Psychological ownership refers to an individual’s perception that a target object (material or non-material) or part of it is experienced as their own. Its core elements are possessiveness and the psychological connection to the target. This psychological state is primarily formed through control over the target, intimate knowledge of it, and self-investment into it. Control over objects is considered capable of generating a sense of ownership. Intimate knowledge arises when the target and the self become integrated through understanding, fostering ownership feelings. Self-investment occurs when greater dedication of time, attention, energy, and effort leads to stronger integration between the self and the target, thereby developing enhanced psychological ownership (Gieseler et al., 2021).

The retirement event itself constitutes a significant resource loss for public officials, stripping them of control over their former position and their proprietary identity. Resource Conservation Theory predicts that individuals experiencing resource loss threats or actual loss will strive to protect remaining resources and adopt a cautious stance toward future resource acquisition. Under ego depletion, the psychological resources available for establishing “intimate knowledge” of and investment into new objects (such as potential reemployment positions) become more limited (Ikeuchi et al., 2021). Regarding cognitive depletion, negative self-perceptions and self-doubt hinder the development of identification with and belongingness to new work roles, making it difficult to generate possessiveness (“this is my job”). In emotional depletion, persistent negative emotional states like tension and exhaustion reduce the willingness and energy required to explore new work environments and establish psychological connections, thereby inhibiting psychological ownership formation (Dang et al., 2017). While behavioral depletion manifests as weakened impulse control, its impact on psychological ownership may be more complex. Due to their professional characteristics, public officials’ deep-seated occupational identity and sense of responsibility may buffer the effect of behavioral depletion on establishing psychological ownership of new work, even when self-control resources are temporarily insufficient. Consequently, the direct negative impact of behavioral depletion on psychological ownership is likely weaker. Therefore, this study proposes the following hypotheses:

Psychological ownership is when an individual feels the object (material or immaterial) or part of the object as if it were his own (Schilke et al., 2021). The core of psychological ownership is the sense of possession and the psychological connection with the object (Wang et al., 2022). Pierce et al. (2003) pointed out that psychological ownership can be formed in three ways: control the goal, close understanding, and self-involvement. Control goal means that the control of the goal can produce the sense of ownership of the goal. Close understanding means that when an individual understands the goal, the goal and the self will merge, resulting in a sense of ownership of the goal. The more time, attention, energy and effort an individual puts into a goal, the more closely his or her self integrates with the goal and develops a stronger psychological ownership. Firstly, psychological ownership of work creates a sense of being the “owner” among retired public officials. This not only enhances their confidence in task competence but also reinforces the positive significance of work for self-realization, strengthening self-efficacy and perceived value. Secondly, psychological ownership fosters a sense of responsibility and affective commitment towards work among retired public officials. This promotes a more positive view of participation’s value and generates intrinsic motivation to maintain and contribute to the work. Finally, strong feelings of psychological connection and responsibility are key intrinsic motivators driving retired public officials to take action. Those with high psychological ownership are more likely to proactively seek reemployment opportunities to achieve sustained contribution and value continuation.

Public sector personnel have been engaged in relevant work for a long time, and have paid a lot of time, energy and effort for their work, so they can skillfully use corresponding knowledge to solve work problems. The single, repeatable work they have been engaged in for a long time makes them have a sense of belonging to work and a strong psychological ownership. Management ownership can make public sector retirees more willing to remain members of an organization and improve individual self-efficacy (Pierce et al., 2003). Retirement, on the other hand, means taking away the sense of ownership. The gap of not being able to continue working will be difficult for some public sector personnel to accept, resulting in a sense of depletion of psychological ownership. Previous studies have found that psychological ownership has a positive correlation with employee retention intention and a negative correlation with turnover intention (Sepeng et al., 2020; Jing and Yan, 2022). Studies have shown that psychological ownership has an impact on employees’ work attitude, such as psychological ownership and employees’ organizational commitment (Peck et al., 2021), job satisfaction, and retention intention (Kavvouris et al., 2020). Psychological ownership can explain the internal motivation of the reemployment intention of the younger public sector retirees. When public sector employees have psychological ownership of work and can obtain satisfaction and sense of gain from work, it will have a positive impact on self-efficacy, attitude and willingness to reemployment. Based on this, the following hypothesis is proposed:

Studies have shown that individuals with high personal subjective cognition, positive emotions and high autonomy will have less ego-depletion (Li et al., 2022). Specifically, individuals with high subjective cognition can mitigate the after-effects of ego-depletion by perceiving current behavior as beneficial for the future (Oeri and Roebers, 2020). If individuals know that their efforts are helpful to others or themselves, they can enhance their commitment to the organization, resulting in altruistic behavior, organizational citizenship behavior, butler behavior and other positive behavioral consequences (Hagger et al., 2019). Individuals with positive emotions can compensate for ego-depletion, promote self-control (She et al., 2022), avoid being influenced by external circumstances, and strengthen their inner choices. Individuals with high autonomy tendency have a strong need to complete personal goals independently, and will make every effort to reduce the possibility of failure, improve individual toughness and enhance the ability to withstand pressure. However, individuals with high psychological ownership will produce a spirit of dedication, and will produce more organizational citizenship behaviors, which will provide rich working resources for individuals (Rodriguez Montalban et al., 2014). Therefore, when public sector retirees experience cognitive, emotional, or behavioral depletion that shake their willingness to re-enter the workforce, if individuals have both rich working resources (autonomy) and rich personal resources (psychological ownership), they can greatly reduce the level of depletion and enhance their personal beliefs. Therefore, psychological ownership can alleviate the negative effects of ego-depletion and generate positive reemployment intentions.

Psychological ownership serves as the key psychological mechanism through which the negative effects of ego-depletion are transmitted. Individuals with higher psychological ownership exhibit greater psychological resilience due to stronger autonomy tendencies, positive affect, and perceived goal value. This enhanced resilience enables more effective buffering against the detrimental impact of ego-depletion, thereby maintaining active reemployment intentions and behaviors. Consequently, the following hypothesis is proposed:

Social support theory posits that environments characterized by high social trust indicate the perception of a more reliable social support network by individuals (Yang, 2019). Such supportive resources can partially offset psychological resources depleted through ego-depletion, providing an essential psychological “safety net” and energy replenishment (Xiao et al., 2021). This facilitates engagement with new environments through understanding, control, and self-investment, thereby contributing to the establishment of psychological ownership.

Social exchange theory suggests that heightened trust reduces perceived complexity and potential risks within re-employment settings (Seo and Park, 2020). When institutions are perceived as fair and others as trustworthy, the perceived risk and effort required to establish psychological connections with new work roles during states of ego-depletion are diminished (Urena et al., 2019). This mitigates the inhibitory effect of depleted states on psychological engagement.

Furthermore, trust in institutions including re-employment policy safeguards and environmental factors such as equitable opportunity provides anticipated external stability and control (Balliet and Van Lange, 2013). This partially compensates for the loss of control stemming from internal depletion, sustaining motivation to develop psychological ownership even when personal resources are temporarily insufficient. Consequently, higher levels of social trust are expected to weaken the negative impact of ego-depletion on psychological ownership (Siegrist, 2021). Hence, it is hypothesized that:

Social trust is shaped by long-term cultural accumulation and results from rational assessments and choices regarding personal gains and losses (Liu and Gong, 2020). It serves as the foundational value for social cohesion and solidarity. The composition and level of social trust influence the degree of societal engagement. Consequently, social trust can be reflected in the level of social quality (Jiang et al., 2020). Cognitive Evaluation Theory posits that individuals in high-trust environments are more likely to attribute ego-depletion states—such as self-doubt or low mood—to temporary, external factors rather than permanent personal deficiencies (Irawan et al., 2022). High institutional trust, including confidence in safeguards like age-discrimination laws, and high moral trust, reflected in societal norms such as respect for elders, are suggested to mitigate the perceived threat of reemployment barriers (Zhou and Guo, 2020). This mitigation buffers the direct negative impact of ego-depletion on reemployment confidence and intention (Huff and Kelley, 2003).

Furthermore, strong institutional trust indicates a belief in effective policy support and legal protections, while strong moral trust signifies an expectation of fair opportunities and an inclusive social environment (Zhang et al., 2018; Rajkumar et al., 2022). These perceptions of trust are theorized to provide alternative resources and positive behavioral cues for coping with ego-depletion (Ichihashi et al., 2020). Consequently, they may directly motivate or sustain the behavioral drive to overcome internal depletion and pursue reemployment. Consequently, the following hypothesis is proposed:

Based on the relevant research conducted by Ying et al. (2023) and Tang et al. (2022), the questionnaire was designed to divide ego depletion into three dimensions: cognitive depletion, behavioral depletion and emotional depletion. There are 11 items.

Psychological ownership was divided into three core variables: psychological ownership, psychological control and psychological identity. The scale mainly draws on the relevant measurement scales in the existing research, and makes appropriate modifications according to the research situation. Sense of ownership refers to the measurement scale of Atatsi et al. (2021), which has good reliability and validity and is the most representative scale for measuring psychological ownership. The items of the sense of Control scale are derived from the sociopolitical control scale in Alexander L’s (Lithopoulos et al., 2023) research on perceptual control in a specific field. The scale of identity is based on the empirical study of Zhang et al. (2023), which has high reliability and validity.

With reference to the relevant studies of Thomassen et al. (2022), this study is deleted and modified according to the actual situation of this study. Indicators were selected from three dimensions: self-efficacy, attitude and intention. There are 14 items in the scale.

According to the relevant studies of Cai et al. (2023) and Kondo et al. (2023), the social trust was measured from two dimensions of institutional trust and interpersonal trust. System trust includes two sub-dimensions: policy trust and legal trust. Interpersonal trust includes two sub-dimensions: urban environment and social atmosphere. There are 16 items in the scale.

Cronbach’s Alpha of the overall scale of ego-depletion was 0.849. Cronbach’s Alpha of cognitive depletion, emotional depletion, and behavioral depletion were 0.834, 0.864, and 0.864, respectively, all of which met the basic standard of greater than 0.7. Cronbach’s Alpha of psychological ownership, reemployment intention and social trust were 0.905, 0.873, and 0.909, respectively, which were greater than 0.7. It can be seen that the questionnaire scale used in this study has good reliability. In addition, the CITC between the observed variable and its latent variable was greater than 0.5, indicating good reliability of the questionnaire.

The KMO test value of the survey data was 0.911 (greater than 0.70), indicating that the questionnaire was suitable for factor analysis. Bartlett sphericity test results showed that the approximate chi-square value was 11191.48, the significance probability was 0.000 (p < 0.01), and the validity structure was good. In the process of factor analysis, 11 common factors with eigenvalues greater than 1 were extracted by principal component analysis, and the total variance explanation rate was all greater than 60%, indicating good validity of the scale. Factor rotation was performed by orthogonal rotation with maximum variance method, and 59 question options were classified into 11 factors. The load of each measurement item was higher than 0.5, and the questionnaire content validity met the requirements. The results of confirmatory factor analysis show that the standardized factor load of each item was greater than 0.5, and the standard error value of S.E was less than 0.5. The results of exploratory factor analysis indicated that the validity of the questionnaire met the requirements. Finally, AVE of each dimension was greater than 0.5, and its square root was greater than the correlation coefficient between variables, indicating that the convergent validity of the scale met the requirements (Carlson and Herdman, 2012).

The results show that the X2/df value was 3.291, less than 5; The RMSEA was 0.047, which was less than the standard level 0.08, indicating a good fit. GFI = 0.962, AGFI = 0.946, NFI = 0.951, IFI = 0.965, CFI = 0.965, TLI = 0.957, all the indicators of goodness of fit reached the general standard. It showed that the structural equation model established in this study was effective and well matched with the recovered data (Figure 2).

In Table 6, Model 1 established a multiple regression model with cognitive depletion and social trust as independent variables and psychological ownership as dependent variables. Model 2 established a multiple regression model with cognitive depletion, social trust and interaction item cognitive depletion * social trust as independent variables and psychological ownership as dependent variables. In model 1, cognitive depletion of independent variable had a significant negative effect on psychological ownership (β = −0.318, t = −10.921, p < 0.001). In Model 2, the regression coefficient of the interaction term between the independent variable and the regulating variable was 0.134 (t = 3.443, p < 0.01), indicating that the interaction term had a significant positive effect on psychological ownership. In addition, the R² of model 1 was 0.133, and that of model 2 was 0.143, which was significantly improved, indicating that the interpretation ability of the model was enhanced. Therefore, it was proved that the moderating variable social trust had a significant positive moderating effect on the relationship between cognitive depletion and psychological ownership. Therefore, it was assumed that H5a was valid (Table 6).

In Table 7, Model 1 established a multiple regression model with emotional depletion and social trust as independent variables and psychological ownership as dependent variables. In Model 2, a multiple regression model was established with emotional depletion, social trust and interaction item emotional depletion * social trust as independent variables and psychological ownership as dependent variables. In model 1, independent variable emotional wear had a significant negative effect on psychological ownership (β = −0.27, t = −9.030, p < 0.001). In model 2, the regression coefficient of the interaction term between the independent variable and the regulating variable was 0.106 (t = 2.811, p < 0.01), indicating that the interaction term had a significant positive effect on psychological ownership. In addition, the R² of model 1 was 0.103, and that of model 2 was 0.110, which was significantly improved, indicating that the interpretation ability of the model was enhanced. Therefore, it was proved that the moderating variable social trust had a significant positive moderating effect on the relationship between emotional depletion and psychological ownership. Therefore, it was assumed that H5b was true (Table 7).

In Table 8, Model 1 established a multiple regression model with behavioral depletion and social trust as independent variables and psychological ownership as dependent variables. In Model 2, a multiple regression model was established with behavior depletion, social trust and interaction item behavior depletion * social trust as independent variables and psychological ownership as dependent variables. In model 1, the behavior depletion of independent variable had a significant negative effect on psychological ownership (β = −0.179, t = −5.928, p < 0.001). In Model 2, the regression coefficient of the interaction term between the independent variable and the regulating variable was 0.099 (t = 2.378, p < 0.01), indicating that the interaction term had a significant positive effect on psychological ownership. In addition, the R² of model 1 was 0.064, and that of model 2 was 0.070, which was significantly increased, indicating that the interpretation ability of the model was enhanced. Therefore, it was proved that the moderating variable social trust had a significant positive moderating effect on the relationship between behavior depletion and psychological ownership. Therefore, it was assumed that H5c was true (Table 8).

In Table 9, Model 1 established a multiple regression model with cognitive depletion and social trust as independent variables and reemployment intention as dependent variables. In Model 2, a multiple regression model was established with cognitive depletion, social trust and interaction item cognitive depletion * social trust as independent variables, and reemployment intention as dependent variables. In model 1, cognitive depletion of independent variable had a significant negative effect on reemployment intention (β = −0.374, t = −13.081, p < 0.001). In Model 2, the regression coefficient of the interaction term between the independent variable and the regulating variable was 0.144 (t = 3.757, p < 0.001), indicating that the interaction term had a significant positive effect on reemployment intention. In addition, the R² of model 1 was 0.161, and that of model 2 was 0.172, which was significantly improved, indicating that the interpretation ability of the model was enhanced. Therefore, it was proved that the moderating variable social trust had a significant positive moderating effect on the relationship between cognitive depletion and reemployment intention. So suppose that H6a was true (Table 9).

In Table 10, Model 1 established a multiple regression model with emotional wear and social trust as independent variables and reemployment intention as dependent variables. In Model 2, a multiple regression model was established with emotional depletion, social trust and interaction item emotional depletion * social trust as independent variables and reemployment intention as dependent variables. In model 1, independent variable emotional wear had a significant negative effect on reemployment intention (β = −0.386, t = −13.413, p < 0.001). In Model 2, the regression coefficient of the interaction term of the independent variable and the regulating variable was 0.066 (t = 1.813, p > 0.05), indicating that the interaction term had no significant effect on reemployment intention. And the R² of model 1 was 0.167, and the R² of model 2 was 0.170, which was significantly improved. Therefore, it was proved that the moderating variable social trust had no significant moderating effect on the relationship between emotional depletion and reemployment intention. So suppose that H6b was not true (Table 10).

In Table 11, Model 1 establishes a multiple regression model with behavioral depletion and social trust as independent variables and reemployment intention as dependent variables. In Model 2, a multiple regression model was established with behavior depletion, social trust and interaction item behavior depletion * social trust as independent variables and reemployment intention as dependent variables. In model 1, the behavior depletion of independent variable had a significant negative effect on reemployment intention (β = −0.340, t = −11.723, p < 0.001). In Model 2, the regression coefficient of the interaction term between the independent variable and the regulating variable was 0.146 (t = 3.671, p < 0.001), indicating that the interaction term had a significant positive effect on reemployment intention. In addition, the R² of model 1 was 0.137, and that of model 2 was 0.148, which was significantly improved, indicating that the interpretation ability of the model was enhanced. Therefore, it was proved that the moderating variable social trust had a significant positive moderating effect on the relationship between behavior depletion and reemployment intention. Therefore, it was assumed that H6c was true (Table 11).

As the leader in promoting active aging, the government should call on society and families to provide support, build an inclusive age-friendly environment, and focus on retirees’ psychological counseling to alleviate concerns and boost their reemployment initiative. Families need to reverse intergenerational support concepts to respond to retirees’ needs for social participation and reemployment. Communities should monitor local retirees’ situations and strengthen cohesion. Retirees themselves should maintain a positive attitude, break free from narrow family boundaries, and actively integrate into society.

Judicial and relevant departments should rely on elderly associations and universities for the elderly to conduct legal popularization and education on elderly rights protection and supporting regulations, promoting standardized and regular legal publicity to make up for retirees’ lack of legal knowledge in rights protection and raise their law-abiding and rights-awareness. The government and judiciary should issue local regulations, clarify legal responsibilities, standardize supporting measures, fill legislative gaps in retirees’ protection, preferential treatment, education and services, and protect their rights from the legislative level.

In line with diversified social governance, human resources and social security departments should incorporate retired human resources into overall planning, establish and improve a retiree information database (classifying data by former professional expertise, age and skills), and build a vertical management network for retiree information to grasp their advantages and trends timely. Communities should set up information service stations for retirees, conduct surveys to clarify their numbers and situations, and use big data to build a community demand feedback platform and an online platform for reemployment intentions, providing convenient two-way information services.

As the Madrid International Plan of Action on Ageing states, “lifelong education is a prerequisite for elderly employment.” Universities for the elderly (retirees’ main knowledge channel) should add courses on law, finance and social welfare, offer special courses on social participation in public administration and education fields, organize practical activities to test learning outcomes, and hold salons or exhibitions to help capable retirees realize self-worth.