Sustainable employability : retaining survivors after downsizing

(1)

–

SUSTAINABLE

EMPLOYABILITY:

Retaining Survivors after Downsizing

Name of Author: Michèle Cox Student Number: 10615369 Date of Submission: 7th_{of July} Version: Final Version

MSc. Business Studies- Leadership and Management Institution: Universiteit van Amsterdam

Faculty of Economics and Business

(2)

(3)

Abstract

As downsizings are no longer an exception, it is essential for the human resources (HR) department to know how to deal with the long-term negative consequences. Especially interesting are survivors of a downsizing, employees who are important for getting the organization back on track. However survivors’ voluntary turnover these days is not uncommon. This thesis focuses on understanding the factors that influence the voluntary turnover intentions of survivors. By examining not only the direct effects of (1) perceived job security, (2) perceived job embeddedness, and (3) trust in management on voluntary turnover intentions, but also testing the mediated effect of (1) affective commitment, (2) continuance commitment, and (3) normative commitment on this relationship, insights to the downsizing literature were provided. Specifically this study found that perceived job embeddedness via normative commitment reduces voluntary turnover intentions of survivors of a downsizing. This result can serve as guidance to the HR department in the financial and insurance sector in dealing with the retention of survivors.

(4)

Table of Contents: Figures and Tables

Fig. 1 The Conceptual Model……….….20

Table 1 Cronbachs Alpha’s of subscales……….……….…….28

Table 2 Scale Means, SD’s, Intercorrelations and Reliabilities…..……….………….31

Table 3 Bootstrap Results for the Conditional Indirect Effects of Model 1…………..….……….……36

Fig. 2 Model 1 Regression Coefficients and Significance………..………36

Table 4 Bootstrap Results for the Conditional Indirect Effects of Model 2 ………..……37

Fig. 3 Model 2 Regression Coefficients and Significance……….37

Table 5 Bootstrap Result for the Conditional Indirect Effects of Model 3 ………..38

(6)

1. Introduction:

“Downsizing, the planned elimination of positions or jobs” (Cascio, 1993, p.95)

In the past years we have seen many examples of firms downsizing in the Dutch banking and insurance sector. The major reason for staff reduction is because of the recession. Due to declines in demand firms are seeking strategies to survive. One strategy is to cut costs. Another factor driving forced leavings is that advances in technology make certain employees, even departments redundant. Hopes are that an organization will see quick improvements in their performance after a downsizing event and will be (even more) competitive or able to survive and thrive again. Unfortunately, results are mixed about the effects of downsizing on organizational performance most likely because the success of a downsizing operation depends on the attitudes and behaviors of ‘survivors’ in the organization. ‘Survivors are those who remain employed subsequent to the downsizing’ (Spreitzer & Mishra, 2002, p.1). These survivors determine whether the organization will become a success or if it suffers majorly in the post-layoff phase (Kumar & Pranjal, 2009). It is therefore essential to understand the consequences of downsizing on the remaining employees and their impact on the performance of the organization in this later stage. This research is about downsizing. Unfortunately Human Resource Management (HRM) literature on how to counter the negative effects of downsizing on voluntary turnover is sparse. In order to introduce the research gap an overview of the current literature on downsizing will be presented below.

(7)

2. Literature Review:

Much of the literature on downsizing is written on either one of two subjects: (1) the causes for downsizing, and (2) the consequences of such downsizing on the persons involved, being either a victim or a survivor. As for the field of HRM the micro-effects of downsizing are of particular importance as the behaviors of the employees eventually determine the success of this operation. We can see that HRM literature is therefore focused mostly on the persons affected by this decision. For this reason many researches have looked at the influence of a downsizing on employees’ attitudes and behaviors as well as the subsequent consequences it has for a firm (Arshad & Sparrow, 2010; Datta, Guthrie, Basuil et al., 2010; Iverson & Zatzick, 2011; Maertz, Wiley, LeRouge et al., 2010). In order to understand workers’ reactions much of the downsizing literature contains the concepts of organizational justice and psychological contract to explain changes in behavior and attitudes following a downsizing

event (Arshad & Sparrow, 2010; Chaudhry, Wayne, & Schalk, 2009). Theories related to

these concepts, used separately or combined, aim to explain the effect of unfairness in the decision making process or by not living up to the expectations of a reciprocal exchange (psychological contract) during or after a negative event on changes in behavior, ultimately resulting in negative consequences for the firm (Chaudhry, Wayne, & Schalk, 2009; Maertz et al., 2010). The effect of a downsizing on important variables directly or indirectly affecting organizational performance is also of interest to research and HRM practitioners. Examples of variables affected by downsizing are: (1) organizational commitment, (2) job satisfaction, (3) job involvement, (4) organizational citizenship behavior (OCB) and as well its influence on (5) voluntary turnover (Arshad & Sparrow, 2010; Datta et al., 2010; Maertz et al., 2010; Trevor & Nyberg, 2008). Studies investigating the relationships between downsizing and the above mentioned outcomes found a negative relation, except for voluntary turnover, for which

(8)

a positive relation was found (Arshad &Sparrow, 2010; Datta et al. 2010; Dierendonck & Jacobs, 2012).

The micro-level processes influencing behaviors and attitudes as well as the consequences for the organization have been investigated quite extensively throughout the years. Datta et al. (2010) even wrote a review establishing an integrative framework of the causes and effects of a downsizing operation. Unfortunately there is not much literature on how to counter the negative effects of a downsizing. However, Campion, Guerrero & Posthuma (2011) conducted a study on HR practices that could benefit and protect both the employee and employer, identifying “8 major reasonable downsizing practices”. These practices are:

1. Identify the business need for downsizing.

2. Communicate with employees throughout the downsizing process. 3. Identify future work.

4. Determine criteria to evaluate employees against future work. 5. Establish fair evaluation procedures.

6. Analyze adverse impact.

7. Evaluate the process and outcomes.

8. Ensure an informed and independent Human Resources staff.

Although this research is somewhat in the direction on how to manage the downsizing process in such a way that both the employee and employer benefit, it mentions few to no HR practices that can counter survivor’s voluntary turnover in the post downsizing phase. Next to that the proposed practices have not been empirically tested, making the practical use of this research limited. More research is therefore needed in this area, as the voluntary turnover of survivors after downsizing and practices to counteract this phenomenon are of major interest to practitioners (Datta et al., 2010). The aim of this master thesis is to fill a gap in both research as well as practice by contributing to a fuller understanding of the aspects related to the retention of survivors after a downsizing.

(9)

This study’s primary objective is therefore to understand the options for financial and insurance organizations to maintain their desired staff in-house in the case of downsizing. As mentioned before, survivors are of great importance to get the organization back on track. Losing employees who have high performance and potential would incur high costs such as decreased productivity and loss of human capital (Iverson & Zatzick, 2011). This would in return lead to costs of recruitment and selection, and training and development of a new or current employee trying to fill the void. As Trevor & Nyberg (2008, p.259) discussed, “voluntary turnover leaves the leaner organization understaffed and thus hindered in its

efficiency”. In order to provide organizations in the financial and insurance sector with

knowledge regarding this subject the research objective in this study will be to understand what factors influence voluntary turnover intentions in the financial and insurance sector after downsizing. This is to prevent the loss of high performing employees and important human capital. The question studied in this master thesis is therefore:

“What factors influence voluntary turnover intentions of survivors in the financial and insurance sector after downsizing?”

Contributions:

This thesis will provide an overview of relevant factors that impact voluntary turnover intentions in the post layoff phase, by establishing a conceptual model based on the voluntary turnover literature. Several of these factors will be tested to see whether they are relevant in real day-to-day business. The study is expected to contribute to the knowledge of HRM on how to counter voluntary turnover in the post downsizing phase by understanding the factors influencing survivors’ voluntary turnover intentions. Comprehending and recognizing these factors will enable HRM to come up with strategies, practices and processes to prevent the voluntary leaving of essential human capital. In this way the HR department will be able to support the continuity of the work within the organization. Furthermore the master thesis will hopefully drive future research on downsizing and its consequences.

(10)

3. Theoretical Background

Downsizing is often used to describe an intended strategy of a firm to let go part of their staff to improve efficiency and/or the effectiveness of the organization. The use of this strategy leads to two groups in an organization: (1) victims who have to leave the organization, and (2) survivors who are allowed to stay in the organization.

In order to establish the theoretical framework for this study a short theoretical background is given. As voluntary turnover is a response to attitudes and behaviors following from a downsizing, it is important to understand how these attitudes and behaviors come about. For this matter we will take a look at the concepts of organizational justice and the psychological contract, which give a good insight into the reactions of survivors.

Organization Justice Theory

In the field of downsizing numerous studies show that downsizing is related to negative effects on survivors’ level of commitment, performance, job satisfaction and positively related to voluntary turnover (Datta et al., 2010; Parzefall 2012). One theory often encountered and discussed in the downsizing literature is the organization justice theory. This theory provides some universal explanations for the different behaviors, emotions and attitudes survivors manifest. Below the three different elements of organizational justice theory will be discussed.

Elements of Organization Justice Theory

Three elements of organizational justice are generally distinguished, namely: (1) distributive justice, (2) procedural justice and (3) interactional justice (Van Dierendonck & Jacobs, 2012). “Distributive justice is concerned with how people make decisions about the distribution of resources” (Sahdev, 2004). To be more explicit: Survivors as well as victims decide whether they find the outcomes of the downsizing fair. Sometimes survivors are not directly affected

(11)

by justice (mis)treatment by their organization, but the observation of the treatment of victims is likely to impact the survivor’s attitudes and behaviors’ (Dierendonck van & Jacobs, 2012, p.97). Distributive justice is thus also determined by the fairness of the outcome for colleagues. Since not only the distributive treatment is used to determine fairness, it is necessary to look at the two other aspects determining justice which are procedural justice and interactional justice.

In times of high uncertainty people attend to the procedures by which certain decisions are made to see if they are fair. For instance when implementing downsizing (Dierendonck, van & Jacobs, 2012). It is about the decision rules used to decide who gets laid off. If survivors feel favoritism is used rather than clearly defined and predictable decision rules, they will feel more threatened (Spreitzer & Mishra, 2002). Also if there is a possibility to express ‘voice’ and if procedures are consistent without any bias regarding people you can talk of procedural justice (Campion, Guerrero, & Posthuma, 2011). In the case the process of the layoff is unfair, having been treated fair distributive is not that important anymore. Therefore in downsizing operations the focus should be more on procedural justice.

The third element discussed in organization justice theory, interactional justice, represents how information is presented to the employees. This not only includes ‘how bad news is delivered to the employees but it also incorporates the explanations given for why certain employees got fired or not’ (Cox, 2012, p.6). The interpersonal treatment and communication by management in this aspect is very important. It should show “dignity, respect and politeness” (Spreitzer & Mishra, 2002).

The three elements of organizational justice are essential in uncertain situation for the process of employee sensemaking. Quite obvious downsizing can be characterized as having high uncertainty, thereby increasing the need to engage in sensemaking processes. Employees try

(12)

to determine the reasons for their own outcomes and in what way they were responsible for it themselves. This sensemaking process also determines whether they still trust their organization. By evaluating the distributive, procedural and interactional justice of a firm’s actions, they can make sense of the situation and determine whether they still feel committed to their job and firm (Dierendonck, van & Jacobs, 2012). Next to that combining the information they receive from this sensemaking process and perceiving justice makes employees more willing to accept change and stop them from engaging in disruptive behavior. This is especially important in the financial and insurance sector in which there is no room for disruptive and negative behavior.

Psychological Contract theory:

Psychological contract theory can be used as another perspective to explain the different attitudinal and behavioral reactions of survivors on downsizing. The concept of psychological contract can be seen as “the beliefs regarding the terms and conditions of a reciprocal exchange agreement between the employee and the employer” (Parzefall, 2012). Many studies have used this concept in explaining employee attitudes and behaviors in the downsizing literature (Chaudhry et al., 2009; Parzefall, 2012). Meeting or failing to meet the terms of this contract lead to various attitudinal and behavioral reactions. When an employer fails to meet these terms, this will result in psychological contract breach in the perception of the employee. According to Datta et al. (2010) survivors might react to a psychological contract breach with voluntary leaving the organization resulting in negative consequences for the firm. In order to understand the factors essential for retaining desired staff in-house, comprehending the influences of psychological contract breach and how to prevent or restore it are essential.

(13)

4. Factors Influencing Voluntary Turnover Intentions:

Below an overview and discussion is given of various factors and their impact on voluntary turnover intentions.

Perceived Job Security

The first and most important aspect - according to past research - influencing a person’s intention to stay is job security. Job security is a person’s assurance about the continuity of his or her job in the organization. “Job security affects trust and perceptions of the psychological contract with the organization” (McDevitt, Giapponi, & Houston, 2013, p.231). In the case of a downsizing perceptions of job security can diminish, due to feelings and perceptions of psychological contract violation. If employees feel their organization cannot assure job security anymore in the future, this can have negative consequences for the firm desiring the employees to stay. In the end it might result in intentions of employees to voluntary leave the firm.

Perceived Job Embeddedness

According to Burton et al. (2010) another factor explaining a lot of variance in voluntary turnover is job embeddedness. The authors state that downsizing can been seen as a push factor making a person think about leaving the organization, while on the other hand job embeddedness can be seen as a pull factor influencing a person to stay. Employees feel embedded in their job when they have good ‘links’ with their colleagues, when they feel they ‘fit’ with the organization, and when they have to ‘sacrifice’ a lot when deciding to leave the firm. Job embeddedness could in this way buffer the negative effects caused by a downsizing. Considering several studies that have already confirmed the predictive capacity of job embeddedness on voluntary turnover, this should be seen as an important factor influencing voluntary turnover intentions (Trevor & Nyberg, 2008; Burton et al., 2010)

(14)

Trust in Management

A third factor described in the literature having impact on voluntary turnover intentions is trust in management. For Spreitzer and Mishra (2002) trust in management is seen as a predictor of voluntary turnover. Trustworthiness is defined as “an individual belief that top management can be trusted, where trust is defined as a willingness to be vulnerable to others” (Spreitzer & Mishra, 2002, p.710). Employees will trust management when they feel management is competent, reliable, concerned about best interests and open and honest about the situation at hand. Employees will be less worried; more committed and less eager to voluntary leave the organization in the presence of employees’ trust in management.

Job Satisfaction

A fourth factor is job satisfaction. Job satisfaction can be seen as “a function of the perceived relationship between what one wants from one’s job and what one perceives it as offering or entailing” (Locke, 1969). When there is a match between what one desires and what one gets, this can result in job satisfaction. If however a negative gap exists, indicating not receiving what one desires, this will lead to lower satisfaction. Researching the downsizing literature reveals that reduced job satisfaction is often an outcome resulting from a layoff. This follows logically from the concept of psychological contract breach, which has often been related to lowered job satisfaction (Zhao, Wayne, Glibkowski, & Bravo, 2007 in Chaudhry, Wayne & Schalk, 2009). Survivor’s perception of their job is often more negative after a downsizing, due to the inability of the firm to fulfill the psychological contract. This leads to lower satisfaction with the job and thus higher intentions to voluntary leave.

(15)

Justice

Justice is of high importance in the context of downsizing. Not only do fair decision rules need to be established and implemented about whom to layoff, also there needs to be distributive and interactional justice. For an organization to be continued to be seen as trustworthy and credible, clear communication, fair outcomes and procedures must be established (Spreitzer & Mishra, 2002). One cannot rely on favoritism or other random rules to decide which person has to leave. Also communication plays a major part in reducing perceptions of unfairness by providing explanations for the downsizing and informing employees about the course of action to follow. Helpful, open, accurate, complete and timely provision of information of “the intent to downsize, the timeframe, how termination decisions will be made, general information about severance packages, and other factors which are important to employees” can reduce the negative reactions to the reorganization (Campion, Guerrero, & Posthuma, 2011, p.175). In this way workers are more likely to accept the change and stay committed to the organization, reducing the chance of voluntary turnovers in the post layoff phase.

Coping Mechanisms

Not only how an organization manages its downsizing is important, but taking care of survivors proves essential. A downsizing is a very threatening situation for survivors. Therefore survivors are in need of some organizational support in order to cope with the uncertainty of the situation. By providing coping mechanisms such as training, giving opportunities to voice opinions and empowering employees, this support can be granted (Campion, Guerrero, & Posthuma, 2011; Maltarich, Nyberg & Reilly, 2010; Spreitzer & Mishra, 2002). Coping mechanisms can in that way be seen as manner to repair the psychological contract.

(16)

After a downsizing has taken place, some of the work previously done by the victims has to be transferred (partially) to the people remaining in the organization. A mismatch between the current cognitive abilities of the survivor and the demands of the current job could be the consequence. Cognitive ability is described as the ‘the general and stable ability to understand, respond to, and learn from the environment’ (Maltarich, Nyberg & Reilly, 2010, p. 1058). For employees willing to stay, it is important that there is a fit between their abilities and the job demands. If there is a mismatch this can lead to the voluntary turnover of survivors not able to cope with these demands. One way to have workers cope with this situation is by providing training. Training can provide and improve the necessary cognitive abilities to handle the new job demands, thereby increasing the chance to retain remaining employees after a downsizing.

Another coping mechanism is empowerment. Empowerment is “a personal sense of control in the workplace as manifested in four beliefs about the person-work relationship: meaning, competence, self-determination, and impact” (Spreitzer & Mishra, 2002, p.714). Empowerment can be seen as a job resource or coping mechanism for survivors to deal with the possible strain encountered after a downsizing, due to increases in job demands or uncertainty. Instead of having to sit, watch and follow orders, employees - by being empowered - can influence the situation at hand and determine themselves when and how to solve the demand (Bakker & Demerouti, 2007). They feel more competent, more determined and in control of the situation, thereby at the same time increasing their internal motivation. Empowerment is therefore expected to reduce voluntary turnover intentions as it helps to reduce the stress of the downsizing and handle the new job demand.

A last important coping mechanism is voice. The ability to express voice during or after the downsizing process might provide a person with a way to change or reverse an unfair

(17)

decision. As mentioned by Datta et al. (2010, p.308) “to manage change allow for employee inputs into the design of changes by providing them with opportunities to express their views”. Partly related to procedural justice above, this coping mechanism can be linked to the fairness rule of ‘correctability’ and would thus ensure increased perceived procedural fairness, thereby enhancing survivor retention.

The three coping mechanisms discussed, provide ways to deal with the new job demands and serve as a resource to deal with the uncertain situations at hand. As Spreitzer and Mishra explain it survivors are then ‘more likely to respond actively, believing they can shape and influence the downsizing situation for the better, giving them a reason to want to stay’ (Spreitzer & Mishra, 2002, p.714). An organization giving this support is likely to receive lower voluntary turnover intentions in return.

In order to get a good understanding of which factors influence voluntary turnover intentions of survivors after a downsizing, the following section will look deeper into several of these factors. A conceptual model is developed for testing to see whether the factors studied really matter in influencing voluntary turnover intentions after a downsizing operation.

(18)

5. The Conceptual Model and Hypotheses

This study draws on the research of Spreitzer & Mishra (2002) in understanding the responses of survivors to a downsizing. According to their research in the case of downsizing employees react to this threatening event in different ways depending on their attachment to the organization. This attachment is influenced again by other factors such as trust in management, empowerment and justice. Unfortunately their research was not able to confirm the mediation effect of organizational attachment. This research will try to test whether organizational commitment - a more inclusive form of organizational attachment focused on three dimension of commitment - does mediate the relationship between variables of perceived job security, perceived job embeddedness and trust in management to voluntary turnover intentions. Below the different hypothesis for testing are presented including at the end a summary of these hypotheses in a conceptual model.

Organizational Commitment and Voluntary Turnover Intentions

Both Trevor & Nyberg (2008) and Dierendonck, van & Jacobs (2012) confirm that empirical work has identified organizational commitment as both a downsizing consequence and a voluntary turnover antecedent. As discussed in the section on organizational justice theory, survivors reassess their willingness to stay in an organization after having experienced the uncertain situation of a downsizing. Whether to remain or not depends on the employee’s commitment to the company. ‘Commitment can broadly be defined as a force that binds an individual to a course of action that is of relevance to a particular target’ (Meyer and Herscovitch, 2001 in Van Dierendonck & Jacobs, 2012, p.99). When looking at organizational commitment we can see from research that it is influenced by a downsizing event due to changes in attitudes and behaviors. Although organizational commitment is comprised out of three components, namely (1) affective commitment, (2) continuance

(19)

commitment, and (3) normative commitment, it is predominantly affective commitment used as a measurement for this construct that is discussed in the downsizing literature in relation to voluntary turnover (Fuller, Hester, Barnett & Relyea, 2006; Spreitzer & Mishra, 2002). However in this research the focus will be on all three components of organizational commitment. Survivors having high affective-, continuance-, and normative commitment are expected to have lower voluntary turnover intentions. The following hypotheses are therefore proposed:

H1a: There is a negative relationship between affective commitment and voluntary turnover

intentions.

H1b: There is a negative relationship between continuance commitment and voluntary

turnover intentions.

H1c: There is a negative relationship between normative commitment and voluntary

turnover intentions.

Perceived Job Security:

Perceived job security once a variable that was viewed as a success factor in the financial and insurance firms for the retention of employees. Nowadays one could question whether job security is still that important as a factor for keeping employees from leaving. For this variable it is important to take into account the context in which this study is done, i.e. the financial and insurance sector. This sector has been seen as one providing high job security and therefore leading to high organizational commitment. However, nowadays this sector shows much reorganization which off course affects perceived job security. One could argue that in order to be committed to an organization the psychological contract needs to be intact. One would expect this to apply to all three components of organizational commitment. If the organization fails to meet this reciprocal exchange by not being able to provide job security

(20)

anymore, there might be a perception of contract violation resulting in less commitment on all components. In return this might cause retaliatory behaviors such as intentions to voluntary leaving the organization.

On the other hand, many changes in the financial and insurance sector took place. Rajan (1997) described it as a change from a paternalistic culture in which employees received lifetime employment to a performance culture in which employment was not guaranteed anymore. Also other trends, not limited to the financial and insurance sector, such as organizations solely providing short-term contracts are observed. Employees entering an organization are therefore likely to go into this exchange relationship with a different attitude than employees years before. Considering these trends, one could question whether employees and also survivors of a downsizing are still that affected by the factor job security when assessing their commitment towards their organization. It might be that they do not see a downsizing as a major violation of the psychological contract anymore, as they already took this into account when starting to work for the company. However as Iverson (1996) and Darwish (1997) already found support in previous studies for a positive relationship between job security and organizational commitment, this study will weigh the positive aspect of job security more heavily, resulting in the following hypotheses:

H2a: There is a positive relationship between perceived job security and affective

commitment.

H2b: There is a positive relationship between perceived job security and continuance

commitment.

H2c: There is a positive relationship between perceived job security and normative

commitment.

(21)

A job can be done everywhere, but what makes a difference are the colleagues one works with, how one as a person fits within the company and what one would have to sacrifice when leaving the organization not limited to the organization itself (Burton, Holtom, Sablynski, Mitchell, & Lee, 2010). As Allen and Meyer (1990) describe: antecedents of the three components of organizational commitment, having peer cohesion and having made organization-relevant as well as community-relevant investments are related to the parts on which job embeddedness is judged (links, fit and sacrifice). The more embedded you are - not only on the job as well as off the job - the more commitment is expected to the organization. Reasons might be more emotional (affective) commitment to the organization as well as to the community, due to many links to colleagues and friends outside work. Also a good fit in the organization and community or because of continuance commitment relating to the sacrifices an employee has to make when leaving both on and off the job can make people more embedded. According to Burton et al. (2010, p.44) “the degree to which one is embedded in a community or organization is likely to influence or provide part of the context in which one interprets a shock”. When considering a downsizing, survivors who are highly embedded in their job and community are very likely to be very involved in these places. Relating to a downsizing and thinking simultaneously of the escalation of commitment literature, when being very involved in a situation people get even more involved and committed to the situation and are thus less influenced by a downsizing (Meyer & Allen, 1991). A logical outcome for a survivor being highly embedded would be higher commitment to the organization. Not only would they have higher affective commitment, but also higher continuance and normative commitment would be expected. Resulting from this reasoning above the following hypotheses are presented on the next page:

(22)

H3a: There is a positive relationship between perceived job embeddedness and affective

commitment.

H3b: There is a positive relationship between perceived job embeddedness and continuance

commitment.

H3c: There is a positive relationship between perceived job embeddedness and normative

commitment.

A fourth factor, trust in management, which has already been studied by Spreitzer & Mishra (2002) was found to be related to organizational attachment. In times of uncertainty trust is essential. This factor is described by the authors as “an individual belief that top management can be trusted, where trust is defined as a willingness to be vulnerable to others” (Spreitzer & Mishra, 2002, p.710). Trust helps survivors to understand why certain things happened and what is expected from them. It helps them to cope with the threat inherent in a downsizing. For employees to trust management, they have to be competent, open, concerned and reliable (Mishra,1996). When looking how trust in management is related to organizational commitment, Spreitzer & Mishra (2002) only considered affective commitment as a measure for organizational attachment to which trust in management was related. This study also considers the other two factors of commitment namely continuance commitment and normative commitment. Continuance commitment is mainly focused on the costs of leaving the organization. Employees stay committed to a course of action when they have made certain investments they forego when leaving (Meyer & Allen, 1991). When survivors still trust management in their ability to ensure the made investments, this is expected to lead to more continuance commitment. Also, when a survivor still feels management is reliable, concerned with their interest and open, this might reduce their perceptions of threat and harm

(23)

(Ozag, 2006). As a result they might feel the obligation to stay and do something in return as management is fighting for the continuity of their organization. Therefore trust in management is also expected to lead to higher normative commitment. For the sake of completeness also the relationship of trust in management to affective commitment is tested. In this matter the following hypotheses are proposed:

H4a: There is a positive relationship between trust in management and affective

commitment.

H4b: There is a positive relationship between trust in management and continuance

commitment.

H4c: There is a positive relationship between trust in management and normative

commitment.

The Mediated Relationship:

From the literature review and the different hypothesized relationships above we can see that the three variables of (1) perceived job security, (2) perceived job embeddedness, and (3) trust in management are shown to act as antecedents of organizational commitment. In return these three components of organizational commitment are proven to be related to voluntary turnover intentions as found in previously done studies. As no direct relationships are found between the (1) perceived job security, (2) perceived job embeddedness, and (3) trust in management and voluntary turnover intentions, mediation might be the case. Logically follows from this that the relationship between the three variables mentioned above and voluntary turnover intentions is expected to be mediated by organizational commitment consisting of the three components of (1) affective commitment, (2) continuance commitment and (3) normative commitment. Employees will after the downsizing try to make sense of the situation by reassessing their commitment to the organization based on the three independent

(24)

variables. By looking at these three factors, they will form a new opinion about their commitment to the organization. This will in return determine the survivor’s choice between staying or leaving the organization, thus resulting in either high or low voluntary turnover intentions. This leads to the final hypothesis presented below:

H5a: The relationship between perceived job security and voluntary turnover intentions will

be mediated by (1) affective commitment, (2) continuance commitment, and (3) normative commitment.

H5b: The relationship between perceived job embeddedness and voluntary turnover

intentions will be mediated by (1) affective commitment, (2) continuance commitment, and (3) normative commitment.

H5c: The relationship between trust in management and voluntary turnover intentions will be

mediated by (1) affective commitment, (2) continuance commitment, and (3) normative commitment.

On the next page the conceptual model with the proposed hypothesis for testing is presented.

(25)

Fig. 1 The Conceptual Model

Note. The mediated hypotheses are not illustrated in this model

(26)

6. Method:

Sample

The survey was conducted in organizations having gone through or being in a downsizing operating in mainly the financial and insurance sector. The population targeted for this study was survivors of a downsizing. The scope of this study allowed responses from multiple sectors; however most responses came from the financial and insurance sector. About 212 employees (the number of people reached via social media is unknown) were contacted to fill out the survey. Response rates, looking at previous studies, varied a lot. The response rates identified ranged from 28 % to 53%, indicating that this research would probably receive a response rate somewhere in between this range (Arshad & Sparrow, 2010; Spreitzer & Mishra, 2002; Trevor & Nyberg, 2008). The actual response rate for this study was 34.7% based on 75 actual responses and 216 send out questionnaires.

The final sample size eventually consisted of n=42 respondents, which is quite small. Looking at the descriptives of the sample we can see that all respondents were employed in The Netherlands of which 66.7% was working in the financial and insurance sector. Most respondents had full-time contracts (57.1%), several had part-time contracts (11.9%) and only one respondent stated that he/she had another type of contract. Reviewing some more demographics the final sample consisted of 16 (38.1%) male employees and 14 (33.3%) female employees. The distribution of males and females of the partial filled in surveys (12) was unfortunately unknown. The average age of the employees in the sample was 43.5 years old. Looking at the years that people worked at their current organizations there was a wide variety ranging from only 0.5 years to 38 years. The average tenure in years however was 14 years.

(27)

Procedure

A cross-sectional research design was used for this master thesis. The research was conducted in organizations having gone through downsizing and operating mainly in the financial and insurance sector. Many firm were contacted using both e-mails and phone calls with a request to take part in the research (see Appendix 1 for the Dutch request e-mail). In all cases permission was asked from HR managers to let their employees take part in the research. Data was collected over 1.5 months using a web-based survey. The hyperlink for the online survey was forwarded to the HR managers of the participating organizations. They were responsible for forwarding the link to the employees. Furthermore communication to the employees, such as sending a reminder, went via the HR managers.

Participants were invited via an e-mail to participate in the research. The initial contact message included a short introduction to the study’s purpose, what information was going to be collected, how they could participate, the value of participating for both parties and the link to the survey. Respondents were asked a control question in order to determine whether they met the requirements of the population to be researched. Participation was voluntary and the anonymity and confidentiality for participating firms and the employees were assured. Data gathered would only be used for research purposes.

Survey

An online survey (see Appendix 2 for the survey) was created using the online survey software & insight platform ‘Qualtrics’ (Qualtrics, n.d.). The survey was administered in English, using existing scales, to ensure the reliability of the data. The questionnaire consisted mainly of closed questions with predetermined answers in order to make comparing more easily across respondents. Each respondent was asked to answer the same set of questions.

(28)

The aim was to gain an understanding of reasons for voluntary turnover intentions after a downsizing, by grasping the different opinions of survivors. For the sake of gathering more data the link of the online questionnaire was also distributed via social media (LinkedIn and Facebook). This was done because the participation of organizations was scarce. However this did not result in an increase of extra responses.

Measures

All construct items came from existing studies in English and were already tested for their reliability. Below these scales are discussed and their related reliabilities for this study are mentioned.

Perceived Job Security: (10 items; α=0.85). Perceived job security was assessed by means of

10 items from the job security scale used by Kraimer, Wayne, Liden & Sparrowe (2005) (10 items; α=0.90). Pearce in Kraimer et al. (2005) define perceived job security as “a psychological state in which workers vary in their expectations of future continuity within an organization”. An example item of the job security scale used is: “I am secure in my job” (+). Items were rated using a 7- point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree). The 10 items measured were averaged to get a single score for perceived job security.

Perceived Job Embeddedness: (18 items; α=0.85). Perceived job embeddedness was

measured by means of the short version of the job embeddedness scale by Holtom, Mitchel, Lee, & Tidd in Felps, Mitchell, Hekman, Lee, Holtom, & Harman (2009) (21 items; α=0.82). Also they provided evidence of predictive validity for the short measurement scale. Job embeddedness is a construct that ‘includes several individual-level factors that enmesh employees in their job’ such as ‘how well people fit in their jobs and community; the interpersonal links they have on and off the job; and what they would have to give up or

(29)

sacrifice in leaving their place of employment or community’ (Felps et al., 2009; Mitchell, Holtom, Lee, Sablynski, & Erez, 2001). 18 items of the job embeddedness scale were required to be answered on a 1 (strongly disagree) to 5 (strongly agree) point scale. The last 3 questions were yes or no answers. After having done a reliability analysis of the scale this study continued only with the first 18 items of the scale, as removing the last three resulted in an improvement of the reliability. (18 items; α=0.85). To develop an individual-level job embeddedness score, the individual’s scores were averaged for each item (Felps, et al., 2009). An example item of the job embeddedness scale used was: “My job utilizes my skills and talents well”.

Trust in Management: (12 items; α=0.95). To measure trust in management trustworthiness

is used, described as “an individual belief that top management can be trusted, where trust is defined as a willingness to be vulnerable to others” (Spreitzer & Mishra, 2002, p.710). The items used to assess trust in management represent the four components of trustworthiness: (1) openness, (2) competence, (3) concern/caring, and (4) reliability. Mishra & Mishra’s (1994) 16 item measurement scale for trustworthiness was used. Items were adapted individually to fit the context of the research; also the size of the scale was reduced. 3 items per component were chosen. The 12 items measured were averaged into a single score for trustworthiness, which was shown appropriate in Mishra & Mishra (1994) giving an α= 0.96. This original reliability does not differ significantly from the reliability found in this study (α=0.95). An example of the trustworthiness scale is: “I believe that management tells the truth”. A 7-point Likert scale is used in which 1 is ‘strongly disagree’ and 7 is ‘strongly agree’.

(30)

Organizational Commitment: Organizational commitment is defined as: “The relative

strength of an individual’s identification with and involvement in a particular organization” (Porter & Smith, note 4 in Mowday, Steers, & Porter, 1979, p226). According to Allen & Meyer (1990) organizational commitment is a construct consisting out of three components: 1) affective commitment, 2) continuance commitment and, 3) normative commitment. For each of the components separate measurements were used. All three used a measurement scale of 7 points consisting of 1 (strongly disagree) to 7 (strongly agree).

Affective commitment is defined by Allen & Meyer (1990) as “an affective or emotional attachment to the organization such that the strongly committed individual identifies with, is involved in, and enjoys membership in, the organization” (8 items; α=0.79). This reliability is slightly lower than the α found in the research of Allen & Meyer (1990) (8 items; α=0.87). An example of an item is: “I would be very happy to spend the rest of my career with this organization”.

Continuance Commitment reflects “the recognition of costs associated with leaving the organization” and therefore the need to stay (Meyer & Allen, 1991). (8 items; α=0.74) As can be seen this original α (8 items; α=0.75) found almost equals the reliability identified for this study. One example of an item used in the continuance commitment scale is: “It would be very hard for me to leave my organization right now, even if I wanted to”.

The third component, normative commitment is described as “the feeling of obligation to remain with an organization”, so an employee feels he ought to stay with the organization (Meyer & Allen, 1991) (8 items; α=0.68). Unfortunately this reliability differs somewhat from the α found for the study of Meyer & Allen (1991) (8 items; α=0.79). An example for the measurement of this component is: “I think that people these days move from company to company too often”.

(31)

Although the scores for the three separate components are mostly averaged to get one organizational commitment score, this research asks for the separate use of each component. Scores will therefore be averaged for each separate component to get one score for each component.

Voluntary Turnover Intentions: Voluntary turnover intentions were defined for this thesis

as: the decision to stay or quit based on an analysis of the utility of one’s current employment relative to the expected utility of alternative employment (Trevor & Nyberg, 2008). In order to measure voluntary turnover intentions, a 4 item turnover intention scale was used from

Arshad & Sparrow (2010). The scale was tested and proved valid showing an α=0.91 which is

slightly higher than the original α=0.89 found in the study of Trevor & Nyberg (2008). An example of an item used is: ‘In the last few months, I have seriously thought about looking for a new job’. Respondents were asked to indicate on a 5-point Likert scale whether they were ‘1= very unlikely’ to ‘5= very likely’ to agree with the statements. The scores for the items were averaged to provide a single score for the dependent variable voluntary turnover intention.

Control variables: Several control variables were accounted for in this research, such as

gender, age, sector, tenure and part-time versus full-time status. These variables were controlled in the analyses as they might give alternative explanations for the relationships tested. They were kept constant during the analyses in order to establish right conclusions regarding the effects of the independent variables on the dependent one.

(32)

Statistical Analysis

In order to examine and analyze the data the Statistical Software Package for Social Sciences (SPSS) (IMB Corp., 2011) was used. First the original raw dataset (n=98) was cleaned from those employees who did not perceive the first screening question. These participants were not able to participate from the beginning and were of no use for the research. Secondly the dataset was screened for employees that did not fulfill the requirements of having experienced and been through a downsizing at their current firm and were allowed to stay at their current organization. By having indicated ‘no’ to this screening question, they were stopped from participating in the research. These cases were removed leaving the dataset with a remaining number of participants of n= 49 having fulfilled the requirements. However the cases left still consisted of some employees only having answered the screen question, but not having continued the rest of the survey. These cases, which were 7, were also removed leaving us with a usable sample of 42 employees. From these 42 employees, 30 had finished the whole survey and 12 had only partially completed the survey. As there is only a small workable sample the partially completed surveys were also used for analyses, providing extra data. After having completed the deletion of not usable cases, the data was checked for errors and missing values. Luckily no errors were found.

As people were allowed an escape option in answering the questions, i.e. filling in ‘n.a’, these cases could not be used for analyses. The ‘n.a.’ values were recoded into 999,00 and were treated as missing values. The percentage of missing values per item was checked using

Frequencies in SPSS. The percentages differed between 2,4% and 45.2% including both user

missing values and system missing values. It was decided that pairwise deletion would be used for dealing with missing values. Continuing the preparation of the dataset for analysis, counter-indicative items were recoded allowing for the right interpretation of the values.

(33)

7. Results

As mentioned in the section on ‘measures’ scale reliabilities were calculated for each of the 7 variables (see Table 1 for a summary). Most indicated good to high reliability (α>0.7) and corrected item-total correlations (>0.3) signaling good internal consistency. This was expected as the scales were taken from previous studies that had already tested and validated these scales. However ‘normative commitment’ showed a somewhat lower reliability of α=0.68. Removing any of the items of this scale would not result in a significant improvement. Although the reliability of the scale was somewhat low, the variable has been used for hypothesis testing. When looking at ‘perceived job embeddedness’ a good reliability was identified. Nonetheless removing the last three items from this scale improved it to α=0.85. This resulted in the decision to remove these three items and to continue with a smaller number of items which made it an easier to use construct.

Table 1. Cronbachs Alpha's of subscales

Cronbachs Alpha

1.Perceived Job Security

α=0.85 2.Perceived Job Embeddedness

α=0.83 α=0.85* 3.Trust in Management α=0.91 4.Affective Commitment α=0.79 5.Continuance Commitment α=0.74 6.Normative Commitment α=0.68 7.Voluntary Turnover Intentions

α=0.91

Note. * This represents the α after removal of PercJobEmb19, 20, and 21

Next preliminary analyses were performed to check for violation of the assumption of normality. To test for normality skewness and kurtosis values were computed. Also histograms, q-q plots and the Kolmogorov-Smirnov test were used. Many items from each variable construct showed deviation from normality. Transformations were performed to these

(34)

deviating items in order to approach normality. All items after the transformation eventually approached normality.

After the preliminary analyses, scale means were calculated for each variable. Some variables had means on the lower end of the scales, such as perceived job security, perceived job embeddedness and voluntary turnover intentions (M=3.57, M=2.56, and M=1.29 respectively). This indicates that employees disagreed to somewhat disagreed with statements made for measuring perceived job security and perceived job embeddedness. These survivors did not feel that their organization was providing a high level of job security, nor did they feel highly embedded in their jobs.

Thinking about voluntary turnover intentions it makes sense that people scored on the lower end of the scale. This lower score indicated that these survivors did not have the intention to leave the organization in the short-term or were considering it. However it is off course questionable whether employees would indeed answer positively to such as statement as - no matter the promise of anonymity – the HR department is still the contact person for communication.

What also can be seen from looking at the mean values is the slightly higher mean value for trust in management (i.e. M=4.6). The reason for this higher value might be that survivors are already glad they survived. They will probably not state that the decision to keep them was the wrong one. They therefore maintain their trust in management.

Furthermore when looking at the mean values for affective commitment, continuance commitment, and normative commitment we see scores that can be rationally understood. Affective commitment out of three is scored lowest (M=2.29). The negative event of a downsizing is clearly impacting how affectively survivors are attached to their organization. Second in place is normative commitment (M=3.19). Survivors in this study show, probably

(35)

due to psychological contract breach, less normative commitment (i.e. feel they ought to stay in the organization). Lastly people show however a moderate need to stay due to continuance commitment (M=4.06).

Correlations

Having calculated scale means, Bivariate Correlations were calculated for both the model’s variables and the five control variables (gender, age, sector, tenure, and type of contract). The correlation matrix is shown in Table 2 containing the means, the standard deviations, and correlations for the variables in the analyses. The relationships between the variables were investigated using the Pearson product-moment correlation coefficient.

Table 2 provides initial support for the hypotheses tested. Studying our mediator variables Affective Commitment, Continuance Commitment and Normative Commitment and their relationship with Voluntary Turnover Intentions’ only 1 significant correlation could be identified. Employees having higher normative commitment showed fewer intentions to voluntary leave the organization, r=-0.459, p<0.01. Hereby initial support was found for H1c.

The relationship between perceived job security and affective commitment shows a significant positive correlation, r=0.34, p<0.05, indicating support for H2a. Higher levels of perceived job security result in higher levels of affective commitment after a downsizing. However no significant correlations were found between Perceived Job Security and Continuance- and Normative commitment.

Support was also found for H3a and H3c. Significant positive correlations were identified between perceived job embeddedness and Affective Commitment, r=0.429, p<0.01 and Normative Commitment, r=0.326, p<0.05. Employees having higher perceived job embeddedness tended to feel more committed in an affective and a normative way towards their company. However no significant correlation was found to support H3b.

(36)

Table 2. Scale Means, SD's, Intercorrelations and Reliabilities

Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 1.Gender 1,43 0,50 (-) 2.Age in Years 43,57 9,11 -0,31* (-) 3.Sector 6,68 1,19 -0,28 0,45** (-) 4.Tenure in Years 13,45 10,32 -0,07 0,59** 0,20 (-) 5.Contract 1,21 0,50 0,23 -0,14 0,13 -0,12 (-) 6.Perceived Job Security 3,55 1,11 0,06 0,16 -0,24 0,43** -0,23 (.85) 7.Perceived Job Embeddeness 2,60 0,36 -0,18 0,05 -0,19 0,13 0,14 0,18 (.85) 8.Trust in Management 4,54 0,85 0,30 0,03 0,03 0,13 0,28 0,13 0,58** (.95) 9.Affective Commitment 2,23 0,34 0,06 -0,04 -0,13 0,34 0,15 0,34* 0,43** 0,56** (.79) 10.Continuance Commitment 4,03 1,07 0,01 0,18 0,11 0,41* -0,22 0,01 -0,04 -0,19 -0,14 (.74) 11.Normative Commitment 3,17 0,82 -0,25 0,30 0,10 0,29 -0,05 0,20 0,33* 0,60** 0,24 -0,10 (.68) 12.Voluntary Turnover Intentions 1,23 0,74 0,33* -0,38* -0,38* -0,31* 0,03 -0,17 -0,21 -0,28 -0,29 -0,30 -0,46** (.91)

Note. * Correlations are significant at p<.05 (one-tailed) ** Correlations are significant at p<.01 (one-tailed)

(37)

Looking at correlations between ‘trust in management’ and Affective Commitment, Continuance Commitment, and Normative Commitment, positive significant correlations can be found for Trust and Afcom, r=0.564, p<0.01, and NorCom, r=0.595, p<0.01. This shows initial support for H4a and H4c. However no support was found for H4b.

The correlations between the independent variables and the outcome variable Voluntary Turnover Intentions showed no significant correlation. This might indicate initial support for the fact that there might be a mediation going on as no direct relationships can be identified between the independent variables and the dependent variable.

Lastly the correlation matrix shows there is no multicollinearity, as none of the independent variables are highly correlated (r=0.9 or above). The correlations are all well below the threshold of r=0.9, indicating there is no violation of this assumption for doing a regression analysis.

Hypotheses Testing

As mentioned above the hypotheses proposed in this study state that perceived job security, perceived job embeddedness, and trust in management and voluntary turnover intentions are mediated by three types of commitment (i.e. affective, continuance, and normative commitment). In order to test these mediation hypotheses, the procedures as proposed by Baron and Kenny (1986) were used. According to Baron and Kenny (1986) and (Kenny, 2014) four conditions have to be met in order to speak of a full mediation relation: 1) the independent variable is a significant predictor of the dependent variable, 2) the independent variable is a significant predictor of the mediator variable, 3) the mediator is a significant predictor of the dependent variable, and is controlled for the independent variable, and 4) the effect of the independent variable on the dependent variable controlling for the mediator should be zero. In the case step 4 is not met, Baron and Kenny (1986) speak of partial

(38)

mediation.

In order to do this the macro PROCESS (A. Hayes) in SPSS was used (Hayes, 2012). This allowed for an easy analysis of the mediation to be tested by only having to choose the right PROCESS model, in this case ‘PROCESS model 4”. This model 4 was run 3 times, as 3 separate mediation models can be identified, each time using one out of three independent variables (i.e. Perceived Job Security, Perceived Job Embeddedness, and Trust in Management). The dependent variable ‘Voluntary Turnover Intentions’ and the mediator variables (1) Affective Commitment, (2) Continuance Commitment and (3) Normative Commitment were filled in each run. Below the results of the mediation analysis for examining the roles of affective, continuance, and normative commitment as mediators are discussed.

Organizational Commitment and Voluntary Turnover

In order to see whether the hypotheses H1a, H1b, and H1c were supported at a 0.05 significance level, a look was taken at the three runs that were executed taking into account the procedure of Baron and Kenny (1986). Looking at the first run with perceived job security as the independent variable, H1b and H1c were found significant. More continuance and

normative commitment towards an organization thus results in lower intentions to voluntarily leave (β=-0.29 and β=-0.43 respectively) the organization. Studying the second run with perceived job embeddedness as the independent variable it can also be seen that 1and H1c

were found significant (β=-0.29 and β=-0.43). The third run with trust in management as the independent variable shows no significant relationships for H1a and H1c. However support

was found (p<0.05) for H1b (β=-0.27). Employees having more continuance commitment

seem to have lower intentions to quit. To conclude, all three runs show a negative significant relationship for continuance commitment and voluntary turnover intentions, giving full support to H1b. Partial support can be granted to H1c showing a significant relationship

(39)

twice. Yet H1a is not supported in all cases.

Perceived Job Security

Hypotheses H2a, H2b, and H2c indicate that employees after a downsizing who perceive higher job security will be more committed in an affective, continuance, and normative way towards their organization. From the outcomes provided by using the procedure of Baron and Kenny (1986) it can be clearly seen that perceived job security shows no significant positive relationships with any of the mediator variables affective commitment, continuance

commitment, and normative commitment (coefficients of respectively β=0.11; β=0.02; and β=0.17). No support can therefore be given to the hypotheses H2a, H2b, and H2c at the 0.05

significance level.

Hypotheses H3a, H3b, and H3c also state that employees with higher levels of perceived job embeddedness after a downsizing have more affective, continuance, and normative

commitment towards their firm. Looking at the results from testing it can be seen that the β

coefficients of H3a and H3c are indeed positive and significant (p<0.05) (β=0.39 and β=0.85

respectively). However H3b has a negative β=-0.11 and was not found significant. The

results indicate that employees having higher perceived job embeddedness experience higher

affective and normative commitment towards their organization. From this can be concluded

that employees who went through a downsizing and survived, and still have good links with their colleagues, continue to feel they fit in the organization and their community, might maintain affective/emotional attachment to their firm and also feel the obligation to stay.

Having trust in management after a downsizing is hypothesized to lead to higher affective,

continuance, and normative commitment of employees towards their organization. The

outcomes indeed indicate significance (p<0.05) for H4a and H4c (β=0.24 and β=0.57

(40)

respectively). Nonetheless no significant relationship was found for trust in management and

continuance commitment. These results illustrate us that employees having trust in management makes them feel more affective and normative committed towards their

organization after a downsizing. An interesting thing to mention is that although no

significant relationship was found for H4b, the β shows a negative relationship between trust

in management and continuance commitment.

The mediated relationship

In order to provide support for the three mediation hypotheses stated below, the first three conditions mentioned by Baron and Kenny (1986) and Kenny (2014) need at least to be fulfilled. Results from Hayes (2012) PROCESS should either indicate significance for only the Indirect Effect (i.e. full mediation) or for both the Indirect Effect as well as the Direct

Effect (i.e. partial mediation).

Hypothesis H5a

Looking at the results it can be directly seen that the first model fails to meet condition 1 of Kenny’s four conditions for mediation (2014). No significant direct effect is found for the first model with perceived job security as the independent variable. Figure 2, as well as the lack of support for hypotheses H2a to H2c, and H1a, illustrate that conditions 2 and 3 are not met either. No significant indirect effect was found at the 0.05 level (see Table 3). From this can be concluded there is no support for either partial or complete mediation, nor is there support for a direct relationship. This suggests that having higher perceived job security is in no way affecting voluntary turnover intentions after downsizing in a direct way, nor is this relation completely or partially mediated by ‘affective commitment, continuance commitment,

and normative commitment’. In return H5a receives no support.

(41)

Table 3. Bootstrap Results for the Conditional Indirect Effects of Model 1

Condition Indirect Effect SE Boot LL 95% CI Boot UL 95% CI

1. AfCom -0.03 0.06 -0.22 0.04

2. ConCom -0.01 0.07 -0.16 0.13

3. NorCom -0.08 0.08 -0.28 0.03

Note. CI = confidence interval; LL = lower limit; UL = upper limit.

Unstandardized regression coefficients are reported. Bootstrap sample size 1.000.

Fig. 2 Model 1 Regression Coefficients and significance

Note. * Significant at 0.05 level

Direct effects are not shown, are found non-significant

Hypothesis H5b

The same analysis was done for H5b. From the results can be identified that although a significant indirect effect (p<0.05) can be found for perceived job embeddedness and

normative commitment, no significant indirect effects were found for affective and continuance commitment (see Table 4). Also no significant direct effect was discovered. This

implies complete mediation for normative commitment (steps 1 to 4 of Kenny’s (2014) are met) and no mediation effects of affective and continuance commitment. Hypothesis 5b is therefore only partially supported. The findings (see Fig. 3) for the supported complete

(42)

mediation suggest that having higher perceived job embeddedness increases the level of

normative commitment, which in return lowers the intentions of employees voluntarily

leaving the organization after a downsizing.

1. AfCom -0.13 0.22 -0.82 0.17

2. ConCom 0.03 0.21 -0.35 0.54

3. NorCom -0.37 0.25 -1.11 -0.03

Fig. 3 Model 2 Regression Coefficients and Significance

Hypothesis H5c.

A last mediation analysis was done for H5c. Here neither significant indirect effects, nor a significant direct effect at a 0.05 significance level were found (see Fig. 4). Hypothesis 5c therefore receives no support at all. Employees having higher trust in management do

(43)

therefore not have lower intentions to quit (direct effect) nor does it influence their levels and is it mediated by ‘affective commitment, continuance commitment, and normative

commitment’ (indirect effect).

1. AfCom -0.07 0.13 -0.34 0.16

2. ConCom 0.06 0.08 -0.04 0.30

3. NorCom -0.22 0.15 -0.56 0.02

Fig. 4 Model 3 Regression Coefficients and Significance