The interrelations between micro-level CSR, Job engagement and organizational engagement during the COVID-19 pandemic: A PLSpredict model assessment

2.1Stakeholder theory, social exchange theory and engagement theory

Three underpinning theories guide the development of this study. Firstly, the Stakeholder Theory [25] offers a deviation from the customary understanding of business as a vehicle to maximize owners’ returns, as it advocates generating value and interest to diverse stakeholders [26]. This theory highlights the importance of building and maintaining sustainable stakeholder relationships for firm performance. In this vein, micro-level CSR practices reflect the essence of the Stakeholder Theory to build and uphold sustainable relationships with highly pertinent stakeholders –employees.

Next, referring to the Social Exchange Theory (SET), the implementation of micro-level CSR practices is rewarded with the Matthew effect of positive employee outcomes. According to Zhang et al. [27], social life often involves a progression of consecutive dealings between two or more parties through a reciprocity process. The reciprocity process results in one party’s tendency to repay the good or bad deeds of the other party, depending on the relationship and trust level [28]. Based on an integrated understanding of the Stakeholder Theory and the SET, when employees receive good deeds through micro-level CSR practices in times of disaster, they would repay their organizations with encouraging attitudes and behaviours.

Hejjas et al. [29] ascertained the role of Stakeholder Theory in providing a valuable foundation to study employee disengagement. The Engagement Theory [30] elucidates disengagement as the lack of commitment and attachment towards the execution of work, while engagement is described as the employees’ motivation and commitment towards their role performance [31, 32]. Therefore, employee engagement is a core priority for every organization, especially during uncertainty.

Taken together, the three aforementioned theories collectively purvey an integrative theoretical foundation for this study. Aguinis et al. [33] articulated that for effective positive outcomes, organizations are recommended to select a CSR strategy that is embedded into the organization’s core competencies. It should be integrated into a firm’s strategy, routines, and operations. Henceforth, the notion of micro-level CSR which aims to create and maintain sustainable employee relationships indicates the execution of internal strategic management as a firm’s strategy. The anticipation of employees’ positive outcomes is advanced by the SET and the Engagement Theory. Specifically, Aguinis et al. [33] explained that a behavioural CSR approach, such as micro-level CSR, benefits both employees and organizations in a mutual exchange process. Moyo [32] pointed out that employee disengagement has heightened during the pandemic, raising the significant need for health and safety to achieve employee engagement. This backdrop motivates the study of micro-level CSR practices for employee engagement in times of crisis.

2.2Micro-level CSR practices, job engagement and organizational engagement

A team of micro-level CSR pioneer researchers (e.g., Al-bdour et al. [34], Turker, [35]) explained that CSR activities discharged by organizations to their employees cover the areas of safety, health, wellbeing, work-family relationships, training and development, and equal opportunities. Ergo, micro-level CSR refers to the psychological and physiological aspects of employees. In Mory et al. [11]’s work, seven dimensions are used to gauge the psychological and physiological aspects of employees. There are employment stability, employee involvement, empowerment, working environment, skill development, workforce diversity, and work-life balance; while in Al-bdour et al. [34]’s research, human rights, health and safety, work-life balance, workplace diversity training, and education were the measures for employees’ welfare. After a careful examination of these two works, we decided to consolidate the overlaps and adopt the more current and complete measurement to address our current research context of the vitality of micro-level CSR in relation to employee engagement in times of crisis. With this, employee empowerment, employee involvement, employment stability, working environment, and work-life balance are incorporated in current study to comprehend its impacts on employee engagement during the COVID-19 crisis.

Nonetheless, the link between CSR perceptions and work engagement is not to be taken as universal. Rupp et al. [36] explicated the potential of individual and contextual factors to act as meaningful boundary conditions for employee engagement. Additionally, Glavas and Agunis [37] provided the nexus of sense making in the significance of micro-level CSR practices; thus, it is worth making the important distinction between job engagement and organizational engagement.

Engagement at work was originally conceptualized by Kahn [30] as harnessing the organizational members to their work roles, hence, it is also known as employee engagement. It refers to the extent to which an employee believes in the mission, purpose and values of the organization and displays that commitment through their actions as an employee and their attitude towards the employer. Engaged employees express themselves physically, cognitively and emotionally during the performance of their roles, which is termed as job engagement [86]. Therewith, there are three degrees of employee engagement, i.e., actively engaged, not engaged and actively disengaged. Commonly, the research uncovered that employees fall into the middle level specifically, not engaged. HR personnel and managers’ opinion that engagement challenge has to do with how the employee feels about the work experience and how he/she is treated in the organization. It is something the employee has to offer; it cannot be explicitly stated in the employment contract. Also, engagement is distinctive on its own and different from satisfaction, and motivation. The original widely used term of employee engagement embraces two attempts of organizations which are to motivate the employees and focus on their commitment to achieving the organizational objectives. These two attempts branched out to job/ work engagement and organizational engagement.

Employee engagement started to gain attention in the late 2000 s after Gallup pointed out the alarming financial losses borne by the US economy (approximately US$450 billion to US$550 billion) due to the declined productivity of disengaged employees. Following this report, employee engagement studies proliferated (e.g., Balliester & Elsheikhi, [38]; Ismail et al. [39]). Gallup’s report also indicated that higher employee engagement is directly related to a greater level of productivity, while Al-Mehrzi and Singh [40] linked it with economic growth and profitability. Subsequently, the positive chain effects of engagement are drivers of innovation and competitive advantage [41]. With the intensified focus on employee engagement levels in business, it is crucial to make distinctions between the different types of employee engagement. Saks and Gruman [23] stated that employee engagement is not a unidimensional concept but a multidimensional notion. Schaufeli and Salanova [42] enlightened that employee engagement covers the relationship between an employee and his/her occupation, work, and organization. This connotation leads to the classification of job/work engagement and organizational engagement. Organizations need to differentiate employees’ engagement with their job from their engagement with their organization. It is possible for employees to be highly engaged in their jobs but feel disconnected from the overall organization or vice versa. Therefore, only by examining job engagement and organizational engagement distinctly can organizations craft appropriate strategies to address the different requirements of both engagement types.

2.3Employee empowerment

Employee empowerment involves employees taking independent actions or decisions in their assigned tasks. It refers to the extent to which the employee is given the opportunity to determine his/her operating context in a self-actualizing way. The adoption of Work from Home (WFH) practices during the COVID-19 pandemic has made the element of employee empowerment more relevant and crucial. Tripathi et al. [43] revealed how psychologically empowered employees to develop a robust consideration of their roles due to perceived organizational support. Employee empowerment will progress into creativity and engagement [44]. Parallel with this discussion, the following direct hypotheses are posited:

H1a: Employee empowerment positively influences job engagement.

H1b: Employee empowerment positively influences organizational engagement.

2.4Employee involvement

Employee involvement refers to the practice of having employees participate in critical operations and management meetings. Carmeli et al. [45] asserted that employee involvement is part of the normative foundation of moral decision-making grounded in the theory of social responsibility. Macey et al. [46] claimed that employee involvement and work engagement are vital to drive organizational success. This assertion is substantial in a crisis context because employee involvement sends a message to the employees that they are valued and trusted [46]. Applied to the understanding of SET, employee involvement necessitates trust between both parties and enhances their reciprocity with positive outcomes. Therefore, the following two direct relationships are put forth:

H2a: Employee involvement positively influences job engagement.

H2b: Employee involvement positively influences organizational engagement.

2.5Employment stability

Employment stability reflects the extent to which organizations provide and secure stable jobs for their employees. In the past, employment stability was analyzed by Lin and Wei [48] in the early CSR context because of aggressive global competition. At present, the battle with COVID-19 has posed a severe economic threat to many businesses. Adopting the Stakeholder Theory’s lens, business organizations have the most fundamental responsibility to their employees, who should be prioritized by granting a secure job. The SET further highlights the occurrence of an obligatory transactional exchange from employees to the organization when this form of organizational support is given. Based on this rationale, two direct hypotheses are proposed:

H3a: Employment stability positively influences job engagement.

H3b: Employment stability positively influences organizational engagement.

2.6Working environment

The working environment carries equal weight in the consideration of micro-level CSR practices. The working environment refers to the issues of health and safety in the workplace. In certain industries and organizations, it is explicitly written in company policies to demonstrate organizational sensitivity towards employees’ safety [91]. In line with the core concept of micro-level CSR, a safe and conducive working environment reflects the physiological well-being of employees. Taking the current pandemic into account, the working environment is extremely important for employees to feel safe before they could engage in their work. Consequently, the following hypotheses are developed:

H4a: The working environment positively influences job engagement.

H4b: The working environment positively influences organizational engagement.

2.7Work-life balance

To cultivate mindfulness for mental health, striking a balance between work-related matters and personal life matters is important as one does not infringe on the other; this is called work-life balance (WLB) [48]. According to Bekir [50], WLB in several countries has severely deteriorated due to the pandemic. The implementation of WFH practices to minimize COVID-19 infection has indeed blurred the line between work life and personal life. Many employees have endured the stress of fulfilling multiple roles’ demands, ranging from high workload requirements to ad-hoc family care arrangements. Such situations cause a deterioration in WLB and detrimentally affect the mental health of employees [51]. These negative consequences do not halt at the individual level but transfer to the work context and organizational level. With this, the following hypotheses are formulated:

H5a: Work-life balance positively influences job engagement.

H5b: Work-life balance positively influences organizational engagement.

2.8Job engagement, organizational engagement and the mediating role of job engagement

Job engagement is described as the relationship of an employee with his/her work. In this context, the core focus is on the employee’s job and roles within the organization. Positive job engagement denotes that the employee perceives meaningfulness in his/her job that extends beyond his/her remuneration package. Job-engaged employees associate their value with the role tied to their own self-perception. When they believe that their job is important and valuable in the organization, they tend to work the extra mile and perform progressively in their job. Commonly, job engagement is interchangeable with work engagement. For this study’s purpose, job engagement is used, to clearly differentiate it from organizational engagement.

Meanwhile, organizational engagement casts a wider scope by going beyond the individual level to encompass employees’ emotional commitment of their complete selves to their organization and its goals [23]. Therefore, organizational engagement is affected by organizational factors such as organizational objectives and values. When organizational engagement is achieved, it benefits organizations substantially as everyone is committed to the firms’ common goals and motivated to work towards them [89]. Henceforth, job engagement reflects self-perception while organizational engagement covers a broader focus. Based on this understanding, we hypothesize that a job-engaged employee will display positive organizational engagement by diligently meeting organizational goals as postulated in H6:

H6: Job engagement positively influences organizational engagement.

Considering the self-originating nature of job engagement, the challenges faced by organizations during this pandemic have more profoundly impacted organizational engagement. Many business organizations are adopting WFH practices during the pandemic, which has landed employees in a vulnerable isolation state. Social distancing and isolation phenomena run the risk of employees perceiving their roles as singular instead of inter-complementary when a job is to be done. Likewise, ensuring organizational engagement at every level has always been a challenge for business organizations. In this regard, we attempt to study the mediating role of job engagement in bridging the relationship between micro-level CSR practices and organizational engagement. Hypotheses H7 to H11 predict a positive mediating effect of job engagement between the five micro-level CSR practices and organizational engagement, as follows:

H7: Job engagement mediates the relationship between employee empowerment and organizational engagement.

H8: Job engagement mediates the relationship between employee involvement and organizational engagement.

H9: Job engagement mediates the relationship between employment stability and organizational engagement.

H10: Job engagement mediates the relationship between wok-life balance and organizational engagement.

H11: Job engagement mediates the relationship between working environment and organizational engagement.

Fig. 2

Measurement model assessment.

The above literature review has underscored the importance of micro-level CSR practices and their interrelations with employee engagement. The hypotheses formulated are further illustrated in Fig. 1: Research Framework. According to Alhozi et al. [86], employee empowerment promotes job engagement by augmenting an organization’s ability to materialize its strategic approach to gain comparative advantage and this positive outcome is cascaded to the organizational level as organization engagement [87]. Thereby, H1a and H1b are developed. Employee involvement is another pertinent element of organizational effectiveness. Dade et al. [88] empirically testified the significant relationship between employee involvement and employee engagement which positively link to organizational effectiveness. With this, H2a and H2b are formulated. In 2015, United Nations envisions to achieve full and productive employment and decent work for all through United Nations Sustainable Development Goal (UNSDG) 8, i.e., Decent Work and Economic Growth. The grand vision has inspired much research to examine the context of decent work. Drawing from The International Labour Organization (ILO), decent work is defined as productive work for individuals in conditions of freedom, equity, security, and human dignity. In essence, work is considered decent when: it pays a fair income; guarantees a secure form of employment and safe working conditions. The definition and work conditions are closely related to the description of micro-level CSR practices [7, 21, 90]. Henceforth, H3 to H5 are formed accordingly to investigate its influence on job engagement. Subsequently, Saks et al. [89] highlighted the need to further explore employee engagement as current literature demonstrated a deficiency in distinguishing between job engagement and organization engagement. H6 to H11 are the indirect relationships that aim to address this absence of enriching findings on job engagement and organization engagement.

Fig. 1

Research framework.

3.1Sampling and data collection

The World Health Organization (WHO) announced that there were over 118,000 COVID-19 cases reported in more than 110 countries as of March 11th, 2020 [84]. No country has escaped from this pandemic and like most countries, the Malaysian government took various precautions to prevent the virus outbreak.

Beginning March 18th, 2020, a series of national quarantines and cordon sanitaire measures were implemented by the Malaysian government. Five phases of the MCO were mandated from March 18th to August 31st, 2020. Data collection began in March 2020 and lasted for three months up to June 2020. We extended our questionnaires to employees who were employed and working throughout that period. Data was collected from eligible candidates throughout Peninsular Malaysia, which is justified as the region contributes 78.6% of the overall Malaysia labor force, as per the Department of Statistics Malaysia.

The judgmental sampling technique was employed wherein researchers exercise their professional judgment in identifying the targeted respondents. This technique enables researchers to reach out to the targeted population and increases the relevance of the sample, given that individuals who do not meet the criteria would not be included. To be eligible to participate in this study, respondents had to be working and attached to an organization but not be self-employed.

3.2Instrument

The measures used in this study were adapted from current literature with minor revision to better suit the research context. The measures of the five micro-level CSR practices dimensions were adopted from Mory et al. [11] while the measure of job engagement was derived from Schaufeli et al.’s [42] five-item scale. All the measures adopted are reflective measures whereby the indicators are consequences of the constructs. A 7-point Likert scale with “1” indicating strongly disagree and “7” indicating strongly agree was used for all the measures. The measures of this study are presented in the Appendix.

3.3Power analysis

G*Power program was utilized to calculate the required sample size for statistical significance. With our research model comes with six predictors, a minimum of 107 observations are required to attain an effect size of 0.15 at 95% power level [52]. Memon et al. [53] informed that there is no one-size-fits-all solution to address the sample size, yet they recommended a sample size of 160–300 is most appropriate for multivariate statistical analysis techniques such as PLS-SEM. With this in mind, we had sent out 800 questionnaires and successfully collected 336 completed responses, yielding a response rate of 42.0%. According to Holbrook et al. [96], a response rate that ranges from 5% to 54% is only marginally less accurate than those that reported a higher response rate. They further explained that while the response rate is informative, it is insufficient to determine the validity of the results. Reporting about the recruitment and furnishing more detailed information about how representative those who agree to participate in comparison with those who do not agree based on the population characteristics is a better risk assessment. Henceforth, the current study’s response rate of 42.0% falls into the indicative acceptable level.

3.4Non-response bias

In conducting a quantitative questionnaire survey, it is not uncommon to have non-response bias. Non-response bias occurs when there is a significant variance between those who responded and those who did not. Therefore, it presents a threat to the findings. Armstrong and Overton [97] offered some approaches to estimate the non-response bias, such as comparison with the know values in the population, subjective estimate, and extrapolation method. For the current study, the subjective estimate was implemented for three core reasons: first, it is a widely recommended basis to cope with the non-response bias; second, Brown [98] claimed that it is a useful approach; and third, Schwirian and Blaine [99] tested and confirmed its’ validity. Subjective estimates state that people who are more interested in the subject of a questionnaire often respond more readily and thereby providing the necessary information for researchers to understand the subject matter. With judgmental sampling technique adopted and the consideration of non-responses bias along with the response rate of 42.0%, we are confident that the detrimental effect of non-response bias is not a concern in this study.

3.5Common method bias

Common method bias (CMB) is a serious methodological phenomenon caused by common variation induced by the measurement method adopted [54]. Data that suffered from CMB potentially lead to artificial inflation of relationships that hinder the measures’ reliability and validity. MacKenzie and Podsakoff [55] proposed statistical and procedural remedies to alleviate the issues of CMB. Current study adopted both procedural and statistical strategies to mitigate CMB issue. Jordan and Troth [56]’s suggestion of procedural remedy was employed. A copy of the detailed research information coversheet was given to all the respondents in order to increase the probability of response accuracy. Also, pilot-test was performed, and comments gathered were incorporated. For the statistical approaches, Harman’s single factor and full collinearity variance inflated factor (AFVIF) approach were executed. Harman’s single factor test results informed that 23.76% of the variance was explained by a single factor which indicates that less than a majority of the variances were explained. Also, the value of 2.208 was obtained for AFVIF which is below the threshold of 3.3 [57]. All in all, CMB did not substantially influence the results of the study as evidenced by the procedural and statistical approaches.

3.6Analytical technique

There are two commonly used Structural Equation Modeling (SEM) to analyze complex inter-relationships between observed and latent variables, namely Covariance based Structural Equation Modeling (CB-SEM) and Partial Least Squares based Structural Equation Modeling (PLS-SEM) [92]. CB-SEM is based on covariance which is also known as factor-based SEM approximates latent variables by common factors, while PLS-SEM is based on variance, called composite-based SEM [93]. Even though factor-based SEM remains prevalent in practice, recent research in psychometrics calls the central principles of the common factor model into question. Rigdon et al. [94] highlighted that common factor proxies cannot be assumed to carry greater significance than composite proxies regarding the existence or nature of conceptual variables. They further expounded that the indeterminacy of common factors creates a band of (metrological) uncertainty in the relationship between the factor inside the model and any variable outside the model including the conceptual variable that the factor seeks to represent [94].

Current study decides on Partial Least Squares Structural Equation Modeling (PLS-SEM) to proceed with the statistical analysis for two core reasons. Firstly, PLS-SEM permits researchers to evaluate complex models [58], which it matches our current research model that comes with many indicators, constructs and structural paths. Secondly, PLS-SEM is a causal predictive approach advancing on regression-based technique in social sciences areas to estimate path relationship with manifest variables [59, 60]. Most recently, Sarstedt and Dank [61] claimed that a model with a certain degree of explanatory power potentially led to immensely different levels of predictive power and contrarywise despite of having identical and different contexts. As the proposed research framework is destined as a frugal fix for organizations during this turbulent time, predictability power is our core priority. These two appealing merits of PLS-SEM justified its’ usage in the current study.

4.1Respondents’ demographic profile

An outline of the total 336 respondents’ demographic profiles is provided in Table 1. The majority of the respondents are females (54.46%) with most of them in the age group of 25 to 30 years old (38.99%), followed by the age group of 18 to 24 year old (32.44%). The married respondents represent many of the responses with 58.63% while the single respondents contributed 38.39% of the responses. It was interesting to find out that a large group of respondents have been working with the existing company for 5 to 10 years. Respondents who are working in the service sector dominated the samples (66.67%), followed by the manufacturing sector (18.45%) and the construction sector (4.17%).

4.2Internal consistency reliability

Following Hair et al. [63]’s recommendation, the first step in evaluating PLS-SEM begins with the measurement model assessment. This assessment consists of the evaluation of internal consistency reliability, convergent validity, and discriminant validity. Internal consistency reliability measures the degree to which the items reflect the latent constructs. It is measured through composite reliability (CR). According to Sarstedt et al. [60], a CR values above 0.70 is considered acceptable. The values of CR for the current study fall in the range of 0.921 to 0.965 which is higher than the recommended threshold of 0.70. The values of Cronbach alpha and rho_A were examined and both values were above the required threshold values. The results in Table 2 signify that internal consistency reliability is achieved.

4.3Convergent validity

Convergent validity examines the extent to which the distinct indicators reflect a construct converging in comparison to indicators measuring other constructs [62]. Hair et al. [63] informed that convergent validity is evaluated based on the factor loadings and average variance extracted (AVE). Hair et al. [63] provided a rule of thumb for the factor loading and AVE respectively. A factor loading value equal to and greater than 0.7 is acceptable while an AVE of 0.5 or higher is considered satisfactory. Most of the items’ loadings are greater than the threshold value of 0.7 except WE1, while the AVE is higher than 0.5.

4.4Discriminant validity

Discriminant validity is then examined to ensure that each latent variable is differentiated from other constructs in the model [63]. Often, it is referred to as the degree to which the indicators are distinct from others across constructs. According to Hair et al. [62], heterotrait-monotrait (HTMT) criterion is recommended to perform the discriminant validity testing. This is in line with Henseler et al. [64]’s suggestion of HTMT ratio of correlation criterion for discriminant validity evaluation. Henseler et al. [64] further explicated that the establishment of discriminant validity is ascertained when the HTMT statistics do not exceed 0.85 or 0.90. Table 3 shows that none of the HTMT values are above 0.90 [64, 65]. Furthermore, all the values of the confidence interval in Table 3 do not display a value of 1 in between, indicating that discriminant validity has been established [64, 65].

4.5Structural model assessment

Upon completion of measurement model assessment, we proceed with the structural model evaluation which involves the hypotheses testing. The structural model assessment entails the examination of path coefficients, coefficients of determinations (R²), the effect size (f²) and predictive relevance (Q²) [62]. R² alludes the overall predictive accuracy of the research model [62]. Cohen [65] provided a guide of R² values, with 0.26 indicating large while 0.13 and 0.02 represent medium and small respectively. The results are available in Table 4. The R² for Job engagement was recorded as 44.4%, which indicates that 44.4% of the variance in Job engagement can be explained by the independent variables. Meanwhile, Organizational engagement achieved a R² of 66.3% indicating 66.3% of its variance are explained by the five dimensions of micro-level CSR practices and job engagement.

For the effect size, Employee empowerment has a medium effect on Job engagement (f² = 0.216), while Employee involvement (f² = 0.063), Work life balance (f² = 0.026) and Working environment (f² = 0.027) have small effect size on Job engagement. Interestingly, Employment stability has no effect on Job engagement. Slightly different effect size was observed on Organizational engagement. For example, Employee involvement has a medium effect on Organizational engagement (f² = 0.193), while Employee empowerment (f² = 0.068), Employment stability (f² = 0.020) and Working environment (f² = 0.030) have small effect size on Organizational engagement but Work life balance displays no effect.

In terms of the naïve predictive relevance, it was assessed using Stone-Geisser Q² [67], all the predictive values are higher than 0, informing the model possesses predictive relevance [68].

According to Diamantopoulos and Siguaw [69], researchers shall address the collinearity issue through variance inflated factor (VIF) before assessing the structural model. They suggested that VIF values below 3.3 for each of the constructs signifies that collinearity is not a threat. Table 4 outlines the collinearity test results with all the VIF values lower than 3.3. With this, it concluded the collinearity is not a threat in the model.

4.6Results of direct hypotheses testing (direct effect)

Next, the bootstrapping procedure using 5,000 resampling was performed to generate the t-values in measuring the statistical significance of the path coefficients. Table 5 illustrates the path co-efficient assessment results informing that there are 8 direct relationships found to be significant. There are H1a, H2a, H2b, H3b, H4a, H5a, H5b and H6. This shows that Employee empowerment is positive affecting Job engagement (β= 0.451, t = 6.338 > 1.645) in H1a. Employee involvement is positively affecting Job engagement and Organizational engagement (β= 0.274, t = 3.329 > 1.645; β= 0.313, t = 4.499 > 1.645) in H2a and H2b. Interestingly, Employment stability only positively affecting Organizational engagement (β= 0.126, t = 2.693 > 1.645) as hypothesized in H3b but not in Job engagement. While Work-life balance is positively affecting Job engagement (β= 0.171, t = 2.561 > 1.645) as hypothesized in H4a but not in Organizational engagement. As for Working environment, it is positively affecting Job Engagement and Organizational Engagement (β= 0.181, t = 2.724 > 1.645; β= 0.101, t = 1.695 > 1.645) in H5a and H5b. Lastly, Job Engagement is found to be positively influencing Organizational Engagement (β= 0.367, t = 5.908 > 1.645) in H6. In short, out of the five dimensions of micro-level CSR practices, only two dimensions, namely Employee involvement and Working environment having significant positive effects on both Job engagement and Organizational engagement. Job engagement displayed significant positive influence on Organizational engagement.

4.7Results of indirect hypotheses testing (indirect effect)

Thereafter, we employed Sarstedt et al. [58] and Preacher and Hayes [70]’s approach in testing the five mediation hypotheses (H7, H8. H9, H10, and H11). We specified the indirect effect of the five dimensions of micro-level CSR practices on Organizational engagement via Job engagement. Again, bootstrapping procedure was conducted to test these mediation hypotheses by generating Biased Corrected and accelerated Confidence Interval (BCa CI). Preacher and Hayes (2008)’s approach for mediation analysis was utilized, whereby 0 signifies no mediation. Table 5 shows that there are four dimensions of micro-level CSR practices (i.e., Employee empowerment, Employee involvement, Working environment and Work-life balance) have positive significant indirect effects on Organizational engagement. This was ascertained through BCa CI with a lower level at 2.5% and an upper level at 97.5% in H7 [LL = 0.111, UL = 0.230], H8 [LL = 0.049, UL = 0.178], H10 [LL = –0.114, UL = –0.022], and H11 [LL = 0.021, UL = 0.110], do not straddle a zero value in between. Hence, H7, H8, H10 and H11 are supported indicating the presence of indirect. A comparable result is established through the variance accounted for (VAF) index [) to decide the size of the indirect effect. We obtained VAF values of larger than 20% and less than 80% for H7, H8, H10 and H11. With VAF values between 20% to 80%, typical partial mediations are observed for H7, H8, H10 and H11 [62].

However, when the Job engagement was introduced, Employment stability is no longer showing significance influence on Organizational engagement in H9. The result was confirmed through BCA CI with a lower level at 2.5% and an upper level at 97.5% with 95% [LL = –0.042, UL = 0.031] which there is a zero value straddles in between. This concludes of H9 is not supported.

4.8PLSpredicts

Following Shmueli et al. [71]’s recent research, PLS-SEM is regarded as a causal-predictive application. In essence, PLS predict works on the concepts of segregating training and holdout samples with the aim to estimate model parameters and evaluate model’s predictive power. With the symmetrical nature of the prediction error of the existing data, root mean squared error (RMSE) and mean absolute error (MAE) were used. Table 6 shows that PLS-SEM<LM for majority of the indicators in the PLS analysis. This is demonstrated in the lower RMSE and MAE values in comparison to the naïve LM benchmark. Besides, the Q² values for the indicators of PLS model performed better when compared to those generated for LM model (Q² > 0). Hence, the results indicate that current research model has a medium predictive power to represent reality.

5.1Theoretical implication

Our first theoretical contribution concerns the applicability of the Stakeholder Theory and the SET to micro-level CSR practices. While prior CSR research (e.g., Miller et al. [72]) has predominantly focused on external stakeholders for firm performance, our study signifies the benefits of focusing on internal stakeholders for firm survival and resilience during crises. With reference to the fundamental concepts of the Stakeholder Theory [25], the generation of value is the core driver of any firm, and this value is to be shared by a group of stakeholders as well as all actors in the society. The Stakeholder Theory further delineates the responsibility of organizations to their stakeholders. In the pandemic situation, it bodes well for organizations to discharge their responsibility to their employees, by implementing micro-level CSR practices. Our findings further reinforce the importance of the Stakeholder Theory in the interconnectivity between business and society. When organizations play their role pertinently through micro-level CSR, such as by ensuring employment stability, they cushion employees’ psychological stress. This is significant as Nordt et al. [73] reported the spike in suicide cases following the rise in the unemployment rate. Hence, it is also part of the governance process of corporate managers and organizations to consider the interest of employees in this critical context. The absence of micro-level CSR practices could potentially result in more severe social issues (e.g., rising suicidal cases and psychological distress) and unfavorable consequences for society. Our research findings provide empirical evidence on the outcome of micro-level CSR practices, i.e., engaged employees, which is much desired in times of crisis. The Stakeholder Theory works hand-in-hand with the SET by expanding the benefits of an engaged workforce from the organizational boundary to societal wellbeing. Using the principles of the Stakeholder Theory, the findings also extend the general perception of CSR to become more relevant to internal stakeholders and showcase the positive effects of micro-level CSR practices on organizations and society.

Our second theoretical implication lies in the detailed reciprocity mechanism of micro-level CSR and its consequences for organizations and employees. We utilized the SET concept and proposed micro-level CSR to create and uphold employees’ engagement in turbulent times. Yet, by examining employee engagement as a single concept, the implications derived would have been limited. As such, we carefully studied the mechanism of micro-level CSR through employees’ job engagement and organizational engagement. The results support the reciprocity approach of the SET. Specifically, when micro-level CSR practices (e.g., employee involvement, employee empowerment, working environment, and work-life balance) are implemented, employees’ job engagement is enhanced. However, for organizational engagement, micro-level CSR practices appear to take a slightly different path. That is, employee empowerment does not influence organizational engagement, while employee involvement, employment stability, working environment, and work-life balance positively affect organizational engagement. With this, we highlight the second theoretical contribution concerning the Engagement Theory as the reciprocity mechanism underlying micro-level CSR practices. Our findings expand on Kahn’s [30] engagement theory, which states that the precursors of job engagement and organizational engagement are not uniform but deviated. For instance, it is interesting to uncover that employment stability does not lead to job engagement but does lead to organizational engagement. Without such in-depth understanding and differentiation of these two types of engagement, organizations and managers may get an ‘A’ for effort yet fail to achieve intended results.

The third theoretical implication emphasizes the applicability of employee-centered CSR research in the unprecedented pandemic context. Despite employee-centered CSR research gaining a foothold (e.g., Low & Spong, [21], Jamali et al. [10]) in the late 2000 s, most studies were mainly conducted in the competitive business landscape. In contrast, the current study was carried out during a disastrous global pandemic. The empirical findings inform us that the execution of micro-level CSR practices does not lead to divergent impacts from the conventional business competitive setting. In fact, the inclusion of micro-level CSR practices is welcome as an internal enabler to weather through the external challenges. Inherently, micro-level CSR is a favorable strategy to be considered in various contexts.

5.2Managerial implication

Many human resource (HR) personnel and managers understand and realize the dampening effects of disengaged employees. However, Clark [74] observed that employees’ engagement levels are consistently low despite the best efforts of organizations. Our study offers some rational explanations for this phenomenon. We believe the HR personnel and managers would be greatly benefited from these findings. We adopted a multifaceted approach by separating employee engagement into job and organizational engagement in order to fully address motivation and engagement in the workplace. Drawing from Saks’s [23] conceptualization, job engagement refers to the extent of which an individual pays attention to the performance of his/her role and is absorbed in his/her work. Meanwhile, organizational engagement denotes an individual’s psychological role in an organization by doing his/her best. Organizational engagement is the emotional commitment employees have to their organizations, which is affected by factors such as organizational objectives, values, and beliefs [23]. When everyone involved is equally committed to the organization’s shared goals and productivity, businesses can reap significant advantages. This understanding guides our deduction that job engagement incorporates an individual’s enthusiasm towards his/her job [42] while organizational engagement delivers the shared values of the organization to employees.

Despite our findings showing a positive relationship between job engagement and organizational engagement, it is organizational engagement that predicts more positive outcomes like organizational commitment, job satisfaction, and organizational citizenship behaviors. For organizations and industries that require a cohesive workforce to stay afloat during the pandemic challenge, we would therefore recommend an emphasis on organizational engagement. Our results show that employee involvement, employment stability, and working environment are the micro-level CSR practices that nurture organizational engagement. The amalgamation of micro-level CSR and organizational engagement instils the organization’s shared values among employees by providing a motivational environment to work in [75]. Additionally, organizationally engaged employees are prepared and willing to devote their time and effort to complete their tasks to achieve organizational goals. This is much needed during an unprecedented and challenging crisis.

Job engagement delivers a different level of outcomes compared to organizational engagement. Since it represents individuals’ enthusiasm towards their job, it improves employees’ ability and performance, and in turn, foster their creativity. Based on our empirical findings, if organizations wish to bring employees’ job engagement to a higher level, micro-level CSR practices under cognitive and physical categories (e.g., employee empowerment, employee involvement, working environment, and work-life balance) would serve this purpose.

Our study highlights the importance for organizations to draw a distinction between the different types of engagement to facilitate targeted, evidence-based interventions that meet their desired outcomes more effectively. This study also offers practical implications to organizations and HR practitioners to vary the use of micro-level CSR practices to achieve the intended engagement outcome. The absence of a clear and distinct understanding in this regard would put most efforts in vain. Nevertheless, micro-level CSR is recommended on a large scale as it does produce an overall Matthew effect on organizations during a crisis.

5.3Methodological implication

This study adopted PLS-SEM owing to its appealing assumptions in the social sciences. Recent developments in PLS-SEM suggest that prediction-driven explanations are unconventional in the social sciences [71, 76]. Hence, framing managerial recommendations as prescriptive statements somewhat diminishes value and credibility. Kaplan [76] concurred that if researchers are unable to successfully predict the empirical outcome of a certain explanation, they will not have solid ground to accept the explanation presented in the managerial implication. Hence, the literature indicates a credential recommendation based on predication in evaluating theoretical falsifiability [83]. The predictive scenario posited herein is substantiated with the assessment of out-of-sample predictive power in PLSpredict analysis. Our predictive assessment results reveal that the proposed research model possessed medium predictive power in symbolizing reality. We confirm the achievement of our initial research objective to examine the influence of micro-level CSR on job engagement and organizational engagement during a crisis period.