PC fulfillment is defined as the “employees’ global perceptions that, overall, the organization has fulfilled its terms of the contract in an equitable manner” (Henderson et al., 2008, p. 1,209). Social exchange theory has been the main theoretical base to explain the effects of PC fulfillment at the relational level (Blau, 1964). As individuals experience positive organizational inducements, they are motivated to continue the development of mutually beneficial relationships by reciprocating the organization (i.e., the exchange counterpart) with positive behaviors. At the individual level, however, self-regulation theory has risen recently as a more fine-tuned theory that explains the dynamics of psychological contracts (Rousseau et al., 2018). Self-regulation theory explains human behavior based on the idea of a feedback loop, where individuals monitor the achievement of personal goals, experience emotions caused by their lived experiences, and modify cognitions and behavior toward goal attainment (Lord et al., 2010).

When employment promises are fulfilled, more propitious conditions are perceived by employees, helping them reach valued personal goals, and experience positive feelings about the employment relationship and the convenience of continuing social exchanges in the future (Castanias and Helfat, 1991; Conway and Briner, 2002). The positive affect caused by PC fulfillment and perceptions of goal conduciveness have implications for the perception of exchange relationships as well (Forgas and George, 2001; Isen, 2001). Similarly, emotions play a significant role in the perceptions of fulfillment. PC fulfillment does not only create positive emotions, but it is more likely to be perceived where discrepancies do not generate negative emotions (Kiefer et al., 2022), suggesting that PC fulfillment involves both perceptions and emotions (Conway and Briner, 2002; Rousseau, 2011). Because of the important mediating role of affect in PC theory, a logical extension of these conclusions is that experiences of PC fulfillment should prevent individuals from experiencing negative emotions and exit cognitions, a result that has received meta-analytic support (e.g., Zhao et al., 2007). As Figure 1 suggests, we expect PC fulfillment to be related to turnover intention through the effects of emotional exhaustion, and we developed those relationships in the following paragraphs.

In longer periods of time, PC fulfillment should prevent individuals from having negative work experiences. For instance, previous research has shown that PC fulfillment was negatively related to emotional exhaustion (Gakovic and Tetrick, 2003; Abdalla et al., 2021), defined as the “feelings of being emotionally overextended and depleted of one’s emotional resources” (Maslach and Goldberg, 1998, p. 64). Emotional exhaustion is a phenomenon that occurs after the depletion, over time, of personal psychological resources. We expect to observe that PC fulfillment deters the occurrence of emotional exhaustion because (i) PC fulfillment injects socio-emotional resources into individuals through the effects of goal attainment and (ii) PC fulfillment reduces the need for individuals to focus on creating strategies to reduce discrepancies and obtain valued goals. To support this point, previous PC research has shown that low past PC fulfillment experiences alter present PC fulfillment (Robinson et al., 1994; Coyle-Shapiro and Kessler, 2002), buffering the effects on important outcomes such as burnout (Zhong et al., 2021).

In this article, we argue that the expected effects of self-regulation in PC theory are less likely to occur for individuals who are less capable of identifying and using the additional resources that PC fulfillment brings about. In particular, we study the moderating effect of emotional intelligence on the relationship between PC fulfillment and emotional exhaustion. Although PC research has already acknowledged the role of individual differences before (e.g., Kickul and Lester, 2001; Coyle-Shapiro and Neuman, 2004; Raja et al., 2004; Hermida and Luchman, 2013), emotional intelligence has been rarely studied as a theoretically relevant moderator.

We understand emotional intelligence (EI) as “the ability to carry out accurate reasoning about emotions and the ability to use emotions and emotional knowledge to enhance thought (Mayer et al., 2008).¹ EI is a between-person individual difference that has been conceptualized as a multi-dimensional ability including the ability to evaluate own emotions, evaluate the emotions of others, regulate own emotions, and use emotions for goal achievement (Mayer et al., 2016). Because of these abilities, an individual’s EI has been positively associated with work performance (Joseph and Newman, 2010; O'Boyle et al., 2011) and other work criteria such as acute stress, leadership emergence, or negotiation value created (e.g., Elfenbein et al., 2007; Côté et al., 2010; Kotsou et al., 2019; Lea et al., 2019).

Moreover, EI has been a significant mechanism of self-regulation theories. It is well-established that emotional intelligence has a positive effect on emotion regulation (e.g., Laborde et al., 2014; Thomas and Zolkoski, 2020), with documented direct effects on emotional exhaustion as high IE individuals are more likely to exert higher efforts on goal striving processes, injecting positive feelings through goal achievement (e.g., Moon and Hur, 2011; Tziner et al., 2020). In addition to the direct effects of EI, individuals are also likely to interact differently with job-specific situations and contextual factors. This distinct theoretical mechanism of the effects of emotional intelligence has been dubbed the “moderation model” by EI scholars (Côté, 2014; Pradhan and Jena, 2018).

We propose that individuals with high rates of EI have a higher capacity for self-regulation that allows them to interact more positively in exchanges with the organization. First, because of their capacity to regulate their own emotions, a high EI makes individuals less vulnerable to negative disruptions, recovering faster from PC breaches. PC research suggests individuals with low self-control are more likely to be affected by PC breaches (Bal et al., 2011; Restubog et al., 2015; Balogun et al., 2018). Considering that breach is more the norm than the exception in many organizations (Robinson and Rousseau, 1994), high EI individuals are less likely to be affected by small breaches of the PC. That is, high EI may use less time in the repair phase of the PC dynamics (Rousseau et al., 2018). On the other side, the level of reactivity of negative events in individuals with low EI is high (Bechtoldt and Schneider, 2016). Lanciano et al. (2012) found that individuals that scored low EI ruminated less about previous stressful experiences.

Second, it is expected that high EI individuals can use the socio-cognitive resources injected by high PC fulfillment in helping them to better regulate their unpleasant feelings in other domains of the working experience and the consequent effect outside the workplace. In addition, high EI individuals are more aware of unpleasant feelings caused by low PC fulfillment, valuing the fulfillment of employment promises, and creating relief in emotional demands when fulfillment occurs. Previous research has shown that individuals with high EI exert increased attention to positive emotions (Szczygieł and Mikolajczak, 2017; Lea et al., 2019). In contrast, individuals with low EI have more difficulties creating strategies for self-regulation and are less likely to perceive PC fulfillment as a relevant factor that determines goal conduciveness. The burden of emphasizing negative feelings in life events and the need to constantly ruminate about strategies to reach personal goals should affect their levels of emotional exhaustion regardless of the perceptions of the fulfillment of the PC.

These arguments are consistent with a moderation effect of emotional intelligence in the negative relationship between PC fulfillment and emotional exhaustion. The PC fulfillment-emotional exhaustion relationship is expected to be observed for individuals with high EI, but not for individuals with low EI (no significant relationship; a neutralization effect). Thus, we hypothesize:

Previous research has provided evidence for a significant relationship between emotional exhaustion and turnover cognitions and behaviors (e.g., Boles et al., 1997; Wright and Cropanzano, 1998; Karatepe and Uludag, 2007; Giao et al., 2020). Emotional exhaustion is a personal experience of energy depletion which in many cases is solved by exiting the situation that provokes that state. At the same time, the PC literature has long sustained the mediating effect of emotional experiences in the relationship between PC evaluations (breach/fulfillment) and turnover intentions. For example, in an SEM meta-analysis, Zhao et al. (2007) showed that a model with emotions as a mediator had the best fit with the data when explaining how PC evaluations impact employee outcomes, such as turnover intentions. The dynamic model of PC processes also conceptualized negative affect as an important mediator of PC evaluations and exit behaviors [See Figure 5 in Rousseau et al. (2018)]. Basically, individuals who experience negative PC evaluations are more likely to search for different employment or occupational options in order to achieve self-set goals, taking emotion as key information to create appropriate behavioral responses (Kiefer and Antoni, 2019). These relationships suggest a mediation effect where PC fulfillment predicts turnover intentions through the effects of emotional exhaustion.

This mediation effect, however, is likely to be affected by the emotional abilities of individuals. As suggested in hypothesis 1, we expect to observe high levels of emotional exhaustion for individuals with low EI in general, regardless of their perceptions of PC fulfillment. Thus, the mediation effect of PC fulfillment, emotional exhaustion, and turnover intentions is expected to occur for individuals high in EI, but not necessarily for individuals low in EI.

To test our hypotheses, we collected quantitative data as part of a larger study about psychological contracts in teams. Participants were recruited from 8 educational institutions and one administrative department of a well-known municipality in Chile. We received support to conduct this study from the municipality as part of a series of organizational initiatives to improve HR management practices. The data was obtained using two-wave self-administered web-based surveys, which were separated by a lag time of 2 months between each survey. This was intended to reduce the potential of common method variance (Podsakoff et al., 2003). In the first survey, we collected PC fulfillment and emotional intelligence, whereas, in the second survey, we collected data about emotional exhaustion and turnover intentions. The data was matched by the research team using employee IDs, following recommendations of our Institutional Ethics Committee to protect participants’ confidentiality.

The HR department provided the updated personnel records of the organization, including emails from 247 employees. Both surveys were distributed to their email addresses through Qualtrics’ mailing tool. The sample size was 181 responses for the first survey and 176 responses for the follow-up survey. In sum, the sample consisted of 150 matched responses, including employees who answered both surveys. The final sample only included subjects who successfully completed the two surveys (137 individuals) and who successfully answered an attention check item (we eliminated 23 responses due to lack of attention). This decision improved the quality of the collected data, providing a final sample of 114 responses. Our final sample size was lower than we expected, although high enough to make a rigorous test of our main hypotheses. All employees were between 25 and 64 years old (Mean = 39.16 years, S.D. = 9.89). Most participants were female (82%) and the tenure ranged from 0.16 to 35 years (Mean = 7.55 years, S.D. = 8.00).

All measures of this study were applied using 5-point Likert scales (1: Completely disagree – 5: Completely agree). The surveys were applied in their Spanish versions, which can all be found in escalas.unegocios.cl. Psychological contract fulfillment was captured using Robinson and Morrison’s (2000) classic scale of global PC fulfillment. To better align our theory to our measure, we only used the fulfillment items, eliminating two negatively worded items, addressing previous concerns that Robinson and Morrison’s measure capture both fulfillment and breach (Conway and Pekcan, 2019). This procedure was already utilized in Laulié (2017) in a similar population, showing excellent psychometric properties. A sample item is “Almost all of the promises made to me by my employer have been kept so far.” Cronbach’s alpha was 0.93. To capture emotional intelligence, we used an ability-based EI instrument developed by Law et al. (2004) and validated in Spanish by Madrid (2020). The instrument contemplates four dimensions, including recognition of own emotions, recognition of emotions in others, regulation of own emotions, and use of emotions. An item example is “I have good control of my emotions.” Cronbach’s alpha was 0.87. Emotional exhaustion was captured using an 8-item subscale of the Oldenburg Burnout Inventory and validated by Halbesleben and Demerouti (2005). A previous Spanish version of this scale was used by Laulié et al. (2022). An item example is “During my work, I often feel emotionally drained.” Cronbach’s alpha was 0.86. Finally, turnover intention was captured by participants using a 3-item scale developed by Cammann et al. (1983). A sample item is “I often seriously think about quitting.” Alpha was 0.85. All descriptive statistics are included in Table 1.

First, we used confirmatory factor analysis (CFA) to evaluate our measurement model. To test model fit, we used typical indexes and some of the commonly used standards in the literature (CFI > 0.90, RMSEA<0.08, SRMR<0.08). Our theoretical model was then compared against alternative measurement models. Second, to test our main hypotheses, we used structural equation modeling (SEM). This approach can reduce bias in the calculation of estimates in a mediation or moderation analysis by managing the effects of measurement error (see Ledgerwood and Shrout, 2011). For hypothesis 1, we used a full indicator method for latent interaction constructs (Marsh et al., 2013; Collier, 2020). For this method, we created interaction terms for each pair of mean-centred indicators of the independent and moderator variables (PC fulfillment and emotional intelligence, respectively), which then serve as indicators of a latent interaction unobserved construct. Third, to test hypothesis 2, we calculated an index of moderated mediation (Hayes, 2015; Hayes et al., 2017) and observed the indirect effects of PC fulfillment on turnover intentions through the effects of emotional intelligence depending on the level of the moderator variable (emotional intelligence). All analyses were carried out using SPSS, AMOS, and the PROCESS macro.

We first examined the measurement model that form our key constructs (PC fulfillment, emotional intelligence, emotional exhaustion, and turnover intention). The hypothesized model (see Table 2) indicates an overall excellent fit of a four-factor model (χ² = 111.402, df = 107, CFI = 0.99; RMSEA = 0.019; SRMR = 0.076). Then, we tested several alternative models, merging some latent constructs (models 2 to 6 in Table 2). All the alternatives resulted in a worse fit to the data. These results suggest that individuals differentiate the four constructs, providing evidence for discriminant validity.

We also created a structural equation model to evaluate our hypothesized relationships. These results are summarized in Table 3. The initial model containing only the predictor, mediator, and outcome shows an excellent fit to the data (χ² = 50.282, df = 56, CFI = 1.00; RMSEA = 0.00; SRMR = 0.055) and provides evidence to confirm some of the expected relationships. PC fulfillment (β = −0.20, p < 0.10) was significantly related to emotional exhaustion, with higher fulfillment reducing the levels of emotional exhaustion. At the same time, emotional exhaustion was significantly related to turnover intentions (β = 0.28, p < 0.01). These results are aligned with PC theory and previous findings in the literature.

Hypothesis 1 predicted a moderation effect of emotional intelligence on the relationship between PC fulfillment and emotional exhaustion. Our results show that, after including emotional intelligence as a predictor of emotional exhaustion (β = −0.88, p < 0.01), the latent interaction construct was significantly related to emotional exhaustion (β = −0.81, p < 0.05), providing evidence to support our hypothesis. Figure 2 plots the shape of the moderation effect. As hypothesized, when emotional intelligence is high, there is a negative relationship between PC fulfillment and emotional exhaustion. In contrast, when emotional intelligence is low, the relationship is neutralized (not significant). Individuals with low emotional intelligence presented on average high levels of emotional exhaustion. This evidence supports hypothesis 1.²³

Hypothesis 2 predicted a moderated mediation effect of emotional intelligence on the mediation of PC fulfillment, emotional exhaustion, and turnover intention. Table 4 shows the main results of this test. When emotional intelligence is low or average, the indirect effects were not significantly different from zero. In contrast, when emotional intelligence was high, the indirect effect was significant. The index of moderated mediation was also significantly different from zero, providing evidence to support hypothesis 2.