1
Amsterdam Business School
A study on the impact of control tightness on individual
performance in professional service firms.
Name: Idriss Maghiouzi Student number: 10591680 Date: June 20, 2016
MSc Accountancy & Control, specialization Control
Faculty of Economics and Business, University of Amsterdam Supervisor: Dr. ir. S.P. van Triest
2 Abstract
In this master thesis I investigate the impact of action control tightness and results control tightness on the individual performance in professional service firms. In
addition, I examine the effect of the interaction between a high human capital intensity and control tightness on the individual performance. With survey data among 321 respondents I find evidence that individual performance is negatively impacted by action control tightness. Whereas no evidence is found that results control tightness influences the individual performance. Furthermore, I don’t find evidence that an interaction of high human capital intensity and results or action control tightness has a negative impact on the individual performance. To my knowledge, this is the first study that looks at those specific types of control and their effects. The results of this study are interesting and more research in this area is needed.
Statement of Originality
This document is written by student Idriss Maghiouzi 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 Preface
This report is written as the final part of the master’s degree Accountancy and control at the University of Amsterdam. I would like to take this early opportunity to thank my supervisor dr. ir. Sander van Triest for his encouragement, help and guidance during the process of writing this thesis. I also would like to thank Helena Kloosterman for giving me the opportunity to join her PSF project. I want to thank everyone who filled in the survey, especially my colleagues at Strategy& who showed support from the beginning. A special thanks goes to my family who has always encouraged me to get the most out of myself.
Idriss Maghiouzi
4 Table of Contents 1. Introduction ... 5 2. Theory ... 7 2.1. Literature review ... 7 2.1.1. Service firms ... 7
2.1.2. Professional service firms... 8
2.1.3. Human capital ... 10
2.1.4. Management control systems ... 11
2.1.5. Control tightness ... 19
2.2. Hypotheses development ... 22
2.2.1. Control tightness and individual performance ... 22
2.2.2. Human capital intensity and individual performance ... 24
3. Research methodology ... 26 3.1. Sample selection... 26 3.2. Survey design ... 26 3.3. Variable measurement ... 27 3.4. Pre-test questionnaire ... 35 4. Results ... 36 4.1. Descriptive statistics ... 36 4.2. Correlation analysis ... 36 4.3. Regression analysis ... 38 4.4. Additional analysis ... 39 5. Concluding discussion ... 41 5.1. Main findings ... 41 5.2. Limitations ... 43 5.3. Future research ... 44 References ... 45 Appendix ... 49
Appendix A: Factor analysis IVs and moderator ... 49
Appendix B: Descriptive statistics ... 51
Appendix C: Additional regression analysis ... 57
5
1. Introduction
Over the past three decades the Professional Service Firm (PSF) sector has emerged as one of the most rapidly growing, profitable, and significant sectors of the global economy (Hinings et al, 2015).
Businesses are competing with complex, rapidly changing, and knowledge-intensive business models driving the need to better understand the role of
performance measure systems (PMS) and how they can better meet managerial needs (Widener 2007). The importance of professional service firms (PSFs) in our economy nowadays justifies understanding them and how they are managed. Despite their significance, little is known of the determinants of their performance (Greenwood et al. 2005). More specifically, getting a better understanding of how the performance is controlled and impacted by control measures, especially when the control is tight, is of great importance. The significance of PSFs in our economy requires us to better
understand how they are managed. Therefore, I believe that my study will be a valuable contribution to the existing literature on management control.
Greenwood et al (2005) state: “PSFs are substantially different from … traditional manufacturing firms (Lowendhal 2000, p. 31)”. They further describe that “PSFs are so different that to apply theories from other firms is not only inapplicable… but may be dangerously wrong (Maister 1993, p. xvi)”. The main difference is in the fact that PSFs’ assets consists for a significant part of human capital. Additionally, managing human capital requires different types of management control systems compared to
manufacturing firms. PSFs are distinctive from other categories of organizations which make it more interesting to research. In this respect, my study will give some insights in the way the PSFs are controlled.
Since PSFs are a different type of organizations, I assume that performance measurement will also be different compared with ‘traditional firms’. I also believe management control choices have impact on the organizational and individual performance. One of these choices is making the management control tight. The combination of tight controls and professional services firms fascinates me and in this
6 study I investigate the impact of control tightness on individual performance in PSFs. I will answer the following question:
How does control tightness impact the individual performance in PSFs?
The remainder of this study is organized as follows. Section ’Theory’ gives an overview of the reviewed literature on service firms, PSFs, MCS and control tightness. Also it provides the developed hypotheses. Section ‘Research methodology’ discusses the research design, sample selection and variable measurement. Fourth section gives an overview of the results and the conclusion can be found in section five.7
2. Theory
In the first part of this section I will start with giving a description of what service firms are. Then, I will discuss the PSFs, MCS and control tightness. In the second part I will, based on the theoretical framework, develop the hypotheses. After some research and consideration I came to the conclusion that there is no theory that would fit and
incorporates all the variables of my proposed models. To support the proposed models I have decided to look for evidence in different sources.
2.1. Literature review 2.1.1. Service firms
Before I discuss what professional services firms are, I will first define service firms as such. Auzair & Langfield-Smith (2005) state that ‘the unique characteristics of pure services include (1) intangibility of services, (2) inseparability of production from consumption, (3) perishability of services, and (4) heterogeneity in service products (Fitzgerald et al., 1991; Hope and Muhlemann, 1997; McColl et al., 1998)’. Service organizations are somehow unique and require special attention (Auzair & Langfield-Smith (2005). Service industries span retail, health care, education, hospitality and professional services. Therefore, limited generalizations can be drawn from such a narrow focus (Modell, 1996, 1998). Services firms are in forms and types. Silvestro et al. (1992) indicates three categories of service firms: professional services, mass services and service shops.
Professional services are organizations with relatively few transactions, highly customized, process oriented, with relatively long contact time, with most value added in the front office and where considerable judgment is applied in meeting customer needs. Management consultancy firms and corporate banking are examples of professional service firms.
Mass services have many customer transactions, involving limited contact time and little customization. Most value is added in the back office and little judgement is needed from the front office. Examples of mass service firms are tobacco and
8 Service shops’ category involves firms that fall between professional and mass
services. Examples of service shops are hotels and rental services. The next section will elaborate on professional services firms since this is the focus of this study.
2.1.2. Professional service firms
Greenwood et al (2005) define PSFs as those whose primary assets are highly educated (professional) workforce and whose outputs are intangibles services encoded with complex knowledge. Van Nordenflycht (2010) states that PSFs are distinctive from other types of organizations. He indicates three PSFs characteristics: professionalized
workforce, knowledge intensive and have low capital intensity. The characteristics are briefly described below and should be seen independently from each other.
Professionalized workforce combines two features: ideology and self-regulation. Leicht & Layman (2006) state: A professional ideology consists of a set norms
manifested both in explicit ethical codes enforced by professional associations and internalized preference often developed during professional training. Self-regulation means that professionalized occupations have strong control over the practice of the occupation (Von Nordenflycht 2010). In other words, the profession has a self-regulated monopoly as it controls the certification that’s required to enter the occupation. Self-regulation is also meant to maintain the trustworthiness of the profession in the eyes of clients. Certification control is for example applicable in accountants and lawyer firms. However, this does not apply for all professional service firms since working experience can also lead to professionalized workforce, without having obtained a specific
certification. This could be the case in consultancy firms for example.
Merchant and Van der Stede (2007, p. 26) state:’ Reengineering guru Michael Hammer even defines a professional as someone who is responsible for achieving a result rather than [for] performing a task.’ Meaning that employees in a professional service firm are expected to achieve certain results rather than performing tasks or taking particular actions. The emphasis here is obviously on the output or outcome of a process and not on the input side. This can be seen as one of most common forms of results controls, which will be discussed later in this paper.
9 Knowledge intensity indicates that production of a firm’s output relies on a
substantial body of complex knowledge (Starbuck, 1992). Knowledge intensity implies that labor has more importance in firms. PSFs are ‘knowledge engines for business’ (Lorsch and Tierney 2002, p. 14). Alvesson (2000, p. 1101), describes
knowledge-intensive firms as those for which “most work can be said to be of an intellectual nature and where well-educated, qualified employees form the major part of the work force.” Greenwood et al. (2005) define the professional workforce as the critical asset of the PSF because it embodies, operates, and translates the knowledge inherent in the firm’s output. Ditillo (2004) states: in general, knowledge intensive firms refer to those firms that provide intangible solutions to customer problems by using mainly the knowledge of their individuals. Typical examples of these companies are law and accounting firms, management, engineering, consultancy firms and research centers. This definition of Ditillo (2004) overlaps with the concept of professional service firms as discussed before. However, whenever the term knowledge intensive firms is used in this study, it is referring to professional service firms.
From the contributions above it can be concluded that knowledge intensity involves the importance of knowledgeable and well educated people in achieving the organization’s objectives.
Low capital intensity indicates that the production in firms doesn’t require significant amounts of nonhuman capital such as inventory and equipment (Von Nordenflycht 2010). This implies a greater importance of employee skills and knowledge. Auzair & Langfield-Smith (2005) argue that one important feature of service firms is the significant human participation in the ‘production process’.
Employees ‘ability to solve complex problems using their knowledge and skills is more valuable to PSFs. Consequently, non-human capital is not the main determinant of the performance and output in those companies.
In summary, PSFs rely more on knowledgeable and professionalized human capital to achieve the organization’s objectives but producing intangible services. Given
10 the fact that that human capital is the most important asset here and that it is one of the variables I am looking at in this study, I will elaborate more on it in the next section. 2.1.3. Human capital
Human capital is described as that which embraces the abilities and know-how of men and women that have been acquired at some cost and that can command a price in the labor market because they are useful in the productive process (Parnes 1984, p. 32). Hsu (2008) states: ‘human capital includes knowledge, skills and capabilities of employees that create performance differentials for organizations’. She further describes human capital as employees’ competencies which can lead to effective job performance. Ditillo (2004) further states: ‘knowledge-intensive firms’ capital consists predominantly of human capital, their critical elements are in the minds of individuals and heavy
demands are made on the knowledge of those who work in them (Ekstedt, 1989, p. 7)’. Skaggs et al. (2004) states: ‘Organizations can increase their human capital levels through selection and/or training (i.e., buy and/or make). Selection refers to an
organization’s efforts to increase its human capital by hiring employees with high levels of education and expertise from the external labor market. Training, on the other hand, refers to an organization’s effort to increase its human capital through internal
developmental activities of its current employees (Snell and Dean, 1992).’ In the same study, Skaggs et al. (2004) conclude that human capital has a very strong positive main effect on organizational performance. This can be explained by the argument that
human capital is vital to the production and delivery processes of service organizations. This can be applied for professional service firms as well since human capital is their main asset.
11
2.1.4. Management control systems
There has been a lot of research conducted on management control systems (MCS) and various definitions are documented by several theorists.
Management control is defined as the control managers exercise over other
managers (Fisher, 1998). Simons (1995) defines MCSs as means used by senior managers to successfully implement their intended strategies. Anthony ( 1965) defines
management control in terms of assuring that organizational objectives are achieved. Merchant (1985) describes management control as the process by which corporate-level manager ensure that mid-level managers carry out organizational objectives and
strategies. Chenhall (2003) states: The definition of MCS has evolved over the years from one focusing on the provision of more formal, financially quantifiable information to assist managerial decision making to one that embraces a much broader scope of information.
Management control systems are put in place to ensure the organization achieves its objectives. Different types of these control systems are described below.
2.1.4.1. Levers of Control of Simons
Simons states (1995):”Management control systems are the formal, information-based routines and procedures managers use to maintain or alter patterns in organizational activities.” Simons distinguishes four interrelated levers of control: Beliefs systems, Boundary systems, Diagnostic and Interactive control systems.
Beliefs system is “the explicit set of organizational definitions that senior
managers communicate formally and reinforce systematically to provide basic values, purpose, and direction for the organization” (Simons, 1995, p. 34). Widener (2007) explains that beliefs systems ensure that core values are communicated in order inspire and motivate employees to search, explore, create, and expend effort engaging in appropriate actions. Ferreira & Otley (2009) state: ‘Beliefs systems guide the creative process of exploring new process of exploring new opportunities and instils widely shared beliefs (Simons, 1995)’. Organizational theory suggests that the beliefs system is
12 critical to high-performing organizations (Calfee, 1993; Analoui & Karami, 2002). In
other words, mission statements for example promote shared expectations among all levels and generations of employees and provides management with a unity of
direction. Beliefs systems correspond with personal controls of Merchant which will be discussed later in this study.
Boundary systems are formally stated rules, limits, and proscriptions tied to defined sanctions and the credible threat of punishment (Simons 1994). Boundary systems put some limitation on the creativity and dynamic environment that is stimulated by the Beliefs systems, this is needed to avoid employee engagement in high-risk behaviors. Widener (2007) states: ‘The boundary system communicates the actions that employees should avoid. Its purpose is to allow employees freedom to innovate and achieve within certain pre-defined areas. The boundary and beliefs systems are similar in that they both are intended to motivate employees to search for new opportunities; however, the boundary system does so in a negative way through the constraint of behavior while the beliefs system does so in a positive way through inspiration (Simons, 1995)’. In other words, Boundary systems make sure employees (do not) take certain actions. They basically control the actions of employees.
Diagnostic control systems are formal feedback systems used to monitor
organizational outcomes and correct deviations from preset standards of performance. Diagnostic controls systems - exemplified by business plans and budgets - are the prototypical feedback systems used to track variances from preset goals and manage by exception (Simons 1994). Measurement is a powerful tool because what gets measured gets managed, and what gets managed, gets done (Anthony and Govindarajan, 2004). Diagnostic controls represent a negative force for two reasons (Henri 2006). On the one hand, diagnostic use focuses on mistakes and negative variances. On the other hand, the sign of the deviation that is derived when outputs and goals are compared is reversed in the feedback signal to adjust the process.
Diagnostic controls are not only powerful tools but also conserve management energy because managers can focus their attention on investigating and correcting deviations to the plan (Sheehan 2006). The diagnostic system is intended to motivate
13 employees to perform and align their behavior with organizational objectives (Simons, 2000). Diagnostic system acts as a constraint on employee behavior according to Simons (2000). This is in line with the conclusion of Van der Stede (2001) who relates diagnostic controls to tight control.
Interactive control systems are formal systems used by top managers to regularly and personally involve themselves in the decision activities of subordinates. Any
diagnostic control system can be made interactive by continuous and frequent attention and interest from top management. The interactive system is intended to help the firm search for new ways to strategically position itself in a dynamic marketplace. Top managers choose which control system (e.g., PM, brand management, budget process) they want to use in an interactive manner (Widener 2007). The purpose of making a control system interactive is to focus attention and force dialogue and learning
throughout the organization (Simons 1994). For true interactive control, managers need to interact, there have to be regular face-to-face meetings with managers at different levels to get their input on what may be driving the changes, and this discussion should lead to new strategies (Simons 1994). Interactive control systems are more forward looking compared with Diagnostic controls which allow managers to manage results on an exception basis. Interactive systems are characterized by active a frequent dialogue among top managers.
Simons (2000) developed the Levers of Control framework. As described above, he identifies four systems: beliefs (e.g., core values), boundary (e.g., behavioral
constraints), diagnostic (e.g., monitoring), and interactive (e.g., forward looking, management involvement). These four systems work together to benefit the firm. Merchant & van der Stede (2003) chose for a different classification for management control systems. Their classification involves four types of control and will be discussed in the next section.
14 2.1.4.2. Merchant’s controls
Merchant & van der Stede (2003) indicate the following four different types of control. Action controls, results controls, cultural and personnel controls. The control types are discussed below.
Action controls
According to Merchant & van der Stede (2003) action controls “are the most direct form of management control because they involve taking steps to ensure that employees act in the organization’s best interest by making their actions themselves the focus of control” (p. 67). Action controls only work in an organization if the managers know what actions are desirable (or undesirable) and have the ability to ensure that the desirable actions occur. Action controls, Merchant & van der Stede (2003, p. 63) further describe, include behavioral constrains, pre-action reviews, action accountability and redundancy.
Behavioral constrains make it more difficult for people to take undesirable actions. The constraints, also seen as a negative form of action control, can be administrative and physical. Administrative constraints include, for instance, restrictions of decision making authority when it comes to expenditure approvals. Another type of
administrative control is segregation of duties, which should make it more difficult for a person to complete a certain task alone. Examples of physical constraints include locks on desks, computer passwords and limits on access to areas where valuable inventories and sensitive information are kept (Merchant & van der Stede 2003, pp. 67-68).
Pre-action reviews involve scrutiny of the action plans of the employees being controlled. Reviewers can approve or disapprove the proposed actions, ask for modifications or ask for a more carefully considered plan before final approval is granted (Merchant & van der Stede 2003, p.69). Examples of pre-action reviews are: requirement of approvals of expenditures of certain types and multiple reviews by different levels during organizational planning and budgeting processes.
Action accountability is about holding employees accountable for the actions they take (Merchant & van der Stede 2003, p. 69). The emphasis here is on the actions and not
15 necessarily the achieved results. Requirements for implementing action accountability
are: defining what actions are acceptable or unacceptable, communicating those actions to employees, observing or tracking what happens and rewarding good actions or punishing actions that deviate from the acceptable.
Redundancy is about having back up employees available which should lead to an increase in the probability that a certain task is satisfactory accomplished. This form of control is mainly used in security functions, computer facilities or other critical
operations.
Merchant & van der Stede (2007, p. 81) link two conditions to the effectiveness of action controls. First, organization can determine what actions are (un)desirable and second: organizations are able to ensure that the (un)desirable actions (do not) occur. Actions controls are seen as the most direct form of management control because they ensure the proper behaviors of the people on whom the organization must rely on (p.92).
Results controls
Results controls “induce employees to behave so as to maximize their chances of producing the results the organizations desires” (Merchant and Van der Stede 2003, p. 26). A prominent example of results controls is pay-for-performance. It involves
rewarding employees for generating good results, or punishing them for poor results. Rewards include, in addition to monetary compensation, job security, promotions, autonomy, and recognition. The combination of rewards linked to results informs or reminds employees what result areas are important and motivates them produce the results the organization rewards (Merchant and Van der Stede 2003, p. 23). Since results controls effect actions, employees are concerned about the consequences of the actions they take. Results controls have a stimulating role as the employees are encouraged to discover and develop their talents and get placed in jobs in which they will be able to perform well.
The authors further state that results controls only can be effectively used in organizations where the desired result areas can be controlled (to a considerable extent)
16 by the employees whose actions are being influenced and where the controllable result areas can be measured effectively (Merchant and Van der Stede 2003, p. 23)
Results controls are necessary for the implementation of decentralized forms of organizations with largely autonomous entities or responsibility centers. Delegated decision rights or task autonomy in combination with effective results controls should ensure that employees do not misuse their discretion and that they are appropriately rewarded (Merchant and Van der Stede 2007, p. 26-27)
One of the preventive-type controls which firms put in place to avoid potential lack of direction is results controls. As well defined results inform employees as to what is expected from them and encourages them to do what they can to produce the desired results (Merchant and Van der Stede 2007, p. 28). The authors define four steps to
implement results controls in an organization: 1) defining the dimensions on which results are desired or not desired, 2) measuring performance on these dimensions, 3) setting performance targets and 4) provide rewards to encourage behaviors that will lead to the desired results.
Moreover, according the Merchant and van der Stede (2007), there are three conditions which need to be met in order to achieve effective results controls: 1) organizations can determine what results are (not) desired, 2) employees whose behaviors are being controlled have significant influence on the results and 3) organizations can measure the results effectively. Although results controls are not effective in every situation, failure to satisfy all three effectiveness conditions will render results controls impotent. In addition results controls are an indirect form of control because they not focus explicitly on the employees’ actions (Merchant and van der Stede, 2007, p. 35).
Personnel controls
Personnel controls are based on the “employees’ natural tendencies to control and/or motivate themselves” (Merchant and Van der Stede 2007, p. 83-85). This form of controls should lead to ‘monitoring’. Management literature refers to
17 Personnel controls involve selecting and placement of people, training and job design
and provision of necessary resources.
Selection and placement are about finding the right people for a particular job. Organizations several techniques and devote considerable time and effort to employee selection and placement. This often involves reference checks on new employees and pre-screening (p. 84).
Once the firms have found the right the people, it trains them and provides them with useful information on what results or actions are expected and how the assigned tasks can be best performed. Professionals are often more interested in performing well in jobs they understand better. Training takes place in several forms such as in
classroom setting or informal talks between employees. The latter should lead to mutual employee monitoring (p.84).
Job design and provision of necessary resources is about designing a job to a person with the right abilities. This should increase the probability of success. Resource needed might include information, equipment, supplies, staff support, decision aids and freedom from interruption (p.85).
Cultural controls
Cultural controls “are designed to encourage mutual monitoring, a powerful form of group pressure on individuals who deviate from the group norms and values” (Merchant and Van der Stede 2003, p. 77). Cultural controls are most effective where members of a group have emotional ties to one another. Merchant and Van der Stede (2003) state that strong cultural controls can be created among co-workers where non-conformers are often pressured to accept group norms. Cultures are built on shared traditions, norms, beliefs, values, ideologies, attitudes and ways of behaving. The authors indicate five important methods to shape culture, and thus effect cultural controls: codes of conduct, group-based rewards, intra-organizational transfers, physical and social arrangements, and tone at the top.
Codes of conduct are formal written documents that provide broad, general statements of organizational values, commitments to stakeholders and the ways in which management would like the organization to function (Merchant and Van der
18 Stede 2007, p. 85). Group rewards are rewards based on collective achievement, which
comes in many forms. Common examples are bonus, profit sharing that provide
compensation based on corporate or entity performance in terms of accounting returns, profits of cost reductions (p. 88).
Intra-organizational transfers involve is also referred to as employee rotation and helps transmit culture by improving the socialization of employees throughout the organization. Transfers also potentially mitigate employee fraud by preventing employees becoming too familiar with entities, activities and colleagues and/or transactions (p. 90).
Physical arrangements involve office plans, architecture, interior décor etc. can also help shape the culture within an organization. Informal cultures with pen office arrangement and casual dress codes deliver the message about the importance of innovation and employee equality (p. 90).
Lastly, tone at the top should be consistent with the type of culture the management tries to create. Their behaviors should be in line with their statements. Managers serve as role models which are cited as a determining factor in creating a culture of integrity in their organizations. Management cannot say one thing and do another (p. 90).
Auzair and Langfield-Smith (2005) distinguish two types of management control systems: more bureaucratic MCS and less bureaucratic MCS. The first one involves actions, formal, tight, restricted, and impersonal controls. Less bureaucratic MCS includes results, informal, loose, flexible, and interpersonal controls. Both control types of Merchant and Simons can be less or more bureaucratic. In the next section I will be discussing one type of control which is seen as more bureaucratic: tight controls. I will first describe control tightness as such, then I will go more in depth on the tight action and results controls.
19
2.1.5. Control tightness
Van der Stede (2001) states: ‘tight control does not have a consistent definition and operationalization across studies. It has troubled theorists for many years (Simons, 1995, p.161)’. However, definitions by several management control experts will be outlined in this section as it’s crucial to understand the concept of control tightness in order to be able to conduct the study.
Tight control system is one in which a manager’s performance is evaluated primarily her ability to attain budgetary objectives during each reporting period
(Anthony and Govindarajan, 1998, pp. 436–437). This definition encompasses budgetary control meaning that control tightness relies on budget information. Merchant (1985) argues that tight control is characterized by its limits of the degrees of freedom the individuals have who are controlled. However, it does not necessarily involve detailed monthly reports and reviews. Merchant (1985) defines tight control as a system that provides high degree of certainty that employees act as the organization wishes. He believes tightening the control system can be done by the following:
1) Definition of goals: Tightening control means defining goals such that they are more complete, more specific, and/or more congruent with the organizational objectives.
2) Communication of goals: Communication is done in a way that employees better understand and accept organizational objectives. This may include
communicating more effectively, timely, frequently, and convincingly.
3) Monitoring: Tightening control involves more frequent, more detailed, and/or more timely monitoring of actions and/or results.
4) Rewarding: Tightening control means enhancing the value of rewards to the employees evaluated and the definition of stricter relationships between rewards and performance.
In summary, tight control is aimed to provide high degree of certainty of achieving the organization’s objectives. This is done through using a detailed, specific,
20 between the individual performance and the action and results controls. Therefore,
the tight form of these two types of control will be further discussed in the next section.
2.1.5.1. Tight action controls
Merchant and van der Stede (2003, p. 122) discuss in which ways tight action controls can be achieved.
Behavioral constrains as discussed before, involve administrative and physical constraints. Both can produce tight controls. Physical constraints include locks on desks to elaborate software and electronic security systems. Administrative constraints can produce tight control when decision making is restricted to higher organizational levels and when it’s guaranteed that those who do not have authority for certain actions cannot violate the constraints that have been established.
Pre-action reviews can be considered as a characteristic of tight control systems. Especially, if the reviews are frequent, detailed, and performed by diligent,
knowledgeable reviewers. Tight pre-action reviews include meticulous scrutiny of business plans and requests for capital by staff personnel and multiple levels of management, including top management.
Action accountability in control can lead to tight control and is dependent on the characteristics of (un)desirable actions, effectiveness of the action-tracking system and the reinforcements (rewards or punishments) provided. Tight control is well suited in organizations where the task related risk is high, meaning little deviations from desired actions can cause harmful effects for clients, employees, or the organization (Demartini, 2013, p. 124). Likewise, tight action accountability is not effective in contexts where there is a high level of uncertainty on the content of the task, so that it is difficult to state what a desirable action is and what is not.
2.1.5.2. Tight results controls
Similarly to action controls, results controls tightness depends on the characteristics of the definition of the desired result areas, the performance measures, and the
21 reinforcement or incentives provided (Merchant & van der Stede, 2003). In order to
achieve this, the following conditions need to be met.
Congruence is about making sure the organization’s objectives are clear and that the management understand them. Additionally, the chosen measurable performance dimensions should be good indicators of the organization’s true objectives.
Specificity and Timeliness, tight results control depends on having performance targets described in specific terms and in relatively short increments of time.
Communication & Internalizationof the performance targets is required in order to achieve tight result controls. High internalization is likely to be achieved when those whose behaviors are being controlled are allowed to participate in the goals setting.
Completeness means that the result areas defined in the management control system include all the areas in which the organization desires good performance and for which the employees involved can have some impact.
Performance measurement effectiveness has impact on results control tightness. Results controls rely on measures that are precise, objective, timely and understandable.
Incentives linked to accomplishment of desired results should lead to tight control. Rewards or punishments that are significant to the employees are directly and definitely linked to achievement of the results.
22 2.2. Hypotheses development
In this study I will answer the following question: How does control tightness impact the individual performance in PSFs? To be able to do this and to facilitate the research structure I have developed several hypotheses.
Building on the taxonomy by Merchant and Van der Stede, research on control tightness and literature on performance in PSFs and human capital, I will formulate my hypotheses.
2.2.1. Control tightness and individual performance
Shields et al. (2000) argue that tighter budget goals are associated with higher levels of individual performance. Merchant (1985) defines tight control as a system that provides high degree of certainty that employees act as the organization wishes. In other words, when an organization uses tight controls, it is more likely that the employees act
according to the organization's strategy and objectives. They will perform in the way it plans. When professionals have a clear picture of what they have to do and what not, they will probably act and perform the tasks as described by the company. Shields et al. (2000) further states that: experimental and survey research in accounting report a positive relationship between standard tightness and individuals' standard-related performance (Chow; Hirst & Lowy; Hofstede, 1967; Lindquist, 1995; Rockness, 1977; Stedry, 1960; Waller & Chow, 1985). Standard tightness is defined as the amount of resources needed to perform at the level of a standard minus the amount of resources provided to perform. Firms provide exact amount of resources needed to achieve certain targets and there is no room for flexibility.
23
Action control tightness, results control tightness, and individual performance. Control tightness involves all forms of control. In my study I am looking at the action and results controls. To search for more evidence for the hypotheses, I will study the relation between the two aforementioned sorts of tight control with individual performance in PSFs.
Merchant and Van der Stede (2003, p. 26) state that results controls induce employees to behave so as to maximize their chances of producing the results the organization desires. Results controls should lead to achieving the organization’s objectives. Auzair & Langfield-Smith (2005) argue that professional service firms are likely to put more emphasis on results controls as they are seen as less bureaucratic controls. Accordingly, PSFs don’t put emphasis on the more bureaucratic controls, in this case action controls.
Considering the contributions by Shields et al. and Merchant and based on other literature in the theory chapter, I predict that individual performance in PSFs is
impacted by the choice of by tight management controls. Assuming that human capital has impact on the results, and that professionals are willing to be assessed on the achieved results rather than what actions they took to achieve these results, I am expecting that the results controls have a positive impact on individual performance, whereas action controls have a negative impact on the individual performance in PSFs. This leads to the following hypotheses:
H1a: There is a positive relationship between results control tightness and individual performance in PSFs.
H1b: There is a negative relationship between action control tightness and individual performance in PSFs.
24
MODEL 1: CONTROL TIGHTNESS AND IMPACT ON INDIVIDUAL PERFORMANCE.
Model above shows that PSFs who put emphasis on results controls are likely to have high individual performance. On the other hand, emphasis on action controls implies low individual performance in PSFs.
2.2.2. Human capital intensity and individual performance
As stated before, knowledge intensity indicates that production of a firm’s output relies on a substantial body of complex knowledge (Starbuck, 1992). Complex work needs to be controlled by the operators who do it. Widener (2003) investigated and concluded that the importance of strategic human capital has influence on the design of
management control systems. Quin et al. (1996) argue that human capital, which enhances the knowledge of the firm, is a primary strategic resource of many firms.
Kehoe & Wright (2013) state: ‘[strategic] human resource management scholars have suggested that organizations can use performance-and commitment-oriented human resource (HR) practices to drive organizational effectiveness (Becker & Gerhart, 1996; Dyer & Reeves, 1995; Wright, Dunford, & Snell, 2001). The scholars have
increasingly recognized the importance of effective human capital management for organizational performance. As stated before, human capital has a very strong effect on the organizational performance.
Based on the research above we can conclude that that human capital intensity has influence on individual performance in organizations. Building on the previous and the fact that PSFs’ human capital consists of knowledgeable and highly skilled individuals, who prefer to be measured on the achieved results and who are valuable to firms, I believe that the intensity of human capital has influence on the relation between
H1a/b Low High Results control tightness x Action control tightness x Individual performance
25 control tightness and individual performance. I expect that the higher the intensity
(highly skilled people dealing with complex knowledge and problems), the more difficult it gets to use tight controls, which would then result in an effect on the individual performance. Therefor my next hypotheses are formulated as follows:
H2a: Results control tightness in combination with high human capital intensity is negatively correlated with individual performance in PSFs.
H2b: Action control tightness in combination with high human capital intensity is negatively correlated with individual performance in PSFs.
MODEL 2: CONTROL TIGHTNESS MODERATOD BY HCI AND IMPACT ON INDIVIDUAL PERFORMANCE.
My expectation, as shown in the model, is that tight action controls negatively
influences the individual performance in PSFs with high human capital intensity. I also predict that, contradictory with H1a, results control tightness combined with a high human capital intensity negatively impacts the individual performance in PSFs.
The next chapter will elaborate on how I will test the hypotheses and how I measure the variables.
H2a Human Capital Intensity +/+ Results control tightness -/-Individual performance +/+ Action control tightness -/-H2b Human Capital Intensity Individual performance
26
3. Research methodology
For my research I will use the thesis survey project of the University of Amsterdam. I will gather responses and add them to the initial database. There are couple of reasons why I decided to participate in the thesis project. Firstly, I find the research project very interesting because of its focus on PSFs as I am working part-time at one of these firms. Secondly, to conduct this research I won’t be able to gather enough responses on my own. The projects makes it possible to obtain a substantial number of questionnaires from high quality respondents. Finally, I have access to data.
To gather enough questionnaires, I use my network of friends and colleagues. All of them are working for professional service firms and meet the criteria of the research project.
3.1. Sample selection
The participants in the PSF project of the University of Amsterdam gathered at least seven complete questionnaires from suitable respondents. By submitting the required amount of surveys access was given to the entire dataset. The respondents have worked in professional service firms and are subjects to the management accounting and control system, and are not owners/partners of the organizations.
The initial dataset consists of 372 responses. In total I excluded 51 responses from the dataset. 48 of them had missing data and 3 respondents had a small standard deviation in their responses. The final dataset consists of 321 responses.
3.2. Survey design
The survey has the aim to look at the design and use of management control and accounting systems in PSFs. The survey covers many aspects of management control and its types. Some of these aspects are: control tightness, human capital intensity, professionalized workforce and customer reliance.
In this study the focus is on the impact of results and action control tightness on the individual performance in PSFs.
27 3.3. Variable measurement
The primary focus of the study is the effects of action control tightness and results control tightness on individual performance in PSFs.
3.3.1.1. Dependent variable: individual performance
The dependent variable in this research is individual performance. The respondent has to answer seven questions related to individual performance while being in the role of the supervisor. In addition the respondent has to answer twelve questions related to individual performance in general. The constructed variables are IND_PER_IR and IND_PER_GE for individual performance in role and in general respectively. In this study, the emphasis is on IND_PER_GE. For the sake of completeness, I include the main results of the variable IND_PER_IR in the additional analysis ‘section.
Individual performance in general
For the dependent variable the respondent has to rate his/her own performance (IND_PER_GE) using a five-point Likert scale. A scale of 1 indicates that a certain item needs much improvement, whereas a scale of 5 indicates an excellent rating. The following items are rated:
Finding improved ways to do things Seeking out career opportunities
Developing skills needed for your future career Coming up with new ideas
Creating better processes and routines Quality of work output
Customer service provided (internal and external) Making progress in your career
Accuracy of work
Quantity of work output
Working to implement new ideas Obtaining personal career goals
28 Before constructing the variables reliability analysis has to be performed. Testing the
reliability is done using Cronbach’s alpha. Cronbach’s alpha has a value between 0 and 1. The rule of thumb commonly accepted is that a minimum level of Cronbach’s alpha is around 0.5 to 0.6 (Nunnally, 1978). The questions related to individual performance in general have a Cronbach’s alpha of 0.88 which is a high level. This number cannot be improved.
Table 1 shows the results of the performed factor analysis, three components are extracted and the item ‘Customer service provided (internal and external) ‘is eliminated as it loads on a different factor. The first component involves four items related to improving ways of working and developing new ideas, the constructed variable is
IND_PER_GE_IMP (improvement). The second component relates to four items about career and personal development. The constructed variable that belongs to this
component is IND_PER_GE_CR (career). The third component includes three items related to work output, the variable constructed for this is IND_PER_GE_WO (work output).
Since individual performance in general is the main focus of this study, the three aforementioned variables will be tested separately and the results will be included in the appendix of this report.
29 Table 1: Factor analysis variable IND_PER_GE.
Component
1 2 3
Coming up with new ideas ,908 ,070 -,136
Working to implement new ideas ,851 ,124 -,104
Finding improved ways to do things ,772 -,046 ,176
Creating better processes and routines ,574 -,049 ,334
Seeking out career opportunities -,014 ,877 -,069
Making progress in your career -,078 ,847 ,090
Obtaining personal career goals ,043 ,827 ,016
Developing skills needed for your future career ,206 ,609 ,105
Accuracy of work -,142 ,075 ,847
Quality of work output ,190 -,044 ,767
Quantity of work output ,017 ,138 ,689
Customer service provided (internal and external) ,259 ,087 ,274 Individual performance in role.
In addition to the above, the respondent has to rate his/her performance in the role of the supervisor (IND_PER_IR). The following items are included and rated on a five point Likert scale. Scale 1 indicates a strong disagreement with the statement, vice versa scale 5 indicates a strong agreement.
This employee always meets all formal performance requirements of the job. This employee always performs all tasks that are expected of him/her.
This employee always completes all duties specified in his/her job description. This employee always engages in all activities that will directly affect his/her
performance evaluation.
This employee always fulfills all responsibilities required by his/her job. This employee always performs all essential duties.
This employee never neglects aspects of the job that he/she is obligated to perform.
30 The reliability analysis show a figure of 0.91 which is a very good level. This is a high
level of reliability and indicates that the questions above measure the same thing. 3.3.1.2. Independent variables: result and action control tightness
The independent variables in this research are results and action control tightness. The constructed variables are RCT (Result Control Tightness) and ACT (Action Control Tightness).
Results control tightness
For the result control tightness, the respondent has to answer eight questions related to results control tightness. The respondent has to rate the following statements on a five-point Likert scale. Whereas scale 1 indicates a strong disagreement, scale 5 indicates a strong agreement with the statement.
In my organization, employees are expected to meet pre-established goals/targets with no exceptions.
Employee attainment of goals/targets is checked constantly.
My supervisor frequently checks to make sure that I am meeting my performance targets.
In my job, there is a performance measure for everything.
My organization sets a large number of performance goals/targets that I am expected to meet.
Responding to new, unforeseen opportunities is considered more important by my supervisor than achieving pre-established goals/targets.
My supervisor is very considerate of my explanations of deviations from pre-established goals/targets.
In our organization, goals/targets are essentially a guideline rather than a true commitment.
There were three negatively worded items which I had to reverse code as they would measure the exact opposite otherwise. Reverse coded items are:
31 Responding to new, unforeseen opportunities is considered more important by
my supervisor than achieving pre-established goals/targets.
My supervisor is very considerate of my explanations of deviations from pre-established goals/targets.
In our organization, goals/targets are essentially a guideline rather than a true commitment.
Before I constructed the RCT variable by averaging the answers of the results control tightness, the reliability had to be determined. The questions related to results control tightness had Cronbach’s alpha of 0.58 which exceeds the minimum level of 0.55. However I decided to delete item: ‘My supervisor is very considerate of my explanations of deviations from pre-established goals/targets’. This improved the Cronbach’s alpha and resulted in a level of 0.66. This number cannot be further improved unless I delete more items or construct two small variables. The factor analysis (appendix) show two
extracted components. The first one contains five items and relates to explicit results control tightness, and is a management control system that is defined as one with a lot of controls in terms of amount and scope (Jaworski et al., 1993). The second component includes three items and measures implicit results control tightness. This is related to a management control system that is defined as one which does not permit any deviation from established goals/targets (Hage & Aiken, 1968; Simons, 1987; Van den Stede, 2001). Since the Cronbach’s alpha is on a reasonable level, I decided the keep RCT is one main construct and not to distinguish between explicit and implicit results control tightness.
Action control tightness
The ACT independent variable is measured by eight questions related to behavior control.
In my organization, we have rules for everything.
My supervisor frequently monitors the extent to which I follow established process, procedures and rules.
32 Established processes, procedures and rules cover all of my job tasks.
Employees in my organization are encouraged to adjust procedures to suit the situation.
Whatever situation arises, we have existing processes, procedures or rules to follow in dealing with it.
Employees in my organization are encouraged to use procedures flexibly. My job allows me to decide how to adjust rules to best perform my job tasks. The organization I work in primarily uses established processes, procedures and
rules to give broad guidelines as to how activities are to be performed.
Like the RCT questions, the respondent has to answer ACT related questions using a five-point Likert scale. Scale 1 indicates a strong disagreement whereas scale 5 indicates a strong agreement with the statement. Following questions were negatively worded and had to be reverse coded:
Employees in my organization are encouraged to adjust procedures to suit the situation.
Employees in my organization are encouraged to use procedures flexibly. My job allows me to decide how to adjust rules to best perform my job tasks. The organization I work in primarily uses established processes, procedures and
rules to give broad guidelines as to how activities are to be performed.
After performing the reliability analysis of the items related to action control tightness and in order to improve the Cronbach’s alpha, I had to delete item: ‘The organization I work in primarily uses established processes, procedures and rules to give broad guidelines as to how activities are to be performed’. The initial figure was 0.54, which improved to 0.68 after deletion of the item mentioned. Factor analysis show that the items load on two
different factors (see appendix for factor analysis). The first component consists of three items related to explicit action control tightness which involves a tight control system with a lot of controls in terms of amount and scope. Whereas the second component, also consisting of three items, has to do with implicit action control tightness and measurers the flexibility of the (tight) management control system.
33 The reported Cronbach’s alpha of 0.68 indicates a good reliability of the items.
Therefore, no further data is deleted nor extracted. I decided to use ACT as main construct to measure the results control tightness.
3.3.1.3. Validity analysis
To test if the variables for results control tightness, action control tightness and
individual performance each have one underlying factor that explains the items scores, I first run convergent validity analysis. This shows that the items for IND_PER_GE load on different factors and I therefor decided to split the dependent variable IND_PER_GE in three sub variables (IND_PER_GE_IMP; IND_PER_GE_WO; IND_PER_GE_CR). One item is left out of the selection as it loads on a fourth factor, I decided not to include it as it measures something that is out of scope of this study.
Second, I run the discriminant validity. The aim is to test whether there are indeed three different factors. This is done by including all questions related to individual performance (in general) to the factor analysis. The analysis show three factors related to individual performance.
3.3.1.4. Moderating variable: human capital intensity
To measure the intensity of human capital, the respondent has to describe the employees in his/her organization by rating the following:
Our employees are creative and bright.
Our employees are widely considered the best in our industry. Our employees are experts in their particular jobs and functions. Our employees are highly skilled.
Our employees develop new ideas and knowledge.
The constructed variable is HCI and summarizes the average of the answers related to human capital intensity .The items are measured using a five-point Likert scale. An item that is rated on scale 1 means the respondent strongly disagrees with the
34 the statement. The questions related to human capital intensity have a Cronbach’s alpha of 0.86 which is a sign of high levels of reliability. Factor analysis is attached in the appendix of this report.
3.3.1.5. Robustness check and control variable
In this study I use education as control variable. I expect that highly educated people perform better. As described in the second chapter, professional service firms are those for which “most work can be said to be of an intellectual nature and where
well-educated, qualified employees form the major part of the work force.” Alvesson (2000, p. 1101). PSFs need well educated and qualified people to achieve the organization’s objectives. Since achievements are determined by performance, I believe individual performance is dependent, amongst others, on how well educated people are.
Education is measured by answering a question related to the highest level of education completed. Possible answers are:
Bachelor degree or lower Master Degree
PhD or other professional doctorate degree (i.e. J.D./M.D./D.D.S)
I created a dummy variable for education: EDU. The recoded dummy is summarized in two categories: score of 0 indicates bachelor degree and a score of 1 indicates master degree or higher. Robustness checks are provided in section 4.4.
3.3.1.6. Model
SPSS will be used to test the hypotheses in this study, it is a software package for statistical analysis. Based on the variables as mentioned in the previous paragraphs, the models for H1a and H1b are:
H1a IND_PER_GE = β 0 + β 1 (RCT) + β 2 (EDU) H1b IND_PER_GE = β 0 + β 1 (ACT) + β 2 (EDU)
35 H2a tests the relation between the action control tightness, moderated by human capital intensity, and individual performance. H2b tests the relation between results control tightness, moderated by human capital intensity, and individual performance. The two models are presented below.
H2a IND_PER_GE = β 0 + β 1 (RCT) + β 2 (HCI) + β 3 (RCT * HCI) + β 4 (EDU) H2b IND_PER_GE = β 0 + β 1 (ACT) + β 2 (HCI) + β 3 (ACT * HCI) + β 4 (EDU) The regression analysis results are presented in the next chapter.
3.4. Pre-test questionnaire
In order to maximize the quality and effectiveness of the survey, two pre-tests were conducted.
The first one aimed to evaluate the quality of the items used to measure the variables the different management control forms. To assess this, an item sort task was used and two paper sheets with subjects were provided. On one paper definitions of the control constructs (types of control) were outlined. On the second paper, 52 statements were formulated to test those constructs. The task for the subjects was to match each statement to the definition they felt it most fit. Out of the twenty professionals asked to complete the pre-test, fourteen have actually completed it and performed the task.
The second pre-test involved the quality assessment of the entire survey. Another 20 subjects have looked at the online survey and answered some questions related to the content of the survey but also the presentation, clarity and the time amount needed to fill in the questionnaire. Some subjects shared written answers and others by telephone. The received feedback involved minor changes (terminology and incorporation of additional answers for some multiple choice questions).
36
4. Results
This section starts with an overview of the descriptive statistics. To better understand the key concepts individual performance and results/action control tightness, I believe it is valuable to present the statistics of the constructed summary variables. Answers to the individual questions can be found in appendix B. Next, regression analysis are presented, followed by additional analysis.
4.1. Descriptive statistics
As presented below, the average scores on individual performance constructs are relatively higher than the scores on the control tightness variables. A possible
explanation for this is that people would rate their performance rather high than low. Also, there is a lot of variety in the answers which makes these results more interesting to study.
4.2. Correlation analysis
Pearson and non-parametric Spearman correlations are shown in the next table. As expected there is a positive correlation between IND_PER_GE and RCT (r = 0,054), however the correlation is not significant (p = 0,334). Also, as predicted, there is a significant negative correlation between IND_PER_GE and ACT (r = -0,119; p = 0,033). The moderator HCI is positively correlated with IND_PER_GE (r = 0,306; p = 0,000). The correlation between IND_PER_GE and Education is positive but not significant (r = 0,044; p = 0,431).
N Minimum Maximum Mean Std. Deviation
ACT 321 1,25 4,38 2,89 ,52179 RCT 321 1,13 4,25 2,85 ,52408 HCI 321 1,20 5,00 3,80 ,71899 IND_PER_IR 321 1,29 5,00 4,12 ,63574 IND_PER_GE 321 1,31 5,00 3,72 ,60803 IND_PER_GE_WO 321 1,67 5,00 3,97 ,61630 IND_PER_GE_CR 321 1,00 5,00 3,57 ,82529 IND_PER_GE_IMP 321 1,00 5,00 3,61 ,76289 Valid N (listwise) 321
37 Co rre latio n tabl e (p val ues in ita lic ) IN D _P ER _GE IN D _P ER _GE _WO IN D _P ER _GE _C R IN D _P ER _GE _I M P IN D _P ER _I R R C T A C T H C I ED U ,7 86 ** ,8 48 ** ,8 07 ** ,4 27 ** ,0 54 -,1 19 * ,3 06 ** ,0 44 ,0 00 ,0 00 ,0 00 ,0 00 ,3 34 ,0 33 ,0 00 ,4 31 ,8 00 ** ,5 58 ** ,5 16 ** ,4 42 ** ,0 31 -,0 84 ,2 60 ** ,0 44 ,0 00 ,0 00 ,0 00 ,0 00 ,5 75 ,1 34 ,0 00 ,4 35 ,8 52 ** ,5 48 ** ,4 90 ** ,3 48 ** ,0 41 -,0 39 ,3 02 ** ,0 90 ,0 00 ,0 00 ,0 00 ,0 00 ,4 60 ,4 83 ,0 00 ,1 09 ,8 23 ** ,5 12 ** ,5 12 ** ,3 09 ** ,0 51 -,1 51 ** ,2 11 ** -,0 27 ,0 00 ,0 00 ,0 00 ,0 00 ,3 67 ,0 07 ,0 00 ,6 36 ,3 39 ** ,3 81 ** ,2 51 ** ,2 32 ** -,0 64 -,1 38 * ,2 50 ** ,0 66 ,0 00 ,0 00 ,0 00 ,0 00 ,2 55 ,0 14 ,0 00 ,2 38 ,0 03 -,0 03 ,0 05 ,0 05 -,0 47 ,3 18 ** ,0 77 -,1 85 ** ,9 50 ,9 52 ,9 26 ,9 22 ,4 05 ,0 00 ,1 71 ,0 01 -,1 09 -,0 85 -,0 28 -,1 62 ** -,0 86 ,3 38 ** -,0 77 -,0 62 ,0 51 ,1 29 ,6 23 ,0 04 ,1 23 ,0 00 ,1 71 ,2 64 ,3 63 ** ,2 95 ** ,3 56 ** ,2 44 ** ,1 81 ** ,0 58 -,0 57 ,1 27 * ,0 00 ,0 00 ,0 00 ,0 00 ,0 01 ,2 99 ,3 12 ,0 22 ,0 63 ,0 50 ,1 04 -,0 02 ,0 45 -,1 68 ** -,0 60 ,1 06 ,2 62 ,3 70 ,0 64 ,9 67 ,4 21 ,0 03 ,2 84 ,0 59 Pe ar so n co rr ela tio ns a pp ea r be lo w th e dia go na l, no n-pa ra me tr ic S pe ar ma n co rr ela tio ns a pp ea r ab ov e th e dia go na l *. C or re la tio n is s ig nif ic an t a t t he 0 .0 5 le ve l ( 2-ta ile d) . ** . C or re la tio n is s ig nif ic an t a t t he 0 .0 1 le ve l ( 2-ta ile d) . A C T H C I ED U IN D _P ER _GE _WO IN D _P ER _GE _C R IN D _P ER _GE _I M P IN D _P ER _I R R C T IN D _P ER _GE
38
4.3. Regression analysis
The relations between the variables are tested using linear regressions, table 1 shows the results.
H1a predicted that the results control tightness (RCT) is positively correlated with individual performance (IND_PER_GE). H1a tests the relation between RCT and IND_PER_GE. As shown in table 1, the model suggests that results control tightness influences individual performance in PSFs (β = 0,065). However, this relation is not significant as the p-value is 0,349. H1a is not supported.
H1b tests the relation between ACT and IND_PER_GE. The results of the regression analysis show a significant negative correlation (β = -0,144, p = 0,037). As predicted, this indicates that action control tightness negatively influences the individual performance in PSFs. H1b is supported.
Table 1: Hypothesized effect of RCT and ACT on the individual performance
Table 2, presents the regression analysis results for H2a and H2b. The hypotheses look at the interaction between results/action control tightness, in combination with high human capital intensity, and individual performance in PSFs. Contrary to what was predicted (H2a), results control tightness has a positive impact but not significant (β = 0,071, p = 0,188). Therefor H2a is not supported.
H2b predicted a negative relation between ACT and individual performance, moderated by high HCI. The results confirm this negative impact, however it is not significant (β = - 0,026, p = 0,624). Therefore H2b is not supported.
Predicted Beta Coefficient Std. Error t-statistic P-value
Intercept 3,899 ,242 16,129 ,000 RCT + ,065 ,069 ,939 ,349 ACT - -,144 ,069 -2,096 ,037 EDU + ,081 ,071 1,148 ,252 F-Statistic: 1,912 P-Value: ,128b Adj R2: ,008
39 The direct effect of high human capital intensity on individual performance is
significant positive (β = 0,358, p = 0,000).
Table 2. Hypothesized effect of HCI on the relation RCT/ACT and individual performance.
4.4. Additional analysis
In this section additional analyzes with some robustness checks are presented. 4.4.1.1. Individual performance in general
As outlined in section 3.3.1.1 is the variable IND_PER_GE based on the factor analysis, divided in three sub constructs: IND_PER_GE_IMP (improvement), IND_PER_GE_CR (career) and IND_PER_GE_WO (work output). This first robust check should lead to similar results as if the main construct is tested (IND_PER_GE). The main findings of the performed regression analysis are shortly reported in this section, more detailed results can be found in appendix C.
The results show strong support for the hypothesized direct effect of ACT on IND_PER_GE_IMP (β = -0,270, p = 0,002). Whereas IND_PER_GE_CR and
IND_PER_GE_WO don’t show any significant results. For RCT, the results are similar to the main construct IND_ PER_GE.
When it comes to moderated interaction, the effect of RCT on Work Output, Improvement and Career is similar to the results of the main construct. There is no
Predicted Beta Coefficient Std. Error t-statistic P-value
Intercept -,026 ,087 -,301 ,764 RCT + -,001 ,059 -,021 ,984 ACT - -,092 ,056 -1,635 ,103 HCI + ,358 ,053 6,768 ,000 RCT * HCI - ,071 ,054 1,319 ,188 ACT * HCI - -,026 ,054 -,490 ,624 EDU + ,033 ,111 ,295 ,769 F-Statistic: 8,865 P-Value: ,000b Adj R2: ,129
40 significant impact. For ACT, a non-significant positive impact is reported on the relation ACT, HCI and Work Output. Improvement and Career show similar results as the main construct.
4.4.1.2. Individual performance in role
I performed a regression analysis on the interaction between RCT, ACT, HCI and IND_PER_IR (individual performance in role). As mentioned earlier in section 3.3.1.1, the respondent has to answer questions on own individual performance in general and in role of the supervisor. The detailed results are included in appendix C.
The results show small negative impacts on the relation of RCT (β = -,013) and ACT (β = -,080) with IND_PER_IR, both effects are not significant.
The same variables as above are tested again with the moderator HCI included. HCI moderation changed the negative impact to a positive one for ACT (β = 0,022) on IND_PER_IR. However, it’s is not significant (p = 0,698). Interaction of RCT and HCI show a beta of 0,128 and p-value of 0,024.
