Theoretical framework and hypotheses

A broad set of theoretical frameworks has been developed to explain political behaviour, several of which have been applied to migration research to understand how migrants adjust to the political environments of their residence countries and engage in political life.

Data sources

The empirical investigation is primarily based on data from the ESS, which includes data from multiple rounds: Round 5 (2010), Round 6 (2012), Round 7 (2014), Round 8 (2016), Round 9 (2018), Round 10 (2020–2022), and Round 11 (2023). As highlighted in the literature, the ESS adheres to rigorous standards, particularly in its questionnaire design, the cross-national equivalence of the instruments used, and the sampling procedures (Jowell et al. Reference Jowell, Roberts, Fitzgerald and Eva2007; Kohler Reference Kohler2007; Reference Kohler, Alber, Fahey and Saraceno2008). The ESS is a high-quality dataset, with a sampling design approximating a simple random sample and a relatively high response rate. It encompasses a wide range of countries and is representative of the general population in these countries. The broad coverage of the ESS makes it particularly valuable for cross-national studies, enabling not only the analysis of cross-country differences but also ensuring that questions are understood consistently by respondents across different languages and cultural contexts. Additionally, the survey provides contextual information about the country in which the respondent resides, allowing the integration of individual-level data with country-level contextual data. Since its inception in 2001, the ESS has become one of the most frequently cited data sources for research on a variety of topics (Malnar Reference Malnar2020).

‘Politics’ is one of the core modules of the ESS, and its questionnaire includes a wide range of political participation items, such as voting in national elections. This aligns closely with the focus of this study, which examines political incorporation. The ESS data are particularly valuable as it contains several key elements discussed in the theoretical framework, such as information on SES variables, political interest, country of birth (for both respondents and their parents), and the year of migration. These variables are critical to understanding electoral behaviour, and the availability of detailed individual-level data on these factors was a priority in the selection of this dataset for our analysis.

While the ESS is representative of the national population in participating countries, it is equally important that it accurately reflects the foreign-born population in the Western countries included in this study. To assess this, the percentage of foreign-born individuals in the ESS sample was calculated and compared with external data from Eurostat, the statistical office of the European Union. A Pearson correlation coefficient of 0.98 indicates a strong match between the foreign-born populations in both datasets, suggesting that the ESS sample closely represents the foreign-born population. It is important to note that the immigrant sample in this study includes both foreign-born respondents and second-generation immigrants – those born in the residence country with foreign-born parents. While it was not possible to directly assess the representativeness of second-generation immigrants in the ESS, the strong correlation between the foreign-born sample in the ESS and external data confirms that the ESS is reliable in representing the foreign-born population in the selected national samples.

In addition to the ESS, this study incorporates macro-level indicators to capture the political context of migrants’ countries of origin using the V-Dem Liberal Democracy Index (v2x_libdem). This index provides a continuous measure of liberal democracy in each country, ranging from 0 (low) to 1 (high). The Liberal Democracy Index has been used in previous research on migrants’ political attitudes and behaviour (Lazarova et al. Reference Lazarova, Saalfeld and Seifert2024; Gülzau et al. Reference Gülzau, Helbling and Morgenstern2025; Quaranta Reference Quaranta2025). It assesses the extent to which liberal democratic principles are realised by integrating several institutional and rights-based dimensions, including electoral democracy, civil liberties, the rule of law, judicial independence, and mechanisms of accountability that constrain executive power (Coppedge Reference Coppedge2023). For this study, we used the V-Dem Liberal Democracy Index scores for the period 2014–2024. To minimise short-term fluctuations and ensure temporal stability, we calculated 10-year average scores for each country, providing a more stable reflection of each country’s political culture.

We acknowledge that V-Dem data are often used to build categorical regime classifications. Following the classificatory approach of Lührmann et al. (Reference Lührmann, Tannenberg and Lindberg2018) and the Regimes of the World (RoW) framework, we derive our dichotomous indicator by applying a threshold to the Liberal Democracy Index: countries with a liberal democracy score ≥ 0.5 are coded as democracies, while those scoring below 0.5 are coded as autocracies. The detailed operationalisation of this variable is presented in Section 4: Measurement of Key Variables.

Measurement of key variables

Our dependent variable, electoral participation, is measured based on respondents’ self-reported voting behaviour in the most recent national election. It is derived from the question: ‘Some people don’t vote nowadays for one reason or another. Did you vote in the last [country] national election in [month/year]?’ Respondents could give the following answers: ‘Yes’; ‘No’; ‘Not eligible to vote’. For the purpose of this analysis, the variable was recoded into a dichotomous format, where respondents who reported voting in the most recent national election were coded as 1 (voted), and those who did not vote were coded as 0 (did not vote). While self-reported voting may be subject to social desirability bias (ie some respondents may overreport participation to conform to perceived norms), it remains a widely used and reliable indicator in survey-based electoral research (Sciarini and Goldberg Reference Sciarini and Goldberg2016).

Central to our analysis are four key independent variables: migratory background; spending formative years in the residence country, which captures early political socialisation in the residence country; length of stay as a measure of cumulative exposure; and the political culture of the country of origin, which captures pre-migration democratic (or authoritarian) cultures.

Migratory background is a key variable, distinguishing between four groups: first-generation immigrants (n = 14,216), who were born abroad and migrated to the residence country after formative years; 1.5-generation immigrants (n = 3,620), who arrived in the residence country during childhood or adolescence and thus spent their formative years there; second-generation immigrants (n = 5,359), who were born in the residence country to two foreign-born parents; and natives (n = 219,202), who were born in the residence country to native-born parents. This classification builds on widely used generational frameworks established in prior research (Rumbaut Reference Rumbaut1997), which divides individuals into different generations, such as 1.5 generation and second generation, based on parental nativity and age at migration. While these classifications reflect the diverse and complex trajectories of immigrant incorporation, they also highlight the nuanced contours of the phenomenon, which may not always be fully captured by standard statistical surveys (Lessard-Phillips Reference Lessard-Phillips2017). To ensure theoretical precision and avoid empirical bias, we apply a stricter operationalisation of second-generation immigrants. This definition excludes those with only one foreign-born parent, such as children of mixed-nationality couples. From a conceptual perspective, individuals with one native-born parent are more likely to be socialised into the dominant political norms of the residence country and tend to exhibit political behaviours more closely aligned with natives. Their inclusion would risk obscuring the unique effects of political socialisation. Empirically, these individuals constitute a substantial portion of the broader second-generation category (over 12,000 out of 17,600 cases), meaning they would disproportionately influence statistical patterns and potentially dilute the distinctiveness of those born to two migrant parents. This refined classification enhances our ability to detect meaningful differences across groups.

Spending formative years in the country of residence is operationalised using respondents’ age at the time of migration. This variable was calculated by subtracting the year of birth from the year of migration, thus capturing the age at which individuals moved to the residence country. Age at migration serves as a proxy for early political socialisation: those who migrate at younger ages are more likely to be socialised into the political norms and institutions of the residence society. We subsequently created a binary variable distinguishing between those who arrived before the age of 13 (often referred to as 1.5-generation immigrants; 3,620 respondents) and those who arrived after the age of 13 (first-generation immigrants; 14,216 respondents). The age 13 threshold was selected because it roughly corresponds to the transition from primary to secondary education in most European school systems, marking a key stage in children’s political learning, where schools play a central role as agents of political socialisation (Hoskins and Janmaat Reference Hoskins and Janmaat2019; Deimel et al. Reference Deimel, Hoskins, Abs, Kuang, Zhu and Kennedy2021). Immigrants who arrive before this stage are more likely to complete primary schooling in the residence country, be exposed to civic education curricula (Campbell Reference Campbell2013; Neundorf et al. Reference Neundorf, Niemi and Smets2016), and socialise with native-born peers during critical developmental years – all of which are central to early political learning.

In addition, the length of stay in the country of residence was derived by subtracting the respondent’s year of arrival from the year of the respective survey round. This calculation was performed separately for each wave of the survey to ensure temporal accuracy. Any resulting negative values were recoded as missing. This variable reflects the number of years an immigrant has resided in the residence country at the time of the interview and is used as a measure of cumulative exposure to the residence society.

To define the political culture of the country of origin, we relied on the V-Dem Liberal Democracy Index and the classification by Lührmann et al. (Reference Lührmann, Tannenberg and Lindberg2018). This index measures the level of democracy within a country, with scores ranging from 0 to 1, where 0 indicates closed autocracies and 1 represents fully liberal democracies, and the scores were derived from the V-Dem-CY-Core-v15 dataset for the period 2014–2024. For first-generation immigrants (ie individuals born outside the residence country), the score of their country of birth is used to capture the political environment they experienced prior to migration. For second-generation immigrants, the political culture is based on the countries of birth of their mother and father – information provided in the ESS dataset. By matching each parent’s country of birth with its respective democracy score, we construct a proxy for the political culture that may have been transmitted intergenerationally. To ensure robustness, we calculated the average V-Dem democracy score for each country across a 10-year period (2014–2024), thereby reducing the influence of short-term political fluctuations.

While it would be ideal to use a time-specific regime measure that matches each migrant’s formative years, data limitations make such an approach difficult to implement consistently. Following Lazarova et al. (Reference Lazarova, Saalfeld and Seifert2024), we therefore adopt a 10-year average of the V-Dem Liberal Democracy Index, which minimises short-term fluctuations and ensures a comparable measure of democracy levels across countries. We acknowledge that this choice may have theoretical implications, especially for first-generation migrants whose formative years pre-date the 2014–2024 reference period, as the political culture measure may not fully capture the context that shaped their early socialisation.

To assess the extent of this potential mismatch, we examined the distribution of year of migration in our sample and found that the vast majority of migrants arrived after 1968. We also inspected V-Dem historical maps (from 1960 onwards) and observed that the colour composition, reflecting liberal democracy scores, changes only gradually over time. Most countries display long-term continuity in their democratic versus autocratic status, which aligns with our final categorisation based on the 10-year average. Taken together, these checks suggest that the 10-year average captures stable and persistent regime characteristics that broadly reflect migrants’ origin-country political environments, even if it does not perfectly measure the exact institutional context experienced during their formative years.

In addition to regime type, our final variable also accounts for the geopolitical area of the individual’s country of birth (for first- and 1.5-generation immigrants) and parents’ countries of birth (for second-generation individuals). The categories based on this approach are as follows:

• • Western Democracies (0): Countries of birth (for individuals or both parents) with a liberal democracy score of 0.5 or higher, located in Western Europe, North America (United States, Canada), and Oceania (Australia, New Zealand).

• • Eastern European Democracies (1): Countries of birth (for individuals or both parents) from Eastern Europe with a liberal democracy score of 0.5 or higher.

• • Non-Western/Eastern European Democracies (2): Countries of birth (for individuals or both parents) located outside Western and Eastern Europe (eg in Latin America, Asia, or Africa) with a liberal democracy score of 0.5 or higher.

• • Electoral/Closed Autocracies (3): Countries of birth (for individuals or both parents) with a liberal democracy score below 0.5.

The final two residual categories are specific to second-generation individuals with parents originating from different regime types.

• • Western + Other Democracies (4): The regime has to be democratic for both parents, but one could come from a non-Western area.

• • Mixed Regimes (5): One parent from a democratic country and one from an autocratic country.

To ensure statistical reliability and interpretability, these categories were merged prior to analysis. Specifically, Mixed Regimes were grouped with Electoral/Closed Autocracies, and Western + Other Democracies with Western Democracies. This choice was both empirically and theoretically justified: small group sizes (eg only 59 second-generation immigrants from ‘Other Western’ countries, as shown in Table A2 in the online Supplementary File) risk unstable estimates, while the merged categories share key institutional features. For robustness checks, we also tested two alternative codings of regime type: one grouping Mixed Regimes and Western + Other Democracies with Western Democracies and another grouping Mixed Regimes with Autocracies while keeping Western + Other Democracies under Western Democracies. The results remained highly consistent across specifications, with only minor variations in coefficients and significance levels, confirming that the findings are robust to alternative classification choices.

Religious affiliation is operationalised based on respondents’ self-reported religious denomination in the ESS. The variable is categorised into four groups: Atheist, Christian, Muslim, and Other. Given the potential overlap between religion and the political culture of the country of origin – particularly since above 90% of Muslim immigrants in the sample originate from autocratic regimes (Table A3, Supplementary File) – religion is not included in the main model to avoid conceptual redundancy and potential confounding. Yet, to assess the robustness of the results, we replicated the main model by replacing the political culture variable with a measure of religious affiliation (the alternative specification is reported in Table A4 of the Supplementary File).

Other control variables considered in this study include demographic characteristics, SES, political interest, and residence country characteristics. Gender is categorised as Male = 0 or Female = 1, which allows for the examination of gender-based differences in political engagement. Age is included as a continuous variable, measured in years. SES variables, namely income and education, are also included, as they are widely recognised as key predictors of political participation. Income is measured subjectively by asking respondents about their household’s economic well-being, with responses ranging from ‘Living comfortably on present income’ to ‘Finding it very difficult on present income’. For analysis, we recategorised Income into three categories: Difficult (respondents who report ‘Finding it very difficult on present income’ or ‘Finding it difficult on present income’), Not difficult (respondents who report ‘Coping on present income’), and Wealthy (respondents who report ‘Living comfortably on present income’). Education is measured according to the International Standard Classification of EducationFootnote ¹ , with categories ranging from ‘Less than lower secondary education’ to ‘Tertiary education’. For simplification, we recategorised Education into three levels: Low (less than lower secondary education), Medium (upper secondary education and advanced vocational education), and High (tertiary education, including university degrees or equivalent qualifications).

Lastly, Political interest is included in the analysis as a key predictor of electoral participation. It is measured using a four-point scale, where respondents indicate how interested they are in politics: ‘Very interested’, ‘Quite interested’, ‘Hardly interested’, or ‘Not at all interested’. This variable reflects individuals’ attention to and engagement with political matters, which are widely recognised as important drivers of electoral behaviour.

Analysis

How much does SES really explain? Comparing the role of SES in electoral participation for natives and immigrants (H1)

Before analysing how the origin political culture shapes political participation, we first examined variations in voting behaviour across migrant generations (as shown in Table A1, Supplementary File). Native-born individuals reported the highest voting rate (mean = 0.80), followed by second-generation immigrants (mean = 0.74) and 1.5-generation immigrants (mean = 0.65). First-generation immigrants had the lowest turnout (mean = 0.52). These descriptive patterns suggest a clear generational gradient in political participation, consistent with theories of political socialisation.

We also examined educational attainment across migrant generations. First-generation immigrants were somewhat overrepresented in the low and high ends of the education spectrum, with 29% holding a high degree compared to 22% among natives. Second-generation immigrants showed a lower prevalence of low education (21%) and slightly higher medium-level attainment than their first-generation counterparts, suggesting a generational convergence in educational outcomes. Yet, first-generation immigrants faced significant economic disadvantages. They reported the highest level of financial difficulty (32%) and were least likely to live comfortably (24%). Second-generation immigrants showed signs of upward socioeconomic mobility, with a higher share living comfortably (27%).

In the following, we present an empirical assessment of our hypotheses in a series of analyses estimated using logistic regression models. The table presents a detailed examination of the relationship between SES and political participation for both immigrants and natives. The models progressively introduce interaction terms between migrant background and education, as well as income, to explore whether SES operates as a similar predictor for political participation across these groups. Concerning the goodness-of-fit statistics, model fit improved substantially from the null model (see Table A5 in Appendix) with country fixed effects (AIC = 241185.4, BIC = 241424.2) to the full model including SES, migrant background, and interactions (AIC = 212024.5, BIC = 212479.5), indicating that these variables contribute meaningfully to explaining electoral participation.

Across all models (available in Table A5 in the Supplementary File), first-generation immigrants show the strongest negative association with political participation. These negative coefficients remain statistically significant even after incorporating interactions with education and income. When interactions with education and income are included, the magnitude of the negative coefficients for first-generation immigrants does diminish slightly, indicating that while SES factors have some impact, they cannot entirely overcome the challenges posed by migration status. In other words, first-generation immigrants benefit less from high education and income in terms of turnout, suggesting persistent barriers to full political inclusion even among socioeconomically advantaged immigrants. One likely explanation is that educational qualifications obtained abroad, particularly from non-Western countries, are not always recognised in the residence countries (Lancee and Bol Reference Lancee and Bol2017), limiting the extent to which education translates into resources in the residence country. This aligns with our expectation that other, non-SES factors, such as length of stay and age at migration as well as the political culture of the country of origin, may play a critical role in shaping political incorporation for immigrants.

In comparison, 1.5-generation immigrants (those who migrated before age 13) show negative associations with political participation as well, but these effects are weaker than those of first-generation immigrants. This suggests that earlier migration may alleviate some of the barriers to political participation, but challenges still persist, especially when compared to second-generation immigrants. Second-generation immigrants, who were born in the residence country, show a participation pattern that is more similar to natives, with only a negligible negative impact. This indicates the potential role of early socialisation and acculturation in the residence country, supporting our hypothesis that this group, which either was born in the country or spent their formative years in the residence country attending school, experiences greater political incorporation due to early exposure to the political culture of the residence country. Similarly, for second-generation immigrants, SES factors such as education and income are more predictive of political participation, aligning with the expectations that, as these individuals are born and raised in the residence country, their SES functions similarly to natives.

Figures 1 and 2 show that while both education and income are positively associated with voting, their political payoffs vary significantly by migrant generation. Natives and second-generation immigrants exhibit strong and consistent returns to SES, while first-generation immigrants remain at a substantial disadvantage – even when highly educated or economically secure. This supports the argument that structural or institutional barriers, not just individual resources, shape political incorporation. In contrast, second-generation immigrants appear better integrated, with SES functioning more similarly to natives.

Figure 1.

Adjusted predictions of voting by migrant background and education.

Figure 2.

Adjusted predictions of voting by migrant background and income.

Overall, these findings underscore the complex relationship between SES and political participation, particularly in the context of migration background. As anticipated, these results indicate that higher levels of SES have a smaller contribution in the case of immigrants and cannot fully explain their political participation.

Evaluating the role of political culture (based on respondent or parental country of origin) (H2)

To test our argument that these differences are related to pre-migration cultural factors, H2, we take into account the ‘political culture’ of the country of origin. The findings support H2, indicating that origin-country political culture continues to shape electoral participation, perhaps contrary to the common presumption that democratic origins necessarily foster higher turnout. Once individual-level differences are controlled for in multivariate models, immigrants from non-democratic regimes (including electoral autocracies and mixed regimes) exhibit higher electoral participation than immigrants from democratic origins.

Descriptive statistics (Table A6, Supplementary File) show that immigrants from Western democratic-origin countries exhibit higher raw rates of electoral participation than those from autocratic regimes. Immigrants from democratic origins are, on average, older (mean age 48 vs. 45.2), more politically interested (13.5% very interested vs. 11.7%), better educated (29.4% with high education vs. 25.6%), and more economically secure (32.2% living comfortably vs. 25.1%). These factors are all strongly associated with higher political participation.

Robustness checks confirm that the positive effect of political culture of origin emerges only once individual-level confounders such as age are controlled for. In models excluding age, differences across regime types become statistically negligible, indicating that observed raw participation differences partly reflect age composition across origin groups rather than regime effects per se (Table A8, Supplementary File). Importantly, similar results are obtained when using an alternative operationalisation of political exposure, namely the proportion of life spent in the residence country, reinforcing the robustness of the findings. Once age is accounted for, however, the mobilising effect of autocratic origin becomes evident, supporting a reactive mobilisation mechanism, whereby exposure to democratic residence country institutions activates political engagement among immigrants from authoritarian contexts.

Moreover, the findings also reveal meaningful variation across the different geopolitical areas within the democratic category. As shown in Table 1, migrants from Eastern European democracies (b = 0.227, P < 0.01) and from non-Western democratic countries (b = 0.345, P < 0.01) have significantly higher electoral participation than migrants from Western democracies. Although these effects are smaller than those associated with autocratic origin, these results suggest that political culture effects are not limited to the autocracy–democracy divide but may also reflect broader regional and historical dimensions of democratic development.

Table 1.

Predictors of electoral participation among immigrants with generational status control (all immigrants)

Note: Robust standard errors in parentheses. Significant at *** P < 0.001, ** P < 0.01, * P < 0.05.

Model 2 further explores whether this pattern varies by generational status. The significant interaction terms indicate that the association between political culture of origin and electoral participation is strongest among first-generation immigrants. Specifically, the interaction term for 1.5-generation immigrants is negative and significant (b = –0.471, P < 0.001) and even more so for second-generation immigrants (b = –0.598, P < 0.001). These results suggest that the mobilising effect of autocratic origin on political participation diminishes across generations. Thus, while the results do not support the expectation that second-generation migrants from autocracies would participate less than their peers from democracies, they reveal a clear pattern of generational assimilation, whereby the influence of pre-migration political culture weakens substantially over generations.

Evaluating the role of formative years, length of stay, and pre-migration political culture (H3–H4)

Migrants born outside the residence country (first-gen and 1.5-gen immigrants)

This model investigates the predictors of electoral participation among first-generation and 1.5-generation immigrants – the latter defined as those who migrated to the residence country during their formative years, before the age of 13. We analyse these groups separately from second-generation individuals for both methodological and theoretical reasons. Methodologically, variables such as length of stay and age at arrival vary only for first- and 1.5-generation immigrants. Theoretically, this separation allows us to test specific hypotheses that focus on the role of timing of arrival – whether individuals arrived during their pre-adult years – and the duration of residence in shaping political socialisation and electoral participation (H3–H4).

Three core variables were included to capture key influences on immigrant political participation: whether the immigrant spent his/her formative years in the residence country, the length of stay, and the political regime of the country of origin. While the first two reflect immigrants’ exposure to the residence country environment and their potential for political socialisation, the third captures the political context in which immigrants were born. The analysis tests the influence of early-life socialisation (H3), length of residence (H4), and interaction effects between political culture of origin and these socialisation variables (H5a, H5b).

The results are shown in Table 2. In line with H3, the findings in Model 1 show that immigrants who arrived in the residence country before the age of 13 are significantly more likely to vote compared to those who arrived later (β = 0.640, P < 0.001). This effect remains statistically significant and even strengthens in Model 2 after introducing interaction terms (β = 0.874, P < 0.001). These results provide robust support for H3, reinforcing the argument that early childhood socialisation in the residence country fosters stronger political integration.

Table 2.

Predictors of electoral participation among foreign-born immigrants (first- and 1.5-generation)

Note: Robust standard errors in parentheses. Significant at *** P < 0.001, ** P < 0.01, * P < 0.05.

Length of stay in the residence country is another positive and significant predictor of electoral participation across both models (Model 1: β = 0.039; Model 2: β = 0.043; both P < 0.001). This supports H4. It should be noted that to avoid collinearity between early arrival and length of stay, the length of stay variable was group-mean centred within early and later arrivers. This strategy allowed us to see how early and late arrivers differ in their likelihood of voting, assuming they have each stayed in the country for an average amount of time. It also shows how additional years of residence in the country affects voting within each group – early and later arrivers – separately. The findings support the view that early exposure to the residence country promotes political incorporation. In our sample of citizens of foreign origin living in European democracies, the findings support the view that early exposure to the residence country promotes political incorporation, particularly among those who arrive before age 13, while it should not be read as an overarching pattern across all migrant-origin groups or institutional contexts. This finding aligns with integration-based theories of political participation, which emphasise the importance of cumulative exposure to democratic institutions in shaping political engagement.

Moreover, the interaction results reveal important heterogeneity, confirming that the impact of pre-migration political culture varies across (re)socialisation pathways, consistent with H2. The interaction between length of stay and pre-migration political culture demonstrates that prolonged residence boosts electoral participation for all immigrants, but the magnitude of this boost varies by regime of origin. The negative and statistically significant interaction term (β = –0.008, P = 0.01) shows that immigrants from autocratic countries experience a smaller participatory gain per additional year of residence compared to those from democratic origins. This supports H5a (Political Culture × Length of Stay) and suggests that political incorporation over time is more challenging for individuals coming from authoritarian rule. As Figure 3 illustrates, both groups benefit from longer residence – immigrants from autocracies start out with a participation advantage at shorter durations of stay, but democratic-origin migrants experience a steeper participatory boost over time. This highlights that while time in the residence country facilitates political incorporation, its effects are conditioned by the pre-migration political socialisation.

Figure 3.

Adjusted predictions of voting based on political culture of origin country and length of stay in residence country (centred).

Moreover, the interaction between spending formative years in the residence country and political culture of origin country reveals that the benefits of early exposure to the residence country institutions and norms are not uniform. The coefficient for autocratic-origin immigrants who arrived before age 13 is negative and significant compared to the reference group (democratic-origin immigrants who arrived after age 13) (β = –0.39, P = 0.001).

Figure 4 reveals that migrants from autocratic regimes who arrived after age 13 exhibit a higher predicted probability of voting than their peers from democratic origins at the same age of arrival. This suggests that migrants from autocracies may experience reactive mobilisation upon settlement, as the migration process itself opens political opportunities unavailable in their country of origin. However, the graph also shows that early arrival (before age 13) is associated with higher electoral participation for both groups – yet this participatory boost is markedly larger for migrants from democracies. Migrants from democratic origins who arrived early have the highest predicted probability of voting overall, indicating that early socialisation in the residence country reinforces democratic-origin political dispositions. By contrast, while early arrival still raises participation among autocratic-origin migrants compared to their late-arriving peers, the size of this increase is smaller. This pattern highlights that the benefits of early exposure to residence country institutions are conditional on prior political socialisation: democratic-origin migrants are better positioned to convert early exposure into political engagement, supporting H5b (Political Culture × Formative Years).

Figure 4.

Adjusted predictions of voting based on political culture of origin country and spending formative years in the country.

In sum, while both early arrival and longer residence in the residence country positively influence immigrant electoral participation (supporting H3 and H4), their effects are significantly moderated by the political culture of the country of origin. Immigrants who were socialised in democratic contexts benefit more from both early childhood exposure and cumulative residence. In contrast, immigrants from authoritarian regimes gain less from these same integration pathways, consistent with H5a and H5b.

Notably, even among immigrants who arrived before age 13, those from autocratic origins show lower predicted probabilities of voting than their peers from democratic backgrounds. This suggests that early arrival alone is insufficient to overcome the deep-rooted effects of authoritarian political socialisation. These findings underscore the enduring influence of regime type in the country of origin – even when immigrants arrive early enough to undergo schooling and socialisation within the residence country’s democratic environment.