Open Access

(Why) are immigrants unhappy?

IZA Journal of Migration20165:3

https://doi.org/10.1186/s40176-016-0052-4

Received: 4 November 2015

Accepted: 14 January 2016

Published: 11 February 2016

Abstract

Recent studies suggest that migrants may be less satisfied with their ‘new’ lives than members of the host population and worry that this may be driven by cultural factors, such as feelings of not belonging. Motivated by this concern, this paper analyses the life satisfaction of immigrants once settled in the host country. We rely on the German Socio-Economic Panel’s immigrant sample for the years 1984–2010 and find that while immigrants are less satisfied than natives, this difference can be explained by factors related to economic integration, such as the details of their employment conditions, rather than cultural factors such as feelings of not belonging, which often loom large in the public mind.

JEL codes: J15, K37, O15

Keywords

Integration Subjective well-being Segregation Citizenship law

1 Introduction

In the popular mind, migration is often associated with an increase in well-being, as embodied in narratives of the ‘promised land’ and a move ‘in hope of a better future’. In recent years, there has been increasing interest in measuring subjective well-being in economics; it has also featured prominently in public discourse and debates (e.g. the Stiglitz-Sen-Fitoussi Report on social progress and well-being, commissioned by Nicholas Sarkozy; the British National Well-Being Project embraced by David Cameron). Recent studies have also increasingly looked at the impact of migration and integration on life satisfaction and have raised concerns that migrants may be less satisfied than the native population and that this could be driven by cultural factors, such as feelings of not belonging. We thus aim to contribute to the literature by examining immigrants’ outcomes from a subjective perspective, looking in particular at whether cultural or economic factors matter more for their life satisfaction. We analyse the immigrant sample of the German Socio-Economic Panel, for the years 1984–2010 (for a detailed discussion of the dataset please see Section 4).

We will examine two questions: (1) we will first look at whether (and if so, why) immigrants are less satisfied than natives (as suggested by some recent papers) and (2) we will then analyse the determinants of the subjective well-being of immigrants, in a sense examining ‘who does better and who does worse’ among the group of migrants.1 Card et al. (2012) recently highlighted that public attitudes towards immigration policy are more influenced by cultural and social concerns than economic ones. Georgiadis and Manning (2013) note the widespread belief that societies will function better if they manage to establish a common sense of identity among the population and contemporary fears in many countries that this common identity is threatened. We will look at various socio-economic and cultural measures that could be associated with integration, trying to shed light on the question which elements matter for subjective well-being.

As Graham (2008) put it: ‘happiness surveys may shed light on how the direction and nature of progress affects well-being’ (p. 85, italics in the original). Such potential policy implications may be interesting in the context of immigration as public discourse often debates the ‘alternatives’ of integration, assimilation or multiculturalism. Research on the subjective well-being of immigrants could thus have broader implications for social cohesion and hence policy.

The paper is structured as follows: Section 2 reviews the empirical literature on the life satisfaction of immigrants, Section 3 discusses our empirical approach and Section 4 introduces the dataset used. Section 5 presents the results including robustness tests and Section 6 concludes.

2 Literature review

In recent years, there has been increasing interest in the migration literature in subjective well-being, primarily linked to the question whether migration increases life satisfaction. Some studies have explicitly compared life satisfaction before and after migration. Stillman et al. (2012) provided a key contribution to this field, relying on a natural experiment to compare successful and unsuccessful applicants to a migration lottery to experimentally estimate the impact of migration on objective and subjective well-being. A recent paper by Melzer and Muffels (2012) used the German reunification as a natural experiment to examine whether migration from East to West increased life satisfaction, focusing on income effects as well as the role of social comparisons. Amit and Riss (2014) compared subjective well-being pre- and post-immigration among North Americans who arrived in Israel during the last two decades and found that social networks, a religious motive for migration and work satisfaction played important roles. Hwang et al. (2011) found negative effects of involuntary migration on migrants’ well-being, despite relative gains in housing quality. Bartram’s (2011) study circumvented the data constraint by contrasting the ‘marginal increase in satisfaction from income’ among natives and migrants and found that the association between income and subjective well-being is indeed stronger for immigrants than for natives—but even for immigrants that association is still relatively weak. Looking at economic migration and life satisfaction around the world, Olgiati et al. (2013) found a distinctive immigrant advantage in translating income into higher life satisfaction in some countries but also found a number of ‘frustrated achievers’ (immigrants who report a negative association between income and life satisfaction) in some other countries.

Graham and Markowitz (2011) took one step back and looked at the selection effect, examining whether migrants are different from the rest of the population in terms of subjective measures of well-being. They found that people who intend to migrate are generally less satisfied than those who do not (as could be expected); however, more broadly, they found that migrants are ‘frustrated achievers’ (as also noted by Olgiati et al. 2013), people who have relatively high levels of objective well-being, such as income, but who are nonetheless dissatisfied with their situations and seek to improve them via migration. Cai et al. (2014) found that individuals with higher subjective well-being have lower international migration desires and that, at the individual level, the subjective well-being-migration relationship appears to be more robust than the income-migration relationship. Knight et al. (2007) examined the question why rural-urban migrant households settled in urban China have an average life satisfaction score lower than that of rural households. Three basic hypotheses were considered: migrants had false expectations about their future urban conditions, about their future urban aspirations or about their future selves. The disparity appears to be driven both by certain features of migrant conditions (in line with the classical literature on the psychology of migration and acculturative stress2) and by high aspirations (in line with Graham and Markowitz’s (2011) notion of ‘frustrated achievers’). In a related vein, Navarro-Azorín and Artal-Tur (2015) analysed the differences in well-being between Spanish municipalities reflected by people’s migratory decisions, and Chindarkar (2014) looked at a sample of Latin American countries and found that life satisfaction is a significant driver of intention to migrate abroad. Lovo (2014) examined the preferences over migration destinations of those revealing a desire to permanently leave their country and found that people’s preferences are better explained by average levels of life satisfaction in the destination country. Bartlett et al. (2010) highlighted the impact of out-migration on life satisfaction in rural areas. Betz and Simpson (2013), Akay et al. (2014) and Longhi (2014) looked at the impact of immigration on the well-being of natives.

A number of recent papers have looked at the well-being of immigrants once settled in the host country and have highlighted the relative dissatisfaction of migrants compared to the native population (e.g. Werkuyten and Nekuee 1999; Baltatescu 2007; Constant et al. 2012; Amit 2010; Bartram 2011; Safi 2010; Gokdemir and Dumludag 2012; Obućina 2013; Hendriks et al. 2014 referred to it as the ‘migrant-local happiness-gap’). Two explanations were often brought forward for this: discrimination and feelings of not belonging or loyalties towards immigrants’ countries of origin. The few papers that explicitly linked immigrants’ ethnic/national identities to their life satisfaction usually found positive effects of strong majority identities and negative effects of strong minority identities on subjective well-being (Boski 1989; Phinney et al. 2001; Bartram 2011).

We hope to contribute to the existing literature on subjective well-being and migration by focusing on the life satisfaction of immigrants once settled in the host country. In particular, we aim to provide value added by looking at whether it is economic or the much-discussed cultural concerns which matter more for immigrants’ life satisfaction. First, we examine whether immigrants are indeed less satisfied than the natives (as suggested by the existing literature). Second, we study the determinants of the life satisfaction of immigrants in greater detail, analysing the role of socio-economic as well as cultural factors. We attempt to deal with estimation difficulties resulting from the endogeneity of current characteristics (such as identity,3 discrimination or social contacts) and omitted variable bias (primarily due to unobserved personality traits) by relying on panel data. We use a fixed effects estimator to account for unobserved individual heterogeneity, examine changes over time and use lags to deal with reverse causality. We also examine restricted subsamples to assess robustness.

3 Estimation

Psychologists argue that life satisfaction is both a ‘trait’ and a ‘state’, i.e. it has a stable, ‘genetic’ component, linked to personality traits such as extraversion, conscientiousness and emotional stability, but it is also affected by circumstances, and these two factors can of course interact in influencing life satisfaction.4 In our empirical approach, we will thus explicitly acknowledge that subjective well-being depends on the interaction of a ‘genetic’ component, linked to personality traits, and circumstances, some of which may be more persistent than others. Most papers have looked at the determinants of life satisfaction relying on simple ordinary least squares (OLS); however, since the ‘trait’ and the ‘state’ components may interact in affecting life satisfaction (i.e. unobserved personality traits may influence subjective well-being, while also being correlated with the characteristics on the right-hand side), not taking into account this unobserved individual heterogeneity leads to biased estimates of the parameters of interest. We thus rely on panel data to compare a pooled OLS estimator with an estimator including individual fixed effects.5 Whereas the pooled OLS estimator allows us to look at the effects of stable as well as changing characteristics or circumstances, but may be confounded by personality traits, the fixed effects estimator uses repeated observations on an individual over time to parse out the ‘trait’ component to give cleaner measures of the ‘state’ component. Although this does not provide value added for the estimation of the effect of the immigrant dummy variable itself (which here by definition is time invariant), and we of course realize that for some variables such as education there is little time variation in our sample, we believe that comparing pooled OLS and fixed effects estimators can give us useful additional insights into whether, for instance, the roles of citizenship or details of employment conditions are driven by a selection effect or changes over time.

As noted before, we are interested in two questions: (1) whether immigrants are less satisfied than natives (as suggested by the existing literature) and (2) what factors influence the life satisfaction of immigrants. We thus start our analysis by looking at the full sample, including both natives and immigrants to examine the first question, and will then restrict the sample to the immigrants only to examine the second.

The existing literature stresses the importance of individual characteristics such as age, gender, marital status and years of education, the harm done by unemployment or by competitive struggles among individuals and the negative effects of serious illness. For a comprehensive early review on the determinants of subjective well-being see, e.g. Diener (1984). While we follow the literature in including controls for these characteristics, we extend this set to include variables which are of special interest in a migration context—please see Table 1 below. For a detailed discussion of each of these variables and available empirical evidence on them, please see Table 7 in the Appendix.
Table 1

Control variables

Personal and parental characteristics

Age, gender, marital status, children, education, employment, home ownership, parental education, rural/urban childhood, health

Migration-related variables

Years since migration, identity, German language skills, ethnic composition of neighbourhood, citizenship

4 Data6

We use data from the German Socio-Economic Panel (GSOEP), a large representative longitudinal survey of private households in Germany. Immigrants are oversampled in the GSOEP: our focus here is on the ‘B Foreigner West sample’, defined as those households where the head of household is Turkish, Italian, Spanish, Greek or from the former Yugoslavia. During the latter half of the 1950s, the German government started actively recruiting guest workers in response to a labour shortage prompted by high economic growth rates. In 1973, the government stopped the recruitment of further guest workers as Germany entered a period of economic recession. In the subsequent years, the inflow of immigrants from the former guest worker countries consisted mainly of family members of those guest workers who remained in Germany (family reunification).

The GSOEP was first administered in 1984; we use data until (and including) 2010. In the GSOEP, all household members are interviewed individually once they reach the age of 16. In principle, all persons who took part in the first wave of the survey as well as their children whenever born are surveyed in the following years. ‘New’ persons thus become part of the GSOEP by being born into a surveyed household/reaching the minimum respondent’s age or by moving into a surveyed household (either from somewhere else in Germany or from abroad). Individuals who move out of a surveyed household or split off into new households are still followed (within Germany) but under a new household identifier. Individuals moving into existing GSOEP households are followed even if they subsequently leave that household. People exit the sample by death, moving abroad or refusal to interview. Individuals or households which could not be successfully interviewed in a given year are followed until there are two consecutive temporary dropouts of all household members or a final refusal. As noted above, compared to the German population, immigrants are oversampled: the ‘B Foreigner West sample’ started with 1393 households, and the sampling probability was about 0.0008. We have not reweighted the data, so our pooled regressions of natives and immigrants should not be seen as representing the German population. We restrict our analysis to first-generation immigrants as we do not wish to pool first- and second-generation immigrants together and the sample of second-generation immigrants who were asked about our variables of interest is unfortunately too small for separate analysis. However, we do not explicitly restrict the age range. Our sample is also not restricted to adults present at the start of the panel. Individuals can re-enter our sample under the same identifier if they drop out only temporarily. Children present in ‘native’ GSOEP households are allowed to enter our sample when they reach the minimum respondent’s age. Our results are robust to explicitly enforcing common support on age (as discussed in Section 5.5) and to excluding them to match the exclusion of second-generation immigrants. We include all adults in the household who answered our questions of interest—our results are robust to clustering standard errors at the household level. Our findings also hold up when controlling for the number of years an individual is present in the panel (the median number of years in our sample is 9 for natives and 11 for immigrants). We also rely on the ‘Microm indicators’, a unique dataset providing information on the ethnic composition of neighbourhoods at various levels of aggregation, which can be merged with the GSOEP.7 These indicators are discussed in detail in Section 5.3.

We use the standard life satisfaction survey question as our dependent variable:

How satisfied are you with your life, all things considered?

with answer categories ranging from ‘completely dissatisfied’ (0) to ‘completely satisfied’ (10). This Satisfaction With Life Scale (SWLS) was developed to assess satisfaction with the respondent’s life as a whole. The scale does not assess satisfaction with life domains such as health or finances separately but allows subjects to integrate and weigh these domains in whatever way they choose. It therefore assesses an individual’s conscious evaluative judgment of his or her life by using the person’s own criteria (Pavot and Diener 1993). In the following, we briefly discuss the reliability of this scale as a measure of life satisfaction.

The SWLS has been shown to be psychometrically sound and well-validated (see, e.g. Diener 1984 or Pavot and Diener 2008 for a review): it shows good convergent validity with other scales and with other types of assessments of subjective well-being. Its scores have been shown to correlate with measures of mental health and to be predictive of future behaviours such as suicide attempts (Diener et al. 1985).8 New evidence from neuroscience also supports the reliability of life satisfaction questions: Davidson (2002, 2004) identified areas in the prefrontal cortex where the level of electrical activity is highly correlated with self-reported happiness, both across people and ‘within’ people over time. Larsen et al. (1983) also found that such single life satisfaction measures did not seem to be highly contaminated by social desirability.

Given our comparison between natives and immigrants, an important concern is culture-based anchoring bias, i.e. cultural differences in response styles, so that people in different cultures might report different answers to the same question, even if in other respects their life quality is the same (see, e.g. Oishi (2002) for a discussion of concepts like ‘happiness’ or ‘well-being’ in different languages and cultures or Senik (2011) on the ‘French unhappiness puzzle’). One basic check is to see to what extent the answers drawn from different countries and cultures appear to be influenced by the same factors: Rose and Özcan (2007) found that the determinants of life satisfaction in Turkey are very similar to those observed in the EU15, increasing our confidence in these measures. Also, overall means are roughly in line according to both the Eurobarometer and World Values Survey results.

There is evidence that momentary mood influences subjects’ responses to subjective well-being questions (Schwarz and Clore 1983). This finding is consistent with memory research (e.g. Natale and Hantas 1982), which shows that people tend to recall past events that are consonant with their current affect. However, despite the influence that current mood can have on subjective well-being measures, Kammann and Flett (1983) and Kammann et al. (1979) presented evidence indicating that this does not substantially distort scores, with both current mood and long-term affect being reflected in life satisfaction measures. Diener and Larsen (1984) found substantial amounts of cross-situational consistency in mean levels of person affect. Life satisfaction as assessed by the SWLS shows a degree of temporal stability (e.g. 0.54 for 4 years), with part of this being explained by personality and part by the stability of conditions in the respondents’ lives. Yet, it has sufficient sensitivity to be able to detect changes in life satisfaction (Pavot and Diener 1993).

Overall, subjective well-being measures seem to contain a substantial amount of valid variance; however, of course, the limitations of using such self-reported data need to be kept in mind in the following analysis.

4.1 Descriptive statistics

Table 2 reports descriptive statistics for the natives and first-generation immigrants in 2010 and looks at economic as well as cultural characteristics. In terms of simple averages, the natives seem slightly more satisfied than the first generation (7.1 versus 6.7).
Table 2

Descriptive statistics, 2010

 

Mean

 

Natives

Immigrants

Life satisfaction

7.112

6.678

Age

44.203

51.392

Male

0.485

0.515

Married

0.583

0.781

Years of education

12.465

9.996

Employed

0.51

0.447

Works in occupation trained for

0.68

0.642

Owns house/apartment

0.589

0.362

Has insecure job

1.636

1.879

Serious illness

0.407

0.365

Years since migration

 

33.934

Minority identity

 

0.58

Majority identity

 

0.202

German citizenship

 

0.245

Speaks German

 

0.837

Writes German

 

0.593

Note: All differences between groups are significant at the 1 % level

The immigrants are somewhat less educated, have less secure jobs, are less likely to work in the occupation they were trained for and are less likely to own the house/apartment they live in (in line with the findings of Constant et al. 2012). Almost 60 % of immigrants in our sample have a strong minority identity, while 20 % have a strong majority identity (in terms of overlaps, 10 % have a dominant majority identity, 43 % have a dominant minority identity, 39 % have two strong identities and 8 % have two weak identities). Around a quarter of the first generation in our sample has German citizenship by 2010.

In terms of life satisfaction, there is almost as much variation ‘within’ individuals over time as there is between them (standard deviations of 1.30 and 1.49, respectively) and there is substantial movement between categories over time; the correlation of this year’s life satisfaction with last year’s is 0.59, for 5 years ago, it is 0.44 (this is in line with the existing literature). We do not observe systematic differences between immigrants and natives in this respect.

5 Results

5.1 Are immigrants less satisfied?

Table 3 below reports fixed effects results as well as pooled OLS for comparison, for a combined sample of natives and first-generation immigrants. As noted before, while most of the literature on subjective well-being relied on simple OLS, we also examine a fixed effects specification to parse out the effects of the ‘trait’ component, in particular unobserved personality traits, to obtain cleaner measures of the effects of circumstances (the ‘state’ component). Again, while we realize that unfortunately this provides little value added for some variables which have no or little time variation in our sample (such as the immigrant dummy variable or education), we hope to gain insights by comparing the pooled OLS and fixed effects estimators, for instance, when looking at the impact of the details of employment conditions or certain cultural factors.
Table 3

Are immigrants less satisfied?

 

POLS

FE

POLS

FE

POLS

FE

 

(1)

(2)

(3)

(4)

(5)

(6)

Age

0

−0.027***

−0.011***

−0.044***

−0.005

−0.029***

 

(0.002)

(0.003)

(0.002)

(0.003)

(0.003)

(0.006)

Male

−0.039

(omitted)

0.125**

(omitted)

0.103

(omitted)

 

(0.043)

 

(0.045)

 

(0.057)

 

Married

0.105

0.046

0.037

0.184**

0.081

0.085

 

(0.059)

(0.091)

(0.064)

(0.065)

(0.088)

(0.139)

Separated/divorced

−0.572***

0.004

−0.492***

−0.006

−0.510***

−0.03

 

(0.103)

(0.157)

(0.1)

(0.103)

(0.138)

(0.219)

Widowed

−0.248

−0.752*

−0.545**

−0.091

−0.848**

−1.898**

 

(0.151)

(0.317)

(0.191)

(0.214)

(0.296)

(0.669)

Years of education

0.033***

0.01

0.008

0.034*

−0.001

0.026

 

(0.008)

(0.029)

(0.01)

(0.017)

(0.012)

(0.057)

Employed (lagged)

0.224***

0

0.098*

0.05

0.182*

0.087

 

(0.041)

(0.049)

(0.043)

(0.035)

(0.078)

(0.093)

Immigrant

−0.276***

(omitted)

0.087

(omitted)

0.072

(omitted)

 

(0.055)

 

(0.061)

 

(0.09)

 

Works in occup.

  

0.131***

0.086**

0.012

0.006

   

(0.037)

(0.029)

(0.051)

(0.067)

Owner

  

0.238***

0.099*

0.178**

0.1

   

(0.044)

(0.048)

(0.062)

(0.12)

Children in household

  

−0.005

0.001

0.001

−0.055

   

(0.039)

(0.038)

(0.057)

(0.082)

Insecure job

    

−0.430***

−0.236***

     

(0.032)

(0.035)

Number of obs.

71,779

71,779

29,835

29,835

28,369

28,369

Number of ind.

8256

8256

3929

3929

3888

3888

Note: Standard errors in parentheses (clustered at the individual level). * denotes significance at 5 %, ** at 1 % and *** at 0.1 %. Columns 3−6 also control for parental education, urban/rural childhood and health. 'POLS' refers to a pooled ordinary least squares specification, 'FE' to fixed effects

The first two columns show a simple specification: in line with much of the earlier literature, we find a negative effect of age and a negative effect of being divorced or widowed. Years of education has a positive effect, though this is no longer significant with fixed effects—unsurprising, since, given our sample of adults, education does not change much over time ‘within’ individuals. As expected, employment has a positive effect (it is lagged here to avoid reverse causality, similar positive effects are found for contemporaneous values or longer lags). In line with the earlier literature, which used similar specifications, we find that the first generation still seems less satisfied than the natives, even controlling for characteristics such as education and employment. In fact, the coefficient on the immigrant dummy (−0.276) roughly corresponds to the difference observed in raw means (6.7 for the migrants versus 7.1 for the natives).9

Columns 3 and 4 examine whether this relative dissatisfaction effect holds up when we introduce further control variables. Note that the large fall in the number of observations between the two specifications is driven by the fact that the health question was not asked each year and that information on parental education and rural/urban childhood is not available for all respondents. Repeating the simple specification of columns 1 and 2 on this smaller sample yields very similar results (please see Table 8 in the Appendix). Our main finding is that the first generation is no longer less satisfied than the natives; the effect is mopped up by the details of employment conditions, in particular whether the respondent works in the occupation they were trained for.10 This question is not usually asked explicitly in other surveys and may capture employment effects that are not picked up in other analyses. The GSOEP first asks respondents about their current position/occupation (‘What is your current position/occupation? Please give the exact title.’) and then asks, ‘Is this position the same as the profession for which you were educated or trained?’, with answer categories ‘yes’, ‘no’, ‘currently in education/ training’ and ‘I have not been trained or educated for a particular profession’ (the last two categories are omitted here; the analysis focuses on the ‘yes’-‘no’ binary distinction). Younger respondents are more likely to work in the occupation they were trained for; this is the case both for immigrants and for natives (64 % of immigrants in the 20–30 age group work in the occupation they were trained for, as opposed to 58 % of the 31–50 year olds and 61 % of 51–65 year olds; the corresponding percentages for natives are 76, 64 and 64 %), though the immigrant-native gap is also largest for this group. Lower educated immigrants and natives are also less likely to work in the occupation they were trained for (58 % of immigrants with up to 12 years of education work in the occupation they were trained for, as opposed to 67 % of those with more than 12 years of education; the corresponding shares are 67 and 76 % for natives). Immigrants are less likely to work in the occupation they were trained for in the first few years after migration (only 54 % of those who migrated up to 5 years ago work in the occupation they were trained for, or 57 % of those who arrived up to 10 years ago, as opposed to 60 % for those who migrated more than 10 years ago) and if they arrived at an older age (64 % of those who arrived under the age of 20, as opposed to 57 % of those who arrived between the ages of 21 and 40, work in their original occupation; ratios are similar splitting the sample at age 14). Immigrants with and without German citizenship do not appear to be systematically different along this dimension; German language skills however matter: 55 (56) percent of those who speak (write) German ‘poorly’ or ‘not at all’ work in the occupation they were trained for, as opposed to 62 (66) percent of those who speak (write) German ‘well’ or ‘very well’—the latter are close to the shares for natives.

This specification looked at the effects of variables, which were as predetermined as possible with respect to current life satisfaction (occupational choice occurred in the past, before migration). Although we would also be interested in the effects of current characteristics, such as the degree of job insecurity, here we are more worried about reverse causality: that people who are less satisfied with their lives as a whole select into more insecure jobs, possibly creating a downward spiral. As it is difficult to find valid instruments for job insecurity, which would not affect subjective well-being, we did not include this measure in our preferred model. Columns 5 and 6 extend the specification of columns 3 and 4 by also including a measure of the insecurity of the respondent’s last job.

Job insecurity is a subjective measure here, assessed on a one-to-three scale,11 and shows a highly significant negative effect. Note that this is likely to be correlated with whether the respondent works in the occupation they were trained for, explaining why this variable is no longer significant. To try to deal with the likely endogeneity of job insecurity (and possible bias due to its subjective nature), we examine robustness to using a predicted measure of job insecurity instead. Relying on a first-stage regression of job insecurity on years with the firm, years of employment and its square (to allow for non-linear effects) and occupational categories, and using these predicted values in the second stage (correcting standard errors using bootstrapping), we find that job insecurity still has a highly significant negative effect, whereas the first-generation immigrant dummy variable is no longer significant. Our results are in line with the findings of Clark et al. (2010) who looked at the impact of job insecurity more generally and found that (at least for men) it significantly reduces well-being and is more important than the simple employment-unemployment distinction. Origo and Pagani (2009) noted that what matters for job satisfaction is not just the type of contract but mainly the perceived job security, which may be independent of the type of contract.

Note that job insecurity (as well as working in the occupation trained for in columns 3 and 4) had significant effects in both the pooled OLS and the fixed effects specifications, thus even once we look at the cleaner measure, removing the effects of unobserved individual heterogeneity. The Hausman test favours the fixed effects specification for all models, pointing to the role of unobserved individual heterogeneity, in particular personality traits.

Overall, our results are in contrast with the conclusions of Bartram (2011), Safi (2010) and Baltatescu (2007), who argued that even after controlling for a number of personal characteristics immigrants are less satisfied than natives; however, they only included measures of education and employment, but did not control for the nature of the job and may thus have been affected by omitted variable bias.

As most of the existing literature found that immigrants are less satisfied than natives, but our results so far do not provide evidence for this, we also examine whether immigrants may be hurt more by adverse shocks such as losing a job or facing decreasing job security. Looking at the effects of job loss (constructed as 1 for those who were unemployed at time t but employed at time t − 1, 0 for those employed at time t and at time t − 1) and increasing job insecurity (job insecurity at t higher than at t-1), these have highly significant effects of the expected signs. Their interactions with the first-generation dummy variable are, however, not significant, suggesting that while becoming unemployed decreases life satisfaction significantly for natives as well as migrants, immigrants are not hurt more (please see Table 9 in the Appendix). While this may seem surprising at first sight, it can be explained by the fact that guest workers in Germany have full access to the German welfare system and having resided in Germany for over 30 years have most probably also built up their informal security nets.

Overall, answering our first question, although in terms of raw means immigrants indeed seem less satisfied than natives, the difference can be explained in terms of observables, in particular details of employment conditions. In fact, similar economic characteristics seem to affect the life satisfaction of immigrants as for the native population.

5.2 What determines the life satisfaction of immigrants?

Turning to the second question, what determines the life satisfaction of immigrants, Table 4 restricts the sample to first-generation immigrants to examine the effects of variables related to ‘integration’. Again we contrast the pooled OLS results as the baseline with a fixed effects estimator, which provides cleaner estimates by removing the effect of personality traits. Note that unfortunately the sample size here is much smaller (both in terms of individuals and in terms of person-years) as some of these ‘cultural’ questions (for instance, on feelings of belonging) were not asked each year and were in part only asked of a random subset of immigrants.
Table 4

What determines the life satisfaction of immigrants?

 

POLS

FE

POLS

FE

 

(1)

(2)

(3)

(4)

Years since migration

−0.014

(omitted)

−0.004

−0.045

 

(0.01)

 

(0.009)

(0.027)

Min. identity (lagged)

0.003

0.032

0.026

0.036

 

(0.063)

(0.08)

(0.058)

(0.065)

Maj. identity (lagged)

0.24

0.007

0.245*

0.051

 

(0.14)

(0.157)

(0.119)

(0.126)

German citizenship

3.986***

(omitted)

2.229**

0.899

 

(0.513)

 

(0.754)

(1.031)

Speaks German

0.221*

0.173

  
 

(0.092)

(0.101)

  

Writes German

−0.012

0.056

  
 

(0.114)

(0.137)

  

Number of obs.

2837

2837

3741

3741

Number of ind.

1170

1170

1231

1231

Note: Standard errors in parentheses (clustered at the individual level). * denotes significance at 5 %, ** at 1 % and *** at 0.1 %. Other control variables are as in columns 3 and 4 of Table 3. 'POLS' refers to a pooled ordinary least squares specification, 'FE' to fixed effects

Looking first at columns 1 and 2, neither majority nor minority identity has a significant effect in either of the specifications—such loyalties/feelings of belonging thus do not seem to play a role per se. This finding is in contrast with the early results of Boski (1989) or more recently Bartram (2011). We believe that this difference can be explained by the fact that we used a lagged measure of identity to avoid reverse causality and accounted for unobserved individual heterogeneity using fixed effects. We also examined alternative identity measures such as ‘feelings of not belonging in Germany/ feeling stateless’ or ‘not feeling at home in the country of origin either’—neither of these were significant once we controlled for individual fixed effects. These results are also robust to allowing for multiple, overlapping identities.12

Better German language skills have the expected positive effect—they bring benefits in the labour market as well as facilitating social contact with Germans. They are no longer significant once we include fixed effects, since, as for education, given our sample of first-generation immigrants who arrived in Germany over 30 years ago, there is little variation in German language skills ‘within’ individuals over time. Results on other control variables are very similar to those reported in Table 3, in particular working in the occupation trained for is still significant (if included job insecurity has a similar significant negative effect as before).

Note the large and highly significant positive effect of German citizenship in the pooled OLS specification—the coefficient would imply an increase in life satisfaction from 6 to almost 10 on the 0–10 scale. In this specification, it drops out when using fixed effects due to collinearity with German language skills—unsurprising since this is a requirement for citizenship. Examining the results without controlling for German language skills (which are unlikely to change much for our sample of first-generation immigrants and are thus mopped up by individual fixed effects anyway), German citizenship is still highly significant in the pooled OLS specification, but is no longer significant when we include individual fixed effects (please see columns 3 and 4 in Table 4). While very few immigrants had citizenship in 1984, this increased to about 25 % by 2010; 299 first-generation immigrants are observed changing citizenship in our sample. In relation to our sample size in Table 4, we would expect this to be large enough to show a detectable effect. This thus suggests that we may be picking up a selection effect in the pooled OLS results rather than a positive effect of obtaining citizenship.13 This also seems more likely given the size of the effect and that German citizenship brings few additional benefits relative to permanent rights of residence. We return to a discussion of this effect in greater detail in Section 5.4 when looking at the impact of changes in the German citizenship law.

5.3 Does segregation affect life satisfaction?

As there is a large literature on the effects of segregation on ‘objective’ measures such as the education or employment outcomes of immigrants, we examine whether the ethnic composition of neighbourhoods matters for their subjective well-being.14 Unfortunately, the GSOEP only contains a rough, self-reported measure of the ethnic composition of neighbourhoods (respondents were asked whether there are ‘no’, ‘few’ or ‘many’ foreigners in their neighbourhood). Looking first at this self-reported measure, we find that while a more ‘immigrant’ neighbourhood seems to have a negative and significant effect on the life satisfaction of immigrants in the pooled OLS results, this effect disappears once we control for fixed effects (please see the first row of Table 5) and is no longer significant in either specification once we control for housing quality (please see the second row in Table 5).15
Table 5

Does segregation affect life satisfaction?

 

POLS

FE

Ethnic composition of neighbourhood

-0.284**

0.011

(subjective measure)

(0.091)

(0.152)

Number of obs.

1552

1552

Number of ind.

1117

1117

Ethnic composition of neighbourhood

-0.18

-0.122

(subjective measure), controlling for housing quality

(0.119)

(0.193)

Number of obs.

771

771

Number of ind.

622

622

Ethnic composition of neighbourhood

-0.025

0.06

(1–9 scale, ‘house’ level)

(0.02)

(0.08)

Number of obs.

1052

1052

Number of ind.

487

487

Ethnic composition of neighbourhood

0.002

 

(percent of foreigners, ‘pl8’ level)

(0.009)

 

Number of obs. (ind.)

540

 

Note: Standard errors in parentheses (clustered at the individual level). * denotes significance at 5 %, ** at 1 % and *** at 0.1 %. Control variables as above. 'POLS' refers to a pooled ordinary least squares specification, 'FE' to fixed effects

However, as we are worried about potential biases in such a self-reported measure, we also analyse the effects of segregation using a unique dataset, which provides information on the ethnic composition of neighbourhoods at various levels of aggregation. We combine information on the ethnic composition of neighbourhoods from the Microm indicators with personal information from the GSOEP. We use two measures of the ethnic composition of neighbourhoods: (1) a one-to-nine scale, which was constructed so that roughly 10 % of the population falls into each category; this is measured at the ‘house’ level, so it includes at least five households, or more if located within the same building (this measure is available for the years 2000–2010) and (2) the fraction of foreigners in the population measured at the ‘pl8’ level—German postcodes consist of five digits; this subdivision adds a further three digits creating areas of roughly equal sizes (this measure is only available for 2010).16

Examining pooled OLS and fixed effects regressions using the same control variables as above, neither of these objective measures of the ethnic composition of the neighbourhood has a significant effect in any of the specifications (please see the third and fourth rows in Table 5). These results hold up when looking at immigrants and natives separately and when looking separately at the employed and unemployed. As the immigrant literature highlighted different effects of segregation for higher/lower educated immigrants,17 we repeated the above analysis for education levels above and below 12 years—the ethnic composition of the neighbourhood still did not have a significant influence on the life satisfaction of either of these groups.

We also restricted the sample to those who have not moved recently (varying the cut-off points) to avoid reverse causality due to life satisfaction affecting residential choices and looked separately at those who do/do not express a wish to move. Results were very much in line with those above, thus increasing our confidence in our findings. Including ‘years since move’ as a control variable did not affect results and was not significant, in line with the psychological evidence that such effects should die out quickly. We also examined whether for those who moved in the years for which we have data on ethnic composition (2000–2010) there is a different effect before and after the move: ethnic composition of the neighbourhood still did not have a significant effect in either period. Including percentages of Turkish or ex-Yugoslav immigrants instead of the general percentage of foreigners did not change the results either.

5.4 Changes in the German citizenship law

Returning to the results in Table 4, we found a large, positive and highly significant effect of having German citizenship. We now examine the source of this effect in greater detail by exploiting the variation coming from changes in the German citizenship law. We rely on two natural experiments: the 2000 change, introducing elements of jus soli into the previously jus sanguinis-based framework and the earlier 1990/1991 changes to naturalization requirements.

In May 1999, the German parliament amended the Citizenship and Nationality Law of 1913. Under the original law, a child born in Germany was granted German citizenship only if at least one parent had German citizenship at the time of its birth. The new reform introduced elements of the birthright citizenship system: a child born to foreign parents on the 1st of January 2000 or after was eligible for citizenship at birth if at least one parent had been ordinarily resident in Germany for 8 years when the child was born and had been granted a permanent right of residence.18 We examine the impact of this reform by comparing households composed of foreign parents whose youngest child was born in Germany between 1990 and 1999 who had resided in Germany for more than 8 years at the time of the child’s birth (the treatment group), with those who have children in the same age group, but who did not satisfy the residence requirement (the control group).19 We examine the effect of eligibility (intention to treat) using the following regression:
$$ {L}_{ijt}={\beta}_0+{\beta}_1{T}_j+{\beta}_2{D}_t+{\beta}_3{T}_j{D}_t+\gamma {X}_{ijt}+{\mu}_t+{u}_{ijt} $$
(4)
where L ijt is the subjective well-being of parent i living in household j at time t; T j is the treatment dummy, differentiating the treatment and control groups defined above; and D t takes the value 1 for surveys after the reform was passed in parliament (May 1999) and is 0 otherwise. X ijt are personal characteristics including age, gender, marital status, years of education and employment status. Year dummy variables, μ t , control for time-specific shocks affecting all individuals. Our parameter of interest is thus β з, measuring the average effect of the introduction of jus soli citizenship on parental life satisfaction. A change in the citizenship law could affect parental well-being as parents’ preferences and attitudes towards the host country might change when their descendants get their ‘new’ status, thus improving their economic opportunities as well as possibly sending a more symbolic message of acceptance.

Unfortunately, the GSOEP does not provide information on children’s citizenship, so we are unable to see what percentage of children in our sample actually changed citizenship as a result of the law. Aggregate numbers, however, suggest that there was a significant increase in naturalizations for the eligible cohorts relative to other cohorts and the change in the citizenship law also received a lot of attention in the media; thus, as far as the ‘symbolic’ message is concerned, we could expect to see an effect on parents of eligible children even if in the end they did not apply for citizenship.

Estimating the above specification using OLS, clustering standard errors at the individual level, we are unable to reject our null hypothesis of no effect of the change in the citizenship law on parents’ life satisfaction (please see the first row of Table 6; 77 % of our sample are in the treatment group). As this could be partly explained by the fact that only own (rather than children’s) citizenship plays a role in subjective well-being, we also examine the changes in naturalization requirements in the early 1990s.20 The 1990 law (enacted in April 1990, effective from the 1st of January 1991) made it easier for young immigrants between the ages of 16 and 23 to obtain citizenship and also liberalized naturalization requirements for those over 23 years, who had resided legally in Germany for 15 years and applied for citizenship before the 31st of December 1995.
Table 6

Effects of changes in the German citizenship law

 

Coef.

Number of obs.

Number of indiv.s

(Std. error)

2000 change

−0.143 (0.104)

6787

1015

1991 change—16–23 year olds, broad treatment group

−0.159 (0.135)

6427

2149

1991 change—16–23 year olds, narrow treatment group

−0.166 (0.133)

6001

1971

1991 change—over 23 years old

0.027 (0.12)

7341

1277

Note: Of the number of individuals reported in rows 2–4, 79, 34 and 63 % fall in the treatment group, respectively

In the following, we examine the effects of the 1990/1991 reform on those aged 16–23 and those over 23. Our first specification compares those aged 16–23 at some point in the years 1990–1995 (i.e. those born 1967–1979) with those born after 1979 and those born 1960–1967, who had not resided in Germany for 15 years yet so were not affected by other changes (immigrated after 1976). As it is unclear ex ante how near- or far-sighted individuals are with respect to such changes, we also examine a narrower treatment group, looking only at those who were immediately affected, i.e. 16–23 in the years 1990–1991 (born 1967–1975), comparing them with the same control group as above. Furthermore, we examine the effect on those who had been in Germany for more than 15 years in 1991, comparing them with those who had been there less than 15 years. We compare those who arrived 1970–1976 with those who arrived 1976–1980 to restrict the difference in age between treatment and control groups. Our results are robust to varying the ‘time windows’ of the treatment and control groups.

Our key finding is that none of these effects was significant (please see rows 2–4 in Table 6). Combined with our earlier result that citizenship did not have a significant effect once we controlled for individual fixed effects, we believe that our analysis looking at changes in the German citizenship law confirms the hypothesis that the pooled OLS specification was most likely picking up a selection effect, that even after controlling for observables those immigrants with German citizenship are systematically different from those without it. In a sense, German citizenship may just be identifying those who are ‘of better types’, i.e. there may be a positive correlation between German citizenship and certain unobserved personality traits, so the effect we are picking up is a selection story not a ‘jump’ in subjective well-being due to obtaining citizenship.

While these changes in the German citizenship law provided an interesting natural experiment as elements of birthright citizenship were introduced into a previously descent-based system and naturalization requirements were liberalized, further research could examine whether and how the effects of having/obtaining citizenship in birthright-based systems such as the US, Canada or France differ from those found here.

5.5 Robustness checks

Although the results presented in Section 5.1 controlled for a number of personal and family characteristics, we may be worried that we are still not ‘comparing like with like’ when contrasting the results for the first-generation immigrants with those for the natives. This may be, for instance, because, although we are controlling for age or employment status, there may be unobserved variables correlated with these, which may bias results if natives and immigrants are systematically different along these lines. To overcome the extrapolation problem of linear regressions, we examine the robustness of our results when explicitly enforcing common support on all of the control variables discussed above, which were applicable to natives as well.21 Relying on this restricted ‘comparable’ subset, we obtain very similar results in terms of signs and significance, thus increasing our confidence in our findings.

Although our dependent variable, life satisfaction, is an ordered categorical variable with response categories ranging from 0 to 10, the above pooled OLS and fixed effects models treated it as continuous and assumed linearity. Unfortunately, as noted earlier, there is no straightforward generalization of the simple logit with fixed effects framework to ordinal variables. Suggested solutions include the Chamberlain estimator, which collapses the outcome to a binary variable and picks a single cut-off point; the Das and van Soest (1999) two-step estimator, which estimates the model for all cut-offs and combines the estimates in a second step; Baetschmann et al.’s (2011) blow-up and cluster estimator, which creates a dataset where each individual is repeated K − 1 times (where the dependent variable is coded 1…K) each time using a different cut-off to collapse the dependent variable; and Ferrer-i-Carbonell and Frijters’ (2004) estimator where an optimal cut-off is defined for each individual, but this is in general inconsistent (Baetschmann, Staub and Winkelmann 2011). In a simulation experiment, Baetschmann et al. (2011) found that the Das and van Soest and Baetschmann estimators generally perform well; Dickerson et al. (2011) found that the difference between the estimators is fairly minor. However, as there is no clear best choice among the above estimators and relying on these would raise difficulties concerning the clustering of standard errors, we chose the above linear models as our preferred specification. Comparing the pooled OLS and fixed effects results obtained above with a simple ordered logit without fixed effects, results are very similar in terms of signs and significance of coefficients.

We also examine the robustness of our results by looking at different subsamples, repeating the above analysis for Turkish and ex-Yugoslav immigrants (the two largest immigrant groups in Germany), employed/unemployed, different education levels and men and women separately. Our results for all subsamples are very much in line with those above. As expected, the citizenship results are driven by the Turkish and ex-Yugoslav groups; very few of the guest workers from Greece, Italy or Spain in our dataset have German citizenship (please see Table 11 in the Appendix).

6 Conclusions

This paper aimed to complement existing studies of migrant integration by looking at the subjective well-being of immigrants once settled in the host country. In particular, we examined two questions: (1) whether immigrants are less satisfied than natives (as suggested by the existing literature) and (2) what determines the life satisfaction of immigrants. Looking at the first question, we found that while in terms of raw means immigrants seem less satisfied than natives, this difference can be accounted for entirely in terms of observable characteristics, in particular details of employment conditions such as whether working in the occupation trained for or the degree of job insecurity. Furthermore, similar factors seem to affect the life satisfaction of immigrants and natives: while immigrants are hurt by adverse shocks such as becoming unemployed, they are not hurt more than the native population.

Turning to the second question, we tried to examine ‘who does better and who does worse’ within the sample of immigrants, looking at socio-economic as well as cultural factors. We found no evidence supporting the concern that feelings of not belonging or loyalties towards immigrants’ countries of origin may have negative effects on their life satisfaction. Relying on a unique dataset including measures of the ethnic composition of neighbourhoods at various levels of disaggregation, we showed that segregation per se does not affect life satisfaction. Although having German citizenship appeared to have a large and highly significant positive effect in the pooled OLS specification, looking at the specification including individual fixed effects as well as examining the effects of changes in the German citizenship law, we believe that this is a selection effect rather than an increase in life satisfaction due to obtaining citizenship.

Overall, we believe that our findings are very encouraging in that although in the short run migration involves ‘acculturative stress’, in the long run (in contrast to the findings of the earlier literature), once immigrants have settled, they are not less satisfied on average than the native population. Furthermore, the effect of ‘integration’ on life satisfaction comes primarily through economic factors—elements that policymakers can influence—rather than ‘cultural’ ones, with details of employment conditions playing a particularly important role.

Footnotes
1

The words subjective well-being and life satisfaction are used here interchangeably, as measured using the standard life satisfaction question: ‘How satisfied are you with your life, all things considered?’

 
2

See Park’s ‘marginal man’ (1928) or Handlin’s ‘uprooted’ and ‘children of the uprooted’ (Handlin 1951, 1966). Concepts like ‘acculturative stress’ have been used to designate this immigration ‘psychopathology’ (Berry et al. 1987; Berry 2001), highlighting migrants’ cultural uprooting and their ‘establishing themselves’ again in the host society (Bourdieu and Sayad 1964; Sayad 1999; see also, e.g. Li 2015).

 
3

Amit and Bar-Lev (2015) proposed a model in which life satisfaction is a major predictor of immigrants’ sense of belonging and found that, as expected, life satisfaction had a significant influence on immigrants’ sense of belonging.

 
4

The ‘trait-state-error’ model was originally proposed by Kenny and Zautra (1995). It has been later used by, e.g. Ehrhardt et al. (2000), Schimmack and Lucas (2007), Lucas and Donnellan (2007) and Schimmack et al. (2010) to look at well-being.

 
5

Ideally, given the discrete nature of our dependent variable, an ordered logit or probit model should be used. However, this raises the issue of how to incorporate individual fixed effects in an ordered logit model; the choice between suggested alternative estimators is not straightforward and raises further difficulties relating to the clustering of standard errors in panel data. Results in the following are thus based on pooled OLS with fixed effects; alternative specifications are examined as a robustness check and are discussed in Section 5.5.

 
6

Data were extracted using PanelWhiz, a Stata add-on written by John P. Haisken-DeNew (Haisken-DeNew and Hahn 2010).

 
7

Due to privacy regulations, this is only accessible on site at DIW Berlin.

 
8

The question also maps underlying thoughts and emotions well, i.e. shows patterns consistent to those which look at answers to questions such as ‘Overall, to what extent do you feel the things you do in your life are worthwhile?’ Positive and negative affect scales showed high independent correlations with a global well-being item (Beiser 1974; Bradburn 1969; Moriwaki 1974). The scales also correlate as expected with happiness ratings made about respondents by others and with other non-self-report data, e.g. Weinstein (1982) found that self-reported happiness was strongly related to an unobtrusive measure of smiling and laughing in an interview.

 
9

Interactions between the first-generation immigrant dummy and employment or education were not significant either, suggesting that while these factors influence life satisfaction, they do not impact immigrants differently.

 
10

In fact, extending the baseline specification of columns 1 and 2 by only adding this variable already makes the first-generation dummy variable lose its significance.

 
11

The question on job insecurity was only asked for currently employed respondents—to avoid such a restriction in the sample, we used a measure of job insecurity of the last job the respondent had. Results are very similar if we only use the currently employed sample, who were actually asked about the insecurity of their current job. Results are also robust to using dummy variables instead of the one-to-three scale.

 
12

Discrimination is not included in the above preferred specification as we are worried that such subjective perceptions may be affected by reverse causality from general life satisfaction. However, as the existing literature suggested discrimination as a possible explanation for the relative dissatisfaction of immigrants, we examined the robustness of our results to its inclusion. Its effect is negative and highly significant as expected, while results on other variables are very similar to those discussed above. In particular, job insecurity and working in the occupation trained for retain their significance. Years since migration has a negative effect on life satisfaction (in line with the findings of Obućina 2013), though in most specifications this effect is not significant.

 
13

For an in-depth study of the determinants of naturalization of non-EU immigrants in Germany, see Zimmermann et al. (2009).

 
14

For the impact of ethnic spatial dispersion on immigrant identity, see Constant et al. (2013).

 
15

Unfortunately as few individuals show up more than once in these subsamples, it should be noted that even the fixed effects estimator cannot be seen as providing completely ‘clean’ measures, parsing out the effect of time-invariant characteristics.

 
16

In terms of raw means, as expected, immigrants live in neighborhoods with a higher proportion of foreigners (6.7 versus 4.4 on the one-to-nine scale, 10 % versus 5 % using the percentage of immigrants measure; please see Table 10). Those living in more immigrant neighborhoods are younger, less educated and are much less likely to own the house/apartment (these differences are significant at the 5 % level). However, they are not more likely to be unemployed; differences in terms of job insecurity and working in the occupation trained for are not significant either at conventional levels. Looking at the ethnic composition of neighborhoods over time, both immigrants and natives seem to be living in less immigrant neighborhoods, suggesting decreasing segregation (data was available for the years 2000–2010).

 
17

See, e.g. Edin et al. (2003), who relied on a natural experiment to identify the effect of ethnic enclaves on the labour market performance of immigrants in Sweden; a similar natural experiment in Denmark was exploited by Damm (2009).

 
18

The law also introduced a transitional provision for the children of foreign residents under the age of 10 on the 1st of January 2000. These children would be naturalized upon application (to be completed before the 31st of December 2000) if at least one parent had been ordinarily a resident in Germany for 8 years at the time of the child’s birth. In order to avoid potential problems of endogeneity related to the child-bearing decisions of immigrants, and variations over time in the composition of immigrant inflows, we focus on the retrospective component of the 2000 reform.

 
19

This approach is in line with the work of Avitabile et al. (2010), who examined the effects of the change in the citizenship law on parental integration. We depart from their approach by using a narrower control group (they also used those parents with children born between 1980 and 1989) as we believe that this makes the treatment and control groups more comparable. Our results are robust to using either control group.

 
20

Dual citizenship was officially not allowed for adults; however, this was subject to a number of exceptions, such as if renouncing citizenship would incur excessive costs (as in the case of ex-Yugoslavs) or would bring serious economic disadvantages or problems with property and assets (as in the case of Turkish citizens). Immigrants could also renounce their Turkish citizenship upon acquiring German citizenship but immediately thereafter to reacquire their Turkish passports. Although unfortunately the data on dual citizenship in the GSOEP is not good enough to see what percentage of immigrants in our sample had not renounced their original citizenship when obtaining German citizenship, anecdotal evidence would suggest that it is unlikely that this was a strongly binding constraint.

 
21

This could, for instance, overcome bias arising from the fact that we are looking at a specific subsample of immigrants who, in terms of age, are at the minimum of their ‘well-being curves’.

 

Declarations

Acknowledgements

I am very grateful to the editor, an anonymous reviewer, Toke Aidt, Paul Baker, Pramila Krishnan, Hamish Low, John Micklewright and Alexander Plekhanov for very helpful discussions and comments on earlier versions of the paper. I also wish to thank Jan Goebel for his kind help at DIW Berlin.

Responsible editor: Amelie Constant

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors’ Affiliations

(1)
King’s College, University of Cambridge
(2)

References

  1. Akay A, Constant A, Giulietti C. The impact of immigration on the well-being of natives. J Econ Behav Organ. 2014;103:72–92.View ArticleGoogle Scholar
  2. Amit K. Determinants of life satisfaction among immigrants from Western Countries and from the FSU in Israel. Soc Indic Res. 2010;96(3):515-34.Google Scholar
  3. Amit K, Bar-Lev S. Immigrants’ sense of belonging to the host country: the role of life satisfaction, language proficiency, and religious motives. Soc Indic Res. 2015;124(3):947-61.Google Scholar
  4. Amit K, Riss I. The subjective well-being of immigrants: Pre- and post-migration. Soc Indic Res. 2014;119(1):247–64.View ArticleGoogle Scholar
  5. Avitabile C, Clots-Figueras I, Masella P. The effect of birthright citizenship on parental integration outcomes. Italy: CSEF Working Paper 246; 2010.Google Scholar
  6. Baetschmann G, Staub K, Winkelmann R. Consistent estimation of the fixed effects ordered logit model. Bonn, Germany: IZA Discussion Paper 5443; 2011.Google Scholar
  7. Baltatescu S. Central and Eastern Europeans migration’ subjective quality of life. A comparative study. J Identity Migr Stud. 2007;1:67–81.Google Scholar
  8. Bartlett W, Cipusheva H, Nikolov M, Shukarov M. The quality of life and regional development in FYR Macedonia. Croat Econ Surv. 2010;12(1):121–62.Google Scholar
  9. Bartram D. Economic migration and happiness: comparing immigrants’ and natives’ happiness gains from income. Soc Indic Res. 2011;103(1):57-76.Google Scholar
  10. Beiser M. Components and correlates of mental well-being. J Health Soc Behav. 1974;15:320–7.View ArticleGoogle Scholar
  11. Berry JW. A psychology of immigration. J Soc Issues. 2001;57:615–31.View ArticleGoogle Scholar
  12. Berry JW, Kim U, Minde T, Mok D. Comparative studies of acculturative stress. Int Migr Rev. 1987;21:491–511.View ArticleGoogle Scholar
  13. Betz W, Simpson NB. The effects of international migration on the well-being of native populations in Europe. IZA J Migr. 2013;2(12), doi:10.1186/2193-9039-2-12.
  14. Boski P. Correlative national self-identity of Polish immigrants in Canada and United States. Amsterdam: Paper presented at the 2nd Regional Conference of IACCP; 1989.Google Scholar
  15. Bourdieu P, Sayad A. Le Déracinement. Minuit, Paris: La Crise de l’agriculture traditionnelle en Algérie; 1964.Google Scholar
  16. Bradburn NM. The structure of psychological well-being. Chicago: Aldine; 1969.Google Scholar
  17. Bruni L, Porta PL. Economics and happiness: framing the analysis. Oxford: Oxford University Press; 2005.View ArticleGoogle Scholar
  18. Cai R, Esipova N, Oppenheimer M, Feng S. International migration desires related to subjective well-being. IZA J Migr. 2014;3(8), doi:10.1186/2193-9039-3-8.Google Scholar
  19. Card D, Dustmann C, Preston I. Immigration, wages, and compositional amenities. J Eur Econ Assoc. 2012;10:78–119.View ArticleGoogle Scholar
  20. Chindarkar N. Is subjective well-being of concern to potential migrants from Latin America? Soc Indic Res. 2014;115(1):159–82.View ArticleGoogle Scholar
  21. Chiswick BR. Hebrew language usage: determinants and effects on earnings among immigrants in Israel. J Popul Econ. 1998;11(2):253–71.View ArticleGoogle Scholar
  22. Chiswick BR, Lee YL, Miller PW. Family matters: the role of the family in immigrants’ destination language acquisition. The University of Western Australia: Department of Economics; 2002.Google Scholar
  23. Clark AE. Unemployment as a social norm: psychological evidence from panel data. J Labor Econ. 2003;21:323–51.View ArticleGoogle Scholar
  24. Clark A, Knabe A, Rätzel S. Boon or bane? Others unemployment, well-being and job insecurity. Labour Econ. 2010;17(1):52–61.View ArticleGoogle Scholar
  25. Constant AF, Nottmeyer O, Zimmermann KF. Cultural Integration in Germany. In Cultural Integration of Immigrants in Europe, Algan, Y., Bisin, A., Manning, A. and Verdier, T. editors. CEPR, Oxford University Press: UK, 2012, Chapter 3, pp. 69-124.Google Scholar
  26. Constant AF, Roberts R, Zimmermann KF. Ethnic identity and immigrant homeownership. Bonn, Germany: Urban Stud; 2009b;46(9):1979–898.Google Scholar
  27. Constant AF, Schüller S, Zimmermann KF. Ethnic spatial dispersion and immigrant identity. Bonn, Germany: IZA DP No. 7868; 2013.Google Scholar
  28. Damm AP. Ethnic enclaves and immigrant labor market outcomes: quasi-experimental evidence. J Labor Econ. 2009;27:281–314.View ArticleGoogle Scholar
  29. Das M, van Soest A. A panel data model for subjective information on household income growth. J Econ Behav Organ. 1999;40:409–26.View ArticleGoogle Scholar
  30. Davidson RJ. Anxiety and affective style: role of prefrontal cortex and amygdala. Biol Psychiatry. 2002;51:68–80.View ArticleGoogle Scholar
  31. Davidson RJ. What does the prefrontal cortex ‘do’ in affect: perspectives on frontal EEG asymmetry research. Biol Psychiatry. 2004;67:219–33.View ArticleGoogle Scholar
  32. Dickerson A, Hole AR, Munford L. A review of estimators for the fixed-effects ordered logit model. United Kingdom Stata Users’ Group Meetings, Stata Users Group; 2011.Google Scholar
  33. Diener E. Subjective well-being. Psychol Bull. 1984;95:542–75.View ArticleGoogle Scholar
  34. Diener E, Larsen RJ. Temporal stability and cross-situational consistency of positive and negative affect. J Soc Psychol. 1984;47(5):1105–17.Google Scholar
  35. Diener E, Emmons RA, Larsen RJ, Grifin S. The satisfaction with life scale. J Pers Assess. 1985;49:71–5.View ArticleGoogle Scholar
  36. Easterlin R. Does economic growth improve human lot? Some empirical evidence. In: Davis PA, Reder MW, editors. Economic growth: essays in honor of Moses Abromowitz. New York and London: Academic; 1974.Google Scholar
  37. Easterlin R. Income and happiness: toward a unified theory. Econ J. 2001;111(473):465–84.View ArticleGoogle Scholar
  38. Edin P-A, Fredriksson P, Åslund O. Ethnic enclaves and the economic success of immigrants - evidence from a natural experiment. Q J Econ. 2003;118:329–57.View ArticleGoogle Scholar
  39. Ehrhardt JJ, Saris WE, Veenhoven R. Stability of life satisfaction over time: analysis of change in ranks in a national population. J Happiness Stud. 2000;1(2):177–205.View ArticleGoogle Scholar
  40. Ferrer-i-Carbonell A, Frijters P. How important is methodology for the estimates of the determinants of happiness? Econ J. 2004;114(497):641–59.View ArticleGoogle Scholar
  41. Georgiadis A, Manning A. One nation under a groove? Understanding national identity. J Econ Behav Organ. 2013;93:166–85.View ArticleGoogle Scholar
  42. Gokdemir O, Dumludag D. Life satisfaction among Turkish and Moroccan immigrants in the Netherlands: the role of absolute and relative income. Soc Indic Res. 2012;106(3):407–17.View ArticleGoogle Scholar
  43. Graham C. Some insights on development from the economics of happiness. In: Bruni L, Comim F, Pugno M, editors. Capabilities and happiness. Oxford: Oxford University Press; 2008.Google Scholar
  44. Graham C, Markowitz J. Aspirations and happiness of potential Latin American immigrants. J Soc Res Policy. 2011;2(2):9–26.Google Scholar
  45. Graham C, Pettinato S. Happiness and hardship: opportunity and insecurity in New market economies. Washington, D.C.: The Brookings Institution Press; 2002.Google Scholar
  46. Haisken-DeNew JP, Hahn MH. PanelWhiz: Efficient data extraction of complex panel data sets: an example using the German SOEP. Journal of Applied Social Science Studies. 2010;130(4):643–654.Google Scholar
  47. Handlin O. The uprooted: the epic story of the great migrations that made the American people. New York: Grosset & Dunlap; 1951.Google Scholar
  48. Handlin O. Children of the uprooted. New York: George Braziller, Inc; 1966.Google Scholar
  49. Helliwell JF. How’s life? Combining individual and national variables to explain subjective well-being. Cambridge, MA: NBER Working Paper 9065; 2003.Google Scholar
  50. Hendriks M, Ludwigs K, Veenhoven R. Why are locals happier than internal migrants? The role of daily life. Socl Indic Res. 2014;1-28. doi:10.1007/s11205-014-0856-7.
  51. Hwang S-S, Cao Y, Xi J. The short-term impact of involuntary migration in China’s Three Gorges: a prospective study. Soc Indic Res. 2011;101(1):73–92.View ArticleGoogle Scholar
  52. Kammann R, Flett R. Affectometer 2: a scale to measure current level of general happiness. Aust J Psychol. 1983;35:257–65.View ArticleGoogle Scholar
  53. Kammann R, Christie D, Irwin R, Dixon G. Properties of an inventory to measure happiness (and psychological health). N Z Psychol. 1979;8:1–9.Google Scholar
  54. Kenny DA, Zautra A. The trait state error model for multiwave data. J Consult Clin Psychol. 1995;63(1):52–9.View ArticleGoogle Scholar
  55. Knight J, Song L, Gunatilaka R. Subjective well-being and its determinants in rural China. Economics series working paper 334, University of Oxford, Oxford, UK: Department of Economics; 2007.Google Scholar
  56. Larsen RJ, Diener E, Emmons RA. An evaluation of subjective well-being measures. Urbana: University of Illinois at Champaign; 1983.Google Scholar
  57. Li M. Pre-migration trauma and post-migration stressors for Asian and Latino American immigrants: transnational stress proliferation. Soc Indic Res. 2015;1-13. doi:10.1007/s11205-015-1090-7.
  58. Longhi S. Cultural diversity and subjective well-being. IZA J Migr. 2014;3(13), doi:10.1186/2193-9039-3-13.
  59. Lovo S. Potential migration and subjective well-being in Europe. IZA J Migr. 2014;3(24), doi:10.1186/s40176-014-0024-5.
  60. Lucas RE, Donnellan MB. How stable is happiness? Using the STARTS model to estimate the stability of life satisfaction. J Res Pers. 2007;41:1091–8.View ArticleGoogle Scholar
  61. Melzer SM and Muffels RJ. Migrants’ pursuit of happiness. The impact of adaptation, social comparison and relative deprivation: evidence from a ‘Natural’ Experiment. Berlin, Germany: SOEP Papers 448; 2012.Google Scholar
  62. Moriwaki SY. The affect balance scale: a validity study with aged samples. J Gerontol. 1974;29:73–8.View ArticleGoogle Scholar
  63. Natale M, Hantas M. Effect of temporary mood states on selective memory about the self. J Pers Soc Psychol. 1982;42:927–34.View ArticleGoogle Scholar
  64. Navarro-Azorín JM, Artal-Tur A. Foot voting in Spain: what do internal migrations say about quality of life in the Spanish municipalities? Soc Indic Res. 2015;124(2):501-15.Google Scholar
  65. Neto F. Satisfaction with life among adolescents from immigrant families in Portugal. J Youth Adolesc. 2001;30:53–6.View ArticleGoogle Scholar
  66. Obućina O. The patterns of satisfaction among immigrants in Germany. Soc Indic Res. 2013;113(3):1105–27.View ArticleGoogle Scholar
  67. Oishi S. The experiencing and remembering of well-being: a cross cultural analysis. Personal Soc Psychol Bull. 2002;28:1398–406.View ArticleGoogle Scholar
  68. Olgiati A, Calvo R, Berkman L. Are migrants going up a blind alley? economic migration and life satisfaction around the world: cross-national evidence from Europe, North America and Australia. Soc Indic Res. 2013;114(2):383–404.View ArticleGoogle Scholar
  69. Origo F, Pagani L. Flexicurity and job satisfaction in Europe: the importance of perceived and actual job stability for well-being at work. Labour Econ. 2009;16(5):547–55.View ArticleGoogle Scholar
  70. Park RE. Human migration and the marginal man. Am J Sociol. 1928;33(6):881–93.View ArticleGoogle Scholar
  71. Pavot W, Diener E. The affective and cognitive context of self-reported measures of subjective well-being. Soc Indic Res. 1993;28:1–20.View ArticleGoogle Scholar
  72. Pavot W, Diener E. The Satisfaction With Life Scale and the emerging construct of life satisfaction. J Posit Psychol. 2008;3:137–52.View ArticleGoogle Scholar
  73. Phinney JS, Horenczyk G, Liebkind K, Vedder P. Ethnic identity, immigration, and well-being: an interactional perspective. J Soc Issues. 2001;57(3):493–510.View ArticleGoogle Scholar
  74. Powdthavee N. How much does money really matter? Estimating the causal effects of income on happiness. York: University of York; 2010.Google Scholar
  75. Rose R and Özcan Y. First European quality of life survey: quality of life in Turkey. Dublin, Ireland: European Foundation for the Improvement of Living and Working Conditions; 2007.Google Scholar
  76. Rumbaut R. The crucible within: ethnic identity, self-esteem, and segmented assimilation among children of immigrants. Int Migr Rev. 1994;28(4):748–94.View ArticleGoogle Scholar
  77. Safi M. Immigrants’ life satisfaction in Europe: between assimilation and discrimination. Eur Sociol Rev. 2010;26(2):159–76.View ArticleGoogle Scholar
  78. Sayad A. The suffering of the immigrant, trans. David Macey: Polity Press, Cambridge; 1999.Google Scholar
  79. Schimmack U, Krause P, Wagner GG, Schupp J. Stability and change of well being: an experimentally enhanced latent state-trait-error analysis. Soc Indic Res. 2010;95:19–31.View ArticleGoogle Scholar
  80. Schimmack U, Lucas RE. Environmental influences on wellbeing: a dyadic latent panel analysis of spousal similarity. Soc Indic Res. 2007;24(1):81–100.Google Scholar
  81. Schwarz N, Clore GL. Mood, misattribution, and judgments of well-being: informative and directive functions of affective states. J Personal Soc Psychol. 1983;45:513–23.View ArticleGoogle Scholar
  82. Senik C. The French unhappiness puzzle: the cultural dimension of happiness. Bonn, Germany: IZA Discussion Paper 6175; 2011.Google Scholar
  83. Stillman S, Gibson J, McKenzie D, Rohorua H. Miserable migrants? Natural experiment evidence on international migration and objective and subjective well-being. Bonn, Germany: IZA Discussion Paper 6871; 2012.Google Scholar
  84. Stutzer A. The role of income aspirations in individual happiness. J Econ Behav Organ. 2004;54(1):89–109.View ArticleGoogle Scholar
  85. Vega WA, Rumbaut RG. Ethnic minorities and mental health. Annu Rev Sociol. 1991;17:351–83.View ArticleGoogle Scholar
  86. Weinstein L. Positive contrast as due to happiness. Bull Psychon Soc. 1982;19:97–8.View ArticleGoogle Scholar
  87. Werkuyten M, Nekuee S. Subjective well-being, discrimination and cultural conflict: Iranians living in the Netherlands. Soc Indic Res. 1999;47(3):281–306.View ArticleGoogle Scholar
  88. Zimmermann KF, Constant AF, Gataullina L. Naturalization proclivities, ethnicity and integration. Int J Manpow. 2009;30(1):70–82.Google Scholar

Copyright

© Kóczán. 2016