, action controls and soft controls. Results and action controls are together seen as hard or formal control mechanisms (Merchant & Van der Stede, 2012). The three type of controls identified by Merchant and Van der Stede are similar to Ouchi’s (1979) fundamental mechanisms through which organizations can seek to cope with this problem of evaluation and control. Ouchi (1979) calls them market mechanisms, bureaucratic mechanisms and clan mechanisms.

Results controls (market mechanisms) measure and reward individuals’ performance (Ouchi, 1979). Thereby, it keeps employees accountable for the results that he/she achieves on a task and are often associated with bonuses or promotion possibilities when achieving the targets (Merchant & Van der Stede, 2012). Results controls do not directly focus on employees’ behavior but they indirectly influence the behavior of employees by measuring the consequences of their actions and decisions (Davila, 2005; Merchant & Van der Stede, 2012). Examples of results controls are performance targets, budgets and different types of incentives. There are three conditions which have to be present for results controls to work effectively. An organization must know what results are desired in the areas being controlled, the employees should be able to influence the results where they are controlled for and the organization must be able to measure the results properly to evoke the desired behaviors (Merchant & Van der Stede, 2012).

Action (behavior) controls (bureaucratic mechanisms) hold employees responsible for certain actions/behavior that must be performed to complete a task (Ouchi, 1979; Merchant & Van der Stede, 2012). In contrast to results controls, do action (behavioral) controls influence organizational actors’ behavior directly by prescribing the actions they should take (Davila, 2005). They operate in “real-time”, during task performance, rather than “ex post”, as with results controls (Abernethy & Brownell, 1997). Action controls are the most direct control mechanisms because they are implemented to ensure that employees perform certain actions that are in organization’s best interest (Merchant & Van der Stede, 2012). Examples of action controls are monitoring activities, reviewing of plans or budgets, standardized processes, physical limitations, policies and procedures, job descriptions, rules and instructions to monitor, evaluate and reward employees’ actions. Action controls can only be used effectively if organizations can determine which behavior/actions is/are desirable and if they

(11)

are able to ensure that the desirable actions are performed (Merchant & Van der Stede, 2012; Ouchi, 1979). In order to do this, organizations should obtain knowledge on what actions are desirable and they must be able to observe the actions of the employees (Merchant & Van der Stede, 2012).

Soft controls (clan mechanisms) refer to the personnel and cultural controls within the MCS of an organization. Personnel controls can help to clarify expectations for the employee, to ensure that each employee is able to do a good job and to increase the employee’s engaging into self-monitoring (Merchant & Van der Stede, 2012). Personnel controls can be implemented by effective personnel recruitment and placement, training, job design and the provision of information and other resources (Merchant & Van der Stede, 2012). Cultural controls are designed to create an organizational culture which encourages mutual-monitoring by employees (Merchant & Van der Stede, 2012). Following Merchant and Van der Stede (2012), an organizational culture which facilitates this can be created by way of codes of conduct and ethics, group rewards, social and physical arrangements, employee rotation and tone at the top and others. Personnel and cultural controls are not used in this research. Therefore, these types of controls are not discussed in further detail.

2.2.2 Control tightness

In addition to choose the types of controls used in the MCS, an organization must decide the tightness of the controls. Control tightness does not have a consistent definition and operationalization in the literature (Van der Stede, 2001; Fisher, 1998; Simons, 2013). Fisher (1998) says that a tight MCS usually has targets that are hard to meet, cannot be revised and where only little target slack is tolerated. Anthony explains control tightness in relation to the budgeting process (Van der Stede, 2001). He defines a tight control system as one in which a manager’s performance is evaluated primarily on her ability to attain budgetary objectives during each reporting period (Anthony & Govindarajan, 1998, pp. 436-437). Thereby, Anthony means that tight controls depend on the properties of the budgeting process, such as the amount of emphasis on meeting the budget, the amount of detail in budget reviews, etc. (Van der Stede, 2001). Anthony describes the tightness of a (budgetary) control system by the following elements (Van der Stede, 2001):

1. Amount of emphasis on attaining budget targets;

2. Degree of budget commitment, i.e. whether revisions are allowed during the year; 3. Amount of detail in interim budget reviews;

(12)

4. Degree of tolerance for interim budget deviations; and,

5. Degree of involvement of top management in the regular activities of business units. Merchant (1981) introduced the concepts of administrative control and interpersonal control as forms of budgetary control systems which closely relate to the tight control concept of Anthony. Merchant (1981) defines administrative control as organizational control systems with (1) greater participation of middle- and lower level management in budget-related activities, (2) the placement of greater importance on achieving budget targets, (3) more formal patterns of communication and (4) greater business system sophistication (use of technology). An administrative control system is more likely to be present in larger, more diverse and more decentralized organizations (Merchant, 1981). Interpersonal control systems have the opposite characters and are more likely to be present in smaller, less diverse and more centralized organizations (Merchant, 1981).

In another study, Merchant (1985) defines tight control from a more outcome-perspective, he defines a tight control system as one that provides a high degree of certainty that employees act as the organization wishes. In his perception, the tighter the control system is, the more likely is the change that the company reaches its objectives. This would mean that tight control systems are better than loose control systems from a company’s perspective. Merchant (1985) argues that tightening control systems can be done by:

1. Defining goals such that they are more complete, more specific and/or more congruent with organizational objectives;

2. Changing communication patterns (effectiveness, timeliness, frequency and convincingness) in such a way that employees will better understand and accept organizational objectives;

3. Monitoring of actions and/or results more frequent, detailed and/or timely; and,

4. Enhancing the value of rewards to employees and defining stricter relationships between rewards and performance.

As seen above, the concept of control tightness does not have a constant definition. For this research, control tightness refers to the degree of flexibility in the MCS. The control tightness depends on the extent or scope of the MCS and the level of tolerance for deviations within the MCS. Tight control means in this case doing things more frequently/timely and allowing less flexibility/deviations. The concept of control tightness used in this research show similarities with the elements of Anthony’s concept of control tightness.

(13)

As mentioned earlier, results controls relate to the measurement of performance targets and incentivizing individuals for their performances. When results control is tight, this means the use of more frequent, objective, precise, complete, financial and timely performance targets and budgets. The monitoring of these targets and budgets are more formal and frequent. Furthermore, when results control is tight, employee participation in target setting is low and targets are not likely to be adjusted. Rewards are mainly monetary in nature when results control is tight. When results control is loose, the MCS should have the opposite characteristics regarding the measurement, monitoring, employee participation in target setting and rewarding of performance.

Action controls relate to the use of monitoring activities, reviewing of plans or budgets, standardized processes, physical limitations, policies and procedures, job descriptions, rules and instructions to monitor, evaluate and reward employees’ actions. Tight action control means the use of many specific standardized processes, policies and procedures, job descriptions and rules which employees must follow. The monitoring of these actions is frequent and formal, the evaluation is objective and rewards are also mainly monetary in nature. Loose action control would allow employees to exercise judgement in determining how to apply procedures and would allow deviation from certain procedures. When action control is loose, the MCS should have the opposite characteristics for the monitoring, evaluation and rewarding of actions.

2.3 Professionals

Professionals mainly work in complex tasks settings (Abernethy & Stoelwinder, 1995). They are highly educated and have been trained to perform these complex tasks independently and solve problems which arise in the execution of these tasks with their specific knowledge and experience (Derber & Schwartz, 1991). Therefore, professionals are usually paid relatively high salaries and are rewarded with special rights in the organization (Raelin, 1985, 1989).

Blau and Scott (1962) set out five characteristics of professionalism. The first characteristic is that professional decisions and actions are governed by universalistic standards which are based on independent objective criteria (Blau and Scott, 1962, p. 60). Blau and Scott (1962) mention that to apply these standards, professionals are required to master specialized knowledge and skills. Von Nordenflycht (2010) mentions this as the central characteristic associated with professionals. The mastering of this body of knowledge and the acquiring of appropriate skills requires a period of specialized training (Blau and

(14)

Scott, 1962, p. 60). The period of required training and education is generally long for professionals. The second characteristic of professionalism Blau and Scott (1962) mention is that professionals owe a specific professional expertness. Professionals’ expertness is limited to deal within a specific area (Blau & Scott, 1962). The third characteristic of professionalism given by Blau & Scott (1962) is that professionals maintain an affective neutral relation with their clients. Professional codes of ethics condemn emotional involvement with the client (Blau & Scott, 1962, p. 61). Fourth, professionals achieve their status and success by their individual performance in accordance with laws and regulations subjected on their profession (Blau & Scott, 1962). Blau and Scott (1962) mention as last characteristic of professionalism that professionals’ decisions must be made without self-interest. This is in line with the third characteristic that professionals must act in accordance to a code of ethics.

Professionals believe they can work independently and autonomously without the need for supervision (Raelin, 1985, p. 155). Professionals have a demand for autonomy and they wish to make their own decisions without external pressure from managers or others (Raelin, 1985; Von Nordenflycht, 2010; Abernethy & Stoelwinder, 1995; Sorensen & Sorensen; 1974; Scott, 1965). This preference for autonomy and aversion for supervision might be due to their expertness obtained in their long educational careers and high salaries. This need for autonomy could lead to difficulties for professionals integrating in organizations with formal organizational control processes, also known as administrative ore bureaucratic organizations (Raelin, 1985). Researchers claim that bureaucratic organizations limits or conflicts with professionals in their autonomy (Hall, 1968; Engel, 1970; Raelin, 1985). Professionals have the knowledge and experience to perform complex tasks and are socialized to perform these independently and autonomously without supervision (Raelin, 1985; Abernethy & Stoelwinder, 1995). It is assumed that social and self-controls and less formal controls are more functional for the management of professionals (Abernethy & Stoelwinder, 1995).

2.4 Professional Service Firms

There is not one clear definition of PSFs but they are often associated with large law-, consulting- and accounting firms (Alvesson, 2004; Von Nordenflycht, 2010). PSFs are often described as Knowledge Intensive Firms (KIFs) with a high presence of intangible assets and where most of its professionals are directly in contact with the market/customers (Alvesson, 2004). PSFs are of interest because researchers assume that they are very different and more

(15)

complex than traditional manufacturing- or service firms (Raelin, 1989; Alvesson, 2004; Løwendahl, Revang & Fosstenløkken, 2001; Løwendahl, 2005; Von Nordenflycht, 2010). This section explains some unique characteristics of PSFs and why there is a need for distinctive theories of management.

Von Nordenflycht (2010) identifies three distinctive characteristics which are often associated with PSFs: knowledge intensity, low capital intensity and a professionalized workforce. The most fundamental characteristic is knowledge intensity, which indicates that the firm relies on a highly intellectual skilled workforce (Alvesson, 2004; Von Nordenflycht, 2010). This workforce is highly educated and typically trained in a standardized body of knowledge which is common to all professionals in that sector and is certified by the relevant professional authority (Løwendahl et al., 2001, p. 913). Low capital intensity means that the firm’s production does not rely on tangible assets, such as production facilities, machines and inventory (Von Nordenflycht, 2010). The third identified distinctive characteristic of PFSs is that they have a highly professionalized workforce (Von Nordenflycht, 2010). The professional services involve a high degree of personal assessment judgement by the experts and are constrained by professional norms of conduct (Løwendahl et al., 2001). Von Nordenflycht (2010) specifically refers to the professionals’ specific expertness (knowledge), the universalistic standards (regulation and control) and the professional codes of ethics (ideology).

Another characteristic of PSFs which is mentioned in the literature is task complexity (Alvesson, 2004; Raelin, 1985; Abernethy & Stoelwinder, 1995). Task complexity, which describes the relationships between task inputs (the required acts and information cues), will be an important determinant of human performance through the demands it places on knowledge, skills and resources individuals need for successful task performance (Wood, 1986, p. 66). Wood (1986), defines three types of task complexity: component, coordinative and dynamic. Component task complexity refers to the number of different tasks that need to be performed and the number of different information cues that need to be processed (Wood, 1986). Coordinative complexity refers to the nature of the relationships between task inputs and task products (Wood, 1986, p. 68). This type of task complexity is related to the required time, sequence, interaction and location to perform the task (Wood, 1986). The last type of task complexity, dynamic complexity refers to the frequency in the changes which have to be made within the set or sequence of required acts and information cues or knowledge and skills requirements for performing the task (Wood, 1986). The last two types of task

(16)

complexity can also be seen as the variability and unpredictability of the tasks in their job which makes them complex (Wood, 1986). Overall, it means that the more components, coordination and dynamics within the task, the more complex they are (Wood, 1986).

Furthermore, PSFs are often subject to a high level of customer involvement during the provision of their services (Maister, 2003; Løwendahl et al., 2001; Cook, Goh & Chung, 1999; Kellogg & Nie, 1995). This relates to the ‘service’ part of PSFs. Most of PSFs’ client relations are complex and require a high degree of customization (Alvesson, 2004). Maister (2003) argues that the high degree of customization and face-to-face interaction with the client are two fundamental dimensions driving the need for distinctive managerial theories of PSFs. Customer involvement is strongly related to the characteristic task complexity because both characteristics make the service provision of professional services complex and unpredictable. Customer involvement is discussed in more detail in section 2.8, after discussing the characteristics of service provisioning in the next section.

2.5 Services

The growth of the service sector in major economies in the last decades shows the importance of this sector and the importance of service firms (Mills & Moberg, 1982; Cook et al., 1999). But still, there is a lack of clear definitions of services. Services and service firms have been defined in different ways that leads to some confusion as to what constitutes a service and service firms (Cook et al., 1999). Quinn, Baruch and Paquette (1987) define services to “include, all economic activities whose output is not a physical product or construction, is generally consumed at the time it is produced, and provides added value in forms (such as convenience, amusement, timeliness, comfort, or health) that are essentially intangible concerns of its first purchaser” (as cited by Cook et al., 1999, pp. 319-320). Service firms are often associated with manufacturing firms since many services are accompanied by facilitating goods (Mills & Moberg, 1982; Cook et al., 1999).

Still, many researchers argue that services have characteristics which cause fundamental differences between manufacturing and services (Mills & Margulies, 1980; Mills & Moberg, 1982; Cook et al., 1999). First, as is clearly mentioned in the definition from Quinn et al, (1987), the output of services production is (mainly) intangible (Mills & Moberg, 1982; Vargo & Lusch, 2004; Cook et al., 1999). This means that services cannot be seen, tasted, felt, heard or smelled before they are bought (Kotler & Armstrong, 2010, p. 269). This brings a lot of uncertainty regarding the quality of the service for the customer and

(17)

makes the measurement of the performance difficult (Løwendahl, 2005). Second, while products are produced at one point in time and consumed in a later point in time, services are produced and consumed at the same time (Sasser & Arbeit, 1976; Vargo & Lusch, 2004). If the output of services is not consumed at the time of production, it is forever lost (Mills & Margulies, 1980, p. 260). Services cannot be stored or possessed as inventory (Mills & Margulies, 1980; Vargo & Lusch, 2004; Kotler & Armstrong, 2010). Furthermore, because of the involvement of humans in the provisioning of service, they are characterized by heterogeneity (Mills & Margulies, 1980; Vargo & Lusch, 2004; Kotler & Armstrong, 2010). Because of the high degree of customer involvement, service delivery is an uncertain process. The production of services is not like a standard production in a manufacturing firm. Heterogeneity in services make it difficult to generalize and standardize (Vargo & Lusch, 2004; Løwendahl, 2005). Finally, services are characterized by inseparability (Vargo & Lusch, 2004). This means that people are involved in a great part of the product and the high level of customer involvement of customers (Vargo & Lusch, 2004; Løwendahl, 2005). The service provider and the customer must interact simultaneously for the service to be received (Vargo & Lusch, 2004). Therefore, services cannot be delivered without close cooperation between the service provider and the client and there is a high degree of uncertainty in terms of what the delivered service is going be exactly (Løwendahl, 2005).

2.6 Customer involvement

In previous studies, authors have used customer involvement, reliance, participation, interaction, contact, co-production and co-creation together. Cermak, File and Prince (1994) define customer participation as “the customer’s behaviors related to specification and delivery of services” (p. 90). Cook et al. (1999) refer to customer involvement as the level of interaction the customer has with the service system and the level to which a customer can actually affect the service delivery process (p. 329). Chase (1981) used the term customer contact what he explains as the degree to which the customer has direct contact with the service provider for the service creation.

Løwendahl et al., (2001) make a classification between two types of services, based on the level of customization. The level of customization is closely linked to the type of interaction with the client, from a hands-off type of interaction which involves a lower level of customization to services who need a deep involvement of both the client and the service provider (Løwendahl et al., 2001). Customized services require a lot of interaction and close

(18)

cooperation between the service supplier and the customer (Løwendahl, 2005). Services which need intense involvement are characterized by complex interdependencies and are difficult to coordinate (Løwendahl et al., 2001). Therefore, it can have a major impact on the performance of the organizational system.

“The high degree of customization in PSFs and innovation in other KIFs (Knowledge Intensive Firms) make traditional management principles such as standardization, routinization, and supervision difficult to apply. This is not to say that these principles are absent in KIFs, but they typically play a less prominent role than in most other organizations: they are typically used more flexibly.” (Alvesson, 2004, pp. 23-24)

Alvesson (2004) indicates that some formal control mechanisms are difficult to apply when there is a high degree of customer involvement. The interaction with, and involvement of customers make the tasks unpredictable. Tasks with a high degree of customer involvement are characterized by complexity, interdependence between tasks and are difficult to coordinate. In order to coordinate the service provision, there is need for more frequent interaction between subordinates and superiors. This could lead to more frequent monitoring and evaluating activities.

2.7 Hypothesis development

In this section, the explained variables in the previous sections are linked together and four main hypotheses are developed. Firstly, the expected relation between action control tightness and performance will be explained. Secondly, the expected relation between results control tightness and performance will be discussed. Both hypotheses are divided into two sub hypotheses, the relation with individual performance and the relation with business unit performance. It could be reasonable to expect that the tightness of a type of control could have a negative effect on individual performance but would be beneficial for the firms’ performance, or that it has a positive effect on individual performance but a negative effect on firm performance. Furthermore, these first two hypotheses between control tightness and performance allows for the impact of customer involvement in these relations to be explored. This leads to the third hypothesis which predicts the effect of customer involvement on the relationship between results control tightness and performance. Finally, the last hypothesis is about the effect of customer involvement on the relationship between action control tightness

(19)

and performance. Just as for the first two hypotheses, the last two hypotheses are divided into two sub hypothesis, one for individual performance and one for business unit performance.

2.7.1 Control tightness & performance

The contingency theory tells that the effectiveness of the design of an MCS is dependent upon contextual variables (Chenhall, 2003). An effective MCS increases the performance or organizations and individuals (Abernethy & Stoelwinder, 1991; Fisher, 1998). An organization can design its MCS by choosing type and tightness of control mechanisms. Large organizations that have more decentralized structures are associated with more formal, traditional MCS (Abernethy, Bouwens & Van Lent, 2004; Chenhall, 2003). PSFs are mainly large organizations with decentralized structures. This would assume that they would rely on formal, traditional MCSs. But PSFs rely on a highly intellectual skilled workforce because they employ mainly professionals with a high level of education (Von Nordenflycht, 2010). The professionals working in PSFs earn typically above average and have high status (Alvesson, 2004). Their demand for autonomy and aversion for supervision and formal organizational processes reduces the willingness to abide by routines and procedures developed by the organization (Maister, 2003; Løwendahl, 2005; Von Nordenflycht, 2010; Raelin, 1985). This indicates that professionals do not prefer tight formal controls.

According to the self-determination theory, autonomy is one of the three intrinsic psychological needs that will lead to increased performance, satisfaction and general well-being (Ryan & Deci, 2000). Furthermore, Hackman and Oldham (1976) linked task autonomy to motivation. They explain with their job characteristic model that employees with high task autonomy should feel a strong personal responsibility for the outcomes of their work (Hackman & Oldham, 1976). The strong personal responsibility for the outcomes employees feel can result in high internal work motivation and high work effectiveness (performance) for those individuals (Hackman & Oldham, 1976). Langfred and Moye (2004) developed a comprehensive model of causal linkages between individual task autonomy and individual performance. They argue that task performance can benefit from task autonomy by increased motivation, capitalization of information asymmetries or by better alignment with task and organizational structures (Langfred & Moye, 2004).

Ouchi (1979) argues that if an organization depends heavily upon monitoring, evaluating and correcting in an explicit manner, it is likely to offend people’s sense of autonomy and self-control, and it will probably result in an unenthusiastic, purely compliant

(20)

response. Because professionals attach a high value towards their task autonomy, it would probably have an even more negative effect on their enthusiasm and response, and therefore, on their performance. So when action control is tight, as in most bureaucratic organizations, professionals do not have the autonomy which they demand in which they decide how to perform a certain task (Engel, 1970). Therefore it is expected that employees in PSFs perform better with loose action control, compared to tight action control. When employees’ performance is negatively affected by tight action control, this would also affect the performance of the business units they work in.

Raelin (1985) indicates, based on the demand for autonomy that professionals desire to be evaluated on the results of their activities rather than on the conformity of their actions to the rules and regulations of the enterprise. They will be challenged and incentivized to meet targets which are in line with organizational objectives (Merchant & Van der Stede, 2012). Maister (2003) also argues that an organization should provide specific targets, measure, evaluate and reward performance and keep them accountable for their results in order to motivate professionals.

“You must provide clear goals; give prompt feedback and reward performance quickly; treat them like winners, involving them in decision-making and seeking their opinion often; give them autonomy in their work, but hold them strictly accountable for results; be tolerant of their impatience, and provide variety in their work experiences, always keeping the next challenging goal out front.” (Maister, 2003, chapter 15)

This could mean that tight results control would not negatively affect professional’s performance, as is expected for tight action control. Professionals prefer to be evaluated on their results instead of their actions. Therefore, it is expected that tight results control has a positive effect on individual performance. This would also positively affect the performance of the business units they work in. Based on the arguments mentioned above, the following is hypothesized:

Hypothesis 1a: There is a negative relation between action control tightness and individual performance in PSFs.

Hypothesis 1b: There is a negative relation between action control tightness and business unit performance for PSFs.

(21)

Hypothesis 2a: There is a positive relation between results control tightness and individual performance in PSFs.

Hypothesis 2b: There is a positive relation between results control tightness and business unit performance for PSFs.

2.7.2 Customer involvement as moderator

As mentioned before, professionals working in PSFs have complex, non-routine and unpredictable tasks because of the close interaction with, and involvement of customers. Therefore, some types of MCSs might not be suitable. In order for results controls to work effectively, an organization must know what results are desired, employees should be able to influence the results and the organization must be able to measure the results properly to evoke the desired behaviors (Merchant & Van der Stede, 2012). The effectiveness of action control mechanisms depends on the ability to determine which actions/behaviors are desirable (Merchant and Van der Stede, 2012; Abernethy & Brownell, 1997). For tasks with a high degree of customer involvement, it is difficult to determine which actions and results are desired. So the criteria for an effective application of action and results controls are not applicable for the tasks which have a high degree of customer involvement.

Abernethy and Brownell (1997) empirically studied the role of results, action and personnel controls on individuals in research & development (R&D) divisions of a large Australian industrial company and a major US scientific organization. This is a good representation of PSFs because R&D settings are often dominated by professionals. Abernethy and Brownell (1997) find when task uncertainty is highest (as it is when there is a high level of customer involvement), reliance on personnel controls is significantly and positively related to performance, while reliance on either accounting or behavior controls has significantly poorer implications for performance (p. 244). This indicates that the use of tight action and results control affects the individual’s performance in PSFs more negatively when the degree of customer involvement is higher.

Hirst (1983) studies the effects of accounting performance measures on individuals in both high and low task uncertainty situations. He finds that tasks that have a high degree of uncertainty and complexity are associated with low reliance on accounting performance measures (Hirst, 1983). Furthermore, Hirst (1983) finds that the use of accounting performance measures in high task uncertainty situations increases the incidence of dysfunctional behavior of subordinates. This finding, together with the results of Abernethy

(22)

and Brownell (1997), suggests that it can be expected that tight action and results controls will have a negative effect on individual performance when the level of customer involvement increases.

Langfred (2005) investigates the moderating effect of task interdependence on the relationship between autonomy and team performance. He uses the definition of Guzzo and Shea (1992) in which task interdependence is the degree to which the interaction and coordination of team members are required to complete tasks (Langfred, 2005). Langfred (2005) argues that task complexity is an important influencer of team performance. He finds empirical evidence that team performance benefits from high levels of individual autonomy when task interdependence is low and that teams performed worse with high levels of individual autonomy when task interdependence is high (Langfred, 2005). This suggests that there is a negative relationship between individual autonomy and team performance under conditions of high task interdependence (Langfred, 2005). In view of the point that services which require a high level customer involvement are characterized by complex interdependent tasks, it could be expected that tight action control (allows for less individual autonomy) has a more positive effect on business unit performance when customer involvement is higher.

Furthermore, because services with a high degree of customer involvement are difficult to coordinate, it could be logic that there needs to be more communication, guidance, procedures and supervision within the teams to perform well. Williams, Macintosh and Moore (1990) argue that in more complex situations with reciprocal interdependencies (mutual exchange of inputs and outputs, which is the case when customer involvement is high) there was more frequent (formal) interaction between superiors and subordinates. They also find that when as less interactions occur with superiors, the performance of the team is negatively affected (Williams et al., 1990). This indicates that tight action and results control are needed to coordinate tasks characterized by a high degree of customer involvement. Therefore, it could be expected that tight action and results controls have a more positive effect on business unit performance for a higher degree of customer involvement. Based on the arguments mentioned above, the following is hypothesized:

Hypothesis 3a: The relation between action control tightness and individual performance in PSFs is more negative for a higher degree of customer involvement.

(23)

Hypothesis 3b: The relation between action control tightness and business unit performance in PSFs is more positive for a higher degree of customer involvement.

Hypothesis 4a: The relation between results control tightness and individual performance in PSFs is more negative for a higher degree of customer involvement.

Hypothesis 4b: The relation between results control tightness and business unit performance in PSFs is more positive for a higher degree of customer involvement.

(24)

3 Methodology

Empirical data used for this study was gathered by conducting a survey. The survey was conducted as part of the PSF Thesis Survey Project by the Faculty of Economics and Business (FEB) of the University of Amsterdam (UvA). This chapter explains the research methodology and the sample used for this research. The first section gives an introduction to the survey project. Next, the sample size and its characteristics are outlined. Furthermore, the constructs used in this research are set out and finally this chapter explains the research model.

3.1 Thesis survey project

The PSF Thesis Survey Project 2015-2016 conducted by the UvA aims to investigate the design and use of management accounting & control systems in PSFs through a pre-developed written questionnaire. More specifically, the project investigates the way in which PSFs design their management accounting & control systems, and the effect that these systems have on the professionals working in these organizations. The questionnaire used for this project contains many different constructs so all participating students were able to develop their own original research question for their theses. In order to participate in the project, students needed to submit at least seven completed questionnaires from suitable respondents. All gathered respondents had to send an e-mail to confirm that they completed the questionnaire. The students had to forward this e-mail to the research coordinator. These e-mails serve as proof that the survey was completed by that individual and increase the reliability of the sample. After the submission of at least seven suitable respondents, the complete response set was made available for the participating students to use in writing their theses.

In order to maximize the quality and effectiveness of the questionnaire, two separate pre-tests have been developed. The first pre-test, carried out with fourteen participants, was developed to assess the quality of the items used to measure the different types of MCSs. Participants had to match different types of explicit or implicit control tightness with statements which they felt the definition was most closely resembled. The four statements for each type of control tightness with least number of incorrect matches were selected for inclusion in the questionnaire. The mismatches in the statements for each type of control tightness ranged from a lowest of zero incorrect matches to a highest of six incorrect matches

(25)

for implicit behavior control tightness. Another pre-test have been carried out with 20 participants to assess the quality of the questionnaire as a whole. The participants were asked to evaluate the questionnaire based on the content, clarity, appearance and amount of time required to the survey. This resulted only in minor changes in wording and the inclusion of additional answers by a few questions.

3.2 Sample

The survey project is based on a joint effort process, the work load of gathering respondents was shared between the participants in the project. Therefore, it was possible to obtain a substantial number of questionnaires from high quality respondents working in PSFs. The respondents should be mid-level employees of medium to large PSFs. The sample population does not include occupations in non-profit, publicly owned organizations such as universities (education/teaching), government organizations, NGOs and social work agencies, but it does include medical doctors which may be employed at public or not-for-profit hospitals. Respondents and the organizational units they manage had to meet the following four criteria: 1. The respondent has worked the field for more than 3 years but probably less than 10 years. The respondent should be out of the apprenticeship/learning phase of their job and should be able to work relatively independently with minimal guidance for his/her supervisor. However, the respondent should not be in a position to exercise power over the direction of the organization.

2. The respondent is not an owner/partner or board member of the organization. In other words, the respondent needs to be subject to the management accounting and control system rather than design it.

3. The respondent works in a medium/large size organization (> 50 employees). 4. The respondent must speak and understand English at a business level.

The complete response set which was gathered by all participants in the project contained 372 respondents. From those 372 respondents, 41 did not finish the questionnaire. This led to a first reduction of the sample size to 331 respondents. One of the first questions in the questionnaire was “Have you read the above definition and remarks?”. In my opinion, there is a chance that people who did not read the definition and remarks did not fully understand the questionnaire and therefore, their answers are untrustworthy. Eleven respondents did not read the definitions and remarks so they are also excluded from the sample size which leads to a sample of 320 respondents. Thirdly, I analyzed the time the

(26)

respondents spent on fulfilling the questionnaire, it takes some time to read and answer all the questions. One respondent spent 2 minutes and 36 seconds on fulfilling the whole questionnaire. This is substantially shorter than the other respondents, which spent at least 6 minutes. I excluded this respondent which leads to a sample size of 319 respondents. Finally, as mentioned as one of the requirements on respondent criteria above, the respondents have to work in the field for more than 3 years to be able to work relatively independently with minimal guidance and to be out of the apprenticeship/learning phase. In the sample of 319 respondents, there were still 36 respondents who answered they have less than three years of work experience in their current field. Excluding those 36 respondents leads to a final sample size of 283 respondents with at least three years of work experience. Table 1 shows characteristics of the respondents included in this sample.

As shown in table 1, the average age of the respondents is 35 years and 62% of the respondents is male. From all respondents, 62.7% have completed a Master degree or higher, the respondents have on average 7.6 years of work experience in their work field and 6.2 years in their current organization. The study is conducted in The Netherlands, however some participants have gathered respondents from other nationalities. From the final sample, 67.7% of the respondents is Dutch (N = 279). Most respondents work as a professional in accounting and thereafter in a consulting firm but the PSFs included in this sample are very diverse. Respondents within the “Other”-category commented for instance that they worked in the field IT management, tax advisory, internal auditing or (financial) auditing. Therefore, we conclude that this category is also a good representation of professionals.

(27)

Table 1 Respondent characteristics

Characteristics

Age N / Mean (SD) / Min / Max 279 / 35 (0.5) / 24 / 63 Years work experience

in field N / Mean (SD) / Min / Max 269 / 7.6 (0.2) / 4 / 11 Years work experience

in organization N / Mean (SD) / Min / Max 281 / 6.2 (3) / 1 / 11

Gender Male 62 %

Female 38 %

Education Bachelor degree or lower 37.3 %

Master Degree 51.6 %

PhD or other professional doctorate degree 11.1 %

Work field Accounting 24.4 %

Consulting management/strategic 9.2 % Medicine/Physician practices 8.5 % Consulting IT/Technology & Software development 7.1 % Marketing, Advertising & Graphic design 5.4 % Consulting HR & Recruiting 5.0 % Investment Banking/Management & Real estate 4.3 % Engineering (Consulting) & Architecture 3.9 %

Law 3.5 %

Financial advising 2.8 %

Risk management services 2.8 %

Project management 2.1 % Research/R&D 1.4 % Pharmaceutical 1.1 % Other 18.7 % 3.3 Survey constructs

The questionnaire used for this project consisted of 22 different constructs. For this study, five of these constructs have been used as variables for answering the research question which are presented in Appendix A. Constructs of individual and business unit performance are used for the two dependent variables. The constructs for action control tightness and results control tightness are used as the independent variables. One construct is used to examine the moderating effect of customer involvement. Furthermore, three items are used as control variables to test if the effects of the independent variables on the dependent variables, and the moderating effect on this relationship, also holds when we take into account the

(28)

effects of these variables. The construct for individual performance is measured on a 5-point likert scale with: (1) needs much improvement, (2) needs some improvement, (3) satisfactory, (4) good and (5) excellent. All other constructs used in this research are measured on the following 5-point likert scale: (1) strongly disagree, (2) disagree, (3) neutral, (4) agree and (5) strongly agree.

Before making a construct of the items to measure variables, all negative worded items were reversed. On all constructs, a varimax principal rotated component analysis (PCA) was conducted. First, a PCA was conducted on all constructs separately. Second, a joint PCA was conducted on all latent variables. The results of this joint PCA are summarized in table 2. To determine which items belong to which underlying factor, we consider all the loadings above .500 to be significant. Field (2013) suggests that a factor loading of minimal .500 is an appropriate threshold for factor loading cut-offs. Next to the PCA, the constructs have been tested on reliability with the Cronbach’s Alpha coefficient in order to decide which items to include in the construct. If constructs consisted of multiple components, the reliability is also tested for each component. The steps taken in making constructs and the results on PCA and reliability testing are described for each construct in the next paragraphs.

3.3.1 Individual Performance

For the measurement of individual performance, the questionnaire used two different perspectives based on Podsakoff and Mackenzie (1989) and Welbourne, Johnson and Erez (1998). First, the respondents are asked to rate their own performance from the perspective of their supervisor, this construct is called Individual Performance In-Role (Podsakoff & Mackenzie, 1989). Questions from this perspective look like: “Imagine that you are in the role of your supervisor, how do you think your supervisor would rate...”. Second, respondents need to rate their own performance based on twelve questions, these questions form the construct Individual Performance General. For this research, the construct Individual Performance General is used because of the consistency with the measurement of business unit performance.

The constructs Individual Performance General measures three components of performance. The first four items relate to doing things specifically related to one's job description, the next four items examine the obtainment of necessary skills to progress through one's organization and the last four items question creativity and innovation in one's job and the organization as a whole (Welbourne et al., 1998). Thus, it would be expected that

(29)

each of those four items load on a different component. When conducting a PCA on this construct and a joint PCA, 3 components were discovered as expected. The items Q36_1, Q36_42, Q36_3 load on one component (C1), Q36_5, Q36_6, Q36_7, and Q36_8 load on a second component (C2) and Q36_9, Q36_10, Q36_11, and Q36_12 load on a third component (C3). Only item Q36_4 does not load. Taking a closer look at the items, it seems understandable that this item does not load on the first component. Consequently, item Q36_4 is deleted. Taking together all eleven items gives a Cronbach’s Alpha of .879 which indicates that this construct is very reliable. Moreover, the components C1, C2 and C3 have separate also a very reliable Cronbach’s Alpha of .746, .846 and .843 respectively. Therefore, it is decided to use the eleven items in the construct of individual performance general to measure the level of individual performance.

3.3.2 Business Unit Performance

Business unit performance was measured in the questionnaire using six items from the construct of King and Clarkson (2015). When conducting a PCA on the construct, four items load on one component and two items also load on another. When having a closer look at the items in the construct, all six items measure similar elements of business unit performance, and the distinction does not seems logical. Furthermore, when testing the joint PCA, all items load on the same component. Therefore, all six items have been included in the construct. The Cronbach’s Alpha of this construct is .819. This indicates that the construct is very reliable.

3.3.1 Action Control Tightness

Action control tightness was measured in the questionnaire using a construct based on studies of Aiken and Hage (1968), Van de Ven and Ferry (1980), Bodewes (2000), Cunningham and Rivera (2001) and Van der Stede (2001). The construct consisted of eight items. Four items from this construct (Q4_10, Q4_11, Q4_12 and Q4_13) have been recoded into new variables due to negative wording.

When conducting a PCA on this construct, the recoded items load on a different component than the other four items (Q4_1, Q4_2, Q4_4 and Q4_8). This could be explained by the distinctions which have been made between explicit action control tightness and implicit action control tightness. The first four items (which have not been recoded) represent explicit action control tightness where a tight system is defined as one with a lot of action controls in terms of amount and scope. The four recoded items represent implicit action

(30)

control tightness where a tight system is defined as one which does not allow any deviation from standard processes, procedures, rules and routines. Taking together all eight items gives a Cronbach’s Alpha of .551 which is below the desired value of at least .700.

Assessing the reliability for the constructs of explicit action control tightness and implicit action control tightness separately gives a Cronbach’s Alpha of .744 for explicit and .616 for implicit. The reliability score for explicit action control tightness is sufficient and all items load on one component in the PCA on the construct. However, when testing this in the joint PCA, only three out of the four items for explicit action control tightness loaded on the same component. Still, given the reliability score for implicit action control tightness is below the desired level and the items within this construct load on two different components it is better to use the construct of explicit action control tightness. Therefore, it is decided to use the construct of explicit action control tightness containing four items in order to measure the level of action control tightness.

3.3.2 Results Control Tightness

For measuring results control tightness, the questionnaire used a construct based on Jaworski, Stathakopoulos and Krishnan (1993), Aiken and Hage (1968), Simons (1987) and Van der Stede (2001) and some items were added. The construct consisted of eight items measured on a 5-point likert scale and two numeric items. The numeric items asked the number of performance targets used in the evaluation of the respondent’s job performance and how often he/she discusses their performance measures with his/her supervisor. Three items (Q5_9, Q5_10 and Q5_11) have been recoded into new variables because the question was negatively worded.

When conducting a PCA on this construct, 2 components were discovered. The items Q5_1, Q5_4, Q5_6, Q5_7 and Q5_12 have a factor loading of >.500 on another component and the former negatively worded items load with >.500 on one component. Both the numerical items did not load. The loading on two components could be explained by the distinction which have been made between explicit results control tightness and implicit results control tightness. The items Q5_1, Q5_4, Q5_6 and Q5_7 represent explicit results control tightness where a tight system is defined as one with a lot of results controls in terms of amount and scope. The three recoded items and Q5_12 represent implicit results control tightness where a tight system is defined as one which does not permit any deviation from established goals, targets and performance measures. Therefore, it would be expected that the