Reduced performance-based horizontal pay dispersion : the effect on fairness perception

(1)

1 Amsterdam Business School

Reduced Performance-based Horizontal Pay Dispersion:

The Effect on Fairness Perception

Name: Simeon Volkers Student number: 10833412 Thesis supervisor: dr. P. Kroos Date: 15 August 2016

Word count: 12200

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

(2)

2 Statement of Originality

This document is written by student Simeon Volkers 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)

3

Abstract

The purpose of this research is to identify a relation between horizontal performance-based pay dispersion and the fairness perceptions of employees. Pay dispersion research shows that the claimed effects are not consistent. With this study I call out to the need for more insight in the effect of pay dispersion from an employee’s perspective. Because fairness perceptions towards the pay system has an important effect on job attitudes and employee motivation it is an important construct which need to be accounted for when designing a pay system. I’ve conducted an experiment where fairness perceptions in different degrees of dispersion is measured. Also in this experiment the position of the individual in the pay distribution is taken into account. This resulted in six different conditions which measure fairness perceptions before and after a manipulation of the pay dispersion. Main analysis was conducted with a repeated measures ANCOVA. There were in total of 89 useful participants used in this study. The results show that when pay dispersion is reduced within a performance based horizontal setting this results in an increase in the fairness perceptions of employees. The strength of the increase of the fairness perceptions is moderated by the position an individual has in the pay distribution. Lower paid individuals have a stronger increase in fairness perceptions when pay dispersion is partially reduced or fully reduced. When pay dispersion stays intact by applying an overall raise for every employee this resulted in a minimal increase of fairness perceptions, but biggest change is caused by the manipulation in pay dispersion. Further a discussion of the results has been made which is followed on by limitations and recommendations for future studies within this field.

Keywords: pay dispersion, fairness perceptions, horizontal pay dispersion, performance-based pay dispersion

(4)

4

1 Introduction

“CEOs make 276 times more than a typical worker (Mishel and Schieder, 2016, p.11)”. This quote from the report of the Economic Policy Institute in the US shows again a reason to open up the debate about pay dispersion within firms. These authors are concerned that this exorbitant CEO pay ratio means that the fruits of economic growth is not going to workers. These type of reports trigger the interest in the pay dispersion of organizations in societal context, but also from an organizations point of view. Pay dispersion within organizations is said to have motivation and sorting effects on employees, which on their turn impact organizational

performance (Downes and Choi, 2014). Human capital is a significant resource for employers for the achievement of organizational goals and objectives (Huselid, 1995) and it is therefore

important for organizations to motivate their employees to produce results that are consistent with firms’ objectives. Bloom (1999) states that one important factor for motivation is the organizations pay spread, because this mechanism has influence on the effort of employees. Important determinant for employees’ motivation is the fairness of the compensation they receive in return for their contribution to the organization (Konovsky 2000; Wade et al. 2006). This research addresses lower and higher paid employee’s fairness perceptions towards their pay in dispersed and compressed systems. I conduct this research from an employee reactions point of view towards pay dispersion because, as it has been posed by Downes and Choi (2014), it provides more insight on what the most effective type of dispersion would be for organizations. According to the literature review of Downes and Choi academics are not in agreement on which design of the pay system (either dispersed or compressed) is the most effective for employee motivation. It seems that managers are looking for a “sweet spot” where the ideal mix of

dispersion is found. With reference to their pay dispersion typology I focus with my experiment on performance based horizontal pay dispersion. The question that I’m trying to answer is whether pay dispersion between similar type of performance based jobs has an effect on the fairness perceptions of the employee.

“Does reducing performance based horizontal pay dispersion have an effect on the fairness perceptions of employees?”

Many studies regarding the pay dispersion field have found different results relating employee turnover, performance and motivation. With this study I want to contribute to the literature in providing extra insight on the effect of pay dispersion between similar hierarchical jobs on fairness perceptions of employees. To my knowledge, fairness perceptions (which is a part of an employee reactions perspective on pay dispersion) have not yet been empirically investigated in a

(7)

7 performance based horizontal pay dispersion setting. I reach out to the call for more insight in this particular field of pay dispersion as proposed by Downes and Choi (2014). This study also contributes to the societal debate whether high dispersed systems are preferable or not. The study gives also practical contributions towards management because it gives them more insight on how to design their pay spread within the organization in order to maximize employee motivation levels.

2 Literature review and theories

2.1 Pay dispersion literature

The phenomenon of pay dispersion has received a growing attention in management literature. With the large amount of studies conducted in this field the results vary on which form of pay dispersion are the most beneficial for organizations. Bloom and Michel (2002) give a definition of pay dispersion. It is defined as “the amount of difference (inequality) in pay created by a firms pay structure” (Bloom & Michel, 2002, p. 33). The amount of difference can also be stated as the pay differential. This pay differential can be either high or low. In a compressed system the difference in pay between the high and low earners is little and relatively close to each other. A high pay dispersion is referred to as a dispersed system. In a dispersed system the difference in pay between the highest and lowest earning individual is relatively largely dispersed.

2.1.1 Benefits of dispersed pay

Some scholars state that high dispersed pay has a positive effect on employee motivation (Lazear and Rosen, 1981; Rosen, 1986; Zenger, 1992). Trevor et al. (1997) show that high pay dispersion is effective for the retention of top performing employees. They argue that when pay spread is considered as low this results in relative low motivation and subsequently performance for employees who are capable to become top performers. Becker and Huselid (1992) studied pay dispersion in a tournament setting where they examined the differences in rewards of contestants. They concluded that a high dispersion in rewards caused for NASCAR drivers to have more motivation to perform better. The drivers took more risk and displayed more aggressive decision making. Other positive effects of dispersed pay were found by Kepes et al. (2009) and Ding et al. (2009). Kepes et al. (2009) show that high pay dispersion had a negative effect on the out-of-service days of a truck company and was positively related with the ROE. Ding et al. (2009) showed that pay dispersion in Chinese manufacturing and service firms had a positive effect on sales growth and product/service quality within the relationship manager –employee. They however did not find a positive effect of the dispersion between groups itself, for example employee-employee or manager-manager.

(8)

8

2.1.2 Negative findings on dispersed pay

Where previous discussed authors find positive effects of dispersed pay there is also literature that finds negative effects of dispersed pay and positive effects for compressed pay. With regard to negative findings of dispersed systems (i.e. positive findings of compressed systems) there are researchers that argue that compressed systems have a positive effect on employee motivation and consequently firm performance (Adams, 1963; Cowherd and Levine, 1992; Milkovich & Newman, 1996). Compressed systems can have advantages for group performance because they include feelings of fairness, common purpose and support team-oriented behavior (Cowherd and Leivne, 1992; Lazear, 1995; Pfeffer, 1994). Kohn (1993) and Pfeffer (1994) suggest that dispersed pay systems provide less room for cooperation, increase dissatisfaction and has a negative impact on performance. Pfeffer and Langton (1993) found that a dispersed system had negative effects on the research productivity and job satisfaction of college employees. Bloom (1999) shows that dispersion was negatively related to the individual and team performance in a Baseball league. In another study Bloom and Michel (2002) found negative effects of a dispersed system on retaining executive managers. Research on the individual pay equity perceptions by Trevor and Wazeter (2006) showed that negative effects are associated with a dispersed system within a survey of public school teachers. Cowherd and Levine (1992) find that more equal pay was positively related to product quality. More negative effects were found of dispersed systems relating

turnover among managers of publicly traded firms (Messersmith et al (2011) and among low level college administrator employees (Pfeffer and Davis-Blake, 1992). These last two findings suggest that the level in the hierarchy in an organization (CEO or employee) has no influence on the turnover in dispersed systems. This finding was further investigated by Shaw et al (2002) which looked at quit rates of good, average and poor performing truck drivers. They find that when firms do not have a performance-based pay system the pay dispersion was positively related to turnover of good performing truck drivers. However, when firms did have a performance-based pay system and this system was communicated well and understood by employees, then the pay dispersion was negatively related to good performer quit rates. These researchers found evidence that pay dispersion can have negative effects on firm outcomes, however they argue that

dispersion is neither good nor bad. The effectiveness of the dispersion depends on the contingencies of the organizations environment.

2.1.3 Typology of pay dispersion

Downes and Choi (2014) find in the literature differences in the type of dispersion. They made a typology of pay dispersion and divided the phenomenon in four different categories. The

(9)

9 researchers make a distinction between performance-based and non-performance based pay dispersion and a distinction between horizontal and vertical pay dispersion.

2.1.4 Vertical versus horizontal pay dispersion

Bloom (1999), Ding et al (2009) and Kepes et al. (2009) mention the difference in vertical and horizontal pay dispersion. This difference relates to where pay dispersion occurs in the

organization, either between or within hierarchical levels of an organization. Vertical dispersion relates to the amount difference in pay between hierarchical levels (manager-employee).

Horizontal pay dispersion refers to the difference in pay within a group of peers which are active in the same horizontal level (employee – employee). Horizontal pay dispersion refers to a group of employees who have the same job or are active in the same hierarchical level. Jobs are defined as a collection of duties and responsibilities (Gupta et al, 2012), which means that jobs can have different titles but in reality perform the same tasks. Bloom (1999) notes that horizontal pay dispersion can be created by performance-based pay or seniority/politics.

2.1.5 Performance-based versus non-performance based

Gupta et al. (2012) mention the distinction between performance-based and non-performance based pay dispersion. Performance-based pay dispersion is a pay system where the dispersion is explained by individual differences in performance. Pay dispersion as a result of human capital factors such as a higher degree are considered as non-performance-based, because human capital factors which partly cause the dispersion are assumed not to have a negative impact on the fairness perceptions or motivation, because they are considered as legitimate. Pay dispersion caused by other factors such as seniority or politics (also non-performance based) would however have a negative impact on individual motivation. Gupta et al. (2012) suggests that pay variation based on performance-based system should motivate higher performance, while other factors such as seniority (time working in a job) and politics would not foster motivation.

2.1.6 Important characteristics of pay dispersion effectiveness

To understand the effectiveness of pay dispersion. Researchers found that the conflicting results can be partially explained by certain characteristics. As Shaw et al. (2002) suggested, the

effectiveness of a pay system depend on certain important contingencies. Downes and Choi (2014) examine the conditions mentioned in literature which influence the effectiveness of the pay system. They identify interdependence, position in the pay distribution, communication, pay basis and human capital as factors which have influence on the outcomes of the effectiveness of the pay system.

(10)

10

2.1.6.1 Interdependence

One of the moderators of pay system effectiveness is the interdependence of the work involved (Shaw et al. 2002). When work is interdependent this will allow individuals to observe each other’s work and must be therefore taken into account when it is about pay dispersion research. In their study Shaw et al. (2002) find that pay dispersion has negative effects on firm performance when the nature of work is highly interdependent.

2.1.6.2 Position in the pay distribution

The position of an individual in the pay system has an impact on the motivation and fairness of the pay distribution scheme (from the individual perspective). This eventually has an effect on the performance of a firm. Trevor and Wazeter (2006) have found evidence that individuals lower in the pay distribution would be more likely to view a highly dispersed pay system as inequitable. On the other hand, employees that are paid considerably high will be more tolerant to the

inequitable pay system, because of their larger reward. Some researchers state that this inequitable effect of the position in pay dispersion can be eliminated by applying higher wages (Brown et al, 2003). According to Brown the negative effects of a dispersed system would be compensated if the average pay is high.

2.1.6.3 Communication and understanding of the pay system

Shaw and Gupta (2007) provide evidence on the moderating role of pay system communication in understanding relationship with pay dispersion, firm performance and pay basis. Firms who communicate the pay system well to employees have a negative relation with high performer quit rates within a performance-based pay situation. When communication is low about the reasons and intent of the pay system this causes employees to weaken their perceptions that pay

differences are the result of legitimate factors.

2.1.6.4 Human capital

Gerhart and Rynes (2003) and Trevor et al. (2012) argue that next to pay basis also human capital is critical for pay dispersion research. When dispersion arises because of human capital factors, such as a higher degree this is considered as legitimate. When measuring pay dispersion, the model should account for human capital variables.

2.1.6.5 Pay basis

Downes and Choi (2014) state this factor as probably the most important factor which influences the effectiveness of the pay distribution system. The pay basis refers to when pay differences are legitimately based on differences in performance outcomes, then employees are likely to consider the system as fair. This moderator is closely tied to fairness perception of employees. With this

(11)

11 perspective in mind this would mean that when employees are paid highly based on their

performance outcomes then the pay dispersion should have a positive effect on motivation, because high performers have the opportunity to earn more money. Previous discussed researchers (Shaw and Gupta, 2007, Kepes et al, 2009) underline this statement with their findings that there are negative outcomes on firm’s performance when pay is non-performance based and that pay dispersion is positively related to return on equity when not based on politics.

2.2 Theories

2.2.1 Equity theory

When pay differences are attached to performance this will result in a perception of a more equitable situation from the employee point of view. This finding can be derived from equity theory of Adams (1963a, 1965). Equity theory considers the nature of inputs and outcomes, the nature of social comparison process, the conditions leading to equity or inequity and possible effects of inequity and the responses of individuals to reduce inequity. Inputs are factors perceived by a person to be relevant for getting return on personal investment. Examples of inputs are effort, education and personal skills (Pritchard, 1969). Outcomes are all factors that are perceived by the person as returns to himself that provide utility or value. The theory states that individuals calculate an input/output ratio and that ratio is weighted (compared to referent others) according to perceived importance to determine the final value of the ratio. If the perceived own ratio is not comparable with the referent other this will cause perceived inequity for the specific individual, for example individual X puts in 8 hours of labor and gets paid 100. Individual Y puts in 4 hours of comparable labor and gets paid 100, and no other factors like human capital or seniority influence this individuals’ inputs, this situation can be perceived inequitable.

The referent other can refer to either a direct relationship with two persons such as in a relationship (friends) or with two persons who are related to a third party for example, two employees paid by the same employer. The theory posits that when individuals perceive their ratio, compared to a referent other, as unfair the individual will take actions to reduce or avoid the inequitable situation. Goal is to make both ratios more equitable. Adams describes four responses to inequity, (1) cognitively distorting his or others inputs or outcomes, (2) make referent other to change their inputs or outcomes, (3) change own inputs or outcomes or (4) change the referent person which the individual compares itself to or leave the profession/field.

(12)

12

2.2.2 Other relevant theories 2.2.2.1 Expectancy theory

Vroom (1964) described the expectancy theory. This theory suggests that that differences in pay are motivating if outcomes are valued high (bonus or high pay) by employees, if employees believe that increased effort results in higher performance thus higher outcomes such as increased pay. In contrast to equity theory, expectancy theory would predict that larger rewards will result in greater motivation effect of the employee, because outcomes are valued high. The theory also implicates that when there is a good relationship with pay and performance this will result in greater motivation because the employee perceives a higher level of performance is tied to a higher level of pay.

2.2.2.2 Tournament theory

A tournament is a contest where actors compete amongst each other for a prize that is based on relative rank and is designed to incentivize an optimal effort level (Becker & Huselid, 1992; Lazear, 1999). The tournament theory originates from the paper of Lazear and Rosen (1981). They focused with their research on optimum labour contracts and promotions. The theory poses that a greater reward (larger prize) for winning (promotion earning) will increase the amount of effort an employee will give to win the prize (achieve the promotion). Promotions within organizations (with attached salary growth) are used as continuous incentive mechanisms to pursue an optimal level of motivation among employees. The higher salary related to a promotion is considered as the “prize” the employee can win. With regard to pay dispersion the theory predicts that large pay differences will result in greater motivation of employees (Trevor & Wazeter, 2006).

2.3 Effects of pay dispersion

Pay dispersion has an effect on individual motivation (effort) and sorting (attraction and

retention effect) (Downes and Choi, 2014). All the accumulated individual responses towards pay dispersion would have an impact on firm-level performance. From an equity theory perspective performance-based systems, either vertical or horizontal, would have positive effects on

motivation and sorting. The effects of non-performance based pay dispersion depends on if the employee considers the dispersion of pay as legitimate. From equity theory we would expect that non-performance based dispersion would create feelings of unfairness (inequity) and presumably this will lead to lower motivation and would make retention more difficult, however if the dispersion is legitimized than negative effects are not present. Shaw and Gupta (2007) find that employees view on the legitimacy of pay differences may moderate the effects of pay dispersion

(13)

13 and Trevor and Wazeter (2006) suggest that the fairness perception of the pay basis influences the negative associations of pay dispersion. For example, when pay dispersion occurs between employees and is based on skills, people may believe that when they improve skills their pay must increase as well. This skill-based dispersion is considered as legitimate and will not result in negative equity perceptions. A dispersion based on politics however may not be as seen as legitimate and will increase negative equity perceptions and thus decrease motivation and favorable sorting effects. Illegitimate non-performance pay dispersion will result a discrepancy with the input and output ratio between two parties and therefor have negative effects on the individual’s motivation (Cohen-Charash & Spector, 2001).

Positive effects on motivation and sorting can also be explained from other theories’ perspective. Expectancy theory predicts that performance-based horizontal dispersion will have positive effects on motivation and sorting. Van Eerde & Thierry (1996) show that by showing that when employees have greater levels of perceived value, greater level of believe that increased effort results in higher performance and higher pay that this has a positive correlation with some

important organizational measures such as effort, individual performance, intention to apply for a job, job attractiveness ratings and preference of jobs. Tournament theory would predict that performance-based vertical dispersion has positive effects on motivation and sorting effects. Larger rewards result in increased motivation to compete for the prize. Sorting effects are present, because employees that will not get into a certain promotion or job level will voluntarily leave the firm or pursue another career. Other sorting effect is attraction of high performing employees. High performers would be attracted to organizations with greater degree of dispersion and would select themselves into positions that earn higher compensation.

2.4 Fairness perceptions towards pay and motivation

According to equity theory fairness perceptions of employees play crucial role in determining whether a pay system is considered as equitable or inequitable. Fairness perceptions of employees are crucial for HR decisions and processes and in particular for compensation decisions such as pay, pay raises and benefits (Jawahar & Stone, 2011). Fairness perceptions have an effect on job attitudes and behavior for example, it is demonstrated that high level of fairness perceptions may result in higher job satisfaction, organizational commitment (Folger and Konovsky 1989;

Greenberg 1994; Hausknecht et al. 2011) and reduces turnover intentions of employees (Hausknecht et al. 2011). Other research states that pay fairness is vital in the determination of employees’ utility and a key aspect of a successful compensation system (Dijke et al. 2009; Hundley and Kim 1997).

(14)

14 Perceptions of fairness can be explained from different point of views. As discussed the fairness perceptions derived from equity theory is called (1) Distributive fairness (or justice), because it looks at the distribution of outcomes and the perceived fairness of this distribution of outcomes (the ratio compared to referent others). Other concepts of fairness are (2) Procedural fairness (Thibaut and Walker, 1975) which refers to the perceived fairness of procedures used leading to the distribution of the outcomes. (3) Interactional fairness which refers to the perceived fairness people get from the treatment of decision makers and (4) informational fairness which focuses on explanations provided to employees to explain why certain procedures were used and why

outcomes are distributed in a certain way (Colquitt et al. 2001).

I will look into the distributive fairness concept as it derives from equity theory. (1) Distributive fairness assesses the fairness perception of one’s outcome based on comparing with the outcome of its peers relatively to their inputs. The equity theory predicts that feelings of inequity might induce people to have negative fairness perceptions towards their pay and subsequently reduce their effort levels and to balance the ratio of economic exchanges (as discussed previously, either adjust own behavior, others or quit).

3 Theoretical Model

3.1 Hypothesis development

In a situation where pay is dispersed feelings of inequity might arise and can result in negative fairness perceptions. This experience of inequity will result in a negative effect on employee motivation and satisfaction (Kepes et al., 2009). To prevent feelings of unfairness and inequity the pay system should be compressed in order to mitigate the potential feelings of inequity (Bloom, 1999). However, the proposed feelings of inequity would in a performance-based horizontal pay dispersion situation be legitimized because employee itself has the power to influence the outcomes by putting in extra effort. For example, when an employee is given a bonus based on his amount of sales this is due to own effort and therefor the pay dispersion is justified and seen as legitimate. However, the effect of performance-based dispersion on motivation may be

nonlinear (Downes and Choi, 2014). There is a possibility that individuals would not exceed more effort if they perceive that differences in pay are too large. The researchers talk about a possible negative curvilinear relationship of performance-based pay dispersion and individual motivation. This means that there is somewhere a limit in pay dispersion which causes positive effects. This remark implicates that when there is too much dispersion in a performance-based situation than the effects at a certain dispersion level will not be positive. This potential negative effect is not yet empirically researched. Downes and Choi (2014) posit this possibility based on the findings of

(15)

15 Gund and Westergaard-Nielsen (2008). These authors find positive effects of wage growth on firm performance and they argue a fairness mechanism operates up to a certain level of wage growth dispersion. Also Downes and Choi (2014) state that potential decrease in motivation caused by perceived feelings of unfairness is generally not considered in pay dispersion research. They argue that for a more complete picture from an equity theory perspective the positive claimed effects of more equitable systems must be tested.

With this thesis I, based on previously discussed knowledge, conduct research in the performance-based horizontal dispersion field. Only relevant research on this type of dispersion was done by Becker and Huselid (1992). This research was done in a very specific setting which was a NASCAR tournament. Their research was tested and supported based on tournament theory. Downes and Choi state that tournament theory can be applied to test the effect of performance-based horizontal dispersion (Gupta et al. 2012) but the relationship is in general more posited by equity and expectancy theory. Downes and Choi state that Becker and Huselids findings have provided excellent evidence regarding the nature of the effect of performance-based horizontal dispersion, but they call out for more research to fully understand the effect of performance-based dispersion on motivation and performance. I address to this concern by exploring fairness perceptions in a performance-based horizontal dispersion setting from an equity theory

perspective.

Based on all information there is a need to get more understanding about performance-based horizontal dispersion and its possible effects on motivation. One mechanism driving motivation are fairness perceptions toward the compensation system and associated dispersion.

Performance-based dispersion would create positive effects on motivation, but it seems that there is a limit somewhere in the dispersion measure. Therefor I formulate the following hypothesis: H1 Reduced performance-based pay dispersion will lead to an increase in fairness perceptions

3.2 Position in the pay distribution

Trevor and Wazeter (2006) mention that employees low in the pay distribution will consider pay dispersion more inequitable than employees high in the pay distribution. I’m testing this finding by assessing participants who are positioned high or low in a horizontal performance-based pay dispersion setting. Reason for testing is that in a horizontal performance based pay dispersion setting there in practice will always be relatively top and low performers (not necessarily bad performers). All of the outputs of the employees together contribute to organizational objectives

(16)

16 and therefor it is important to test the effect of dispersion on both groups. Therefor I formulate the following hypothesis:

H2 The position in the pay distribution ladder has a moderating effect on the fairness perceptions

H2a Fairness perceptions of employees with a high position in the pay distribution ladder will decrease when performance-based pay dispersion is reduced

H2b Fairness perceptions of employees with a low position in the pay distribution will increase when performance-based pay dispersion is reduced

3.3 Variables and relations

The figure below shows the research model which demonstrates the relationships that are tested. The focus will be on horizontal performance-based pay dispersion. Reasons for this focus is because there has been done relatively lots of research on vertical performance and

non-performance pay dispersion. More understanding is needed for the specific type of dispersion. I use fairness perception as dependent variable and a dispersed system (no reduction), a semi-dispersed system (partial reduction) and compressed system (full reduction) as one of the independent experimental variable. Also position in the pay distribution is an independent variable, where employees are either relatively high or low in the pay distribution. Fairness perceptions are an important construct for employee motivation and also fairness perceptions have not been researched yet in this context.

Figure 1 Research Model

(17)

17

4 Methodology

4.1 General overview

To test whether pay dispersion has an effect on the fairness perceptions of employees I have conducted an experiment vignette study. I test H1 and H2 using an experiment where participants assume the role of a sales representative for a fictional company. Sales representatives earn a fixed wage and face a bonus compensation system (in the form of

commissions) in each of two periods. In each period the sales representative assesses the fairness of the bonus compensation system. In the first period dispersion occurs between sales

representatives who work in different regions. The participant will be either work in the North or South region. If the participant is placed in the North region he or she will be in the placed in the lower paid level of the pay distribution, if placed in the South region he or she is placed in the higher paid level of the pay distribution. In the second period I introduce three manipulations which changes the level of the dispersion between different sales representatives. These manipulations are discussed in the “Conditions” section.

4.2 Participants

I’ve recruited participants through the social media channel of Facebook. 100 participants initially responded to the inquiry for completing the questionnaire. The Facebook profile where the questionnaire was send out to had a reach of approximately 700 persons. The link which guided to the survey was shared twice. In total 89 participants completed the questionnaire and were mostly highly educated. For more background see the results section (under Descriptive statistics)

4.3 Procedures

Each participant in the experiment assumes the role of a sales representative at a fictional

company. The vignette (see appendix) tells the participant that initially there is a sales department within the company which is important for the performance of the company. It is told that the department is divided in region North and South. The sales representative is informed that employees in both regions initially earned an equal fixed compensation and have equal bonus scheme terms. The case further explains the nature of the sales employees’ job which is engaging and recruiting customers, providing quotes and answering customer questions. The participant is than informed that due to an external report the bonus scheme of region South is increased from 10 bonus per sale to 20 bonus per sale. Because the report only was applicable to the South region (report initiated adjustment of bonus scheme) the North region bonus scheme remained unchanged. The participant then assessed the fairness of their pay by rating their agreement with

(18)

18 several statements regarding overall and distributive fairness. These statements are measured on a 9-point likert scale with 1 as strongly disagree and 9 as strongly agree.

Further validity questions are asked relating the presented case in order to make sure the participant fully understands the setting and knows in which place of the distribution they are ordered.

In the second period the participants learn that general management again made changes in the bonus compensation system. These manipulations are considered in the conditions section. After this manipulation the sales representative assesses the fairness of their period 2 bonus system using the same scale as in period 1. Also extra questions regarding distributive fairness

perceptions are considered as used by Colquit and Rodel (2015) handbook for measuring fairness and justice (see section measuring fairness perceptions).

Finally, in the post experiment questionnaire questions are asked regarding the sales

representative perceptions of legitimacy of the initial pay dispersion in period 1 (dispersion after the external report) and various questions to collect demographic data. It took participants around six minutes to complete the experiment and no incentives were offered.

4.4 Influencing factors

To test whether the dispersion has an effect on the fairness perceptions various variables need to be taken into account which can according to literature possibly influence the dependent variable of fairness perceptions.

First of all, I take interdependence of work into account in the experiment. Dispersed systems in interdependent work settings are said to have detrimental effects on motivation. So in the vignette design the condition relating to interdependence is held constant. It will be explicitly mentioned that individuals work independent of each other in order to mitigate potential extra negative fairness perceptions caused by possible interdependence of work.

Next to interdependence, human capital is taken into account. Some dispersion can be perceived by participants as a product of higher degree or experience. I need to rule out that fairness perceptions are derived based on human capital explanations. Therefor the case will mention that each individual’s colleague is at a same level regarding degree and experience (homogeneity) in both conditions.

Communication/understanding about the pay system also are taken into account in the

experimental design. Literature tells us when communication is low about the reasons and intent of the pay system this causes employees to weaken their perceptions that the dispersion is a result

(19)

19 of legitimate factors. Therefor a reason will be given for the dispersion in the horizontal

performance-based setting which is the external report. The legitimacy of this reason which causes the pay dispersion is measured, because when participants would consider the reasons of the dispersion as illegitimate this will have a negative effect on their possible fairness perceptions. Goal of taking into account the discussed influencing factors is to be certain that possible

changes of fairness perceptions toward the dispersion is purely caused by the change in dispersion (towards more compressed or dispersed). Additional validity questions are asked to participants in order to make sure participants understand that they work independent and have the same human capital factors.

4.5 Conditions

4.5.1 Position in pay distribution

The experiment will indicate if someone is a sales representative of region North or region South. Employees in region North are considered as the low-paid employees and employees in region South are considered as the high-paid employees. Because there is a difference between the bonus compensation between these groups the fairness perceptions of participants in North and South will be measured, because it is expected that fairness perceptions differ based on the position in the distribution.

4.5.2 Dispersion manipulation

In the first four experimental conditions I reduce the initial pay dispersion between the sales associates within the different regions.

Partial reduction Full reduction No reduction

High in distribution 1 2 3

Low in distribution 4 5 6

In condition 1 and 4 (Low position/Partial reduction and High position/Partial reduction) the bonus scheme of the North region will increase from 10 to 15, while the South region bonus scheme stays at 20. With this increase the difference in the performance -based compensation through the bonus between the different regions will be reduced in the second period. The two conditions will have participants who are operating in the high position of the pay distribution (South) and who are operating in the low position of the pay distribution (North).

In condition 2 and 5 (Low position / Full reduction and High Position/Full reduction) the bonus scheme of the North region will increase from 10 to 20, making the North and South bonus scheme

(20)

20 compensation leveled in the second period. Also just as in the first two conditions participants are either in the low or high pay distribution.

In condition 1, 2, 4 and 5 the pay dispersion is reduced by increasing the lower paid sales associate bonus scheme. It is possible that an increase of the bonus scheme can increase the fairness perception because of the employees consider the increase bonus scheme as a gift. Thus an alternative explanation for a possible increase in fairness perceptions towards pay can be explained by the gift bonus increase from the employer. I want to rule out this possible

opportunity that fairness perceptions only derive from increase in pay resulting from the bonus scheme. Therefor two additional conditions (3 and 6) are added whereby the dispersion

manipulation is held constant over the two different regions (Low position – No reduction and High position / Low reduction). The bonus commission for North will be raised from 10 to 20 and the bonus commission for South will be raised from 20 to 30 (equal raise which results in same absolute dispersion). When I find no increase in the fairness perceptions of the sales

representative between the first and second period in the Low position – No reduction and High position / Low reduction conditions this would demonstrate that any observed increase in fairness perceptions in the other conditions is not explained by the increase in the bonus scheme compensation, but rather by the relative dispersion. If fairness perceptions in the No Reduction condition increase this would be consistent with being a response to increased wage rather than the reduced pay dispersion. When comparing results of the No Reduction conditions with the Partial Reduction conditions allows to distinguish whether any observed increase in fairness perceptions is due to reducing pay dispersion or explained as a gift exchange.

4.5.3 Depended Variable

The primary depended variable are the measured fairness perceptions. Fairness perceptions are measured using Colquitt and Rodell (2015) handbook for measuring justice and fairness. Questions are asked to participants relating distributive justice. Procedural justice questions would be relevant to measure when participants would have influence on the decision making process of the bonus system, which here is not the case and difficult to measure in an

experiment. I purely focus on the fairness perceptions on the distributive outcomes. Since pay dispersion results in difference in outcomes. Next to the distributive fairness perceptions also an overall fairness perception question is asked with the question: “My bonus is fair given the work that I do for Elix Inc.” before the manipulation and after the manipulation “Based on management new decision, my bonus is fair given the work that I do for Elix Inc.” The effect of the manipulation is measured as the change in fairness perceptions between these periods. So Fairness change is

(21)

21 equal to the sales representative fairness in period 2 (after manipulation) minus their fairness perceptions in period 1 (before manipulation).

Figure 1 Expectations

Figure 2 shows the initial expectations regarding the fairness perceptions within the different tested conditions. As shown, expected was that position in the distribution plays an important role in the fairness perceptions of the employee regarding their bonus pay. Further expectations were that full reduction of pay dispersion is considered as fair for employees low in the

distribution and the less the dispersion is reduced (partial reduction) the weaker the effect on fairness perceptions will be.

4.6 Legitimacy

Pay dispersion has a negative effect on employees low in the pay distribution when they do not believe the pay dispersion results from legitimate reasons (Pfeffer and Langton, 1993). Therefor an additional question is asked relating the perceived legitimacy for the initial decision that caused the pay dispersion. Further analysis will be done whether perceived legitimacy is associated with the fairness perceptions of the participants.

I further test the initial perceived legitimacy of the dispersion within the case and see if this has an impact on the perceived distributive fairness perceptions of the participants.

Fairness perception Partial reduction No reduction Position in distribution Low High Full reduction Low High 1 2 3 4 5 6

(22)

22

5 Results

Following are an illustration of the results of this study. I’ve conducted a repeated measures ANCOVA because the experiment consists of two moments of measurement of fairness perceptions within subjects and I’m comparing between groups. Further additional analysis is done regarding certain control variables such as age, gender and education level. To conclude I’ve checked the results for robustness.

5.1 Descriptive statistics

The research sample in this study amounted to, included with correction after manipulation checks and non-complete questionnaires, 89 participants (from 100). 45,6 percent were male and 54,4 female. Average age was 27,31 (SD = 5,574) and 94,4 percent had a higher vocational study or university degree. All participants were from the Netherlands. For all respondents, 68,9 percent has had working experience with bonus schemes and 31,1 percent did not. Participants were placed randomly in different conditions with an online survey system. Following table shows the distribution of the participants across the different conditions. Table 1

Descriptive statistics in the 6 conditions (N = 89)

Condition N Male Female Average age HBO/WO

(Bachelor/Mas ter) Education level MBO secondary vocational education level Bonus experience (yes) High in pay distribution / Partial reduction 13 23% 77% 27,285 100% 0% 71,43% High in pay distribution / Full reduction 17 41% 59% 28,473 89,47% 10,53% 57,89% High in pay distribution / No reduction 16 44% 56% 27 100% 0% 50% Low in distribution / Partial reduction 14 57% 43% 27,375 93,75% 6,25% 81,25% Low in distribution / Full reduction 18 50% 50% 25,052 100% 0% 73,68% Low in distribution / No reduction 11 55% 45% 27,181 81,81% 18,19% 90,09%

(23)

23

5.2 Factor analysis

First of all, fairness perceptions are measured using Colquit & Rodell measurement scales. A 9-point likert scale is used to assess overall fairness perceptions relating distributive fairness. First fairness was measured before the manipulation (T0) and then fairness was measured after manipulation (T1).

I’ve computed overall fairness measures for the two periods by first conducting a factor analysis regarding the fairness questions. I’ve found with factor analysis is that all fairness related

questions (before and after manipulation) consisted combined of two underlying factors. Variables Dist_fairness_T0 (before manipulation) and Dist_Fairness_T1 (after manipulation) were computed by computing mean scores for each respondent. Reliability measure of the Dist_fairness_T0 has a Cronbach’s Alpha of ,89 and Dist_Fairness_T1 has a Cronbach’s Alpha of ,907. Because deleting items would not result in a big increase of Cronbach’s Alpha and Cronbach’s Alpha is considered as high (>0.7), there for all 5 items (before and after) are included in the new measures.

Table 2

Pattern matrix of the factor analysis over 10 items measuring the fairness construct before and after the manipulation

Factor 1 “T1” 2 “T0” Fairness -,124 ,946 Reflectionwork -,018 ,787 Approriatenessreward ,010 ,785 Contribution ,071 ,611 Justification ,218 ,762 Fairness1 ,926 -,047 Reflectionwork1 ,844 ,029 Appropriatenessreward1 ,764 ,079 Contribution1 ,631 ,053 Justification1 ,945 -,067 R2 52,77% 14,45% eigenvalue 5,580 1,745 Cronbach’s alpha ,907 ,89 5.2.1 Fairness perceptions

Table 3 and 4 show the means of the measured distributive fairness perceptions in different conditions. Initial fairness perceptions before the manipulation differ between the high and low group, t(98) = -9,02, p < ,001. Fairness perceptions initially are higher when a participant is in the

(24)

24 high end of the pay distribution with an average mean of 7,33, whereas participants low in the distribution show an average mean of fairness perceptions of 5,81.

After the manipulation the average fairness perceptions change slightly in the high group from 7,33 to 7,30, t(48) = 0,39, p = ,70. Average fairness perceptions in the low group increase significantly from 5,81 to 7,06, t(45) = -9,66, p < ,001. This is however all manipulations

combined. When we look at each individual manipulation we see in the Low / Partial condition an increased mean from 5,92 to 7,03, t(15) = -6,06, p < ,001. Also in the Low / Full group an increase of mean occurred after manipulation from 5,78 to 7,60, t(18) = -11,93, p < ,001. The Low/No Reduction group fairness perceptions increased little but significantly from 5,71 to 6,20, t(10) = -2,36, p = ,04.

Table 3

Fairness perceptions before manipulation (Dist_fairness_T0)

Pay distribution Reduction Mean Std. Deviation N

Low Partial 5,92 ,53 14

Full 5,78 ,65 18

No 5,70 ,52 11

Total 5,81 ,57 43

High Partial 7,50 ,33 13

Full 7,15 ,85 17

No 7,38 ,66 16

Total 7,33 ,67 46

Pay distribution Reduction Mean Std. Deviation N

High + Low Partial 6,69 0,92 27

Full 6,45 1,01 35

No 6,2 1,03 27

Total 6,6 0,99 89

Table 4

Table 4 Fairness after manipulation (Dist_fairness_T1)

Low Partial 7,03 ,30 14

Full 7,60 ,36 18

No 6,20 ,70 11

Total 7,06 ,72 43

(25)

25

Full 7,11 1,09 17

No 7,36 ,97 16

Total 7,30 ,92 46

Pay distribution Reduction Mean Std. Deviation N

High + Low Partial 7,24 0,51 27

Full 7,36 0,83 35

No 6,89 1,03 27

Total 7,18 0,84 89

I’ve included in the model a covariate legitimacy because legitimacy has a correlation with fairness perceptions before (Pearson correlation = ,72, p<0,05) and after the manipulation (Pearson correlation= ,49, p<0,05). Therefor legitimacy is included in the model. Other control variables are not included in the model because they do not impact the dependent variable (see for more information control variables section).

Table 5 Correlations

Correlations Dist_Fairness_T0 Dist_Fairness_T1 Age Legitimacy

Dist_fairness_T0 1 Dist_fairness_T1 ,528** 1 Age -,030 ,021 1 Legitimacy ,719** ,487** -,051 1 * p <,05 **p < ,01 *** p < ,001 5.3 Hypothesis testing

Hypothesis 1 and 2 were tested with a repeated measures ANCOVA to see if the different conditions have an effect on the fairness perceptions. The goal is to measure if the different conditions have influence on the change in fairness perceptions. Because the data of different conditions come from the same entities and are related I have to assume that the relationship between the pairs of experimental conditions are similar. The assumption of sphericity is therefore tested. A violation of this assumption can lead to an increased probability of a Type II error (Field, 2009). Mauchly’s test is nonsignificant and the assumption of sphericity is not violated.

The results from the repeated measures ANCOVA (see appendix) show the position in the pay distribution (distribution*fairness) has a significant impact (F(1,82) = 100,564 and p <0,05) on the fairness perception change variable. Partial eta squared is ,551 which means that 55,1% of the fairness perception can be explained by the position in the pay distribution. When looked at the

(26)

26 5.5 6 6.5 7 7.5 8 D I S T _ F A I R N E S S _ T 0 D I S T _ F A I R N E S S _ T 1

LOW IN

DISTRIBUTION

Partial Full No

manipulations we see that the manipulation (reduction*fairness) has a significant impact (F(2,82) = 11,937 , p <0,05) on the fairness perception variable. As can be seen from table 3 and 4, the distributive fairness increases from 0.55 point for partial, 0.91 for full and 0.19 for no reduction (for both distributions), and results indicate this change is different across groups: the more reduction, the more positive change in fairness. Thus hypothesis 1 is confirmed: a reduced pay dispersion will lead to an increase in fairness perception . The partial eta squared is 0,225. When we combine the position in the distribution and reduction manipulations

(distribution*reduction*fairness) in a new model with legitimacy as covariate we see a significant result (F = 14,656, p < 0,05) with a partial eta squared of ,263. Thus hypothesis 2 is confirmed: the distribution has a moderating effect of the reduction on fairness perception. The output of the ANCOVA (see appendix) shows that the position in pay distribution whether high or low has an effect on the dependent variable fairness perception. Also it shows that the manipulation of the pay dispersion measure has an effect on the fairness perceptions. To see what way this effect goes post-hoc analysis is done to see the specific results which are discussed in the plots.

5.4 Effect on high pay distribution employees

The estimated marginal means model plot shows that when participants were placed in the high end of the pay distribution fairness perceptions changed little. When higher paid employees get faced with a manipulation where the pay dispersion is reduced this has negative effect on their fairness perceptions which gives rise for support of H2a, but the effect on fairness perception is minimal. In each manipulation the effect is minimal where in the partial reduction the fairness perception change from 7,24 to 7,10, in the full reduction manipulation the variable decreased from 6,88 to 6,71 and in the no reduction the fairness perception almost stayed the same from 7,21 to 7,15.

Figure 2 Fairness before and after manipulation

5.5 6 6.5 7 7.5 8 D I S T _ F A I R N E S S _ T 0D I S T _ F A I R N E S S _ T 1

HIGH IN

DISTRIBUTION

Partial Full No

(27)

27 -1 -0.5 0 0.5 1 1.5 2 2.5 N O R E D U C T I O N P A R T I A L R E D U C T I O N F U L L R E D U C T I O N

CHANGE IN FAIRNESS

Low in distribution High in distribution

5.5 Effect on low pay distribution employees

The estimated marginal means model plot shows that when participants are low in the distribution the manipulations have a positive effect on the fairness perceptions. This finding gives support to H2b. There is a difference in fairness perception change when looking at the amount of reduction of pay dispersion. When pay dispersion is partially reduced this results in positive fairness perceptions rising from 6,15 to 7,25. The effect on fairness perceptions if even stronger when pay dispersion is fully eliminated this gives rise from 6,00 to 8,00. When looked at the no reduction manipulation we see that this has the weakest effect on increase in fairness perceptions (from 6,00 to 6,50.).

5.6 Interaction effects

Following plot shows the interaction effect of the change distributive fairness coefficient for the employees in the low distribution and in the high distribution. The plot shows that reduction of pay dispersion has minimal effect on the high-paid bonus employees. In other words,

participant’s ratings for fairness perceptions for high-paid bonus employees is unaffected by how much the pay dispersion was reduced. Fairness perceptions decreased slightly regardless the amount of reduction of pay dispersion. However, for employees low in the distribution the pay dispersion reduction has an impact. When dispersion is fully reduced to a leveled pay situation (compared to high) than the fairness perceptions change stronger than when pay dispersion is reduced partially.

Figure 3

(28)

28

5.7 Response to increase bonus

When looking at the fairness change in the no reduction condition I find an increase in fairness perceptions for the participants low in the distribution (+0,50.) but a slight decrease in fairness perceptions for participants in the high distribution (-0,02.). The increase in fairness for the low participants can be explained by the raise in bonus they get. Participants low in the pay

distribution got in this condition an equal bonus raise as the participants high in the pay distribution. Therefor the finding of increased fairness for participants in the low distribution cannot be solely attributed to the pay dispersion change, but rather a response to the bonus increase. However, the raise in this condition was 10 and in the partial reduction the raise was 5 and in full reduction the raise was 10. So when comparing the no condition group to the other conditions we can say that some of the portion of increased fairness is caused by the pay dispersion reduction (but not everything), because the increase in fairness perceptions for the partial and full reduction group is higher than the increase in fairness perceptions in the no reduction group. For extra insight the plot “combined groups” shows the difference in fairness perceptions change within the different conditions.

Figure 4

Combined groups fairness before and after manipulation

5.8 Robustness

The data consisted of two extreme outliers I’ve conducted additional analysis to check the robustness of the results by eliminating the observed outliers. Mauchly’s test of sphericity is still

6 6.2 6.4 6.6 6.8 7 7.2 7.4 D I S T _ F A I R N E S S _ T 0 D I S T _ F A I R N E S S _ T 1

COMBINED GROUPS

(29)

29 not significant. The results (see appendix) of the test show that pay distribution and the

manipulations are still significant, but legitimacy is not significant anymore(F(1,80) = 1.082, p=0,301). The model combining fairness with pay distribution position and the manipulations is still significant, so the same conclusions can be drawn from the hypotheses. Only the effect size of the interaction effect between distribution and reduction (partial eta squared) changes from 0,263 to 0,239 which means that when correcting for outliers the model explains less effect on the fairness perceptions than the previous model.

5.9 Additional analysis

5.9.1 Correlation gender

I’ve conducted additional analysis to see whether the control variables such as age, gender, education and bonus experience level have influence on the previously discussed items. I find that there is no significant correlation between age and the fairness measurements (before (r = -.03, p=0,78) and after manipulation (r = .02, p=0,844)).

5.9.2 Gender and fairness

Analysis of gender (conducted with a t-test) on the before fairness measure and the after fairness measure shows that there is no significant result (p=,42 and t=-0,82 for before fairness

measurement and p=,23 and t=1,21 for the after fairness measurement). To check if gender has an influence in the specific conditions a chi-square test was conducted and shows a value of 3,99 (p = ,55). This indicates that there is no significant dependence between the groups with

regarding the effect of gender on the fairness perceptions.

5.9.3 Age and fairness

Analysis whether the mean age within the different conditions is different show that this is not the case because the ANOVA is insignificant (F = 0,024, p = 1.00).

5.9.4 Education level and fairness

Education level was mostly the same over the population (mainly highly educated) and therefore not taken into account in the additional analysis.

5.9.5 Bonus experience and fairness

Bonus experience also did not show a significant result. T-test conducted for bonus experience resulted for Dist_Fairness_T0, t(87) = -1.93, p = .057,. Also Dist_Fairness_T1 shows no significant result, t(87) = -0,956, p =,34.

(30)

30

6 Discussion

The aim of this study was to investigate whether pay dispersion in a performance-based

horizontal system had an impact on the fairness perceptions. Further there was an aim to make a distinction between employees who face a high position in the pay setting and employees who face a low position in the pay setting. The findings make the suggestion that there is a significant effect of the pay dispersion on the fairness perceptions of employees. It does however makes a difference whether the employee is in the higher or lower end of the pay distribution. The results support the hypothesis that low in pay distribution employees have higher fairness perceptions when the firm reduces the pay dispersion. This finding is consistent with Downes and Choi (2014) finding in their literature study where they say that individuals low in the pay hierarchy will likely to view dispersed pay as inequitable. This can be derived from the initial fairness

perceptions that are measures which appear to be lower for the lower bonus employees than the fairness perceptions for the higher bonus employees. Trevor and Wazeter (2006) also point out that negative pay equity perceptions are particularly persuasive for participants with low pay standings in dispersed pay systems. From equity theory point of view we can explain the results because of a possible perceived inequity of the distributive outcomes of the initial pay dispersion. Lower employees experienced less fairness than high paid employees, probably because the perceived legitimacy towards the decision which caused the dispersion was lower. As such perceived legitimacy influences the fairness decisions of the participants this could be one of the reasons why lower level employees’ fairness perceptions were way higher after reducing the pay dispersion. It could be possible that the reducing of the pay dispersion (raising lower-paid

employees’ bonus) gave the participants a feeling of a more equitable situation. In the experiment the inputs each employee gives to earn the bonus was not specifically specified. Only the fact that there was a more competitive environment in the high-paid bonus employees’ region could mean that those have to work harder, but this does not necessarily have to be the case from the lower-level employee’s perspective.

6.1 Contribution to literature

This research makes contributions towards literature because, to my knowledge, no research was conducted with pay dispersion and fairness perceptions. It seems that fairness perceptions are influenced highly based on the distributive outcomes.

6.2 Limitations

Several limitations of this study should be acknowledged. First the methodology that is used makes use of a hypothetical situation. The results might not reflect results that would be drawn

(31)

31 from the true work field and therefor has concerns regarding the external validity of this research. Also the simplified operationalization of the bonus system in the experiment might not reflect real life situations bonus systems, which can be more complex.

The design of the experiment shows transparency among employees regarding their

compensation structure. In practice this could not be the case, where employees do not know how much their direct colleague in different regions earn. If not known the fairness perceptions could have shown different results.

Also the respondents of the study are mostly from a single sub-culture population, which might be skewed because of the education levels of most respondents (mostly high educated). Because the sample did not consist of lower educated participants the results might create an image which is not representative for lower educated participants.

Other important limitation is the number of respondents in the experiment. Some conditions consisted of 11 participants while other consisted of 18. Nevertheless, the statistic test shows us a significant relation of the effect of reducing pay dispersion on fairness an increase in respondents would make the results more robust.

6.3 Recommendations for further research

Regarding for further research I advise scholars to explore the effect of horizontal-performance based pay dispersion further in relation to other elements that impact organizational

performance. This research indicates that dispersion in a horizontal-performance based setting has a negative effect on the fairness perceptions of employees. However, fairness perceptions are one of the underlying constructs of employee motivation and effort. There is a need of more in depth research of the effects of this type of dispersion on the motivation and effort of

employees.

Also an interesting subject to investigate in the performance-based horizontal pay dispersion field is the effect of transparency on perceived fairness and motivation. In business’ where internal competition between employees (to motivate each other to perform better) is fostered (with for example monthly rankings) it would be probable that fairness perceptions and motivation differ than in companies where bonus rewards are more discrete.

Also as Shaw et al (2002) point out that pay dispersion effectiveness depends on the

contingencies of the organization. It would be interesting if scholars could provide more insight or make a distinction between for example public and private organizations. More public organizations try to introduce performance-based pay systems among their employees. It could

(32)

32 be possible that employees in these type of organizations respond differently to pay dispersion than for example profit organizations.

6.4 Recommendations for practice

This study shows that dispersion in a horizontal field can have negative effects on the fairness perceptions of employees. Firms should consider in order to create an effective system to increase the lower bonus paid employees and therefor compressing the performance-based pay system. This increase in fairness perceptions can lead to more motivation among the workforce.

7 Conclusion

This experiments studied a sample of 89 participants mainly students and young professionals in different working fields and their response of fairness perceptions towards changing pay

dispersion. A fictional study was conducted where participants filled out an online questionnaire where they first read a fictional case where they are a sales employee of a firm. In the case management decided to introduce pay dispersion between different sales team (before there was no dispersion). Hypotheses focused on whether reducing this initial pay dispersion had an effect on the fairness perceptions. Reason for testing fairness perceptions is that fairness perceptions are an underlying construct of employee motivation and therefor important regarding pay dispersion. Scholars were not in agreement whether dispersed or compressed pay is better for organizations. Participants were placed either in the high or low end of the pay distribution. The main finding is that reducing pay dispersion has a positive effect on the fairness perceptions of employees. To be sure that the increase of fairness perceptions was not caused by increase in bonus compensation an additional condition shows that indeed fairness partially increases because of increase in wage, but relatively to the other conditions this increase in fairness is lower. So therefore the change in fairness perceptions can be also explained as caused by the change in pay dispersion. The findings in this thesis make an argument for more compressed pay systems in organizations in order to foster more positive fairness perceptions. The effect of reducing pay dispersion has the biggest effect on employees who are in the lower level of the pay distribution. This can be explained with equity theory whereas the distributive outcomes are more equal after a reduction the pay dispersion. The study also shows that perceived legitimacy of initial pay dispersion decisions influence the fairness perceptions towards pay dispersion. This finding is consistent with Shaw and Gupta (2007) suggestion that the legitimacy of pay

differences may moderate the effects of pay dispersion. Further research is needed in the field of performance-based pay dispersion for example whether transparency of compensation (between

(33)

33 colleagues) and pay dispersion has an effect on the fairness perceptions. Also more detailed research is needed to measure employee motivation regarding performance-based pay dispersion.

(34)

34

References

Adams, J. S. (1963). Towards an understanding of inequity. The Journal of Abnormal and Social Psychology, 67(5), 422.

Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 267–299). New York: Academic Press.

Backes‐Gellner, U., & Pull, K. (2013). Tournament compensation systems, employee heterogeneity, and firm performance. Human Resource Management, 52(3), 375-398. Baumol, W. J., & Fischer, D. (1986). Superfairness: applications and theory. Mit Press.

Becker, B. E., & Huselid, M. A.. (1992). The Incentive Effects of Tournament Compensation Systems. Administrative Science Quarterly, 37(2), 336–350

Bloom, M. (1999). The performance effects of pay dispersion on individuals and organizations. Academy of Management Journal, 42(1), 25-40.

Bloom, M., & Michel, J. G. (2002). The relationships among organizational context, pay dispersion, and among managerial turnover. Academy of Management Journal, 45(1), 33-42. Bognanno, M. L. (2001). Corporate tournaments. Journal of Labor Economics, 19(2), 290-315. Bothner, M. S., Kang, J., & Stuart, T. E. 2007. Competitive crowding and risk taking in a tournament: Evidence from NASCAR racing. Administrative Science Quarterly, 52: 208-247. Brown, M. P., Sturman, M. C., & Simmering, M. J. (2003). Compensation policy and

organizational performance: The efficiency, operational, and financial implications of pay levels and pay structure. Academy of Management Journal, 46(6), 752-762.

Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A meta-analysis. Organizational behavior and human decision processes, 86(2), 278-321.

Colquitt, J. A., D. E. Conlon, M. J. Wesson, C. O. Porter, and K. Y. Ng. 2001. Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology 86 (3): 425-45.

Colquitt, Jason A., and Jessica B. Rodell. "Measuring justice and fairness 8."The Oxford handbook of justice in the workplace 1 (2015): 187.

Connelly, B. L., Tihanyi, L., Crook, T. R., & Gangloff, K. A. (2014). Tournament theory thirty years of contests and competitions. Journal of Management, 40(1), 16-47.

Reduced performance-based horizontal pay dispersion : the effect on fairness perception