The societal and political effects of job insecurity remain poorly understood, despite paid work being a central aspect of most people’s daily life (Klug et al., 2024; Wroe, 2014). Job insecurity has increased in recent decades due to economic and societal changes, such as the rise in temporary contracts and part-time employment (Azedi, 2023; Klug et al., 2024). This development may not only harm employees’ health and well-being (Rudolph et al., 2025) but also influence their societal and political attitudes and behavior. As such, work-related factors like job insecurity may drive people toward more extreme political options (Geishecker & Siedler, 2012). Employees may perceive their fear of losing their job as a breach of their psychological-democratic contract with the governing parties. As a reaction to this perceived contract breach, they may then decide to identify with and vote for parties that promise security, which extreme parties usually do. This influence may be especially strong in structurally weak regions with high unemployment and lower income levels, such as the eastern German states (Koschatzky & Kroll, 2019). In such structurally weak regions, fear of job loss may significantly impact voting behavior.
One predictor of extreme right-wing identification and voting that has been discussed is economic insecurity and unemployment at both the individual and structural levels (e.g., Hernández & Kriesi, 2016). Only a few studies have focused on job insecurity. Research on job insecurity has mainly examined its personal impacts, such as negative effects on mental health (Rudolph et al., 2025) or reduced job and life satisfaction (Jiang & Lavaysse, 2018). Some evidence suggests it may also influence political attitudes (e.g., Fossati, 2014). One study found that individuals with higher perceived job insecurity were more likely to identify with extreme right-wing parties in former East and West Germany (Geishecker & Siedler, 2012). In contrast, another study across four European countries linked job insecurity to support for the radical left, but not the radical right (Gomez et al., 2024). However, the first study focused on party identification, while the second looked at voting intentions. Hence, research examining the relationship between job insecurity and voting behavior remains limited and inconclusive (Klug et al., 2024).
This study leveraged a unique opportunity to collect data on job insecurity and actual voting behavior following the 2024 state elections in Brandenburg, Saxony, and Thuringia (Germany). All three were part of the former socialist German Democratic Republic (GDR) and thus are still influenced by the concept of a “workers’ and peasants’ state” adopted by the GDR (Major & Osmond, 2002). We investigated whether job insecurity is linked to voting for right-wing parties and party identification and whether these relationships are mediated by lower trust in politicians. If employees perceive job insecurity as a breach of the psychological-democratic contract they have with the governing parties, their trust in politicians might decrease, which in turn could affect their voting behavior and party identification. Since the governments of the three federal states were led by political parties ranging from the far left to conservative prior to the elections, we tested whether the federal state moderates these relationships. Additionally, prior research suggests unions may buffer against far-right voting (Rydgren & Tyrberg, 2024) and reduce job insecurity (Sousa-Poza & Sousa-Poza, 2000). However, unions are historically weaker in East Germany, and collective bargaining coverage remains lower than in the West (Lübker & Schulten, 2024). Therefore, we examined whether union attitudes, that is, the perception that unions are essential to represent employees’ interests and protect their rights, moderate the relationships between job insecurity and radical voting and party identification. Our conceptual model and pre-registered hypotheses are shown in Figure 1.
Figure 1
Hypothesized Model
Note. Voting Behavior and Party Identification are both operationalized from extreme left to extreme right in the later analyses.
Our study offers key contributions to the literatures on job insecurity and political psychology. While research has largely overlooked the societal impacts of perceived job insecurity, our findings do not support a link between job insecurity and voting behavior across federal states, union status, or employment. Contributing to political psychology research, we show that trust in politicians and union attitudes are stronger predictors of political participation, including voting behavior. Overall, this study enhances understanding of the associations between political attitudes and behavior, emphasizing the critical role of trust in politicians and positive union attitudes in reducing support for (far) right parties, whereas job insecurity does not seem to be directly associated with voting behavior or party identification.
Theoretical Background
Fear of unemployment has often been discussed as a potential predictor for right-wing political affinity (Salmela & von Scheve, 2017). Accordingly, we focus on job insecurity as focal predictor in our study. Job insecurity refers to a feeling of uncertainty about future employment (Azedi, 2023) or the perceived threat to one’s job or important aspects of it (Shoss, 2017). Though definitions vary, they share the idea that individuals who feel insecure about their jobs worry about job continuity or finding new employment (De Witte, 2005). The broader concept consists of two dimensions: worries about job loss (quantitative job insecurity) and worries about significant job changes (qualitative insecurity; Shoss, 2017). This study focuses on the former, which has been the most widely researched. A key characteristic of job insecurity is its subjectivity—employees may perceive the same objective situation differently (De Witte, 2005). Thus, colleagues in similar situations may experience varying levels of job insecurity based on personal circumstances and interpretation. As people act based on their perceptions rather than objective reality, job insecurity can influence decision-making and behavior.
Individual factors influencing job insecurity include performing manual labor or having lower levels of education (Klug et al., 2024). However, job insecurity can also stem from country-level factors like economic instability or unemployment rates, as well as meso-level influences in one's social environment (Debus et al., 2012; Debus & Unger, 2024). Generally, job insecurity is seen as harmful, negatively affecting heterogeneous outcomes such as well-being, behavior, and attitudes (Klug et al., 2024). A meta-analysis showed that job insecurity is negatively associated with job and life satisfaction, organizational commitment, and safety behaviors, and positively related to emotional exhaustion, cynicism, and turnover intentions (Jiang & Lavaysse, 2018). These outcomes are often explained through the transactional stress model (Lazarus & Folkman, 1984), which sees job insecurity as a threat and highlights the stress caused by uncertainty (Klug et al., 2024). For instance, a longitudinal study with employees in Germany found that job insecurity reduced mental health but had no significant effect on physical health at the within-person level (Rudolph et al., 2025). It also showed that impaired physical and mental health increased perceived job insecurity, as poor health can impair workability.
While most research has focused on individual and organizational impacts, recent work by Klug et al. (2024) suggests that job insecurity may also trigger bottom-up processes influencing political attitudes, viewing it as shaped by both personal factors and the broader context. Research in this area is still considered “in its infancy” (Klug et al., 2024, p. 1980), and existing studies have produced mixed and inconclusive results (Gomez et al., 2024). Furthermore, much of the research has focused on objective indicators, such as unemployment rates (Sipma & Lubbers, 2020) or job status (Oesch, 2008), rather than subjective perceptions. Hence, there is a lack of research on how subjective experiences of job insecurity influence political attitudes and behavior (Klug et al., 2024).
Since paid work is a central part of most people’s lives, the threat of job loss may shape attitudes toward political actors and society by signaling a threat to one’s social status and a breach to the psychological-democratic contract employees might have with governing parties (Wroe, 2014). Wroe (2014) proposed that individuals support democracy in exchange for economic stability and secure employment. When job insecurity arises—often due to economic underperformance—it may be seen as a contract breach, potentially reducing political trust and influencing party support (Wroe, 2014, 2016). Party support can occur in different forms, such as voting behavior or identification with parties (Swigart et al., 2020). Voting behavior represents party affiliation through concrete actions, whereas party identification describes party support as a social identity and a form of self-concept.
The sparse research on job insecurity and political attitudes has shown that job insecurity can influence political attitudes. For instance, job insecurity heightened anti-egalitarian views by threatening individuals’ identity as employed persons (Selenko & De Witte, 2021), with such attitudes being central to political conservatism. Similarly, job insecurity has been linked to racist attitudes, including distrust and fear of immigrants (De Witte & Meuleman, 2007). In a 17-year German study using data from the German Socio-Economic Panel Study (SOEP; Geishecker & Siedler, 2012), fear of job loss was associated with stronger support for far-right parties. This aligns with relative deprivation theory (Merton, 1957), suggesting that perceived threats to social status may drive anti-foreign and far-right sentiments. Among participants in East Germany, unemployment increased the likelihood of identifying with extreme right parties by 113% compared to those who felt secure in their jobs (Geishecker & Siedler, 2012). Relatedly, a study among employees in East German federal states observed that a higher perceived threat to Germany’s economic situation in general as well as a perceived threat to one’s personal economic situation were associated with stronger right-wing extremist attitudes (Kiess et al., 2023).
Another study found a contrasting effect: job insecurity increased support for radical left, not radical right, parties (Gomez et al., 2024). Based on data from France, Germany, Greece, and Spain, the authors argued that job insecurity represents a material rather than a status threat. Material threats are more likely to drive support for redistribution and economic equality—key values of radical left parties. Similarly, Sipma et al. (2023) suggested that insecurity may promote support for the radical left by increasing agreement with redistributive policies. However, their findings were mixed. Objective insecurity indicators did boost support for the radical left among working-class voters, but subjective job insecurity was linked to greater support for radical right parties, highlighting that objective and subjective measures may influence political behavior differently.
Job insecurity can fuel resentment toward mainstream democratic parties, which are seen as failing to deliver on economic promises (Wroe, 2014). This disillusionment may lead individuals to support parties that oppose economic globalization and modernization (Geishecker & Siedler, 2012). Feelings of deprivation and fear of losing social status can heighten anti-foreign sentiments, right-wing extremist attitudes, and strengthen ties to extremist right-wing parties (Geishecker & Siedler, 2012; Kiess et al., 2023; Klug et al., 2024). De Witte and Meuleman (2007), showed that relative deprivation mediated the relationship between job insecurity and anti-immigrant attitudes. According to relative deprivation theory, perceived job insecurity may particularly drive support for far-right parties due to the associated social threat.
Hypothesis 1: Job insecurity is associated with a higher identification with right-wing parties and right-wing voting.
The economy in former East German federal states remains weaker than in the West, with higher unemployment and lower per capita income (Koschatzky & Kroll, 2019). These structural weaknesses may lead residents to feel more strongly that the psychological-democratic contract has been violated. This perception may potentially contribute to strong support for the far-right AfD and the rapid rise of the new party “Bündnis Soziale Gerechtigkeit und Wirtschaftliche Vernunft” (BSW; formerly “Bündnis Sarah Wagenknecht”; “Alliance for Social Justice and Economic Reason,” formerly “Alliance Sarah Wagenknecht”) in East Germany. Both parties are populist parties that position themselves against “the elites,” against “foreigners,” and gender minorities and share authoritarian habits (Thomeczek et al., 2024). While the AfD is clearly right wing, the BSW has split up from the Left Party. In an earlier study, East Germans perceiving job insecurity were 90% more likely to identify with right-wing extremist parties, measured by asking participants to which party they are inclined towards, than those without such insecurity (Geishecker & Siedler, 2012).
The perceived economic disadvantages and collective deprivation in East German states may further erode trust in politics, thereby increasing support for extreme parties. Trust can broadly be defined as “the willingness of one party to be vulnerable to the actions of another party” (Mayer et al., 1995, p. 112). Globally, political trust has declined, not only in authoritarian regimes but also in established democracies (Wroe, 2014). Job insecurity is negatively associated with trust in political leaders, institutions, and parties, and with satisfaction with democratic performance (Van Hootegem et al., 2022; Wroe, 2014). Wroe (2014, 2016) conceptualized job insecurity as a psychological-democratic contract breach, where economic insecurity signals a failure of the state, reducing political trust. Furthermore, the effect of job insecurity on radical right voting was mediated by political distrust and anti-immigration attitudes (Sipma et al., 2023). Political distrust also increased the likelihood of voting for both radical left and radical right parties. Similarly, job insecurity was associated with stronger perceived distributive injustice, which heightened political cynicism (Van Hootegem et al., 2022). These findings highlight mechanisms that may explain how job insecurity affects political trust and voting.
Hypothesis 2: Trust in politicians mediates the effects of job insecurity on party identification and voting decisions.
However, the three federal states differed in their lead governing parties before the 2024 elections: The government of Thuringia was led by the Left Party (Die Linke), Brandenburg was led by the Social Democratic Party (SPD), and Saxony was led by the Christian Democratic Union (CDU). This range—from far-left to conservative—suggests that the relations between job insecurity, voting behavior, and party identification may be moderated by state context, as citizens might assign blame for job insecurity differently. Thuringia is particularly notable: despite its government being led by a far-left party, the far-right AfD received most votes in the election, signaling a shift from far-left to far-right preferences. Therefore, the argument that lower wages and less secure economic positions would increase support for radical left parties due to perceptions of unfairness and inequality (Ramiro, 2016) does not seem to be supported for this federal state. This challenges the idea that economic insecurity leads primarily to radical left support due to perceived inequality. Accordingly, we propose a moderating role of federal state, but we do not specify the direction of these effects.
Hypothesis 3: The federal state moderates the relationships between job insecurity, right-wing party identification, and right-wing voting behavior.
Increases in job insecurity have sometimes been linked to the decline in union membership (Ebbinghaus & Visser, 1999). Research shows that union members experience less job insecurity (Anderson & Pontusson, 2007; Sousa-Poza & Sousa-Poza, 2000). Unions represent workers and aim to improve working conditions collectively, fostering solidarity by protecting group interests (Carriere, 2020). Hence, job insecurity has been interpreted as a motivator for employees to unionize (Bender & Sloane, 1999), offering a group for workers to identify with and helping them cope with job insecurity.
Furthermore, unions are viewed as tools to revive workplace democracy (Weber et al., 2020), potentially strengthening democracy in society. Union members are more politically engaged, knowledgeable, and active in discussions than non-members (Kerrissey & Schofer, 2018). Therefore, the relationship between job insecurity and far-right voting could be weaker among those with positive attitudes toward unions. Union membership has been shown to reduce support for the radical right in Sweden (Rydgren & Tyrberg, 2024) and for populist radical right parties (Arndt & Rennwald, 2017). As such, it could be argued that unions might be perceived as an alternative tool to address the perceived psychological-democratic contract breach elicited by job insecurity. Hence, the effects of job insecurity on voting behavior and party identification might be less pronounced for employees with positive union attitudes.
However, in the three federal states of this study, unions are historically weak, with fewer employees covered by collective agreements than the German average (Lübker & Schulten, 2024). Additionally, unions have faced setbacks, such as the 2003 eastern German metalworkers' strike that led to a 10% drop in membership (Dribbusch, 2007). This suggests that union attitudes may not influence, or may even strengthen, the link between job insecurity and far-right voting or party identification. This aligns with research showing no effect of union membership on support for radical right parties (Hövermann et al., 2025; Rennwald & Pontusson, 2021). However, job insecurity can also reduce union loyalty and lead to membership withdrawal (De Witte et al., 2008). Therefore, we assume that union attitudes moderate the relationship between job insecurity and voting behavior or party identification in all three federal states. However, due to the partially contradictory results regarding the direction and strength of the effects observed to date, we are not proposing a hypothesis but rather formulating a research question.
Research Question 1: Do union attitudes moderate the relationships between job insecurity, voting behavior, and party identification?
In addition to examining these associations, we conducted exploratory analyses to determine if the associations differed between union members and non-members, as union membership was found to reduce populist radical right parties support (Arndt & Rennwald, 2017). We also examined whether job satisfaction predicted the outcomes, assuming that dissatisfied individuals would be more likely to vote for and identify with extreme parties (Hövermann et al., 2025). Furthermore, we examined U-shaped relations between job insecurity and political orientation, assuming that extreme political attitudes are more strongly related to job insecurity than moderate ones. Finally, we repeated all analyses with the inclusion of income and job satisfaction as control variables, to account for participants’ living and work conditions as potential confounding variables.
Method
Study Context
In Germany, political parties can be placed within a coordinate system based on left-wing versus right-wing, liberal versus conservative, and state versus market freedom. According to Decker (2018), the Left Party and the Social Democratic Party (SPD) are associated with state/social justice and are considered left-wing. Alliance 90/The Greens are also associated with state/social justice but are considered more liberal/libertarian than the Left Party and the SPD. The other parties (the Christian Democratic Union, CDU; the Free Democratic Party, FDP; and the AfD) are considered more right-wing. The FDP advocates more strongly for market freedom. The AfD is considered more authoritarian, and the CDU is in-between (see Zacher et al., 2026, for an overview). The BSW is more difficult to place within this coordinate system because it combines traditionally conservative and progressive standpoints.
Sample
Data was collected using online questionnaires through the public opinion research institute Infas-Quo. This study's data collection was part of a larger study examining various associations between voting intentions, voting behavior, and variables such as trust in politics during the federal state elections. Further details are available in the Supplementary Materials (Table S1) and Finn et al. (2024). The sample is not representative, but was sampled in heterogeneous terms regarding age, gender, education, and political orientation. The timing of data collection for T2 varied by federal state, as elections were held on different dates, but all data was collected between 28. October 2024 and 08. November 2024. A follow-up survey for T2 was conducted from 25. November to 13. December 2024, due to missing responses. A total of 1,162 participants took part in the data collection. Demographic characteristics for all samples are provided in Table S1.
The final sample for our analysis included only employed participants, leading to a final sample size of 710 participants. The sample was nearly evenly distributed across the three federal states: 253 participants from Brandenburg, 251 from Saxony, and 206 from Thuringia. It was almost evenly split between female (49.15%) and male (50.85%) participants, with an average age of 45.47 years (SD = 13.12, range 17 to 73), slightly above the average age in East Germany (47.3 years; Destatis, 2025). Of the participants, 18.59% were union members, 39.86% lived in two-person households, and 64.37% had no children living with them. Almost all were German citizens (98.87%) and the most were born in East Germany (91.13%). The most common party affiliation was the far-right Alternative for Germany (AfD), with 22.39% identifying with it and 28.17% voting for it.
Transparency
This study was preregistered at https://osf.io/jeqzk. RQ1 was originally framed as a hypothesis but was later rephrased as a research question during the writing process. Due to a technical issue, the focal variables, job insecurity and union attitudes, were measured after the elections (T2 of the bigger data collection) instead of before the elections (T1 of the bigger data collection). As a result, we deviate from preregistration and use only data collected at T2. All study materials and data are available at https://osf.io/69qvt/. The study was approved by the Friedrich-Schiller-University Jena University ethics committee and is part of a larger data collection by the Center for Right-Wing Extremism Research, Democracy Education and Social Integration (KomRex) at the University of Jena and the Chair of Social Psychology at University of Trier.
Measures
Our predictor variable job insecurity was measured using three items from the German Version of the Copenhagen Psychological Questionnaire (COPSOQ; Burr et al., 2019). The items were rated on a 6-point scale (1 = to a very small extent, 6 = to a very large extent) and preceded by “Are you worried that …”. The items were “… you become unemployed?”, “… new technologies make you superfluous?”, and “… it would be difficult for you to find a new job if you became unemployed?” The reliability of this scale was satisfactory, McDonald’s ω = .81.
Trust in politicians was measured using three items where participants indicated their agreement with statements about currently elected politicians in their respective federal state. Responses were given on a 6-point scale from 1 = completely disagree to 6 = completely agree. The items were: “The currently elected state politicians look after the interests of people like me,” “The currently elected state politicians act responsibly,” and “The currently elected state politicians have sufficient expertise.” The scale showed high reliability, McDonald’s ω = .92.
Union attitudes were measured using four items from Liepmann et al. (1984). The items were rated on a 6-point scale from 1 = strongly disagree to 6 = strongly agree. Participants indicated their agreement with the following items: “Individuals cannot represent their interests without the union,” “Without the activities of unions in the past, the situation of employees would be significantly worse,” “To uphold employees’ rights, we still need unions,” and “Unions are necessary for the enforcement of employees' interests.” The scale demonstrated good reliability, McDonald’s ω = .91.
For voting behavior, participants were asked which party they voted for with their list vote. Possible responses included the Social Democratic Party (SPD), Christian Democratic Union (CDU), Green Party, Free Democratic Party (FDP), Alternative for Germany (AfD), Left Party, Volt, Bündnis Sahra Wagenknecht (BSW), Other, and No vote. Party identification was assessed by asking if participants generally leaned toward a specific party, with the same response options, except “No vote” was replaced by “no specific party.” The distribution of votes in our sample closely mirrored actual voting results (Table S2). Two approaches were used to recode the dependent variables: (a) we recoded them as 1 = right-wing voting/identification (AfD) and 0 = no right-wing voting (all other parties), and (b) we recoded the parties on a left-right continuum using values from the German Parties before the European Elections (GEPARTEE; Thomeczek et al., 2024), where extreme left was scored as 1 and extreme right as 11 (similar to Jolly et al. (2022) using the Chapel Hill Expert Survey). The values for each party are shown in Table S3; with the order of parties from left to right being Left Party, Green Party Volt, SPD; BSW, FDP, CDU, and AfD. These values are referred to as GEPARTEE outcomes in the analysis with participants selecting "Other" or "No vote/No specific party" excluded for the latter approach.
For exploratory purposes, participants were asked whether they are currently union members using a single item with response options 1 = yes and 2 = no, which was recoded to 1 = union member and 0 = not a union member. They were also asked about their overall job satisfaction using an item from the COPSOQ (Burr et al., 2019): “If you look at your work situation as a whole, how satisfied are you with your work overall, considering all the circumstances?“ (1 = very dissatisfied, 6 = very satisfied). Further exploratory variables are the left-right political orientation single item (Breyer, 2015) with a scale from 1 = left to 11 = right, and net household income categorized from 1 = below 500 € per month to 11 = more than 5.000 € per month.
Results
Table 1 presents the correlations, means, and standard deviations for the focal variables in the final sample. Correlation tables by federal state and excluding the income variable (due to 38 missing responses), are available in the Supplementary Materials (S2). To assess model fit, we followed McDonald and Ho's (2002) criteria, considering the root mean square error of approximation (RMSEA) < .05 as good fit and < .08 as acceptable. We used bootstrapping and diagonally weighted least squares estimation, as WLSMV does not support bootstrapping in mediation analyses. All analyses were conducted in R (R Core Team, 2023).
Table 1
Means, Standard Deviations, and Correlations With Confidence Intervals (N = 672)
| Variable | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1. Party Identification (Dichotomized) | 0.23 | 0.42 | – | ||||||||||
| 2. Party Identification (GEPARTEE) | 7.22 | 2.65 | .85** | – | |||||||||
| [.82, .87] | |||||||||||||
| 3. Voting Behavior (Dichotomized) | 0.28 | 0.45 | .76** | .75** | – | ||||||||
| [.73, .79] | [.71, .79] | ||||||||||||
| 4. Voting Behavior (GEPARTEE) | 7.27 | 2.69 | .67** | .89** | .86** | – | |||||||
| [.63, .71] | [.87, .90] | [.83, .88] | |||||||||||
| 5. Job Insecurity | 2.55 | 1.34 | .03 | .07 | .02 | .04 | – | ||||||
| [-.05, .10] | [-.02, .15] | [-.06, .09] | [-.04, .11] | ||||||||||
| 6. Trust in Politicians | 3.09 | 1.29 | -.29** | -.35** | -.32** | -.31** | -.01 | – | |||||
| [-.36, -.22] | [-.43, -.28] | [-.39, -.25] | [-.38, -.24] | [-.08, .07] | |||||||||
| 7. Union Attitudes | 4.18 | 1.24 | -.09* | -.14** | -.10* | -.13** | .12** | .15** | – | ||||
| [-.16, -.01] | [-.22, -.05] | [-.17, -.02] | [-.21, -.05] | [.04, .19] | [.08, .23] | ||||||||
| 8. Political Orientation | 5.63 | 2.09 | .43** | .56** | .46** | .54** | .02 | -.23** | -.13** | – | |||
| [.37, .49] | [.50, .61] | [.40, .52] | [.48, .59] | [-.05, .10] | [-.30, -.16] | [-.21, -.06] | |||||||
| 9. Union Membership | 0.19 | 0.39 | .01 | .03 | -.03 | .01 | .09* | .08* | .25** | -.01 | – | ||
| [-.07, .09] | [-.06, .11] | [-.11, .05] | [-.07, .09] | [.02, .17] | [.01, .16] | [.18, .32] | [-.08, .07] | ||||||
| 10. Job Satisfaction | 4.44 | 1.17 | -.02 | -.02 | -.02 | .02 | -.24** | .19** | .02 | -.02 | .11** | – | |
| [-.10, .05] | [-.11, .06] | [-.09, .06] | [-.06, .10] | [-.31, -.17] | [.11, .26] | [-.06, .09] | [-.10, .05] | [.03, .18] | |||||
| 11. Income | 7.02 | 2.59 | -.06 | -.05 | -.06 | -.05 | -.17** | .17** | .01 | .03 | .10** | .21** | – |
| [-.14, .02] | [-.14, .03] | [-.13, .02] | [-.12, .03] | [-.24, -.10] | [.10, .24] | [-.07, .09] | [-.05, .10] | [.03, .18] | [.13, .28] |
Note. Party Identification and Voting Behavior were coded with 1 = right-wing voting (AfD) and 0 = other. Party Identification GEPARTEE and Voting Behavior GEPARTEE were coded on a scale from 1 = extreme left to 11 = extreme right. Job Insecurity, Trust in Politicians, Union Attitudes, Job Satisfaction were rated on a scale from 1 to 6. Union membership was coded with 1 = union member, 0 = not a union member. Political Orientation was rated on a scale from 1 = left to 11 = right. Income was measured using 11 categories from 1 = less than 500 € per month to 11 = 5.000 € and more per month. Those with missing responses to the Income Variable were screened out.
*p < .05. **p < .01.
Hypotheses Testing
We first tested a SEM with job insecurity as a predictor, trust in politicians as a mediator, and union attitudes as a moderator to predict dichotomized voting behavior and party identification (both coded with 1 = right-wing, 0 = other). The model fit was close to acceptable, RMSEA = .096, 90% CI [.091, .101], standardized root mean square residual (SRMR) = .074. It explained 11.7% of the variance in voting behavior and 10.0% in party identification. Job insecurity did not significantly predict voting behavior, b = .02, p = .358, or party identification, b = .02, p = .345, offering no support for Hypothesis 1 (Table 2 and Figure 2). Additionally, job insecurity did not predict trust in politicians, b = -.01, p = .854.
Figure 2
Results of the SEM
Note. Standardized path coefficients are reported. The order of the values is dichotomized outcomes/ GEPARTEE outcomes.
Trust in politicians was a significant predictor for both outcomes: voting behavior (b = -.12, p < .001) and party identification (b = -.10, p < .001). Lower trust in politicians increased the likelihood of voting for and identifying with a right-wing party. However, there was no support for the assumption that trust in politicians mediates the relationship between job insecurity and these outcomes. Both indirect paths were not significant: voting behavior estimate = .00, p = .856; party identification estimate = .00, p = .857. Hence, Hypothesis 2 was not supported.
Positive union attitudes were associated with a higher likelihood of not voting for right-wing parties, b = -.05, p = .007, and not identifying with a right-wing party, b = -.04, p = .052. However, the interaction between union attitudes and job insecurity did not predict voting behavior b = -.00, p = .734, or party identification b = .01, p = .554. Thus, we can answer the research question (RQ1) by concluding that union attitudes do not act as a moderator.
We then tested a SEM using the GEPARTEE outcomes (1 = extreme left to 11 = extreme right), excluding participants with no party identification/no vote or “other” responses, resulting in n = 529. The model fit was close to acceptable, RMSEA = .095, 90% CI [.090, .100], SRMR = .075. This model explained more variance, with 15.4% in voting behavior and 16.9% in party identification, compared to the dichotomized outcomes model. Job insecurity did not predict voting behavior (b = .11, p = .389), nor party identification (b = .22, p = .073), lending no support for Hypothesis 1 (Table 2 and Figure 2).
Table 2
Results for the SEM
| Path | Dichotomized Outcomes (N = 710) | GEPARTEE Outcomes (N = 529) | ||||||
|---|---|---|---|---|---|---|---|---|
| b (SE) | z | p | β | b (SE) | z | p | β | |
| Job Insecurity → Trust in Politicians | -0.01 (0.06) | -0.18 | .854 | -0.01 | 0.01 (0.09) | 0.13 | .894 | 0.01 |
| R2 = .00 | R2 = .00 | |||||||
| Job Insecurity → Voting Behavior | 0.02 (0.02) | 0.92 | .358 | 0.05 | 0.11 (0.13) | 0.86 | .389 | 0.05 |
| Trust in Politicians → Voting Behavior | -0.12 (0.01) | -8.38 | < .001 | -0.32 | -0.77 (0.10) | -8.00 | < .001 | -0.35 |
| Union Attitudes → Voting Behavior | -0.05 (0.02) | -2.67 | .007 | -0.11 | -0.45 (0.15) | -3.11 | .002 | -0.15 |
| Job Insecurity × Union Attitudes → Voting Behavior | -0.00 (0.01) | -0.34 | .734 | -0.02 | 0.08 (0.08) | 0.99 | .321 | 0.05 |
| R2 = .12 | R2 = .15 | |||||||
| Job Insecurity → Party Identification | 0.02 (0.02) | 0.95 | .345 | 0.05 | 0.22 (0.13) | 1.79 | .073 | 0.11 |
| Trust in Politicians → Party Identification | -0.10 (0.01) | -7.58 | < .001 | -0.30 | -0.77 (0.09) | -8.22 | < .001 | -0.36 |
| Union Attitudes → Party Identification | -0.04 (0.02) | -1.95 | .052 | -0.08 | -0.46 (0.14) | -3.23 | .001 | -0.15 |
| Job Insecurity × Union Attitudes → Party Identification | 0.01 (0.01) | 0.59 | .554 | 0.03 | 0.08 (0.08) | 1.03 | .302 | 0.05 |
| R2 = .10 | R2 = .17 | |||||||
Note. β = standardized path coefficient. Values in bold highlight significant predictors. SE = standard error, CI = confidence interval. For dichotomized outcomes, Voting Behavior and Party Identification were both coded with 1 = right-wing (AfD) and 0 = other. For GEPARTEEE outcomes, Voting Behavior and Party Identification were both coded from 1 = extreme left to 11 = extreme right
Job insecurity was also no significant predictor of trust in politicians, (b = .01, p = .894). As in the previous model, lower trust and less positive union attitudes increased the likelihood of identifying with and voting for right-wing parties. However, neither trust in politicians acting as a mediator nor union attitudes moderating these relationships were supported.
To test if federal states moderate the associations between job insecurity, voting behavior, and party identification (H3), we conducted two multigroup SEM and compared the results using the Wald χ2 test. The fit for the multigroup SEM with dichotomized outcomes was acceptable, at least regarding RMSEA and SRMR (RMSEA = .097, 90% CI [.093, .101], SRMR = .08), as well as for the GEPARTEE outcomes (RMSEA = .094, 90% CI [.089, .099], SRMR = .085). The results for each federal state can be found in the Supplementary Materials (S3). To test whether the paths from job insecurity to voting behavior and party identification differed between federal states, we conducted pairwise comparisons of the paths for two federal states using Wald tests. No significant differences were found in any comparisons: Brandenburg vs. Saxony (dichotomized outcomes: χ2(2) = 0.00, p = .998, GEPARTEE outcomes: χ2(2) = 0.00, p = .999), Brandenburg vs. Thuringia (dichotomized outcomes: χ2(2) = 0.01, p = .996, GEPARTEE outcomes: χ2(2) = 0.00, p = .999), and Saxony vs. Thuringia (dichotomized outcomes: χ2(2) = 0.00, p = .998, GEPARTEE outcomes: χ2(2) = 0.01, p = .999). Thus, we conclude that federal state does not moderate the relationships between job insecurity, voting behavior, and party identification and Hypothesis 2 was not supported.
Exploratory Analyses and Control Variables
Since our job insecurity measure could be considered a combination of qualitative job insecurity (fear of losing one’s job) and perceived employability (difficulty finding a new job), we ran the SEM again. This time, we used only the item assessing fear of losing one’s job as the predictor. The model fitted the data well (dichotomized outcomes: RMSEA = .061, 90% CI [.053, .069], SRMR = .06; GEPARTEE outcomes: RMSEA = .065, 90% CI [.056, .075], SRMR = .06). For the dichotomized outcomes, trust in politicians predicted both voting behavior (b = -.12, p < .001) and party identification (b = -.10, p <.001), and union attitudes predicted voting behavior (b = -.05, p = .011). Regarding the GEPARTEE outcomes, both trust in politicians and union attitudes predicted both outcomes significantly. The indirect paths were not significant, neither for the dichotomized outcome (voting behavior: b = .00, p = .396; party identification: b = .00, p = .400) nor for the GEPARTEE outcomes (voting behavior: b = -.00, p = .897; party identification: b = -.00, p = .897). Hence, the results remain consistent across operationalizations of job insecurity. The full results are reported in Table S14.
To examine whether union members and non-members differed, we conducted two additional multigroup SEM using union membership as grouping variable (dichotomized outcomes: 132 union members, 578 non-union members, GEPARTEE outcomes: 108 union members and 421 non-union members). The model fit for the dichotomized outcomes was close to acceptable (RMSEA = .096, 90% CI [.092, .050], SRMR = .08), and the model fit using the GEPARTEE outcomes was also acceptable (RMSEA = .093, 90% CI [.088, .099], SRMR = .08). For non-union members, trust in politicians and union attitudes were significant predictors of voting behavior and party identification with lower trust and less positive union attitudes increasing the likelihood of voting for and identifying with right-wing political parties. For union members, only low trust in politicians predicted identifying with a right-wing party. The Wald χ2 results showed no significant differences between the two groups (dichotomized: χ2(9) = 2.88, p = .959; GEPARTEE: χ2(9) = 3.81, p = .923). All results are provided in the SOM (S5).
The regression analyses to examine whether job satisfaction and right-wing voting were negatively associated showed no significant relationship between the two: dichotomized outcome: b = -.01, p = .596 and GEPARTEE outcome: b = .04, p = .670. Figure 3 displays the distribution of job insecurity across votes.
Figure 3
Relationship Between Job Insecurity and Voting Behavior
We also found no significant results when conducting two regression analyses to test for a U-shaped relationship between job insecurity and political orientation. Neither for dichotomized voting behavior nor for GEPARTEE outcomes were job insecurity (dichotomized: b = .03, p = .636, GEPARTEE: b = -.33, p = .318) and squared job insecurity (dichotomized: b = -.00, p = .774; GEPARTEE: b = .07, p = .198) significant predictors. Similarly, for political orientation, neither job insecurity (b = .14, p = .583) nor squared job insecurity (b = -.02, p = .657), were significant predictors.
We repeated all the analyses described earlier including income and job satisfaction as control variables (see S6). The results remained unchanged for the SEM (Tables S19 and S20) and for the multigroup-SEM with federal state as grouping variable (Tables S21-26), with one exception: income was a significant negative predictor for dichotomized party identification in Brandenburg (b = -.00, β = -.10, p = .025), meaning those with higher incomes were less likely to identify with the right-wing party AfD in Brandenburg. However, this was a very small effect. For the comparison between union and non-union members (Tables S27 – S30), all associations remained unchanged. The only significant effect of the control variables was found in the union-member group where job satisfaction was positively associated with right-wing voting behavior for the dichotomized outcomes. However, as this was a very small sample (n = 132), we would urge for caution in interpreting this effect.
Finally, we re-ran all analyses with the full sample consisting of employed and not employed participants (n = 1,123; all results are provided in S7). The SEM showed good fit for both dichotomized outcomes (RMSEA = .049, 90% CI [.046, .053], SRMR = .08, CFI = .95) and GEPARTEE outcomes (RMSEA = .045, 90% CI [.041, .050], SRMR = .08, CFI = .96). The path coefficients and significances remained unchanged (Tables S30 and S31), with the exception that job insecurity significantly predicted party identification (b = .25, p = .011). In the multigroup SEM with federal state as grouping variable, we observed three changes in the results compared to the overall results which we describe in the SOM (pp. 39–46, Tables S32-S37). The Wald tests indicated no significant differences in the paths between the three federal states. For the union vs. non-union member comparison, the multigroup SEM results were also consistent with the previous analyses (Tables S38-S41).
Discussion
This study explored how job insecurity is related to voting for right-wing parties, considering the potential mediating role of trust in politicians, the role of federal state context, and moderating effect of union attitudes. We found that lower trust in politicians and positive union attitudes were consistently linked to a lower likelihood of voting for or identifying with the far-right AfD and right-wing parties in general. However, contrary to expectations, job insecurity was not significantly associated with these political outcomes. There was no mediation effect by trust in politicians or moderation by union attitudes. These findings were consistent across federal states, union members, non-union members, after controlling income and job satisfaction, and among employed and non-employed participants. Furthermore, job satisfaction did not predict political outcomes, and no U-shaped relationship was found between job insecurity and political ideology.
Building on the concept of a psychological-democratic contract breach (Wroe, 2014, 2016), we argued that job insecurity might signal such a breach to employees, potentially reducing political trust and fostering more extreme voting behavior. However, our findings suggest that job insecurity may be too distant and narrow a construct to influence trust in politicians, political identification, or voting behavior directly. While job insecurity signifies uncertainty in the work domain, this uncertainty might be mitigated by stability in other life areas, preventing it from affecting political outcomes. Rather than individually perceived job insecurity, collective job insecurity perceptions might elicit stronger effects on political outcomes. For example, Shoss et al. (2023) observed that perceiving a higher prevalence of job insecurity in one’s country was related to stronger breach perceptions regarding government psychological contracts. This also aligns with the findings of Kiess et al. (2023), who found that perceptions of the economic state of the country are associated with right-wing extremist attitudes.
Additionally, there may be other mediating factors and boundary conditions in the relationship between job insecurity and political behavior. For example, job insecurity may also be related to emotional states, such as anger or contempt which, in turn, could predict extreme voting. Other moderators, like household financial stability (e.g., a partner's income or accumulated wealth), could also influence this dynamic, alongside union attitudes. For instance, Van Hootegem et al. (2025) observed that employees with high income inadequacy and job insecurity, in which is the most precarious profile in their study, reported lowest trust in government, democracy, and the economy.
However, our results can be considered to align with research showing that not economic factors, but rather factors related to social aspects, such as anti-immigration sentiments, or emotional states, such as anger, contribute to right-wing support (Erisen & Vasilopoulou, 2022; May, 2025). Furthermore, other factors such as the perceived social acceptability of voting for far-right parties might also play a role. In the federal states that were included in our analysis, the AfD received around 30% of the votes (Landeswahlleiter Brandenburg, 2024; Statistisches Landesamt des Freistaates Sachsen, 2024; Thüringer Landesamt für Statistik, 2024). Future research could examine whether this also influences voting behavior. Relatedly, when examining perceived job insecurity among voters of different parties (Figure 3), we found that voters of the liberal Free Democratic Party (FDP) reported higher levels of job insecurity than voters of other parties. However, since only 23 participants in our sample voted for the FDP, further research is needed to determine if differences in job security exist among voters. Ultimately, more theoretical exploration is needed to understand the links between job insecurity and political outcomes, the mechanisms behind these relationships, and potential boundary conditions.
Overall, this research expands our understanding of the relationship between job insecurity and political behavior, particularly in contexts of rising support for far-right parties (for a recent position paper, see also Vesper et al., 2026). Although research suggested that job insecurity might drive individuals toward extreme voting due to economic uncertainty (Geishecker & Siedler, 2012; Gomez et al., 2024), our results challenge this assumption. We did not find significant evidence that job insecurity directly predicts voting for right-wing parties. This suggests that job insecurity may not have the broad societal effects previously anticipated, suggesting a need for more nuanced models that account for multiple mediators like trust in political institutions or emotional responses like anger and frustration. Additionally, this study emphasizes the importance of political trust and union attitudes as predictors of radical voting behavior. Trust in politicians and positive union attitudes, rather than job insecurity, were predictors of voting for right-wing parties, indicating that broader dissatisfaction with political leadership and union attitudes may play a crucial role in shaping electoral choices. This finding underscores the need for further exploration of trust dynamics and union attitudes in political psychology. However, the overall effects and variance explained in both outcomes were rather small, highlighting that further research on explanatory mechanisms for right-wing voting and identification is needed.
Practical Implications
Our findings have several practical implications. First, trust in politicians is associated with a lower likelihood of voting for radical right parties, suggesting that politicians should prioritize building trust with citizens. Trust can be fostered through three key steps (Osterloh & Weibel, 2006). These include establishing clear rules and accountability to ensure predictability, demonstrating integrity, competence, and benevolence to show politicians act in the public's interest, and fostering a sense of identification, where citizens feel seen and represented by their leaders.
Second, our findings suggest that unions should focus on cultivating positive attitudes, particularly among non-members. This can be done through both cognitive and emotional pathways. Union representatives should clearly explain how the union's actions lead to desirable outcomes for individuals. Additionally, fostering emotional identification with union representatives will help people feel more represented.
Overall, the results suggest that reducing the appeal of far-right parties requires restoring trust in political institutions and unions. Politicians should prioritize transparent, accountable governance to build trust. Additionally, positive union attitudes were found to reduce far-right voting, highlighting the need to strengthen unions as a safeguard against extremist voting. Union leaders should focus on raising awareness of the collective benefits of membership, especially in regions with weaker union presence.
Limitations and Future Research
This study has some limitations that suggest avenues for future research. First, job insecurity and union attitudes were measured at T2 instead of T1, diverging from our preregistration. As a result, we can only discuss associations, not causal relationships. While we hypothesized that job insecurity influences voting behavior, it's possible that political attitudes also shape perceived job insecurity. Future research should use longitudinal designs to track individuals over time, providing clearer insights into the directionality of the relationship between job insecurity and voting behavior, particularly in the context of federal or state elections.
Furthermore, our job insecurity measure may not have been ideal despite being a validated measure, as it combines items examining fear of job loss (i.e., quantitative job insecurity; Shoss, 2017) and items asking how easy it would be to find a different job (i.e., employability; De Cuyper et al., 2008). However, our exploratory results using only the item assessing quantitative job insecurity yielded similar results. Nonetheless, future research should consider using scales that align better with definitions of quantitative job insecurity.
Another limitation is the small sample size in some analyses, particularly for union members. While we proceeded with the analyses, we advise caution in interpreting these results due to the sample size being below the recommended 250 participants per group. However, given that the results were consistent across groups, we are optimistic that they are generalizable. Our sample was also limited to participants from three federal states in Eastern Germany, which, while offering valuable insights, may not be generalizable to other national or regional contexts. The political, economic, and cultural landscape in these states may differ from other parts of Germany, Europe, or the world. Future research should examine these relationships in a broader range of settings to better understand the role of context in shaping the political consequences of job insecurity. Hence, we would encourage future research to conduct multi-level analysis, either comparing associations across countries or in smaller units, such as voting districts. This would also allow to compare regions with high and low unemployment rates and thus include objective as well as subjective job insecurity as potential predictors.
Finally, all data were self-reported, which may introduce concerns about common method bias, social desirability, recall error, and self-enhancement bias. To address these concerns, we used diverse response formats for focal variables, such as Likert-type scales and political party preferences, and applied two approaches to code the outcome variables. However, future research should aim to collect data from multiple sources, such as peer ratings and archival measures (e.g., union membership or voting records at the population level), to enhance the robustness of the findings.
Conclusion
This study explored the link between job insecurity and voting for right-wing parties, with a focus on trust in politicians as a potential mediator, as well as federal state context and union attitudes as a moderator. Our findings reveal several important conclusions. First, job insecurity did not significantly predict voting behavior or political identification with right-wing parties across the three federal states. This suggests that while job insecurity is important in domains like well-being, it may not directly influence political behavior, possibly because it is too narrowly focused on employment and does not extend to broader political attitudes, especially if other areas of life provide stability. Second and as in other studies, trust in politicians emerged as a significant predictor of voting behavior and political identification; lower trust was associated with a greater likelihood of supporting right-wing parties. This highlights the importance of political trust in shaping electoral choices and suggests that increasing trust in political institutions could help mitigate far-right voting. Third, union attitudes were associated with a lower likelihood of voting for far-right parties, though did not moderate the relationship between job insecurity and voting behavior. This indicates that positive union attitudes can serve as a protective factor against far-right preferences, independently of job insecurity. Overall, the study underscores the crucial roles of trust in politicians and union attitudes in shaping political behavior, suggesting that these factors may be more influential than job insecurity in determining electoral outcomes. Further research is needed to explore additional mechanisms linking employment-related insecurity to political participation and voting behavior.
This is an open access article distributed under the terms of the