Flexibility I-deals and prosocial motives: a trickle-down perspective

Abstract Growing concerns of maintaining the best talent have contributed to the rising number of idiosyncratic deals (i-deals) at the workplace. I-deals refer to the personalised work arrangements between employees and their employers where the terms benefit both parties. Despite the acknowledgment that supervisors are key in creating i-deals, research to date has overlooked their role. Drawing on prosocial motives and social learning theory, we explore an overall model of what triggers employee flexibility i-deals and the consequences of such i-deals on employee outcomes. In so doing, we explore one of the key yet untested assumptions of i-deals theory: that they are intended to be mutually beneficial. We investigate our model with matched supervisor – employee data (n = 186) collected in El Salvador and Chile. Findings reveal that there is a positive association between supervisors’ prosocial motives and employees’ flexibility i-deals. Moreover, prosocial motives of supervisors trickle-down and shape employees’ functioning at work (i.e. work performance and deviant behaviours) and lead them to be more prosocially motivated through employees’ flexibility i-deals.


Introduction
In recent years, there has been a dramatic increase in the prevalence of idiosyncratic deals (i-deals), which are defined as individually negotiated personalized employment arrangements between individual employees and their employers (Rousseau, 2005). The high volatility of markets (Benko & Weisberg, 2007;Greenhaus et al., 2010), lower technology costs, deregulation of employment relations (Leede et al., 2004) and intensifying competition for attracting and retaining talent have led organisations to adopt i-deals as a means for ensuring efficient use of human resources ). According to Liao et al. (2016), i-deals are mostly negotiated around five key areas: (i) career development (i.e. providing challenging work or training opportunities); (ii) task significance (i.e. altering job content to make it more satisfying); (iii) flexibility (i.e. working during a preferred schedule and/or from a preferred location); (iv) reduced work; and (v) customized financial arrangements.
Research to date on i-deals predominantly focuses on i-deals from the vantage point of the employees (Conway & Coyle-Shapiro, 2015). The limited attention given to the role of the supervisors neglects the dispositional conditions that make them more likely to grant flexibility i-deals. This is an important gap to fill in the literature for two main reasons. Firstly, supervisors are known to be gatekeepers of employees' use of flexibility (Kossek et al., 2016). Secondly, they often act as role models within organisations in which employees are encouraged to help others (Ambrose et al., 2013;Grant & Patil, 2012). Thus, we seek to contribute new insights to the i-deals literature by examining the relationship between the prosocial motives of supervisors, which are defined as the desire to benefit others, (Bolino & Grant, 2016) and flexibility i-deals. Specifically, we propose that when supervisors are motivated to help others, they are more likely to grant flexibility i-deals to their employees. Our focus on flexibility i-deals is inspired by a recent growth in studies on this unique type of i-deals (e.g. Las Wang et al., 2019) and their indispensable role in the ever-changing workspace, where flexibility regarding where and when one works will gain significant momentum in the coming decades.
Our second goal is to explore how prosocial motives of supervisors trickle-down to impact employee outcomes through flexibility i-deals. To explain the trickle-down effect of prosocial motives, we use social learning theory (Bandura & Walters, 1977;Bandura, 1986) and the reciprocity principle from the social exchange theory (Blau, 1964). We argue that employees who are provided flexible i-deals will perceive this as a result of their supervisors' prosocial motives and will reciprocate by engaging in similar positive behavioural patterns, as these employees will perceive such supervisors as credible role models (Lemoine et al. 2019). Thus, we expect that employees will acquire similar prosocial motives through role modelling and reciprocate by displaying prosocial motives to others.
Prior research suggests that flexible work arrangements are associated with improved individual effectiveness within the work domain (Guerrero et al., 2014;Las Heras et al., 2017). Robinson and Bennett (1995) suggest it is important to address employees' desirable behaviours, such as work performance, but it is also equally important to approach the dark side of employees' behaviours that produce organisational losses, such as deviant behaviours (Kelly et al., 2020). To address this recognised need, in this study we explore the associations between employee flexibility i-deals and two key work outcomes: work performance and employees' deviant behaviour.
The contributions of this study are twofold. Drawing on social learning and social exchange theories, first, we seek to explore the role of a dispositional condition that enables i-deals creation, namely supervisors' prosocial motives. Social learning theory is a core, yet often overlooked mechanism that is relevant to understanding i-deal creation (e.g. Stollberger et al., 2020). The theory suggests that individuals tend to observe and model behaviours of positive leadership, as such leaders are attractive and trustworthy. In a similar vein, we contend that social learning and social exchange are relevant theoretical lenses that can explain the dyadic processes related to prosocial motives and how they may drive better functioning at work (Lemoine et al., 2019). Such a perspective will be one of the few steps to explore a key yet untested assumption of how supervisors grant flexibility i-deals and whether these deals generate mutually beneficial outcomes for both parties involved in this process.
Second, we explore the role of flexibility i-deals as an explanatory mechanism that links the supervisor and employee prosocial motives. Given the important role of supervisors within organisations (Grant & Patil, 2012), we posit that employees who are provided with flexible i-deals will perceive this as a result of their supervisors' prosocial motives and ultimately become more willing to reciprocate similar behaviours. In doing so, we contribute to the better understanding of the flexibility i-deal creation process, namely the fabric of prosociality interwoven in this process, and to recent debates where it has been demonstrated that supervisors are likely to grant i-deals to employees who engage in socially connecting behaviours towards their co-workers (e.g. Rofcanin et al., 2017 for a study on how supervisors grant task i-deals to employees). We test our conceptual model, which is seen in Figure 1, in the under-studied contexts of Chile and El Salvador. In the following section, we develop our hypotheses.

I-Deals theory
Idiosyncratic deals are personalised agreements of a non-standard nature which individual employees seek out and negotiate with their employers (Rousseau, 2005). These arrangements are intended to benefit both employee and employer . The scope of i-deals include a range of flexible arrangements (e.g. flexible work hours) such as remuneration, development opportunities, flexibility and job content (Rosen et al., 2013;Rousseau, 2005). A recent review on i-deals (Liao et al., 2016) underlines that the provision of i-deals is beneficial for employees as recipients and for the organisations. The underlying assumption of i-deals theory has been that changes in the external environment, such as the increased competition for talented employees and individuals differing needs, has led employers to exhibit greater initiative to their employees while negotiating personal work arrangements (Liao et al., 2016). From this perspective, it becomes imperative to understand what triggers i-deals to generate the maximum benefits for both the employee and the management in organisations.
Flexibility i-deals are becoming increasingly common across organisations (Rosen et al. 2013;Las Heras et al., 2017). As boundaries of work and life have become increasingly fluid and blurred, flexibility i-deals have gained increasing momentum and research attention because they provide employees with discretion over where and how to work Wang et al., 2019). Furthermore, a growing body of research has begun to explore the antecedents of flexibility i-deals, mainly focusing on employee, leader and structural/contextual characteristics. Among the antecedents at employee level, employees' personal initiative, achievement and status striving, political skills, tenure and firm-specific human capital have been shown to be positively related to flexibility i-deals Lee et al., 2015;Ng & Lucianetti, 2016;Rosen et al., 2013). Furthermore, recent research has been increasingly interested in the influence of leaders and managers in the context of i-deals. For example, a high-quality leader-member exchange relationship is consistently shown to increase the chances of obtaining flexibility i-deals at manager level (Hornung et al., 2014;Rosen et al., 2013). Also, structural/contextual characteristics (e.g. the internal career management system, team heterogeneity and size) were found to influence the room and necessity for such i-deal negotiations (Bal, 2017;Hornung et al., 2009;Lee et al., 2015). In relation to the above, we expand the debate on the role of the supervisor by examining how supervisors' prosocial motivation may influence employees initiating an i-deal. Furthermore, we respond to calls for research (e.g. Vidyarthi et al., 2014;Las Heras et al., 2017) and study specific types of i-deals by exploring a trickle-down model in which supervisors' prosocial motivation impacts employees' prosocial motivation and work outcomes via flexibility i-deals. Here, we propose that flexibility i-deals are mechanisms that explain how this positive trickle-down effect occurs between supervisors and employees.

Supervisor prosocial motives, flexibility i-Deals and employee outcomes
I-deal negotiations unfold within a dyadic pattern of relationships in which the end result of the negotiation is likely to depend on the supervisors' granting authority of this deal. Supervisors play a critical role in granting flexibility i-deals because they are usually equipped with the resources to render them effective and establish a pattern of relations where other employees do not feel alienated as a result of flexibility i-deals (Bal & Rousseau, 2015). Furthermore, as part of leader-follower interactions, supervisors often act as role models and provide a foundation for employees in terms of providing cues for acceptable behaviours and norms within the organisation (Ambrose et al., 2013;Bandura & Walters, 1977).
Prosocial motives are described as the willingness to help co-workers and the desire to create a positive impact on others through one's own work (Grant, 2008). However, the evidence on the relationship between prosocial motives and work-related outcomes are mixed. Previous studies highlight that prosocial motives are positively related to individual-level outcomes such as creativity, as prosocial motivation may encourage employees to take the perspective of third parties and support them develop novel ideas (Grant & Berry, 2011). Furthermore, studies have shown that prosocially motivated employees have less inflated self-assessments (Meglino & Korsgaard, 2004), avoid complacency (Grant & Wrzesniewski, 2010), and react more constructively to unfair events (Bobocel, 2013).
Nevertheless, other studies have shown that there may be significant drawbacks of prosocial motivation, as it requires substantial sacrifice of individuals' own resources, mainly in the form of time and energy to help co-workers (Bolino & Klotz 2015;Rofcanin et al., 2018). Thus, the negative and positive experiences related to prosocial motives hint that prosocial motivation requires further exploration (Kelly et al., 2020).
Furthermore, despite acknowledging that supervisors are key in the process of i-deals creation, it is only recently that researchers have started to explore the role of supervisors in such process (e.g. Rofcanin et al., 2017). Among the few studies which investigate the antecedents at the supervisory level, Hornung et al. (2011) show that when a supervisor is perceived as considerate, they are more likely to grant i-deals. Thus, drawing upon i-deals theory and prosocial motives research, we expect that there will be a positive association between the prosocial motives of the supervisor and the creation of flexibility i-deals for the employees. This is because prosocially motivated supervisors will see granting flexibility i-deals as an opportunity to support employees achieve work-life balance (Bolino & Grant, 2016), as well as working to the benefit of the organisation. Recent research on i-deals has shown that caring for the well-being of co-workers and sharing the benefits of i-deals with them influences the success of an i-deal creation process .
Our first hypothesis is: H1. There is a positive association between supervisors' prosocial motives (measured by the supervisor) and employees' flexibility i-deals.
Flexibility i-deals are opportunities designed to benefit employees by balancing their work and non-work demands (Bal & Rousseau 2015). Given the significant role of the supervisor within organisations, the obtainment of flexibility i-deals by employees most likely depends on the prosocial motives of the supervisors. As such, it has been demonstrated that supervisors are more likely to grant i-deals to employees who demonstrate care and concern for others (e.g. Rofcanin et al., 2017). Drawing on this tenet of i-deals theory, we turn to the role of supervisors and explore the impact of their prosocial motives on employee flexibility i-deals. The next question is then as follows: What is the impact of flexibility i-deals on employee prosocial motives at work?
Research suggests that while processing information, individuals either tend to use a central route, where they engage in an elaborative cognitive process, such as social learning and social exchange or a peripheral route, where they respond with their feelings or heuristic cues (Wo et al. 2019). Social learning theory proposes that people learn from observing or modelling others (Bandura, 1986). By observing others, an employee can learn the acceptable norm and will act accordingly (Bakker et al., 2016). According to social learning theory (Bandura & Walters, 1977;Bandura, 1986;Gibson, 2004), observational learning occurs through four steps: (i) Attention: In order to learn through observation, the person first must pay attention to the model. The attention can be influenced by a number of factors including the status of the model and the nature of the interactions between the individual and the model; (ii) Retention: The person should be able to remember the behaviour that has been observed; (iii) Behaviour production: The person should be able to replicate the behaviour that the model has demonstrated; and (iv) Motivation: People are more likely to demonstrate and adopt a modelled behaviour if they expect this behaviour will create positive outcomes. Drawing upon social learning theory, we expect that employees will perceive supervisors as role models in their work environment and will be inclined to adopt their behaviours. Receiving i-deals will trigger an individual to think about the benefit received and the appropriate means of reciprocation, reflecting the central route of information processing (Wo et al., 2019).
Indirectly supporting our arguments, the findings in Stollberger et al. (2020) demonstrate a positive association between servant leadership of supervisors and moral behaviours of employees. This finding is relevant as it supports the notion of how concern and care for others trickles-down to shape the behaviours of employees. A recent review study on moral leadership also demonstrates that the positive impact of moral leadership trickles-down to employees via helping and concern for others, which the authors define as prosociality (Lemoine et al., 2019). Drawing on this research, we expect that receiving flexible i-deals may trigger a cognitive approach of considering the relationship with the supervisor and the required obligations and responses toward third parties.
Social exchange theory postulates that in social exchange relationships, actions of one party serve as a 'staring mechanism' , facilitating reciprocal actions from the other party (Blau, 1964). According to social exchange theory, when individuals receive a favour, the norm of reciprocity will propel them to return the favour, due to the expectation that continuing this pattern in the relationship will bring additional valued benefits (Blau, 1964) This reciprocity element has also been used in i-deals research as the main motive behind an employer's choice to grant i-deals (Anand et al., 2010;Vidyarthi et al., 2014). Furthermore, trickle-down research suggests that supervisors who receive positive behaviours such as fair treatment will feel obligated to reciprocate such behaviour to employees to discharge the obligations and repay the source (i.e. the manager or organisation) (Stollberger et al., 2020;Wo et al., 2019).
There is evidence that underpins that employees are able to infer others' motives (Maierhofer et al., 2000). This is mainly because (i) motives are observable through verbal statements and behavioural patterns and (ii) by working together, employees are able to observe supervisors' patterns of behaviour across a range of situations, which increases employees' perceptions of trust and engagement with their supervisors (Kenrick & Funder, 1988). Indeed, we reason that receiving flexibility i-deals can serve as a display of a supervisor's prosocial motives, a type of motivation that is also supported in a given organisation (Bandura & Walters, 1977). Thus, we contend that employees who observe such motives from their supervisor will feel inclined to adopt and reciprocate similar motives to help others.
Based on these premises of social learning (Bandura, 1986) and social exchange (Blau, 1964) theories, we contend that, supervisors' authorisation of flexibility i-deals will have significant value for employees, as it clearly demonstrates that helping others (e.g. colleagues) is an accepted and legitimate norm within the organisation. Furthermore, we expect that employees will perceive receiving flexible i-deals as a favour and a basis for reciprocity, propelling them to become motivated to help others.
We thus argue that supervisors' prosocial motives will have a positive impact on employees' prosocial motives via its impact on flexibility i-deals which become a mechanism to support and realise this transition effect. Our second hypothesis is: H2. Employee flexibility i-deals mediate the positive association between supervisors' and employees' prosocial motives. We extend our model by delineating the role of employees' flexibility i-deals as a mechanism for explaining how supervisors' prosocial motives may impact employees' work performance and deviant behaviours. By doing so, we explore one of the key propositions of i-deals theory, that is, whether i-deals are beneficial for both employers and employees. Initiatives that facilitate work-life integration are likely to have a positive impact on employees' perception of their work environment, which in turn improves individual and organisational outcomes (Kelly et al., 2008). Drawing on social learning and exchange theories, we argue that flexibility i-deal creation may act as a bridging mechanism that will reinforce positive behaviour and reduce negative outcomes, such as individual level deviance.
In a growing body of research, relatively few studies have focused on flexibility i-deals and their impact on employee outcomes. Among these, the findings in Las  show that flexibility i-deals improve work performance of employees by enabling these employees to engage with their family domain effectively. The findings in Vidyarthi et al. (2014) support the notion that flexibility i-deals enable employees to conduct their work in their own discretion, which is likely to impact their career satisfaction positively. Focusing on flexibility i-deals, Wang et al. (2019) demonstrate that the recipients of flexibility i-deals experience less conflict and more enrichment at home. These studies demonstrate that receiving flexibility in terms of where and when an employee works is likely to impact employee work performance positively. Furthermore, little research exists on the potential role of i-deals on employee negative outcomes, such as deviant behaviours (Liao et al., 2016, Kelly et al., 2020. A recent study by Kelly et al. (2020) focuses on schedule flexibility i-deals, and shows that the recipients of such deals are less likely to engage in deviant behaviours. This is because the provision of schedule-related flexibilities facilitates the employee's balance of competing demands arising from work and home simultaneously.
Building on these recent studies, the theoretical foundations of i-deals, social learning theory and the norm of reciprocity of social exchange theory, we further propose that flexibility i-deals of employees are likely to impact employee work performance positively. Considering the supervisor as a credible role model, we expect the employee will feel obligated to adhere to this norm and reciprocate by demonstrating prosocial motives, enhanced work performance and less deviant behaviour to third parties.
In sum, we propose that the positive impact of supervisors' prosocial motives on employees' work outcomes depends on the obtainment of flexibility i-ideals. Our third hypothesis is: H3 (a) Employees' flexibility i-deals are positively associated with their work performance and negatively associated with their deviant work behaviours; (b) employees' flexibility i-deals mediate the association between supervisors' prosocial motives and employees' work outcomes (i.e., work performance and deviant behaviours).

Context and procedure
We investigated i-deals in the under-studied context of Chile and El Salvador. Most studies of i-deals have been conducted in English-speaking countries (Anand et al., 2010) and European contexts (Hornung et al., 2010;Ng & Feldman, 2015), but studying i-deals in other contexts is important in order to determine their potential relevance under different conditions (Las Heras et al., 2015).
The economies of Chile and El Salvador reflect macro-economic volatility and, as a result, companies face unexpected shocks (Ross, 1999;Frankel, 2010). Research has shown that both Chile and El Salvador have suffered from resource curse effects, by which exploitation of a specific resource leads to temporary wealth at first, but in the long run is associated with social and economic challenges (Arrowsmith & Parker, 2020). These challenges include, shortages of skills, innovation and other capabilities (Frankel, 2010). Moreover, the business contexts in both countries are characterized by non-market hierarchical relations in terms of capital and technology allocations support and labour regulation. In addition, there is a tendency to under-invest in skills and research capabilities. Thus, it can be argued that the business contexts in these two under-researched countries provide an interesting platform to explore the prevalence of i-deals.
The participants in this study were full-time employees of two large companies. We accessed the companies in El Salvador through non-academic partners in the country, and through one of the co-authors in Chile. UTINORTH is a utility company in El Salvador. It is publicly listed and employs over 700 employees across the country. FINAN is a financial company operating in Chile. It is a private company and employs over 600 employees, mainly in Santiago de Chile.
Companies participating in the project benefited from the research by receiving an in-depth company-specific executive report. We conducted power analyses to determine the final sample size (Ellis, 2010). Following the official consents of the companies, we collected data from employees and their supervisors. Specifically, employees evaluated supervisors' prosocial motives, their own prosocial motives and flexibility i-deals. Supervisors evaluated the work outcomes of employees.
The average age of employees was 39.2 years (SD = 9.8 years), and 42 percent were male. Most had undergraduate degrees (57 percent), while 24 percent had postgraduate degrees, 12 percent had other types of degrees and six percent had high school qualifications. The average age of supervisors was 39 years (SD = 8.1 years), and 58 percent were male. Most supervisors had undergraduate degrees (58 percent), 40 percent had postgraduate degrees, and 2 percent had other types of degrees. Before the study began, the company supervisors and employees were briefed regarding the purpose, procedure and confidentiality of the study.
We used online surveys and granted all participants strict confidentiality, so that only the researchers had access to their responses, which were kept in a server in Canada. We back-translated the survey items to increase face validity (Brislin, 1986;Prieto, 1992). Using online tools, at separate times, we collected data from employees and their supervisors, who evaluated employees' work outcomes. We used e-mails as IDs to match the data from the employees and their direct supervisors. We invited 423 employees to participate in the study as employees, and obtained 201 fully usable responses (48 percent). We invited 143 employees to participate as supervisors, and obtained 76 responses (53 percent). Due to missing data, our final sample included a matched data of 186 employees and 59 supervisors.

Measures
Unless otherwise stated, all items were measured on a 7-point Likert scale (from 1 = strongly disagree to 7 = strongly agree).

Prosocial motives
Using a four-item scale by Grant (2008), employees evaluated supervisors' and their own prosocial motives. Employees were prompted to say what motivates them to work. An example item for them is 'Because I care about benefiting my co-workers through my work' 'α = .93, and they were asked 'In your opinion: What motivates your supervisor at work?' . One sample item includes perceptions of supervisors' prosocial motives. It is as follows: 'In my opinion, my supervisor wants to benefit others through his/her work' , α = .90.

Flexibility I-deals
We used the five items of flexibility i-deals to measure the degree to which focal employees received flexibilities in their schedules (3 items) and location (2 items) that were different from those of their co-workers (Rosen et al., 2013). An example item was: ' At my request, my supervisor has accommodated my off-the-job demands when assigning my work hours' (α = .76).

Work performance
Supervisors evaluated the performance of each employee using four items from a scale by Williams and Anderson (1991). We selected these four items to measure the in-role work performance of focal employees. In particular, these items evaluate the extent to which employees meet the expectations of supervisors as well as the job (ranging from 1 = below average to 7 = above average). One example was: 'He/she meets the formal performance requirements of the job' (α = .89).

Deviant behaviours
Supervisors evaluated the deviant behaviours of focal employees using the items that refer to harmful behaviours toward the organisation developed by Bennett and Robinson (2000). The original scale was developed to measure self-reported accounts, yet we adapted it for supervisors to report employees' behaviours. The supervisor is asked to 'indicate to what extent this employee has engaged in each of the following behaviours in the last year' A sample items is 'Taken additional or longer breaks than it is acceptable' and 'Neglected to follow my (or other bosses') instructions' . The scale is 1 = Never 4 = Frequently 7 = Daily. (α = .92).

Analytical strategy
We tested all hypotheses simultaneously by conducting a path analysis using Mplus 8.3 (Muthén & Muthén, 1998 using the robust maximum-likelihood estimator (MLR). To test the mediation hypotheses, we adopted the bootstrapping approach (using 10,000 bootstrap samples) recommended by Preacher and Hayes (2008). Moreover, in order to determine the distinctiveness of all variables, we conducted confirmatory factor analysis using M-plus 8.3.
We followed recommendations to minimize common method bias in the design of our study (e.g. Podsakoff et al., 2003), assuring participants that their responses would be treated confidentially, randomized items within question blocks, separated independent and moderator variables in the survey and used different response scales for different variables. Moreover, in line with suggestions (Podsakoff et al., 2012) and recent research (e.g. Bal et al., 2012) we conducted various statistical analyses (Lindell & Whitney, 2001) and results demonstrated that CMB was not an issue in our analyses.
Supervisors evaluated employee work outcomes of performance and deviant behaviours. To control for the nested structure of our data, we applied multi-level regression analyses using MLwiN software (Hox, 2002). To determine whether multi-level analysis was appropriate, we calculated ICC (1) to account for the proportion of the total variance attributable to differences between supervisors (Level 2): The ICC (1) for work performance was 78 percent and 32 percent for deviant behaviours. Our results therefore supported the use of multi-level regression analysis. To adequately control for both within-group and between-group variances, we used grand-mean centred estimates for all Level 1 predictors, and unit-level mean-centred estimates for all Level 2 predictors (Raudenbush & Bryk, 2002). To test our mediation hypothesis, consistent with recent research on multi-level mediation analysis (e.g. MacKinnon & Fairchild, 2009;Preacher, 2015), we used the Monte Carlo method for assessing mediation (MCMAM). 1 We used an online tool developed by Selig and Preacher (2008) to calculate confidence intervals. When confidence intervals do not contain zero, the indirect association is significant. Table 1 displays the means, standard deviations, correlations and internal reliability values of our study variables. We ran several multi-level CFAs to test the factorial validity of our proposed conceptual model. The hypothesised five-factor model (including supervisor prosocial motives, employee flexibility i-deals, employee prosocial motives, employee work performance, and employee deviant behaviours) demonstrated a satisfactory fit to the data: χ2 ( Hypothesis 1 postulated that there is a positive association between supervisors' prosocial motives and employees' flexibility i-deals. Results supported this hypothesis: β = .43, S.E. = .09, p < .001; Table 2). Hypothesis 2 postulated that employees' flexibility i-deals would mediate the positive association between supervisors' and employees' prosocial motives. Bootstrapping with 20,000 iterations supported this hypothesis, as the confidence intervals did not include the value of zero (95% CI [0.01, 0.14]), (see the footnotes in Table 2). Hypothesis 3 (a) proposed that employees' flexibility i-deals are positively associated with employees' work performance and negatively associated with employees' deviant behaviours. Results supported this hypothesis (β = .19, S.E. = .04, p < .001; See Table 2 for work performance; (β = −.17, S.E. = .04, p < .001; Table 2 for deviant behaviours). Hypothesis 3 (b) proposed that employees' flexibility i-deals would mediate the association between supervisors' prosocial motives and employee work outcomes. Findings supported the role of employees' flexibility i-deals as a mechanism linking supervisors' prosocial motives to employees' work performance (at 90% CI [0.01, 0.08]) is 95% CI [0, 0.09]) and their deviant behaviours (95% CI [-0.1, −0.01]). We also re-ran and tested the overall model as a post-hoc analysis to see if the model holds true when all the constructs are in the model. Figure 2 demonstrates the results of the path analyses and offers support for the overall paths, which are significant.

Results
We ran several alternative models to test and support the validity of our proposed associations. In alternative model 1, we tested a model The indirect effect is calculated using an online interactive tool that generates an r score: http://quantpsy. org/medmc/medmc.htm. The first path of the indirect relationship relates to the association between supervisors' prosocial motives and employees' flexibility i-deals (0.25) and the second path of the indirect relationship relates to the association between employees' flexibility i-deals and outcomes (0.19 for work performance; -0.12 for deviant behaviours; 0.15 for prosocial motives) when supervisors' prosocial motives is present in the equations. for all values, gamma coefficients, their corresponding standard error values are reported. n = 186 employees; 59 supervisors. *p < 0.05; **p < 0.01; ***p < 0.001.
where flexibility i-deals of employees lead to enhanced work performance, decreased deviant behaviours, and improved prosocial motivation of employees. We argued that these outcomes then increased supervisor prosocial motivation. Only the associations between employee flexibility i-deals and work outcomes were significant (β = .09, S.E. = .03, p < .001; for work performance; (β = −.11, S.E. = .05, p < .001; Table 2 for deviant behaviours) and the rest of the proposed associations were not significant. In alternative model 2, we tested a path model in which supervisor prosocial motivation was associated with employee work outcomes and prosocial motivation, which then led to employee flexibility i-deals. This model did not converge, leading to inconclusive findings.

Discussion
Global trends including competition and individualised careers have necessitated an increasing need for i-deals (Bal & Dorenbosch, 2015;Call et al., 2015). We can observe this need, for instance, in the decrease of collective bargaining in many countries across the globe, such as in the U.S. (U.S. Bureau of Labour Statistics) and U.K. (Hoque & Bacon, 2014). Informed by recently growing research on flexibility i-deals and drawing from social learning and social exchange theories, we aimed to understand (i) the role of supervisors' prosocial motives as antecedents to employees' flexibility i-deals, (ii) the role of employees' flexibility i-deals as mechanisms for explaining how supervisors' prosocial motives trickle-down to shape employee outcomes of work performance, deviant behaviours and prosocial motives. We discuss our theoretical contributions below.
Our first contribution relates to our focus on the role of the supervisors' dispositional factor (particularly their prosocial motive) as an enabler of flexibility i-deals' creation for their employees. The definition of i-deals emphasises that these deals unfold between an employee and their employer, the latter of which is usually represented by a supervisor (Rousseau, 2005). Although previous literature has emphasised that i-deals unfold in a dyadic relationship involving employees and their supervisors (Bal & Rousseau, 2015), only recently have researchers started to pay attention to the perspective of supervisors regarding the employee behaviour (e.g. . Our understanding of what triggers supervisors to grant flexibility i-deals to their employees is limited. This is an important gap given that supervisors are ultimately responsible for the implementation of i-deals, as they have the power and the resources to either grant them or not. Furthermore, these supervisors are often aware of the unique work needs of their employees, putting them in a unique position to authorise these deals and manage the implications and consequences in a team setting. Contributing to this gap, our findings reveal that supervisors' prosocial motives are positively associated with flexibility i-deals of employees. This finding enhances a recent discussion on the underlying motives of why supervisors grant i-deals and broadly contributes to debates on the motives behind i-deal negotiation: In his conceptual model, Bal (2017) demonstrates that one of the key motives describing why employees negotiate i-deals is to generate prosocial impact. In a subsequent study, Bal and Vossaert (2019) develop an i-deals motivation perspective to show how this negotiation unfolds. Our contribution lies in re-focusing on the role of supervisors' prosocial motives in realising these deals.
Secondly, our study contributes to the accumulated research that examines the trickle-down effect of supervisor dispositions on employee outcomes (Ambrose et al., 2013). In particular, our findings demonstrate evidence for the trickle-down effect driven by flexibility ideal creation. We proposed that supervisors who are more prosocially motivated are more likely to grant flexibility i-deals to their employees. Employees in turn, may reflect a similar pattern of prosociality and helping behaviours. This finding is consistent with our theorising based on social learning and social exchange theories, as employees are likely to adapt and reciprocate similar motives that are in line with what they observe through their social interactions with their supervisors. In other words, supervisors' authorisation of flexible i-deals will set an example and show employees that helping others is the social norm of the organisational culture and will facilitate employees to reciprocate accordingly. Our findings contribute to recent research that explores a similar phenomenon in the context of studying trickle-down models. The findings in Stollberger et al. (2020) show that employees look up to and emulate their moral leaders and engage in prosocial behaviours and demonstrate prosocial motivation. The results of the current study add to this research by integrating flexibility i-deals as a potential mechanism in moral leadership literature which could explain how and why this trickle-down model occurs. Additional research that focuses on trickle-down models of i-deals reveals that a similar perspective of studying trickle-down models is needed in research pertaining to flexibility i-deals (Rofcanin et al., 2018). Put together, despite the recent burgeoning stream of trickle-down studies (Wo et al., 2015), to our knowledge this is the first study to integrate the role of prosocial motives of supervisors and investigate the trickle-down effect on employees driven by flexibility i-deals as an underpinning mechanism.
While prior research has revealed that supervisors are likely to facilitate the provision of task i-deals to employees who exhibit socially connecting behaviours such as helping co-workers , for flexibility i-deals, this may not be the case, as the content of flexibility i-deals concerns the timing and location of where one works and is more relevant to focal employees' non-work lives (Bal & Rousseau, 2015). Echoing the findings of most recent research on flexibility i-deals (e.g. Las Wang et al., 2019) and theory (Bal & Rousseau, 2015), when deciding to grant flexibility i-deals, supervisors may focus more on the non-work needs of the focal employee, such as achieving a better balance between work and home and less on how the focal employee is likely to use these i-deals at work and the implications of these types of i-deals in the workplace as a whole. We thus recommend future studies unpack the impact of trickle-down models of flexibility i-deals in their relevance to the family domain by focusing on family-related outcomes such as family engagement and functioning.
As a further contribution, our findings support that flexibility i-deals are positively associated with enhanced work performance and reduced deviant behaviours, both of which are supervisor evaluated. Prior research on i-deals tends to focus on employee outcomes that are self-rated and this could be an issue from the perspective of creating and boosting bias in one's own evaluation of employee outcomes. With regards to deviant behaviours, this is especially critical as, due to social desirability bias, employees might refrain from reporting the extent to which they engage in deviant behaviours (e.g. Rioux & Penner, 2001).
The findings regarding the impact on employee outcomes are statistically significant when all the paths of our proposed model are run simultaneously. Regarding the former, we add to current understanding of how flexibility i-deals, which are mainly granted to increase family performance, may indeed be beneficial for employees' work domain outcomes (Wang et al., 2019). We use social learning and social exchange theories to argue that employees granted flexibility i-deals via supervisor authorisation will learn by observing firsthand that helping others is the acceptable social norm in the organisation, and therefore will be motivated to reciprocate and help others.
To our knowledge, our findings represent the first attempt to explore the extent to which flexibility i-deals relate to fewer behaviours that are harmful to the organisation. A recent study on schedule flexibility i-deals (a specific sub-type of flexibility) demonstrated that such deals reduce deviant behaviours targeted at organisations. Our study is likely to expand these findings by adopting a broader measurement approach of flexibility i-deals (e.g. both location and schedule flexibility are measured) and the impact of deviant behaviours targeted at the individual as well as the organisation. Future research should further explore the influence of flexibility i-deals' spillover on the non-work domain and then back to work performance. It might be that flexibility i-deals enable employees' family engagement and work-life satisfaction, and that, in turn, enables better work performance. Another interesting area will be to explore whether and why the recipients of flexibility i-deals actually engage in socially connecting behaviours versus disconnecting behaviours towards their co-workers and the implications these behaviours have for i-deals theory.
This research also contributes to debates in individualization of HRM practices. The focus of strategic HRM has been on team or organisational levels (Datta et al., 2005). Only a few recent studies have adopted a cross-level approach to explore the effects of macro-level (e.g. team or organisational) HR practices on employee outcomes (e.g. Kehoe & Wright, 2013;Snape & Redman, 2010). Therefore, the essence of differentiated HR practices, where the focus is placed on the employee, has been overlooked. From this perspective, i-deals are examples of differentiated HR practices. Therefore, this study contributes to research on HR differentiation in two ways. First, by highlighting its positive effects on employees' work outcomes and their prosocial motives and second, by shedding light on the circumstances which deals are granted to focal employees through exploring the role of employees' and supervisors' perceptions of their prosocial motives.
The findings of this research also contribute to the debate on the contingency approach of HRM (Kaufman & Miller, 2011;Marescaux, De Winne, & Sels, 2019). This debate mainly revolves around whether more HR practices associate with better employee work performance, or whether employees perform better when they are provided with individualized HR practices, such as flexibility i-deals. While we cannot compare these two cases, our results reveal that flexibility i-deals lead to better work performance, fewer deviant behaviours, and increased prosocial motives. From this perspective, the findings of this study expand recent research on individualised HRM that adopts a contingency angle. For example, Bal and Dorenbosch (2015) show that the effect of using individualised HR practices on performance and turnover depends on employees' age. Similarly, Clinton and Guest (2013) reveal that the effects of differentiated HR practices vary across different job groups. To our knowledge, this is the first study to go beyond individual-and team-level contingencies to understand when differentiated and individualized HR practices may be effective and beneficial.

Limitations and suggestions for further research
Despite its strengths, some limitations of this study must be noted. The first limitation relates to the cross-sectional design of the study, which prevents us from drawing causal inferences from the variables we researched. We suggest the use of a longitudinal design for future research, with a pre-determined time lag between each variable (e.g. six months to a year, which would be consistent with research on i-deals; Ng & Feldman, 2015). This would enable researchers to explore the processual nature of the proposed model in sequence.
In relation to the point above, researchers might consider a within-person design to explore the effects of flexibility i-deals. For example, future research might explore the effects of flexibility i-deals on family and work performance on a weekly basis. Such a design would be a novel approach, as recent conceptual discussions on i-deals have emphasised that employees may strike micro i-deals that might show variation over time (Rofcanin & Stollberger, 2015). This might be particularly interesting for flexibility i-deals that vary from week to week.
A third limitation relates to the sample context. The unique cultural contexts of El Salvador and Chile might have impeded the extent to which it is part of the norm for employees to approach their supervisors to negotiate for individualised deals. For example, paternalism is a predominant cultural value in El Salvador and Chilean business settings (House et al., 2004), and it is therefore expected from supervisors to care about and show concern for their employees' family lives. This might explain the prevalence of flexibility i-deals in this context. Given the absence of a culture-related variable in this study, it is difficult to attribute differences in the findings to cultural contexts. Hence, it will be interesting to incorporate the impact of culture, particularly from an under-studied context, as research on i-deals has been predominantly carried out in the Western context.
The focus of this study was on flexibility i-deals. Due to the different nature of types of i-deals, such as work responsibility and task i-deals, future studies might explore the extent to which the same findings can be observed for other types of i-deals. In relation to our focus on flexibility i-deals, it might be that the availability of the flexible work practices in general (i.e. FWPs) might influence the extent to which flexibility i-deals prevail in these companies.
Measurement of prosocial motives has been adopted and utilised in research to date. While its use has been empirically validated, we recommend future studies to adopt a more fine-tuned and nuanced approach of measuring prosociality at work. This measure is likely to remove the social desirability biases inherent in the measurement developed by Grant (2007) and is likely to bring a fresh set of eyes on this growing body of research.
We focused on work performance and deviant behaviours as employee outcomes. Future research on flexibility i-deals could explore more proximal outcomes of employees in the family domain, including family engagement and functioning as well as enrichment and conflict.
Our model explored the mediating role of flexibility i-deals, which implies a change in the outcome variables of interest. A convenient way of analysing this model would be by adopting a latent curve analysis. Our data did not allow for this type of analysis to be carried out and we recommend further research to explore similar research questions by adopting a change-modelling approach.

Practical implications
The key implication of the findings is that the implementation of individualised HR practices, such as early leave and flexitime, to address employees' unique needs and work preferences should be supported by human resources departments. This implementation needs senior management involvement, commitment and the establishment of related policies, practices and procedures at the HR level. As such, flexibility i-deals may constitute important parts of HR strategy of organisations; however, contextual conditions must be considered when designing and implementing them. Previous research has found perceived leader support to be a key determinant for individualised HRM practices (Behson, 2005). Companies should foster supportive environments, for example, by facilitating employees to leave when they need to attend non-workrelated issues and emphasizing the role of prosocially oriented behaviours (Thompson et al., 1999). By providing support, organisations might also avoid negative attributions from other employees, keeping employees motivated and committed to the organisation (Cook, 2009). In order to maximize the effectiveness of i-deals and receive supervisors' support, we suggest that senior management and HR units should (1) invest in and carry out periodic interventions aimed at evaluating employees' resources and needs; (2) help employees fine-tune their job demands and resources, for example, by giving them more autonomy to balance family and work demands; and (3) at the individual level, provide employees with individualized support through mentoring and coaching to help them optimize their use flexibility i-deals.

Note
1. This approach relies on the parameter estimates and their associated asymptotic variances and co-variances. In particular, this method draws randomly from the joint distributions of the parameter estimates, calculates the product value of the two parameter estimates and repeats this a very large number of times. In the end, a confidence interval is estimated to test indirect associations (Bauer et al., 2006).

Disclosure statement
No potential conflict of interest was reported by the authors.

Data availability statement
Data available on request due to privacy/ethical restrictions. The data that support the findings of this study are available on request from the corresponding author, [DT]. The data are not publicly available due to [restrictions e.g. their containing information that could compromise the privacy of research participants].