The impact of management control choices on team learning within Professional Service Firms
Name: Didy Kromhout Student number: 11163011
Thesis supervisor: dr ir. S.P. van Triest Date: 20 January 2018
Word count: 12.349
MSc Accountancy & Control, specialization Control
Statement of Originality
This document is written by student Didy D. Kromhout 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.
Abstract
The purpose of this paper is to investigate the effect of choices in the management control system on team learning. The study was conducted by a survey which was completed by 148 managers of departments within Professional Service Firms (PSF’s). The survey was
predesigned and part of a research project conducted by the University of Amsterdam. Using agency and self-determination theory, I hypothesized that formal controls (results and action controls) will lead to a lower level of team learning, while informal controls
(personnel and cultural controls) will lead to a higher level of team learning.
Additionally, I investigate whether the extent of knowledge sharing that is already in place in a firm moderates the relationship between the control mechanisms and learning
activities. A high measure of knowledge sharing is expected to have a positively moderating effect on the relationship between the formal controls and team learning and a negatively moderating effect concerning the relationship with the informal controls. I find that only the direct positive effect of the informal controls is supported, while there is no significant effect for the formal controls. Knowledge sharing therefor does has a positive direct effect but does not moderate the impact of the control mechanisms on team learning.
Acknowledgments
I acknowledge the support, patience and willingness to answer my numerous questions of dr. ir. Sander van Triest. I also want to thank dhr. dr. S.W. Bissessur for teaching the research seminar for Accounting and Control in such a way, which writing this thesis, almost, seemed “doable”. Also, I want to give a warm “thank you” to the study advisors, Anne Boverhof and Irina Hofstee, whose kind attention and inquiries gave just that nudge to make me actually pick up and finish my degree after a year’s absence. You made a real difference.
Finally, I would like to thank those who supported me throughout the whole process: for listening to my sighs and sometimes complaints and for not complaining when having to eat pizza…again!
Table of Contents
1 Introduction ... 5
2 Theory ... 8
2.1 Context of the PSF ... 8
2.2 Knowledge based theory of the firm ... 9
2.3 Agency Theory ... 11
2.4 Self-determination Theory ... 12
2.5 Team learning ... 13
2.6 Management Control ... 13
2.7 Result control and team learning ... 16
2.8 Action control and team learning ... 17
2.9 Personnel control and team learning ... 17
2.10 Cultural control and team learning ... 18
2.11 The effect of the level of knowledge sharing ... 19
3 METHOD ... 22 3.1 Survey ... 22 3.2 Construct ... 24 3.2.1 Result control; ... 24 3.2.2 Action control; ... 25 3.2.3 Personnel control; ... 26 3.2.4 Cultural control; ... 26 3.2.5 Team learning ... 27 3.2.6 Knowledge sharing ... 27 3.2.1 Control variables ... 28 4 RESULTS... 30
4.1 Results per hypothesis ... 33
4.1.1 Hypothesis 1 ... 33
4.1.2 Hypothesis 2 ... 33
4.1.3 Hypothesis 3 ... 33
4.1.4 Hypothesis 4 ... 33
4.1.5 Hypothesis 5a, 5b, 5c and 5d ... 33
4.1.6 Control variable ... 34 4.2 Robustness test ... 34 4.3 Summary ... 37 5 DISCUSSION ... 38 5.1 Discussion... 38 5.2 Theoretical implications ... 40
5.3 Limitations of the research ... 41
5.4 Practical implications ... 41
1 Introduction
Competitive advantages have long been recognized to allow firms to create and maintain above average performances (Porter, 1985; Barney, 1991; Argote & Fahrenkopf, 2016). A unique source for a competitive advantage is knowledge and the ability of a firm to integrate new knowledge (Porter & Millar, 1985), as it may lead to cost reduction and enhance differentiation through the development of new products (Gupta & Govindarajan, 2000). Knowledge is however not automatically transferred (Grant, 1996). And when it is, it is not always picked up on or fully taken advantage of (Bunderson & Reagans, 2011; Bock, Zmud, & Kim, 2005; Fey & Patrick, 2008) as both the act of sharing and learning comes at a cost to the individual. The consequences of not sharing knowledge and of not learning range from missing out on potential cost reductions to missed innovations and could lead to large otherwise unnecessary expenses and missed income (Cabrera, Collins, & Salgad, 2006; Gupta & Govindarajan, 2000).
“…the primary role of the firm is integrating the specialist knowledge resident in individuals into goods and services. The primary task of management is establishing the coordination
necessary for this knowledge integration” (Grant, 1996).
In this paper I will answer the following research question: what is the impact of Management Control choices on the level of team learning, taking into account the
moderating effect of the level of knowledge sharing already occurring within a team? I will explain, in line with the agency-theory and self-determination theory that the installment of a management control system is necessary since team learning is essential for the success of a firm, but comes at a cost for the employees (Cabrera, Collins, & Salgad, 2006; Bock, Zmud, & Kim, 2005). I will follow the definition of management control defined by Merchant and Van der Stede, that management control are all devices or systems managers used to ensure that the behaviors and decisions of their employees are consistent with the organization’s objectives and strategies and structure the research using the different control mechanisms, namely: result controls, action controls, personnel controls and cultural controls. (Merchant & Van der Stede, 2012). In line with the knowledge based theory of the firm, I argue that the firm has an important and specific role in providing
coordination to integrate the specialized knowledge of their members and eventually transfer that knowledge into goods and services (Grant, 1996).
Team learning, as an important factor in the innovation process (Stata, 1989; Berman, Down, & Hill, 2002), is defined as a continues process of reflection, action and integration of new knowledge, leading to changes and improvements (Edmondson, 1999; Argote, 2015). The learning process however does not always, immediately, lead to results, therefore team learning is not measured by its results, but by the learning behavior activities that are necessary to obtain and process data such as seeking feedback, sharing information, asking for help, talking about errors and experimenting (Edmondson, 1999).
Knowledge Sharing in this research is defined as the act of making knowledge available to others within the organization (Ipe, 2003) which is measured by the level of Knowledge Sharing taking place, as perceived by management.
Team learning and knowledge sharing are closely related in two important ways. First of all, barriers for sharing knowledge and barriers for engaging in team learning overlap (Fey & Patrick, 2008; Bock, Zmud, & Kim, 2005; Gupta & Govindarajan, 2000; Grant, 1996), and secondly both knowledge sharing and team learning require incentives and coordination. In this research I argue that the level of knowledge sharing already taking place in an
organization has a positive effect on the level of the level of team learning where it concerns the effect of formal control mechanisms since a firm with a high level of knowledge sharing has already successfully overcome barriers that are also considered barriers to the process of team learning. Concerning informal controls, I argue that a high level of knowledge sharing has a moderating effect on the relationship, since knowledge sharing takes over some of the effects of the informal controls.
Although research in the field of knowledge management has taken a spur the past 20 years, a vast majority of the research has focused on different types of knowledge,
organizational culture, the effects of information technology and human attitudes towards knowledge sharing. It is however not clear yet how knowledge is actually managed in organizations (Gupta & Govindarajan, 2000; Fey & Patrick, 2008) or which specific variables support the flow of ideas and experiences, and what management control practices can
affect those variables (Cabrera, Collins, & Salgad, 2006). Research that has investigated management control effects on knowledge sharing and organizational learning has often focused on the effect of one control mechanism, such as the effect of reward systems on knowledge sharing, with often contradictory results. Research on the effect of management control choices on organizational Learning is hard to find. In view of this gap in research, this paper provides a contribution to existing literature, as it tries to investigate whether
management control systems can be used to align the interests of the firm and the professionals when it comes to sharing knowledge and organizational learning.
The method of research is through a survey amongst managers of Professional Service Firms (PSF’s), an interesting setting as it is thought that the knowledge intensity of a PSF, provides an example that will become more and more relevant to non-PSFs and are therefore viewed as models for an increasingly knowledge-based economy (Nordenflycht, 2010).
This paper is structured as follows: Chapter 2 will present the research framework and an overview of up to date underlying literature regarding the dependent variables knowledge sharing and team learning as well as the independent variables control mechanisms.
Subsequently the relationship between the variables is explained through the Agency theory and the Self-Determination theory, followed by the development of the hypothesis. The research method will be described in chapter 3, explaining the collection of data through a survey and the development of the constructs used for measurement. Chapter 4 will provide an overview of the statistical results, followed by a conclusion, restrictions and implications both on theory as in practice and suggestions for further research in chapter 5.
2 Theory
Figure 1 illustrates the framework of the research and the hypothesis relating the various variables. The development of the hypothesis will be discussed in this chapter.
Figure 1; Hypothesis model
First I will give a literature overview of PSF’s, the context in which this research is
conducted. Secondly I will provide a short literature overview of the relevant theories and the independent variable team learning. Thirdly I will elaborate on Management Control types and the relationship between knowledge sharing and team learning.
2.1 Context of the PSF
PSF’s are known to be knowledge intensive organizations with an important share of the organization’s knowledge embedded within the individual professionals (Starbuck, 1992). A literature review does not give a consensus of the definition of a PSF but often a list of examples is given to give an impression of what entails a PSF. However, three central characteristics are commonly noted in relation to PSF’s (Nordenflycht, 2010) that are of interest and importance in the context of this research. Firstly, PSF’s are knowledge intense and standardization is difficult as activities are often non-routine. Secondly, PSF’s are low on capital intensity, meaning that the professional can easily transfer from one firm to another. Taking with him the knowledge that he has obtained. And thirdly, PSF’s rely on a
professionalized workforce. These characteristics indicate that the services provided heavily rely on a substantial base of knowledge. Besides hiring highly educated people,
professionals also learn on the job, from the clients they work for and the projects they engage in. This last source of experience is highly valuable as professionals are often
educated in a standardized body of knowledge, known to all professionals in that sector. It is therefore in the interest of the firm to collect this unique knowledge and incorporate it in the existing knowledge base of the firm. Another characteristic is the often high level of autonomy of the professional, as personal judgement is often required to service the need of the client in a customized manner. In order to do so, the professional also has a high degree of interaction with the client. (Løwendahl & Revang, 2001)
Working with a professionalized workforce also means dealing with a set of norms and ethical codes, such as a preference for autonomy, but also the sense of responsibility the professional experiences to protect the interests of the client, or even society in general (Nordenflycht, 2010). This norm, or trusteeship, could also entail that client needs are set higher than those of the firm. (Løwendahl & Revang, 2001)
2.2 Knowledge based theory of the firm
According to the knowledge based theory of the firm, firms exist primarily to integrate the specialized knowledge of their members and transfer that knowledge into goods and services. Its primary task is therefore the establishing of the coordination necessary for this knowledge integration, not only between firms, but even more critically, within the firm (Grant, 1996); a role that the market is unable to fulfill efficiently.
Conform the resource-based view of the firm, the transferability of a firm's resources, that is knowledge, is regarded as critical to creating a sustainable competitive advantage, which have long been recognized to allow firms to create and maintain above average
performances (Porter, 1985; Barney, 1991; Argote & Fahrenkopf, 2016). A unique source for a competitive advantage is knowledge as it may lead to cost reduction and enhance
differentiation through the development of new products (Porter & Millar, 1985; Gupta & Govindarajan, 2000). Knowledge in general is regarded as a highly valuable asset, but especially firm-specific knowledge, as it is created within the organization and is therefore
difficult to imitate by competitors is considered a precious source for a competitive advantage (Hitt, Bierman, Shimizu, & Kochhar, 2001; Grant, 1996; Linda Argote, 2000). In fact; knowledge is seen as one of the most important strategic resources of an organization, and the management of this knowledge is considered critical to organizational success. (Ipe, 2003).
To transfer knowledge involves both the sending and receiving of that knowledge. The level of receiving knowledge is dependent on the recipient's ability to add new knowledge to existing knowledge. A process called aggregation. The level of sending or sharing
knowledge, that is the act of making knowledge available to others within the organization (Ipe, 2003) is related to the ability of the owner of the knowledge to receive a return that is valued equal to the value of the knowledge shared, a process called appropriability. (Grant, 1996)
The process of appropriability, when it concerns firm specific knowledge however is very complex. F.e. since the knowledge has at least partially been created within the firm, while working for the firm and is firm specific, so who then is the ‘owner’ of the knowledge and to whom should the returns be allocated to?
The process of appropriability is important as it indicates that sharing knowledge comes at, at least a perceived, cost to the owner of the knowledge, who will want to receive a return for sharing it.
These and other complexities of knowledge sharing, point out that coordination aimed at integrating knowledge is both necessary and no small matter, even if there is no goal
conflict. There are however ways to stimulate the integration of knowledge and a number of mechanisms can be identified to integrating specialized knowledge, such as in a PSF: rules and directives, sequencing and setting up Routines in certain situations. A more expensive and intensive mechanism to integrate knowledge is done through Group problem solving and decision making. This process is both more personal and requires intense
2.3 Agency Theory
Agency theory is concerned with problems that may occur when there is a formal
relationship between two entities, the principal and the agent. The theory is based on the assumption that the agent will foremost act out of self-interest as opposed to acting first in the best interest of the principal. An agency-problem arises when it is difficult or expensive for the principal to verify what the agent is actually doing, due to information asymmetry. When a principal and an agent enter into a formal relationship often monitoring costs and bonding costs, such as rewards and other incentives will occur. Despite these positive efforts and costs, the agent will most likely still make decisions that will not maximize the welfare of the principal. The income that is lost due to these decisions is referred to as “residual loss”, a negative cost. The sum of these positive and negative costs are called agency costs (Jensen & Meckling, 1976).
Agency theory also assumes that agents are foremost rational actors who will act risk averse (Fey & Patrick, 2008) in which case the principal will have to take measures to make sure the interest of the agent is aligned with the interest and strategy of the principal. (Eisenhardt, 1989). The theory is highly applicable in the context of a PSF, where professionals often have a lot of autonomy and direct contact with the client, often at the location of the client. Furthermore, besides putting self-interest before the interest of the firm, ‘trusteeship’ will push the interest of the firm further back as a priority as the needs of the client are deemed more important (Løwendahl & Revang, 2001). As a result, the professional could f.e. decide not to make time available to share knowledge of participate in organizational learning activities because of time pressure to meet client’s needs. In the context of knowledge sharing, information asymmetry poses another problem. Since professionals learn from the clients they work for and the projects they are engaged in (Løwendahl & Revang, 2001) the professional gains new knowledge. In advance the firm cannot foresee what knowledge will be gained or could be useful for the organization, (Fey & Patrick, 2008) risking the loss of valuable information if the professional does not engage in knowledge sharing. In other words, even though it is desirable at an organizational level to contribute to the shared knowledge base, it may not be rational for the professional who may want to protect individual expertise, client’s interest or earning more money by moving on immediately to the next project and new ‘billable hours’ (Løwendahl & Revang, 2001).
2.4 Self-determination Theory
The self-determination theory distinguishes between two types of motivation, or rather on two sources of motivation, namely intrinsic and extrinsic regulations. (Gagne & Deci, 2005; Ryan & Deci, 2000).
Intrinsic motivation will drive people to do an activity because they find it interesting and derive satisfaction from doing it and as such, intrinsic motivation is a good reflection of the positive potential of human nature and its tendency to seek out novelty and challenges, to extend and exercise one's capacities, to explore, and to learn. Besides these natural borne tendencies, two other ‘criteria’ must be met. First, there must also be a sense of causality; that the choices made and action taken have a positive influence on the results obtained. And secondly, that there is a sense of autonomy; a sense that the choices made and actions taken are self-determined. Intrinsic motivation is therefore also called autonomous
motivation.
Extrinsic motivation, on the other hand, requires an external source of ‘reward’ or regulation, as it is not the activity itself that gives satisfaction, but the obtained reward. Extrinsic motivation is therefore also called controlled motivation. Extrinsic motivation can vary in the degree to which external control and regulations are necessary to stimulate desired actions, depending on the degree to which the person has internalized the external regulation, a process called the autonomy continuum. This process of internalization varies from external regulations such as rewards to interjected regulations that appeal f.e. to a sense of self control of the employee, to identified regulations that have become personal importance to the employee to fully integrated regulations.
For example, in the context of this research, a professional could be persuaded to share knowledge and participate in organizational learning because it will result in a promotion or improve the chances of participation in a highly interesting project. Another motivator could also be the avoidance of a demotion or a negative appraisal. This is an example of fully regulated external motivation. On the other hand, a professional could also participate in knowledge sharing and team learning because he identifies with the strategy of the firm to create a strong competitive advantage based on the unique knowledge present within the firm. Even though the regulation is external, the professional has internalized it to a high degree and motivation has become, near, autonomous.
2.5 Team learning
Firms rely more and more on teams as the environment around firms changes faster and becomes more uncertain. Understanding what factors influence team learning therefore becomes more important, as it through the act of learning that changes in the environment are detected, we learn about customers' requirements, improve members' collective
understanding of a situation, or discover unexpected consequences of their previous actions (Edmondson, 1999). Team learning is therefore an important factor in the innovation
process (Stata, 1989; Berman, Down, & Hill, 2002). Team learning can be defined as the process of detecting faults and correcting them (Edmondson, 1999) and as the process of integrating new knowledge, using it to come to new insights (Argote, 2015). In this research, team learning is not measured by its effects, but by the learning behavior activities that are necessary to obtain and process data such as seeking feedback, sharing information, asking for help, talking about errors and experimenting.
New knowlegde or information is not always pick up or fully taken advantage of as the ability and willingness of people to engage in the process of learning is influenced by a number of factors such as status, perceived personal risk, power, the willingness to take risks and a natural guardedness to offered knowledge (Bunderson & Reagans, 2011; Bock, Zmud, & Kim, 2005). Since many of us tend to think our own work is best, we have a tendency to either explicitly or implicitly not fully consider work done by others (Fey & Patrick, 2008). Research has shown that the ability to learn varies dramatically (Kozlowski & Argote, 2012) and it is therefore important to carefully manage the process of learning. The following paragraphs will elaborate on the impact control choices can have on team learning, followed by the development of hypothesis.
2.6 Management Control
Management control is about making sure that an organization reaches its objectives (Merchant & Van der Stede, 2012) through implementing and executing plans to reach long-term strategic objectives, through measurement and if necessary adjusting activities to reach those objectives. Management control systems then are the tools that management uses to achieve its objectives͚; it includes “all the devices or systems managers use to ensure that the behaviors and decisions of their employees are consistent with the organizations
objectives and strategies” (Merchant & Van der Stede, 2012, p. 5). Other approaches, such as the approach by Robert Simons (1987) view management control as the formal,
information-based routines and procedures used by managers to maintain or alter patterns in organizational activities, or as tools to implement strategies by influencing human
behavior. Following the framework of Merchant & vd Steede, 4 control variables are
identified, which can be divided into formal control variables and informal control variables.
Formal control variables;
Result controls; result controls are focussed on identifying performance measures and rewarding obtained results. These performance measures can be measured ex ante; clearly determined and communicated in advance as a way of stimulation, or ex poste; measured afterwards allowing external factors to be taken into account. The theory of result control is based on expectancy theory comprising two factors. First, the believe/expectancy of an employee that he can influence the outcome with his own behavior and choices and, secondly that the level of how much the employee desires to reach the outcome (valence) can be influences. Rewarding the outcome it is assumed will increase valence. Rewards can be both extrinsic as intrinsic and range from financial rewards to promotion, training, status, a higher level of autonomy or a sense of accomplishment. For result controls to be effective it is important that the desired result is clarified, that the employee can affect the result and that results can be measured effectively.
Action controls; action controls are controls that focus on the actions themselves, ensuring that employees perform certain actions or don’t perform undesirable actions by rewarding of punishing behavior. Ways to install action controls are f.e. through behavioral constraints such as limited access to assets or data or decision making, limited access to assets or data, limited decision making authority, separation of duties (also an internal control mechanism) or through action accountability measures such as work rules, procedures and codes of conduct. Action controls have a number of advantages such as clarity to the employees on what to do in certain situations and the improvements of actions through codification of best practices. An important downfall of action controls in the setting of a PSF in which situations are often unique, however could be that following the procedure becomes more important than the goal and that innovation and creativity are stifled. For action controls to
be effective the manager both needs to be clear on what behavior is or is not desirable and needs to be able to observe the behavior of his employees.
Informal control variables
Personnel controls; personnel controls build on natural tendencies of people to motivate themselves, by clarifying expectations, ensuring that each employee is able to do a good job by making sure they have the right capabilities and resources, and to increase the chance of self-monitoring, stimulating people to do a good job and increase commitment, including intrinsic motivation and loyalty. Personnel controls focus on three areas: Selection and placement of personnel, Training and job design. Especially training can act as a
motivational tool as it gives the professional an even greater sense of professionalism.
Cultural controls; cultural controls are designed to encourage mutual monitoring; the controls are built on shared traditions, norms, beliefs, values, ideologies, attitudes and ways of behaving and norms can be both written and unwritten. In fact, cultural controls are a powerful form of group pressure on individuals. Ways to attempt to shape organizational culture could be through; codes of conduct, group rewards that can encourage teamwork, on-the-job-training of employees, rotation of employees, physical arrangement and importantly, ‘tone at the top’.
Some control issues encountered specifically in the services industry and therefore in PSF’s arise from the difficulty to measure both the volume and quality of the service provided. This is made even harder since the client itself is partially responsible for the result of the service provided, since the professional often works in close cooperation with the client. The quality of the input from the client determines therefore, at least partially, the quality of the output as it is the result of a co-operation so to say. Since services often cannot be formalized in procedures it is also difficult to pin point exactly what has led to a positive customer experience or, more relevant to this research, what has led to new learning and what exactly has been learned. As a consequence formal controls are difficult to apply; result controls because of the difficulty of measuring both volume and quality and
behavioral controls because the service provided is often customer specific and cannot be “prescribed” in procedures or routines.
2.7 Result control and team learning
Result controls are a formal type of control that provide some important advantages specifically applicable to the context of a PSF, for they are often effective when it is not clear, in advance, what behaviors are most desirable and can yield good control while allowing the professional to maintain a high level of autonomy. In the context of a PSF, the professional often has a high level of autonomy as he has to rely on personal judgement to service the need of the client in a customized manner and has a high degree of interaction with the client (Nordenflycht, 2010; Løwendahl & Revang, 2001). If the end result can be clearly specified and measured, result controls can be highly effective in managing professionals and aligning their choices with the best interest of the firm.
The goal and results specifically of a learning process however are difficult to clarify or measured as, in fact, the learning process does not always yield results (Edmondson, 1999). As clarity, influence and measurement are requirements for the use of result controls, this type of control is not an effective measure to motivate team learning, even though research has provided tentative evidence that incentive structures can have a positive effect on knowledge sharing (Fey & Patrick, 2008; Jimenez-Jimenez & Sanz-Valle, 2012).
In this paper I argue that learning results cannot be specified clearly and therefore cannot be stimulated by the use of result controls. The use of result controls, which can be highly effective in a PSF, will put the focus of the professional on obtaining other, specified result; therefore taking away the focus on engaging in team learning, resulting in the following hypothesis:
Hypothesis 1; a higher level of result control will result in a lower level of team learning
2.8 Action control and team learning
In the context of working with professionals it is often not possible to clearly define
desirable actions (Nordenflycht, 2010; Løwendahl & Revang, 2001) or to observe and track actions. Rewarding or punishing behavior is therefore difficult. However, even though desirable actions of professionals generally cannot be clearly defined in advance,
professionals are held accountable under the premise that they should “act professionally”. (Merchant & Van der Stede, 2012). Further, based on the self- determination theory, stipulating that the highly educated professional should have a sense of autonomy to be motived (Løwendahl & Revang, 2001), action controls could decrease the motivation of the professional (Gagne & Deci, 2005; Ryan & Deci, 2000).
In this paper I argue that action controls demotivate the professional, who needs autonomy to be motivated, to participate in activities such as team learning activities. Therefore the Hypothesis is as follows:
Hypothesis 2; a higher level of action control will result in a lower level of team learning
Figure 3; Hypothesis 2 the negative effect of action control on team learning
2.9 Personnel control and team learning
According to the self-determination theory, people have a natural born tendency to explore and learn. However professionals also have a tendency to hoard information and a natural guardedness to offered knowledge and the tendency to either explicitly or implicitly not to consider work done by others fully (Fey & Patrick, 2008; Bock, Zmud, & Kim, 2005),
indicating that external regulation is needed to control the motivation to engage in team learning.
Personnel controls such as the addition of knowledge through hiring highly qualified personnel and by putting emphasis on training, will act as external regulators to increase motivation. Training provides a good opportunity to share knowledge in and informal setting and helps to provide a common language which enables the transfer of knowledge and a way to communicate a shared vision (Jimenez-Jimenez & Sanz-Valle, 2012; Gagne & Deci, 2005; Ryan & Deci, 2000; Løwendahl & Revang, 2001). Additionally training, increases individual capabilities and helps to maintain a learning-oriented organizational culture. Further, research has shown that training has a positive effect on the interpretation of knowledge. The same process of training however also holds a potential risk to the effectiveness of Learning as it leads to a convergence of organizational and individual believes, decreasing the variability of individual beliefs that improve organizational and average individual knowledge in the long run. Emphasis on hiring new personnel could sustain the level of variability and level of effectiveness of learning (March, 1991).
Based on the above argumentation, I expect a positive relationship between personnel control and the level of team learning, leading to the following hypothesis.
Hypothesis 3; a higher level of personnel control will result in a higher level of team learning
Figure 4; Hypothesis 3 the positive effect of personnel control on team learning
2.10 Cultural control and team learning
Based on the self-determination theory, achieving shared values and a shared vision can be a very strong motivator to act in the interest of the firm in two ways; that is to participate in organizational learning. On the one hand, shared values and vision create group pressure to act in a way that is most beneficial to the firm (Merchant & Van der Stede, 2012) and
could also become internalized, as mutual learning leads to a convergence of organizational and individual believes, and become like an intrinsic motivator to act accordingly (March, 1991).
Based on the above argumentation, I expect a positive relationship between the two variables, leading to the following hypothesis.
Hypothesis 4; a higher level of cultural control will result in a higher level of team learning
Figure 5; Hypothesis 4 the positive effect of cultural control on team learning
2.11 The effect of the level of knowledge sharing
Knowledge in general is regarded as a highly valuable asset and the management of this knowledge is considered critical to organizational success. (Ipe, 2003; Hitt, Bierman, Shimizu, & Kochhar, 2001; Grant, 1996). It is however the act of sharing knowledge that opens a window of opportunity to maximizing a firm’s ability to react to customer needs and to generate solutions and efficiencies that create a competitive advantage, even after individuals have left the organization. (Reid, 2003; Argote & Ingram, 2000; Lin, 2007). Yet the transfer of knowledge is often limited and below potential (Gupta & Govindarajan, 2000) for various reasons. From an organizational point of view, the organizational culture, difficulty in clarifying the benefits of knowledge sharing, lack of communication are
considered some of the barriers for effective knowledge management (Cabrera, Collins, & Salgad, 2006). From an individual’s point of view, reasons range from a person’s natural tendency to hoard knowledge (Bock, Zmud, & Kim, 2005) to the personal costs experienced in relation to knowledge sharing such as the time involved, effort it consumes and the loss of expert power to trust and personality (Cabrera, Collins, & Salgad, 2006).
Team learning and knowledge sharing are closely related in a number of ways; for example barriers for engaging in team learning knowledge overlap with above mentionned barriers such as the need for trust and personal barriers such as a natural guardedness to offered knowledge and the tendency to either explicitly or implicitly not to consider work done by others fully (Fey & Patrick, 2008; Bock, Zmud, & Kim, 2005). Also, both knowledge sharing and team learning require incentives and coordination (Gupta & Govindarajan, 2000). In this thesis I argue that the level of knowledge sharing taking place in an organization has a positively moderating effect on the relationship between the formal variables and the level of team learning, as a firm with a high level of knowledge sharing has already successfully overcome relevant barriers to team learning, the effect of which will not be fully undone by a high level of formal controls. Further, the level of knowledge sharing has a negatively moderating effect on the relationship between the informal variables and the level of team learning, as informal controls will add to the effect of knowledge sharing on team learning already in place. The effect of informal controls will lessen as the desired level of team learning is approached. Based on the above argumentation, I deduct to the following hypothesis:
Hypothesis 5a; a higher level of knowledge sharing will positively moderate the negative effect of a higher level of result control on the level of team learning
Figure 6; Hypothesis 5a the positively moderating effect of knowledge sharing on the effect of result control on team learning.
Hypothesis 5b; a higher level of knowledge sharing will positively moderate the negative effect of a higher level of action control on the level of team learning
Figure 7; Hypothesis 5b the positively moderating effect of knowledge sharing on the effect of action control on team learning.
Hypothesis 5c; a higher level of knowledge sharing will negatively moderate the positive effect of a higher level of personnel control on the level of team learning
Figure 8; Hypothesis 5c the negatively moderating effect of knowledge sharing on the effect of personnel control on team learning.
Hypothesis 5d; a higher level of knowledge sharing will moderate the positive effect of a higher level of cultural control on the level of team learning
Figure 3; Hypothesis 5d the negatively moderating effect of knowledge sharing on the effect of cultural control on team learning.
3 METHOD
Research in this paper will be based on the results of a research project, executed by the “Accounting Section” of the UvA. The data was obtained through a survey and allows for research on how professional service organizations address the challenges of managing highly skilled professionals who are working for – and together with – highly demanding clients. How do firms balance the trust they have in their professionals’ skills and values with the need to react to their clients’ targets and deadlines? Each respondents of the surveys is a manager of a team of professionals, with responsibility for the people and projects in their team.
Each student has had to contribute at least 6 completed surveys and proof that the participants filled in the survey themselves. The data in the survey has been anonymized.
3.1 Survey
A total 148 of respondents filled in a survey, which consisted of parts; the substantive part, consisting of 77 questions and a general part, consisting of questions.
Concerning the substantive part; 76% of the surveys were filled in completely, leaving 35 questions with missing values, ranging from 1 to 3 missing values per question. Two respondents failed to fill in respectively 9 and 12 adjacent questions, however the missing values show no pattern and are completely random. 29 Substantive questions of the 77 have been used in this research of which 9 show either 1 (7questions) or 2 (2 questions) missing values. Missing values were deleted pairwise and are not part of the calculation.
22 Respondents failed to fill in all general questions of the survey. The missing values show a clear pattern of gaps. Questions related to nationality, location and work experience were often all skipped by these respondents. Table 1 gives a clear overview of the characteristics of the respondent and table 2 gives and overview of the characteristics of the PSF.
Since these respondents did fill in the substantive questions and the general questions are unrelated to the research questions, these surveys have not been deleted listwise.
Table 1; characteristics of the respondents
The survey contains questions on the following 10 subjects: • Innovativeness: 1–9 (Capon et al., 1992)
• Client relationships: 10-21 (Nobel et al. 1998, Andersson et al. 2002, Aulakh et al. 1996) • Knowledge sharing: 22–25 (Faraj and Sproull 2000)
• Team learning behavior: 26–32 (Edmondson 1999)
• Task characteristics: 33–41 (Withey et al. 1983, Gresov 1989, Abernethy and Brownell 1997) • Client interactions: 42–48 (Homburg and Stebel, 2009)
• Personnel control: 49–61 (mostly self-developed, also Snell 1992) • Autonomy: 62–65 (Hartline and Ferrell 1996, Breaugh 1989)
• Decentralization: 66–70 (Gordon and Narayanan 1984, Abernethy et al., 2004) • Performance assessment: 71–77
Respondents were generally offered a choice of seven pre-coded responses with the neutral point being neither agree nor disagree. Respondents could answer questions using this 7-point Likert scale to express how much they agree or disagree with a particular statement. The pre-codes answers ranged from: (1) Strongly disagree, (2). Disagree, (3) Somewhat disagree, (4) Neither agree nor disagree, (5) Somewhat agree, (6) Agree and (7) Strongly agree. Questions 23 and 24 were negative questions within the construct of knowledge sharing, the answers given have been reversed coded.
3.2 Construct
The questionnaire used for this study was a pre-developed questionnaire. The survey consisted of 11 constructs of which the construct for ‘personnel control” consists of 13 questions of which 6 questions are concerned with result control, 1 question with action control, 3 questions with personnel control and 3 with cultural control. The assignment of relevant questions to the four types of control is based on the theory on management control by Merchant & Van der Stede (Merchant & Van der Stede, 2012).
The pre-set construct on knowledge sharing and team learning have been used without alteration from construct used in the survey to answer the research question.
Appendix 1 gives an overview of items in the constructs used; the items excluded from and added to the preset constructs of the survey, are clearly marked.
A Varimax principal rotated component analysis (PCA) was conducted on all constructs. This was done in three stages; first per construct, second a joint PCA on all variables without limitations on the number of factors and third a joint PCA was calculated on all variables with a limitation of 6 component, equal to the number of constructs tested. For this analysis only values exceeding 0,50 were selected. Further, only components with an eigenvalue larger than 1 were selected. The results of the PCA with 6 components is shown in table 2.
Each one of the six construct was expected to be related in one components, however the construct of result control, action control and team learning were related to multiple components. Additionally the Cronbach Alpha has been calculated on each construct separately and jointly. In the following paragraph the reliability results and consequences are discussed per construct.
3.2.1 Result control;
For this research, control mechanisms are measured by the scale of Snell (1992). This scale measures the level of input, behavioral and output controls that managers exercise to align the actions of individuals with the interests of their employing firm. The level of result control is measured by the use of output controls only, as defined by Snell and could be described as decentralized controls that focus on achieving set goals. For this research the scale is adapted to the context of a PSF in which it is difficult to set fixed goals or link
rewards to concrete results. Questions related to fixed targets and results and consequent results are left out. Instead, focus is put on the effort that is put into the evaluation of practice members performances. The pre-set construct in the survey contained 6 items.
The conducted joint PCA loads with 6 components load 4 items of the pre-set construct to 1 component. However in the joint PCA without restrictions on the number of components, 3 items (Q49, Q50 and Q51) are loaded on one component and 2 items on another. The latter result is confirmed in the PCA per construct.
A closer analysis of the items in the construct show that Q49, Q50, Q51 and Q55 are concerned with either the process or the possibility of evaluating employees, whereas Q53 and Q54 are focused on the content of the evaluation and concerned with on what the employee is evaluated.
Since this research is aimed at measuring the effects of the level of result controls used, regardless of on what the employee is evaluated, Q53 and Q54 have been withdrawn from the pre-set construct.
The Cronbach Alpha has been calculated to assert reliability of the construct, to assert that the selected items actually measure this single construct. The Cronbach Alpha for this construct measures 0.713; that is above the preferred minimum of 0,7 and well above the threshold of 0,5.
3.2.2 Action control;
For this research, action control is measured by a formal behavioral control defined by Snell (1992) as codified/articulated operating procedures, assuming a centralized hierarchy. For this research the scale was adapted to the perspective of providing services in which actions are customized to the need of the client and the professional is in close contact with the client. The construct exists of one item.
Additionally to the pre-set item in the survey to measure action control, two survey questions were added from the pre-set construct on task characteristics.
Both question 39 and 40 are based on the scale of Withey, Daft & Cooper (Withey, Daft, & Cooper, 1983) and were used by Abernathy to measure task uncertainty (Abernathy & Brownell, 1997 ). Both questions have been adapted for this survey and zoom in on both
the codification of processes into procedures and communication of expected behavior concerning recurrent tasks. Based on the theory of Merchant & vd Stede, these are clear indicators to measure the use of action controls.
3.2.3 Personnel control;
For this research, personnel control is measures by input controls as defined by Snell (1992) as both selection and training. The questions used have been limited and focus primarily on the effort put into training and developing the professional and the emphasis put on staffing procedures. The questions have been adapted to the context of a PSF.
In the factor analysis (PCA) all items are loaded on one component and the Cronbach Alpha is 0,823.
3.2.4 Cultural control;
For this research, cultural control is defined by the effort put into shaping culture and atmosphere in the organization.
In contrast to the above mentioned control mechanisms, cultural control is not measured using the scale of Snell, as these control items are considered “socialization controls”. The pre-set construct in the survey contained 3 items.
The conducted joint PCA loads loaded 1 item (Q59) of the pre-set construct to 1 component and two items (Q60 & Q61) to others. Also in the singular PCA, the division of items over multiple components is confirmed.
A closer look at the items in the construct show that Q60 focusses on a work context outside the practice in which the client and his opinion play a major role. Due to this third-party involvement this item is not suitable to be part of a construct that is both informal and internal. Q61 is a very general item which is subject to the interpretation of the respondent on what is practice culture. For this reason I have also excluded this item.
On the other hand, I added 3 questions from the scale of Capon et al. (1992) to the
questions in the pre-set construct for control. The question in the scale of Capon are aimed to investigate the ‘degree of openness’ in a firm to capture the informal organization
of Capon et al. (1992) in a study to investigate the relationship between environment, strategy, formal organization and the informal organization on innovation. The question added to the construct of cultural control items to measure the informal organization. Based on the theory of Merchant & vd Stede, these items also represent clear indicators to
measure the use of cultural controls.
The construct consists of 4 items, each focusing on building culture in a demonstrative ways. All items, including the added items load to one component. The Cronbach Alpha of the construct is 0,775.
3.2.5 Team learning
The measurement of team learning based on the scale of Edmondson (1999)
The scale of Edmonson measures the level of team learning as a combination of both the effects of the team’s believed interpersonal context and the effect of team behavior. This paper focusses on the effects of team behavior and examining under what conditions learning occurs in organizational work groups (1999). Questions about the interpersonal context are left out of this survey.
The pre-set construct consisted of 7 items. 6 Out of these items were loaded on one component in the PCA. 1 Item (Q27) was loaded on another item.
Closer examination of the construct showed that 6 items were clearly related to facets of a learning process, such as gather information, reflection and discussion (Argote, 2015). Item 27 however is concerned with dealing with differences of opinion, without relating that difference to the learning process. It is therefore too general and has for that reason been taken out of the construct. The remaining construct has a solid Cronbach Alpha of 0,809.
3.2.6 Knowledge sharing
The measurement of knowledge sharing is bases on part of the scale of Faraj and Sproull (2000) on measuring the coordination of expertise on an individual level. The questions selected from this measure concern the level of which information is brought to bear, leaving questions on the location of the expertise and the need of expertise aside. Two items in this construct (Q23 & Q24) have been reverse coded.
3.2.1 Control variables
The control variable have been added to the analysis because they might lead to alternative insights or explanations. The effect of the control variables is measured by both the
correlation of the control variable and the dependent variable and by identifying changes in the calculated regression when adding the control variable.
Three control variables have been added:
- the variable ‘practice size”; the number of FTE employees of the practice, The results ranging from 1; less than 10 fte, 2; between 10 and 49 fte, 3; between 50 and 99 fte and 4; more than 100 fte.
- The variable ‘accountancy firm’; a so called dummy variable was computed to indicate the respondents working for an accountancy firm. Appendix 2 gives an overview of the labels that were marked as ‘accountancy firm’
- The variable ‘consultancy firm’; a so called dummy variable was computed to indicate the respondents working for a consultancy firm. Appendix 2 gives an overview of the labels that were marked as ‘consultancy firm’.
Organizational size may impact the control system design (Chenhall, 2003), but the
relationship with team learning behavior is not clear. By controlling for size, I make sure that any results can be attributed to the MCS rather than to size effects. Additionally, the PSF sector may impact both MCS and team learning. I control for the two largest sectors in the sample, namely accountancy and consultancy.
After adjusting the pre-set constructs of the survey, the joined principal component analysis shows the following, solid, results:
4 RESULTS
Based on the previous chapter, using the constructs as described in appendix 1 in the following chapter focusses on the results per hypothesis. First the various variables are described in table 3, followed by both Pearson’s and Spearman’s correlations tests to establish the existence, nature and strength of a relationship. The results are displayed in table 4 and 5. Secondly, a multiple linear regression was calculated comprising both
dependent and independent variables as well as interactive variables to calculate the effect of interaction of the control mechanisms. Since firms use multiple control mechanisms and rarely use only one mechanism (Merchant & Van der Stede, 2012). Table 6 shows the results of the multiple regression which have been used to test the hypothesis. Finally, to test the robustness of the choices make in establishing the constructs, both the correlation tests and regression test have been computed using the predesigned constructs of the survey for formal and informal control. The results are displayed in table 7, 8 and 9.
Table 4; construct correlation
4.1 Results per hypothesis 4.1.1 Hypothesis 1
The bivariate correlation between result control and team learning is positive and
significant, but mediocre (Mukaka, 2012). The coefficients in all regression models (2-4) are not significant. Therefore hypothesis 1 is nog supported.
4.1.2 Hypothesis 2
Similar to the above results, the bivariate correlation between action control and team learning is also positive and significant, but the coefficients in all regression models (2-4) are not significant. Therefore hypothesis 2 is also not supported.
4.1.3 Hypothesis 3
The relationship between personnel control and team learning is supported by the bivariate correlation test. Further, the coefficients in all regression models (2-4) are significant. Therefor hypothesis 3 is supported.
4.1.4 Hypothesis 4
The relationship between cultural control and team learning is supported by the bivariate correlation test. Further, the coefficients in all regression models (2-4) are significant. Therefor hypothesis 4 is supported.
4.1.5 Hypothesis 5a, 5b, 5c and 5d
The direct effect of knowledge sharing on the relationship between the independent formal control variables and team learning in model 3 is in part positively significant; the
relationships between the formal controls and team learning remain insignificant (sig. = 0,460 and sig. = 0,179) whereas the relationship between the informal controls and team learning remain significant (sig. = 0,001 and sig. = 0,000) when adding the variable
knowledge sharing. All interaction variables have an insignificant coefficient, and hypotheses 5a through 5d are therefore not supported.
4.1.6 Control variables
When examining the impact of the control variables on both the dependent and
independent variables in the bivariate correlation test, only practice size shows a mediocre correlation with the variables: team learning, result control, action control and personnel control. The variable accountancy further shows a mediocre correlation with personnel control, however the correlation between the other variables is negligent in both
correlation tests. In the regression analyses, the consultancy dummy is significant at p < .10. This suggests that team learning happens more in consultancy firms. Consulting firms typically solve problems or implement new systems, which may necessitate a larger emphasis on learning behavior.
4.2 Robustness test
In order to test the effect of deviating from the preset constructs in the survey in making up the constructs used in this research, a robustness test was conducted, using the original constructs from the survey in making up the variables formal control and informal control Formal control consists of both the result and action control mechanisms and informal control of personnel and cultural control mechanisms. The above mentioned control variables are also used in the test.
Table 7; descriptive analysis
Table 8; construct correlations on construct formal/informal control
In line with the results calculated in the above paragraphs, the bivariate correlation between both formal and informal control and team learning is positive and significant. Further, regarding the relationship between formal control and team learning, the coefficients in all regression models (2-4) are not significant,.
The coefficients in all regression models (2-4) regarding the relationship between informal control and team learning are significant.
The outcomes of the robustness test continue to support the changes made to the
constructs when the moderating variable of knowledge sharing is added; the coefficients in regression model 4 remain insignificant for both formal and informal control.
4.3 Summary
It can be summarized that the formal controls have no significant effect on team learning, independent of the level of knowledge sharing within the firm. The significant effect of informal controls is slightly moderated by the level of knowledge sharing, resulting a lower unstandardized coefficient B. An overview of the results is shown in table 12.
5 DISCUSSION
5.1 Discussion
With this research I have tried to gain insight in the effect of choices firms make in designing a control system on team learning behavior; Contrary to my expectation, an emphasis on result controls, will not lead to a decrease in the level of team learning. The relationship was not significant. The assumption that the use of result controls would take away the focus of learning activities because the results of learning behavior are difficult to clarify, if there are any results at all (Edmondson, 1999) is not supported.
Even though the process of learning does not always yield a, direct, measurable result, the effort put into the learning process can, to a certain degree, be measured. And rewarded. Stipulating a desirable amount of effort put into learning activities as a measurable result, could explain the absence of a negative relationship and even support the tentative evidence that incentive structures can have a positive effect on knowledge sharing (Fey & Patrick, 2008; Jimenez-Jimenez & Sanz-Valle, 2012).
Equal to the relationship between result control and team learning, I also expected an emphasis on action controls to result in a negative relationship with team learning. The relationship however is not significant.
Although a professional should have a sense of autonomy to be motived (Løwendahl & Revang, 2001), evidence does not support the expectancy that stipulating desirable actions would in fact demotivate the professional to engage in learning activities.
Although it is difficult to clearly define desirable behavior in advance within a PSF, a possible explanation could be found in a broader view of what could be described as the expectancy that the professional “acts professionally” (Merchant & Van der Stede, 2012), including not only externally focused activities, toward the client, but also internally focused activities, towards the best interest of the firm.
The expectations on the informal control mechanisms of personnel control and cultural control were confirmed by the results. Both mechanism have a positive relationship with
team learning. The moderating effect of effort already put into knowledge sharing is confirmed, although only slightly, in both the relationship with personnel and cultural control.
The evidence seems to support previous research indicating that personnel controls such as training provides a good opportunity to share knowledge in and informal setting and helps to provide a common language which enables the transfer of knowledge and a way to communicate a shared vision (Jimenez-Jimenez & Sanz-Valle, 2012; Gagne & Deci, 2005; Ryan & Deci, 2000; Løwendahl & Revang, 2001) as well as on the interpretation of
knowledge (Jimenez-Jimenez & Sanz-Valle, 2012). Additionally, this research gives support to the idea that emphasis on hiring new personnel could sustain the level of variability and level of effectiveness of learning (March, 1991).
The results also seem to support the motivational strength of f.e. shared values and a shared vision to act in the best interest of the firm (Merchant & Van der Stede, 2012) and to function as an external motivator (Gagne & Deci, 2005) to participate in learning behavior.
The expected influence of knowledge sharing activities that are already present in a firm is not supported. Knowledge sharing does not interact with the management control system and does not moderate the impact of the control mechanisms on team learning. There is a positive direct effect on the relationship between the informal control mechanisms and team learning that could be explained by the similarities between hurdles towards
knowledge sharing and team learning (Fey & Patrick, 2008; Bock, Zmud, & Kim, 2005, Gupta & Govindarajan, 2000). In A firm that already invests in taking down hurdles towards
knowledge sharing will only need to eliminate a small number of hurdles to maximize team learning.
On the other hand, the evidence shows only a limited relationship, indicating that even though sharing knowledge and learning behavior are related and part of knowledge management (Argote, 2015; Alavi & Leidner, 2001; Grant, 1996), they are in fact separate activities (Gupta & Govindarajan, 2000).
5.2 Theoretical implications
The competitive advantages of knowledge and firm- specific knowledge have long been recognized to allow firms to create and maintain above average performances (Porter M. , 1985; Barney, 1991; Argote & Fahrenkopf, 2016). Although research in the field of
knowledge management, knowledge sharing and organizational learning has increased over the last decades, the integral effect of the choices made and emphasis put on the different management control mechanisms is still understudied. Literature covers a vast area of topics concerned with organizational learning and knowledge management, ranging from the various natures of knowledge such as collective knowledge versus individual knowledge (Alavi & Leidner, 2001) and explicit versus tacit knowledge (Grant, 1996), hurdles in
transferring, receiving and interpreting knowledge (Bunderson & Reagans, 2011; Bock, Zmud, & Kim, 2005; Fey & Patrick, 2008) and factors influencing knowledge transfer
(Fullwood & Rowley, 2017). It is however not clear yet how knowledge is actually managed in organizations (Gupta & Govindarajan, 2000; Fey & Patrick, 2008) or which specific variables support the flow of ideas and experiences, and what management control practices can affect those variables (Cabrera, Collins, & Salgad, 2006). Research that has investigated management control effects on knowledge sharing and organizational learning has often focused on the effect of one control mechanism, such as the effect of reward systems on knowledge sharing, with often contradictory results (Jimenez-Jimenez & Sanz-Valle, 2012). Furthermore, a recurring difficulty in comparing existing literature were the overlapping, but not quite comparable, definitions of knowledge sharing, knowledge
management, organizational learning etc. In this research I have come to a clear, supportive but separate definition of knowledge sharing and team learning to investigate the effect team learning. Knowledge sharing is defined as the act of making knowledge available to others within the organization (Ipe, 2003) and team learning activities are defined as the process of detecting faults and correcting them (Edmondson, 1999) or as the process of integrating new knowledge, using it to come to new insights (Argote, 2015).
This empirical study into the effect of management control on team learning therefore makes a contribution on existing literature in several ways. As it investigates whether management control systems can be used to align the interests of the firm and the professionals when it comes to sharing knowledge and organizational learning.
5.3 Limitations of the research
There are a number of limitations in this research. Since the survey was designed to cover a broader range of subjects, existing of 11 constructs, the constructs used in this research were at times to limited for the purpose of this research. The construct for action control in fact was limited to one item. Secondly, the number of respondents (n=148) did not support the differentation of the results per branche, or the usage of the different branches as a control variable. Both from an interest point of view as from a robustness point of view, a larger respondents population would have supplied additional opportunities in this area. Thirdly and finally, the number of respondents also makes it hard to generalize the outcome for “all” Professional Service Firms. This research is therefor of explotatory nature.
5.4 Practical implications
Through their day-to-day experiences professionals gain diverse knowledge that is valuable to the firm. Capitalizing on this asset however requires attention and input from
management. In fact it is one of the primary purposes of management to coordinate learning activities (Grant, 1996), especially in firms that rely so heavily on the expertise and knowledge of its employees such as PSF’s (Starbuck, 1992). This research shows that investing in personnel and cultural controls such as codes of conduct, shared traditions, norms, beliefs, values, ideologies, attitudes and ways of behaving are a powerful way to motivate employees to share knowledge and learn. Confirming existing literature (Jimenez-Jimenez & Sanz-Valle, 2012), this research also confirms that investing in personnel controls is also influential in stimulating the process of learning.
The results also show that focusing on results and actions controls does not hamper team learning. In other words, team learning as well as focusing on financial goals and targets can coexist. Finally, literature does indicate that the learning process needs to be specifically facilitated and coordinated as it does not take place without managements attentention (Grant, 1996)
5.5 Future research
This research is exploratory and gives and impression of the effects of control choices on learning activities. However, to research the relationship between the various control mechanisms more in depth, more research is necessary. F.e. as this research shows that the effort put into cultural control has the strongest impact on the level of team learning
activities occurring within the firm it would be interesting to research the effect of choices on formal control relate to, given a solid investment in cultural control. Also, the context of the PSF provides unique opportunities for research; as new knowledge is created through working with clients, possibly at the location of the client, but is shared and learned from internally. Questions arise such as the effect of the level of working on location on the relationship between the management control mechanisms and the learning process? Additionally, in this research I was not able to specify the results by the various branches that PSF’s operate in. This is however, also from a practical point of view, desirable and additional research using a larger respondent population could give valuable insights.
