The impact of job insecurity on job burnout among hospitality employees during COVID-19 pandemic: the moderating role of supervisor and co-worker support

Purpose – Thisstudyevaluatesthemoderatingroleofwork-basedsocialsupports(i.e.supervisorsupportand co-worker support) in the relationship between job insecurity and job burnout among hospitality employees in Malaysia. Besides, the direct effect between job insecurity and job burnout is examined. Design/methodology/approach – The cross-sectional data of this study were based on a total of 220 self-administered questionnaires that have been completed by hospitality employees from three different states in Malaysia.Respondentswererecruitedbasedonasnowballsamplingapproach.Thedatawerecollectedduring theCOVID-19pandemic,whichwasfromOctober2020toJanuary2021. Findings – Partial least square-structural equation modeling (PLS-SEM) was performed via SmartPLS software.Thefindingconfirmedthatjobinsecuritysignificantlyintensifiesemployees ’ jobburnout.Supervisor support and co-worker support were found to moderate the link between job insecurity and burnout. As anticipated,the relationshipbetween jobinsecurityandjob burnoutincreasedwhensupervisorsupportislow. But high co-worker support was found to strengthen the impact of job insecurity on job burnout instead of the reverse. Originality/value – This study supplements the existing literature by clarifying which sources of work- based social support (i.e. co-worker support or supervisor) is more salient in alleviating the adverse impact of job insecurity on job burnout during the COVID-19 pandemic among hospitality employees in Malaysia.


Introduction
The hospitality industry is an important revenue contributor to many countries, inclusive of Malaysia. Nonetheless, the outbreak of unprecedented, highly contagious and life-threatening new coronavirus, COVID-19, has severely affected the hospitality industry throughout the world (Karatepe, Saydam, & Okumus, 2021). Given the concern for public health, the The impact of job insecurity on job burnout evaluated the moderating effects of co-worker support and supervisor support between job insecurity and burnout.
Moreover, the assertion about the buffering effect of social support between stressors/job demands and psychological strain was not always consistent, which deserve further evaluation in different scenario (Cooper, Dewe, & O'Driscoll, 2001;Jolly, Kong, & Kim, 2021). The unprojected reverse buffering effects of social support were observed in past studies (Beehr, Bowling, & Bennett, 2010;Cooper et al., 2001;Kokoroko & Sanda, 2019). A recent study (Abbas, Malik, & Sarwat, 2021) showed that the negative impacts of job insecurity on several outcomes (i.e. self-esteem, mental health, life satisfaction and economic self-efficacy) were stronger when supports from family and friends were low among hospitality employees. Paradoxically, job insecurity leads to an increase in economic deprivation even though high social supports were available (Abbas et al., 2021). Their study, however, focused on the interaction between job insecurity and overall nonwork social support. As the perception of job insecurity among employees stems from the uncertainty about the continuity of their employment, supportive workplace relationships, such as supervisor support and co-worker support, are expected to be useful resources in ameliorating employees' burnout experience. Employees who obtain work-related assistance are likely to perform better in their job and enable goal attainment (Bakker & de Vries, 2021).
Nevertheless, Cooper et al. (2001) noted that the extent to which social support will buffer the impacts of stressors/demands-strain relationships are likely to vary across situations and individuals. People may place a higher value on a specific form of social support as a key resource that they would preserve and strive to obtain in different situations (Jolly et al., 2021). As such, there is a need to clarify which source of work-based social support (supervisor support and supervisor support) is more salient in mitigating job burnout caused by job insecurity during the pandemic, which remains unclear thus far.
To extend the knowledge in the existing hospitality literature, we aim to analyze the direct effect of job insecurity on job burnout, as well as the moderating effect of supervisor support and co-worker support on this relationship among hospitality employees in Malaysia. Accordingly, this research seeks to answer two research questions: RQ1. Does job insecurity substantially affect the level of job burnout among hospitality employees? and RQ2. Do supervisor support and co-worker support buffer the impact of job insecurity on job burnout?
The primary contribution of our study is we provide a more comprehensive view of the impact of job insecurity on job burnout by incorporating the moderating effect of supervisor support and co-worker support in this relationship. Specifically, this study was conducted during the COVID-19 pandemic; thus, the results of this study would provide a better understanding of the relationship among the variables in a such unique situation. Besides, this study also enables the examination of the level of intensity of job insecurity on job burnout among hospitality employees during the pandemic in Malaysia context, a developing country.

Theoretical background and hypothesis development 2.1 Underlying theories
The occurrence of job burnout can be explained through two major underlying theories: Job Demands-Resources (JD-R) model and COR theory. JD-R model (Demerouti, Bakker, Nachreiner, Schaufeli, & Wilmar, 2001;Schaufeli & Taris, 2014) explains that job demands erode a person's physical and mental resources that lead to health impairment process, The impact of job insecurity on job burnout thereby eliciting adverse responses, such as burnout that can result in undesired performance. Both job demands and job resources encompass psychological, physical, organizational or social aspects of the job. However, job resources are essential in fulfilling employees' psychological needs, and they are useful in attaining goals at work, fostering personal growth and development, as well as reduce psychological and physiological costs that cause by job demands (Demerouti et al., 2001;Bakker & de Vries, 2021). Accordingly, job burnout is frequently experienced by individuals who encountered high job demands (e.g. job insecurity, role conflict, workload, time pressure and hostile working environment) and poor resources. Job demands required sustained mental, emotional and physical efforts that can exhaust employees. On the contrary, individuals who are equipped with job resources (e.g. autonomy, social support and feedback) are more engaged in their work and demonstrate less tendency of being burnout (Schaufeli, 2017). Furthermore, the preceding study also showed that resources (job resources or/and personal resources) ease in lowering the negative implications of job demands on burnout (Bakker & de Vries, 2021;Usman et al., 2021).
Besides, the theoretical foundation on the salient role of resources in affecting the level of strain experienced by a person also can be found in COR theory (Hobfoll, 2002). From the COR perspective, "individuals strive to obtain, retain, foster and protect those things they centrally value" (Hobfoll, Halbesleben, Neveu, & Westman, 2018, p. 104). Four types of resources are described in COR theory (Hobfoll, 2002;Hobfoll et al., 2018): objects (e.g. tools, home and car), conditions (e.g. autonomy and social supports), personal characteristics (e.g. resilience and self-efficacy) and energies (e.g. knowledge and time). COR theory connoted that stress occurs when (1) there is a threat of the loss of key resources, (2) face the actual loss of key resources or (3) fail to obtain the valued resources after substantial efforts or investments (Hobfoll et al., 2018).

Job burnout
Burnout, which is the result of prolonged chronic stress (Maslach, 2003), has turned out to be a global phenomenon that affects employees across occupations. The scientific inquiries in this subject matter remain topical in response to various socioeconomic challenges faced by the organization, inclusive of the recent public health crisis that post tremendous pressure on hospitality operators and employees. Job burnout, based on Maslach, Schaufeli, and Leiter (2001)'s conceptualization, is characterized by three key elements, namely emotional exhaustion (i.e. loss of energy and a sense of resource depletion), depersonalization or cynicism (i.e. detachment from work, treating others like an object or with an impersonal response) and reduced personal accomplishment (i.e. feelings of incompetence, incapable and not productive). A myriad of studies reported the negative outcomes of job burnout, such as health problems, depression and anxiety (Halbesleben, 2010;Koutsimani, Montgomery, & Georganta, 2019), declining well-being (Adil & Baig, 2018), job satisfaction (Cheng & O-Yang, 2018), commitment and performance (Halbesleben, 2010;Goering, Shimazu, Zhou, Wada, & Sakai, 2017).
As depicted in COR theory (Hobfoll, 2002) and the JD-R model (Demerouti et al., 2001), resources scarcity and heightened job demands predict job burnout. On the other hand, the opposing concept of burnout, which is known as boreout, is characterized by boredom, limited growth opportunities and perceived that work is meaningless (Stock, 2015). Boreout arises due to constant low job demands, such as unchallenging work, as well as repetitive and standardized work practices, thus demotivating and resulting in dissatisfaction (Abubakar, Rezapouraghdam, Behravesh, & Megeirhi, 2022). Meanwhile, job insecurity signifies a situation with high demand that will inflate employees' burnout (Jiang & Lavaysse, 2018).

Job insecurity and hospitality research during the COVID-19 pandemic
To date, job insecurity remains a menace to employees across different organizations, particularly when there are significant changes or uncertain situation that prevails in a company (Etehadi & Karatepe, 2019;Shin, Hur, &Choi, 2020). Vander Elst, De Witte, and concluded that various job insecurity definitions found in the literature share certain common features, such that it is subjective (i.e. relies on one's interpretation of his/her actual working environment), involuntary in nature and entail employees' feeling of helplessness and uncertainty about future.
On the other hand, some authors (e.g. Borg & Elizur, 1992;Jiang & Lavaysse, 2018;Sverke et al., 2002) distinguish between the cognitive component (threat of job loss of employment) and affective component (fear or worry of job loss) of job insecurity. In this regard, Vander Elst et al. (2014) explained that despite the two being distinct components from a theoretical perspective, they are practically closely connected for employees who are struggling with potential job loss. Accordingly, Vander Elst et al. (2014) defined job insecurity as the "subjectively perceived and undesired possibility to lose the present job in the future, as well as the fear or worries related to this possibility of job loss" (p. 365). Vander Elst et al.'s (2014) job insecurity scale was validated in five European nations. In their study, Vander Elst's (2014) reported that job insecurity was negatively related to affective organizational commitment, perceived general health and self-reported performance across these countries.
As indicated earlier, perceived job insecurity among hospitality employees escalates during the COVID-19 pandemic (Jung, Jung, & Yoon, 2021) in view of the massive ravage to the industry (Jafari et al., 2021;Ozbay et al., 2022). Hospitality operators, including those who involve in gambling-related hospitality businesses, need to cut costs via human resources practices (e.g. unpaid vacation and layoff) as part of the crisis management during the pandemic (Ghaharian, Abarbanel, Soligo, & Bernhard, 2021).
Hospitality researchers have contributed several interesting studies on job insecurity during the exceptionally challenging period of the pandemic. For instance, Karatepe, Okumus, and Saydam (2022) revealed that job insecurity escalated job tension, which in turn resulted to the erosion of hotel employees' trust in the organization and raise the inclination to be late at work and leave work earlier. On the other hand, Khan, Niaze, Nazir, Hussain, and Khan (2021) pointed out that economic crisis and mental health were mediated by job insecurity and such linkages were stronger when the fear of COVID-19 increased. Furthermore, both Jung et al. (2021) and Koo, Curtis, and Ryan (2021) showed that the influence of job insecurity on turnover intention was mediated by job engagement, utilizing samples of hospitality employees from Korea and the USA, respectively. Further, the effect of job insecurity on job engagement was stronger among Generation Y during the pandemic (Jung et al., 2021).
2.4 Hypotheses 2.4.1 Job insecurity and job burnout. Based on the assertion of COR theory (Hobfoll, 2002), job insecurity can be considered a threat of possible resource loss (e.g. income and other benefits that obtain from employment). The belief that the COVID-19 pandemic is expected to affect the industry for quite a long-term exaggerated uncertainty among hospitality employees (Ozbay, Sariisik, Ceylan, & Çakmak, 2022Ozbay et al, 2022. Individuals who have limited resources and lack of ability to gain additional resources to cope with such uncertainty are likely to experience chronic stress or burnout (Hobfoll et al., 2018). Similarly, the JD-R model denoted that when perceived job insecurity (a form of job demand) is high, employees need to devote extra efforts to safeguard their employment, which leads to the exhaustion of physical and psychological resources. Consequently, job insecurity impairs employees' well-being, health, job satisfaction, commitment, work engagement and performance (Charkhabi, 2019;Darvishmotevali & Ali, 2020;Jiang & Lavaysse, 2018).

The impact of job insecurity on job burnout
Several lines of evidence have identified job insecurity as one of the key hindrance demands that exacerbate job burnout among employees (De Cuyper, M€ akikangas, Kinnunen, Mauno, & Witte, 2012;De Witte et al., 2016;Jiang & Lavaysse, 2018;Jiang & Probst, 2019;Vo-Thanh et al., 2022). Moreover, Aguiar-Quintana, Nguyen, Araujo-Cabrera, and Sanabria-Diaz (2021) noted that perceived job insecurity was followed by a substantial level of anxiety and depression among hotel employees in Spain. Besides, € Ung€ uren et al. (2021) also found that COVID-19 anxiety augmented perceived job insecurity among Turkish hotel employees, and the linkage between job insecurity and burnout was more apparent among those with low financial well-being. Likewise, Vo-Thanh et al. (2022) revealed that health risks associated with COVID-19 resulted in the rise of job insecurity and exhaust the Vietnamese frontline hotel employees emotionally. With the reviews, we envisage that job insecurity is a salient antecedent that will intensify job burnout among hospitality employees. The hypothesis is formed as follows: H1. Job insecurity is positively related to job burnout.

2.4.2
Moderating effects of supervisor support and co-worker support. Social support, in general, is defined as "social interactions or relationships that provide individuals with actual assistance or with a feeling of attachment to a person or group that is perceived as caring or loving" (Hobfoll & Stokes, 1988, p. 499). In most instances, social supports act as a coping resource that has a stress-amelioration effect (Cooper et al., 2001). Two valuable sources of work-based social support are supervisor support and co-worker support. Susskind, Michele, and Borchgrevink (2003) defined supervisor support as "individuals' beliefs that supervisors offer them work-related assistance to aid in the performance of their job" (p. 181). On the other hand, co-worker support can be explained as employees' belief that their co-workers/ colleagues are willing to extend their help in facilitating task execution (Susskind et al., 2003). As employees spend most of their time in the workplace, supportive work relation is viewed as important for effective delivery of services and completion of tasks.
Based on COR theory, Hobfoll, Freedy, Lane, and Geller (1990) posited that social support is an important vehicle that helps individuals to expand their existing resources and to protect against future resource loss. According to Hobfoll et al. (1990), "people will strive to maintain social support both to meet their needs to preserve particular resources and to protect and maintain their identify" (Hobfoll, 1990, p. 467). Past studies showed that both supervisor support and co-worker support are beneficial for better work engagement, reduce job burnout or its dimension (e.g. emotional exhaustion) and spur job satisfaction (Charoensukmongkol, Moqbel, & Gutierrez-Wirsching, 2016;Grobelna, 2021;Pienaar, Sieberhagen, & Mostert, 2007). Nevertheless, it also needs to note that these two sources of work-based social support do not always produce similar impacts on work outcomes (e.g. Talebzadeh & Karatepe, 2020).
In a highly unpredictable environment during the COVID-19 pandemic, fear of potential job loss is intense among hospitality employees; hence, they easily become a victim of job burnout. Earlier research has indicated that social support has a role in buffering the link between stressors and psychological strain (Cooper et al., 2001). As postulated in COR theory, employees are expected to make use of the available social support (e.g. co-worker and supervisor support) to secure their job or to protect existing resources. Indeed, several empirical studies demonstrated that social support help to shield against the impacts of job demands on job burnout and other work outcomes, such as noncompliance behavior and job dissatisfaction (e.g. Brough, Drummond, & Biggs, 2018;Shin et al., 2020).
Undeniable that the extensive review delineates a few inconsistent findings on the moderating effect of social support. Jolly et al. (2021) contended that the difference in the measurement used (e.g. a combination of several sources of support in one scale vs specific sources of social support) is among the reasons that lead to the differing results. Nevertheless, IHR prior studies have lent support to the positive impacts of both supervisor support and co-worker support in providing tangible resources, information and guidance that are useful to employees in reducing the pressure of constant high job demands (Lim, 1996;Day, Crown, & Ivany, 2017). Additionally, Usman et al. (2021) indicated that task-related support from co-workers was found to be effective in dealing with emotional exhaustion as it reduces perceived uncertainties among employees during the COVID-19 pandemic. As such, we anticipate that a high level of work-based social support will reduce job burnout when one is experiencing job insecurity. Therefore, relevant hypotheses have been formed as follows: H2a. Supervisor support significantly moderates the relationship between job insecurity and job burnout, such that a higher level of supervisor support would weaken the negative impact of job insecurity on job burnout.
H2b. Co-worker support significantly moderates the relationship between job insecurity and job burnout, such that a higher level of co-worker support would weaken the negative influence of job insecurity on job burnout.
Based on the above reviews, a research model is presented (see Figure 1).

Research methodology 3.1 Data collection procedure and participants
The cross-sectional data used in this study were collected from October 2020 to January 2021 by distributing self-administered questionnaires to the hospitality sector employees who work in three states in Peninsular Malaysia, namely Perak, Federal Territory of Kuala Lumpur and Selangor. A snowball sampling technique was used in the recruitment of the participants. Initially, the questionnaires were distributed personally to the targeted respondents, but the online survey was used subsequently due to the implementation of movement control order in which there are restrictions for interstate and interdistrict traveling. The total number of completed questionnaires received was 220. The questionnaires were prepared in dual language (English and Malay) to facilitate those who may not have a good mastery of the English language. The back translation process was performed with the assistance of an expert in both languages. The sample size of the present study is adequate based on the analysis conducted via G*Power (Faul, Erdfelder, Lang, & Buchner, 2007) for an 80% statistical power, a medium effect size of 0.15 and type-one error probabilities determined at 0.05. A total of 113 (51.4 %) male employees participated in this study, which is slightly higher than the female (n 5 107 or 48.6%). A total of 49 respondents (22.3%) are below the age of 25, 34.5% (n 5 76) are between the ages of 25 and 34, 24.5% (n 5 54) are between the ages of 34 and 44, 15.5% (n 5 34) are between the ages of 45 and 54 and 3.2% (n 5 7) are 55 and older.

Job insecurity Job burnout
Supervisor support Co-worker support Figure 1.

Research model
The impact of job insecurity on job burnout On the other hand, the majority of the employees have 1 to 3 years of working experience (n 5 70 or 31.8 %). The rest of the participants have less than a year (n 5 53 or 24.1%), 4 to 6 years (n 5 52 or 23.6 %), 7 to 9 years (n 5 22 or 10%) and more than 10 years (n 5 23 or 10.5%) working experience, respectively. Out of the 220 respondents, 27.3% (n 5 60) have high school qualification, and others have obtained their diploma (n 5 50 or 22.7%), bachelor's degree (n 5 100 or 45.5%) and master's degree (n 5 10 or 4.5%). On the other hand, 50.5 % (n 5 111) of the respondents are single, while 49.5 % (n 5 109) are married. Finally, most of the respondents are working in cafes/restaurants (n 5 105 or 47.7%), followed by those working in hotels/lodging (n 5 76 or 34.5%), airlines (n 5 14, 6.4%) and other organizations (n 5 25 or 11.4%).

Measures
All the key constructs were operationalization based on the well-validated scale from the past studies. The respondents rated on a Likert scale with response options from "1 5 strongly disagree" to "5 5 strongly agree" for all the measures in this study.
Job insecurity (JI) is a four-item scale adopted from Vander Elst et al. (2014), whereby the original items were developed by De Witte (2000). The sample item for the scale is "Chances are, I will soon lose my job." The measurement for job burnout was adopted from the Maslach Burnout Inventory (Maslach, Jackson, & Leiter, 1996), which consisted of three core dimensions of 22 items covering emotional exhaustion (nine items, e.g. "I feel frustrated by my job"), depersonalization (five items, e.g. "I don't really care what happens to some recipients") and reduced personal accomplishment (eight items, e.g. "I have accomplished many worthwhile things in this job").
Supervisor support (four items) and co-worker support (three items) are two sources of work-based social support covered in this study; both were adapted from Susskind et al. (2003). Items for co-worker support (CS) are "When performing my duties, I rely heavily on my co-workers," "I find my co-workers very helpful in performing my duties" and "My co-workers provide me with important work-related information and advice that make performing my job easier." On the other hand, items included for supervisor support (SS) are "I find my supervisor very helpful in performing my duties," "When performing my duties, I rely heavily on my supervisor," "My supervisor provides me with important work-related information and advice that make performing my job easier" and "I can count on my supervisor to solve the problems that relate to my job." Control variables: Dummy coding was used for the two control variables in the present research, namely gender and marital status. We controlled the two variables due to their potential association with job burnout. A few studies revealed that marital status affects job burnout (Cañadas-De la Fuente et al., 2018). Meanwhile, female was found to be more prone to burnout as compared to their male counterparts (Artz, Kaya, & Kaya, 2022;Yan et al., 2022). On the other hand, a significant impact of marital status on job burnout has been reported in a few studies (e.g. Cañadas-De la Fuente et al., 2018; Mousavy & Nimehchisalem, 2014).

Results
All the data collected were entered into IBM-SPSS Statistics version 26, and no missing data were detected during the data screening process. Partial least squares-structural equation modelling (PLS-SEM) was performed by using SmartPLS version 3.3.9 (Ringle, Wende, & Becker, 2015).
As the survey was self-report, whereby each respondent completed all the data required in the questionnaire, Harman's single factor analysis was used to detect the potential problem of IHR common method bias. The results from the unrotated exploratory factor analysis showed no evidence of common method bias as the first factor explained 24.43% of the total variance, which is less than the 50% threshold specified by Podsakoff, MacKenzie, Lee, and Podsakoff (2003). Besides, a full collinearity assessment (Kock, 2015) was conducted, and variance inflation factor (VIF) values for all latent variables were well below the value of 3.3 (see Table 1); thus, it can be concluded that the data are free from common method bias.
All the measurement models in this study are reflective constructs. As for job burnout, it is a reflective-reflective hierarchical component model (HCM). In this study, an embedded twostage approach was used in estimating the job burnout higher-order construct, as such the latent variable scores of its three dimensions were saved during the first stage and become the indicators of the higher-order construct in the next stage (Sarstedt, Hair, Cheah, Becker, & Ringle, 2019). Table 2 reveals the convergent validity and reliability results for all the constructs in the measurement model. The indicators with loadings between 0.4 and 0.7 were retained as long as the average variance extracted (AVE) of the construct surpassed 0.5 and composite reliability (CR) exceeded 0.7 (Hair, Risher, Sarstedt, & Ringle, 2019). Besides, all the measures indicate good reliability as Cronbach's alpha (α) values were well above 0.7 (Nunnally, 1978). One item of reduced personal accomplishment (PA22) was removed from the respective construct during the inspection process of the measurement model as it has a low factor loading. The elimination of this item is required to meet the convergent validity requirement. Likewise, Table 2 also indicates that the factor loadings for all the dimensions or the firstorder of job burnout fulfill the criterion of convergent validity (AVE 5 0.604; factor loadings range from 0.518 to 0.897) and internal consistency (CR 5 0.813; α 5 0.723). Table 3 illustrates that the discriminant validity for the multi-item constructs in this study is well established since the square root of AVE for every latent variable is higher than the correlation values between the variables as found in the same row and same column of the correlation matrix (Fornell & Larcker, 1981). Besides, none of the heterotrait-monotrait (HTMT) values indicated in Table 4 were beyond the threshold of 0.85 (Kline, 2011) or 0.90 (Gold, Malhotra, & Segars, 2001). As such, the results confirmed that the key latent variables in this study are distinct concepts.
Next, a bootstrapping procedure with 5000 resamples was performed to assess the structural model. Among the control variables, gender is not a significant predictor of job burnout (β 5 0.044; p > 0.05), while marital status was found to have a significant impact on employees' job burnout (β 5 0.150; p < 0.01). Nevertheless, the control variables are not the main interest of this study Table 5 demonstrates the results of hypothesized relationship, that is job insecurity (β 5 0.480; p < 0.001) was found to have a significant positive influence on job burnout. Hence, Hypothesis 1 is supported. To test the moderating effect of the two sources of social support, interaction terms were created: supervisor support 3 job insecurity (SS 3 JI) and co-worker support 3 job insecurity (CS 3 JI). As displayed in Table 5, a significant path can be found between SS 3 JI and job burnout (β 5 À0.257; p < 0.05). This shows that supervisor support moderates the relationship between job insecurity and job burnout. The significant interaction result was presented graphically. As shown in Figure 2,  The impact of job insecurity on job burnout to greater job burnout, especially when supervisor support is low. On the other hand, when there is a high level of supervisor support, the effect of job insecurity on job burnout is weaker. Further simple slope test demonstrated that both positive slopes for low supervisor support (β 5 0.536; p < 0.001) and high supervisor support (β 5 0.267; p < 0.001) are significant. As such, Hypothesis 2a is supported. Besides, the interaction between CS 3 JI and job burnout was found to be significant as well (β 5 0.436; p < 0.001). However, the graph as illustrated in Figure 3 reveals that when co-worker support is low, perceived job insecurity has no impact on employees' job burnout (β 5 0.068; p > 0.05). Conversely, the positive relationship between job insecurity and job burnout becomes stronger when co-worker support is high, and the slope is found to be significant (β 5 0.569; p < 0.001). In short, this result shows that the impact of job insecurity on job burnout increases due to the higher level of co-worker support instead of the reverse. Hence, Hypothesis 2b is not fully supported.  Table 5 indicates the effect size of each proposed relationship between the variables. In accordance with the Cohen's (1988) rule of thumb, the effect size value of 0.02, 0.15 and 0.35 represents small, medium and substantial relative impact of predictors on an endogenous variable. Hence, the impact of job insecurity on job burnout is medium (f 2 5 0.314).   The impact of job insecurity on job burnout Besides, both interaction terms, SS 3 JI (f 2 5 0.048) and CS 3 JI (f 2 5 0.109), exert small effects on job burnout. Once the interaction terms were included in the model, the R 2 value of job burnout increased from 0.314 to 0.385 (R 2 change 5 0.071). As such, the inclusion of the interaction terms explained an additional 7.1% of the variance in job burnout. The Q 2 value for job burnout is 0.209, which is more than zero signifying the predictive relevance of the model is adequate.
Lastly, model fit was assessed. Henseler, Hubona, and Ray (2016) introduced standardized root mean square residual (SRMR) as a model fit measure for PLS-SEM. The SRMR values for both saturated (measurement) model and estimated (structural) model were 0.08, respectively. The SRMR value of not more than 0.10 or of 0.08 for a more stringent requirement is regarded as a good fit (Hu and Bentler, 1998;Ramayah, Cheah, Chuah, Ting, & Memon, 2016). Thus, the goodness of fit for the present model is satisfactory. Besides, the exact model fit tests were performed via the squared Euclidean distance (dULS) and the geodesic distance (dG). The results showed that the discrepancies for both saturated and estimated models were less than the 95% quartile (HI 95 ) of their bootstrap distribution (dULS saturated model 5 0.488, HI 95 5 0.504; dULS estimated model 5 0.487, HI 95 5 0.505; dG saturated model 5 0.334, HI 95 5 0.366; dG saturated model 5 0.333, HI 95 5 0.367). Evaluations of SRMR and bootstrap-based overall fit test (dULS and dG) are adequate to examine the goodness of fit for PLS-SEM (Benitez, 2020).

Discussion
Our study empirically investigates the effect of job insecurity on job burnout. Besides, this study also sought to ascertain the moderating effect of supervisor support and co-worker support between job insecurity and job burnout. As anticipated, job insecurity results in a substantial increase in job burnout among hospitality employees, and the finding coincides with the studies by several researchers (Jiang & Probst, 2019;Kinnunen, M€ akikangas, Mauno, De Cuyper, & De Witte, 2014;Vo-Thanh et al., 2022). The study was conducted during the most vulnerable time of the hospitality industry, and job losses were widely reported since the onset of COVID-19 pandemic (Chin, 2021). Thus, perceived high job insecurity among the employees is inevitable, which eventually evoke job burnout. The result was aligned with insights from the JD-R model (Demerouti et al., 2001), in which job insecurity drains employees' energy and emotion. Moderating effect of coworker support (CS) between job insecurity (JI) and job burnout IHR Next, we found that the positive relationship between job insecurity and job burnout was more pronounced when supervisor support was low. In contrast, job insecurity cause less harmful effect on hospitality employees when supervisor support was high. Hence, the result affirmed the buffering effect of supervisor support between job insecurity and burnout. Based on the tenet of COR theory (Hobfoll et al., 1990), individuals will strive to minimize or protect the loss of resources by obtaining additional resources. As such, supervisor support is an effective job resource that can be utilized by employees to attenuate burnout that arises due to high job insecurity. Adair et al. (2008) noted that supervisor support is important in the downsizing process as it complements the perception of a lack of support from the organization. This is particularly relevant among hospitality employees during the COVID-19 pandemic.
An interesting finding in this study is the reverse buffering effect of co-worker support between job insecurity and job burnout. That is, hospitality employees who perceived high job insecurity suffer from more burnout when high co-worker support is available rather than the opposite direction. Despite unexpected, the reverse buffering effect of co-worker support between stressors and psychological outcomes were observed in the past (e.g. Deelstra et al., 2003;Kokoroko & Sanda, 2019). Besides, Shin et al. (2020) showed that co-worker instrumental support showed an opposite moderating effect between job crafting and work engagement.
We viewed the possible reasons for the present finding from two perspectives. First, the co-worker support measure used in this study, which is relevant to instrumental/task-related support (e.g. providing assistance and guidance related to the job), is likely to affect the result. Jolly et al. (2021) expounded that the diversity of social support measures used in the literature is among the possible reasons for the mixed findings. They further argued that a specific source and type of support may better suit with a specific demand (Jolly et al., 2021). Instrumental support, in particular, is more effective to be provided by the supervisor instead of a co-worker as the supervisor is responsible for job allocation and performance evaluation of the employees (Mathieu, Eschleman, & Cheng, 2019). Second, previous studies indicated that co-worker support is not always helpful in all situations, such as when the support is not needed by the employee (Beehr et al., 2010;Kokoroko & Sanda, 2019). In an experimental study, Deelstra et al. (2003) showed that instrumental support at work negatively affects employees' self-esteem and exacerbates stress if the support cannot be rejected by the employees. Accordingly, in the context of high perceived job insecurity during the pandemic, the hospitality employees may have the tendency to show their ability in completing the tasks rather than relying on the help from their co-workers, which can be an indicator that they are relatively weak or make them feel uneasy and worry that they are viewed as incompetent and become a target of layoff in the future.

Theoretical implications
Our study provides some additional theoretical contributions to the existing literature. First, this study shows that job insecurity has a paramount impact on hospitality sector employees' job burnout during the public health crisis in the Malaysia context. The pathway between job insecurity and burnout conforms to the assumption as postulated in the JD-R model (Demerouti et al., 2001) that job demands trigger burnout when individuals are unable to adapt.
Second, our study adds to the existing literature by revealing the different magnitude of influence of two key work-based social supports (supervisor support and co-worker support) between job insecurity and job burnout among hospitality employees during the outbreak of the COVID-19 pandemic. Besides, distinguishing different sources of support, rather than utilizing a combined measure of various supports, helps to clarify the mixed results in past The impact of job insecurity on job burnout studies (Jolly et al., 2021). This study provides an insight that supervisor support is a salient buffering agent between job insecurity and job burnout in contrast to co-worker support that acts in an opposite direction in this relationship. A reverse buffering effect occurs when greater social support worsens the effects of demands on an individual's stress or burnout level (Beehr et al., 2010;Cooper et al., 2001).
Based on the findings, we demonstrated that co-worker support that typically viewed as a job resource from the perspective of JD-R theory can become a form of demand that drain an employee. On the contrary, supervisor support acts as a more consistent and valuable job resource in this study, thus fitting the assumption in JD-R theory as well as the COR theory, which posits that "people strive to maintain and gain resources" (Hobfoll et al., 1990, p. 472) to minimize or avoid the loss of valuable resources.
In this regard, we showed a different perspective about the moderating role of co-worker support. It may be oversimplified to view different sources of social support have similar buffering effects between a particular job demand and burnout. The effectiveness of a specific type of social support depends on the extent to which it matches the demand that encountered by the employees (Jolly et al., 2021). In sum, this study provides a more nuanced understanding that co-worker support, especially task-related support, is not favorable in reducing burnout when one is facing with a high threat of job insecurity, such as during the pandemic. In contrast, employees can look for supervisor support when encountering difficulties at work. Notably, not all forms of social support will buffer the impacts of a demanding environment in every situation (Mathieu et al., 2019), specifically this study points to the complex nature of co-worker support.

Managerial implications
The practical implications that can be derived from the findings as follows: first, management should have effective communication with employees to reduce their worry and uncertainty about their current employment. Anxiety about the possibility of losing a job is a painful process for employees. As such, the management should be transparent and honest about the development of the company. Notably, organizations risk losing talented employees if they fail to communicate well as many employees are potentially looking for other alternatives or even switching to another career during this period of high uncertainty. While this study was conducted during the peak of the pandemic, hospitality operators are likely going to face a shortage of experienced workforce when the demands in the sector recover. Moreover, there is competition for scarce talent among the operators in the hospitality industry as well as cross-industry for the same skills and competencies (Lytle, 2020).
Besides, the hospitality management should provide early and frequent communication to explain the problems faced and the future direction of the company to the employees, especially in a crisis. The human resource department can provide relevant training to the managers of different departments on how to better communicate and support employees who are encountering job burnout. On the other hand, the voluntary participation in corporate social responsible activities among hospitality operators during a public health crisis (e.g. hotels take part as a temporary quarantine center) not only help to secure income sources, protect employees' job security and retain domestic customers, but also enhance employees' work engagement (Choi & Choi, 2021). Besides, business operators can provide free meals and/or raise fund to their community and hospitality employees who were affected by the pandemic or any future crisis (Norris, Taylor, & Taylor, 2021). Additionally, the hospitality management should involve more actively in internal marketing approaches, such as prompt responds, positive feedback, concern for employees' welfare and assistance by supervisors helps to improve employees' engagement and customer-oriented behavior and guest satisfaction (Olorunsola, Saydam, Ogunmokun, & Ozturen, 2022).

IHR
Aside from this, employees may perceive job insecurity as a serious threat of resource loss due to a sense of powerlessness, which may derive unclear expectations about job performance, and they are unsure of the appropriate corrective actions should be made (Greenhalgh & Rosenblatt, 1984;Gordon, Tang, Day, & Adler, 2019). Hence, supervisors not only need to set a clear performance appraisal criterion but also assist employees with their job-related problems. Hospitality organizations can implement a system that supports employee cooperation and teamwork and help to reduce the sense of job insecurity among staff.
Social support from supervisors and co-workers is generally viewed as important to most employees, especially in a highly competitive industry. Nevertheless, it is important for the support providers, especially the co-workers to be more sensitive to the nature of the support given, such that avoiding task-related support that is not desired by their counterparts and potentially become a threat to their self-image (Beehr et al., 2010). The unnecessary assistance would result in frustration and negative emotion as well as intensify the level of burnout among the employees.

Limitations and recommendations
Several limitations need to be addressed in this study. First, the measure of supervisor support and co-worker used in this study focuses more on instrumental or task-related support; therefore, we suggest that future studies should include measures that clearly distinguish among the four types of support as described by House (1981): emotional support (e.g. caring and sympathy), instrumental (e.g. direct and practical help), information support and appraisal support (e.g. providing feedback). Additionally, nonwork social supports (e.g. support from friends and family members) can be included in the model for future research. Second, the 220 samples of this study consist of employees from various sub-segments within the hospitality industry. Though the sample size is adequate in meeting the criteria for data analysis in the present study, expansion of the sample size for each sub-segment within this industry (e.g. airlines, hotels and restaurants) will allow clearer understanding, and comparison can be made on the hypothesized relationships in this study.
Third, the research model of the present study is limited to assessing job insecurity as an antecedent of job burnout while supervisor support and co-worker support are the moderators, and the consequences of job burnout have not been further explored. Henceforth, future studies should advance the current model by incorporating potential outcome variables, such as physical and mental health, job performance, service quality, organizational citizenship behavior and absenteeism. Another avenue for future study is by having an additional antecedent, such as job design in the model to predict burnout, as well as an emerging conceptboreout (Abubakar et al., 2022).
Besides, we examined job burnout as a higher-order construct for a more parsimonious model, and future studies can examine the moderating effects of both supervisor support and co-worker support between job insecurity on the three dimensions of job burnout.
Next, the sample of our study primarily consists of hospitality employees in Malaysia, a multi-ethnic country (i.e. the major ethnic groups are Malay, Chinese and Indian). Malaysia is a collectivistic society in contrast to the USA, which is high in individualism (Hoftstede Insights, 2022). Though social supports generally are viewed as an important protective factor when faced with adversity (Abbas et al., 2021), Glazer (2006) accentuated that perception and responses to social support vary based on national culture. In a collectivist culture, employees are more cautious in getting social support to avoid burdening their social network as compared to those in individualistic cultures (Kim, Sherman, & Taylor, 2008). Besides, Rojas-M endez (2021) reported that people from a long-term orientation society (e.g. South Korea) were less worried about the effects of the pandemic as compared to a short-term orientation society (e.g. the USA). Malaysian society score in long-term orientation The impact of job insecurity on job burnout is moderate (Hoftstede Insights, 2022); the differences may affect one's perceived job insecurity and tendency of experiencing chronic stress. Considering the above, the generalization of the present findings may be limited to countries with similar national cultural characteristics. As such, future researchers can take a precautionary step by incorporating the national cultural dimensions as control variables in their research framework.
Further, the generalization of the present result may be restricted by the diversity of financial support given by the different governments to the employees during the pandemic, especially between developed versus developing nations. For example, the coronavirus relief package approved by the US Senate in early 2020 enabled an unemployed worker to gain an extra relief of $600 per week for four months in addition to standard state benefits of $385 (Iacurci, 2020). Such lavish unemployment compensation may affect employees' perception of job insecurity, and it was reported that some are even reluctant to return to work to enjoy the benefit (Smith, 2021). Despite the limitation, job insecurity has been described as the utmost concern among employees in developing countries, especially during an unexpected crisis (Abbas et al., 2021). Besides, the coverage of unemployment protection benefits in Asia and Pacific countries was only around 37% (International Labour Organization, 2021). Taken together, the future study can consider the effect of government financial aid on employees in evaluating the linkages between the variables in this study, especially in an unexpected crisis.
Lastly, this study was conducted during the COVID-19 pandemic; hence, the results may be unique to this specific condition. A comparative study can be a direction for future research, especially to clarify the role of co-worker support which showed a contradictory result in the present study.