What Drives Youth’s Intention to Migrate Abroad? Evidence from International Survey Data

Abstract Despite the bulk of international migrants being youth, little is known about the factors driving young people’s migration behavior at the global level. Using the individual-level survey data from Gallup World Poll across 139 countries over the period 2010–2016, this study contributes to the literature by exploring a wide range of factors potentially shaping young people’s (aged 15–34) desire, and a more concrete plan, to migrate abroad permanently. Results show that factors, such as holding post-secondary education, being unemployed, and working part-time involuntary, are increasing the desire of youth to migrate abroad as well as the probability that they turn this aspiration into a more concrete plan over the following year. Similarly, having negative expectations about the economic outlook, the number of available job opportunities, and the prospects for upward career mobility are found to increase the propensity to migrate abroad, both among unemployed and employed youth. Results also show that material deprivation may represent a significant push factor behind youth migration, although budgetary constraints may prevent youth from transforming their migration desires into actual plans in low-income countries. Moreover, findings suggest that contextual factors, such as discontent with local amenities and national governments, increase the desire of youth to migrate abroad, but they have little or no influence on the probability that these dreams are turned into more concrete plans. Finally, this study shows that while youth’s and adults’ migration propensities are often driven by the same motives, the influence of education and labor market-related factors on migration intentions is considerably stronger among youth than adults.


Introduction
The available data suggest that most people decide to migrate when they are young. Global estimates on migrants' net inflows show that between 2010 and 2015, the net inflows of young migrants (15-29) were five times higher than those of adults (30-64), equaling 14.8 million and reaching 18.4 million when youth aged 30-34 are taken into account (S4YE, 2017). In addition to these global estimates, the few available data on actual migrant flows by age group also suggest that young people are highly mobile. For instance, data on migration inflows for Europe show that people aged 15-29 were by far the most mobile group during the period 2013-2017, with mobility peaking at the age of 25-29 (Belmonte and McMahon, 2019). Similar findings also apply to several other countries outside Europe (Nawrotzki and Jiang, 2015). Yet, despite young people arguably account for the bulk of international migrant flows, there are still large gaps in the evidence-base concerning the factors driving young people's international migration at the global level. This is largely due to the limited availability and inconsistency of age-disaggregated data on migrants' flows, which impedes large cross-country analysis of the drivers of youth migration flows (Belmonte and McMahon, 2019). A relatively recent strand of research has tried to overcome this limitation by focusing on the drivers of potential migration rather than on actual migration (see Section 2). Exploring the drivers of potential migration, though relevant in its own right, may help to better understand the factors that shape youth actual migration behavior. In fact, while intentions to migrate do not necessarily translate into actual migration behavior, several studies show that intentions are often driven by the same factors triggering actual migration and can, therefore, represent a good predictor of actual emigration trends in the future Henkens, 2008 and2013;Creighton, 2013;Ajzen, 2005;De Jong, 2000;Tjaden et al., 2019). 1 Moreover, from a policy perspective, exploring the migration incentives of those still residing in origin countries is just as important as investigating migration motives of people who already moved (Gubert and Senne, 2016;Fouarge and Ester, 2007). In fact, as data on migration intention are collected from the sending country, they are not affected by self-selection bias that allows having a broader picture of the propensity to migrate across different socioeconomic groups (Fouarge and Ester, 2007). Against this background, research on youth potential migration has grown in recent years, pointing to a wide range of demographic, socioeconomic, institutional, and contextual drivers of migration intention among young people (see Section 2). Yet, evidence remains relatively fragmented (Belmonte and McMahon, 2019;Dibeh et al., 2018), at times drawing insights from relatively small survey samples focusing on individual countries or groups of students (Williams et al., 2018). This study aims to contribute to the literature by exploring the factors that are driving intention to migrate abroad permanently among 228,802 youth aged between 15 and 34 years old surveyed across 139 countries during seven different waves of the Gallup World Poll (GWP, henceforth) conducted between 2010 and 2016. 2 The added value of this study is threefold. First, 1 Tjaden et al. (2019) find that: "On average, a 1 per cent increase in emigration plans (preparations) increases a migration flow from country S to R by 0.75 per cent. Importantly, this does not mean that 10 potential migrants make 7 actual migrants. As a rule of thumb, one can divide the number of potential migrants by 10 and the number of people who have already prepared their departure by 3 to reach roughly the number of actual migrants." 2 GWP data have been already used to analyze whether migration intentions are associated with factors such as the existence of social network (Manchin and Orazbayev, 2018), gender discrimination (Ruyssen and Salomone, 2018), subjective well-being (Nikolova and Graham, 2015;Cai et al. 2014), food insecurity (Smith and Floro, 2017;Sadiddin et al., 2019) the wealth of information contained in the GWP allows complementing existing evidence on the drivers of youth potential emigration from developing and emerging countries, which still remains rather scattered. This will also help us to examine to what extent the drivers of youth potential migration play out differently across countries at different stages of economic development. Second, this study explores not only the factors that drive young people's desire to migrate abroad, but also those that turn desires into more concrete migration plans. Disentangling more concrete migration plans from pure migration desires may be important to understand which triggers may transform emigration aspirations into actual migration and target adequate policy responses. Finally, this study also explores whether the drivers of migration intentions among young people also play a role in steering migration propensity across the rest of the population. This comparison should allow to better understand the differences in migration incentives at different stages of the life cycle. Results from probit estimations show that individuals' employment status and level of education are among the most important drivers of young people's propensity to migrate abroad permanently. In particular, being unemployed, working involuntary in part-time employment, and holding post-secondary education are all factors increasing the desire of youth to migrate abroad as well as the probability that they turn this desire into more concrete migration plans over the following year. Interestingly, there appear to be little differences in how these factors influence youth migration propensity across countries at different stages of development. In addition to individual-level labor market factors, having negative expectations about the economic outlook, the number of available job opportunities, and the prospects for upward career mobility are all factors increasing the propensity to migrate abroad, both among unemployed and employed youth. Moreover, results show that, in developing and emerging countries, youth with low income, while showing higher migration desire, are less likely to transform their desires into more concrete migration plans.
This reinforces earlier findings suggesting that income constraints can represent an important barrier to the realization of migration aspiration in poorer countries. Beyond socioeconomic factors, other traditional push factors, such as discontent with local amenities, weak social networks in the origin country, or lack of trust in government, do also increase the probability that youth desire to migrate abroad, although they play little or no role in determining whether those desires are turned into more concrete plans. Finally, this study shows that, while young people's and adults' potential migration are often driven by the same factors, youth propensity to migrate is more strongly driven by adverse labor market conditions and individuals' level of education. The remainder of this study is structured as follows. Section 2 reviews the literature studying the drivers of youth potential migration. Section 3 describes the data used and presents the empirical estimation strategy. Section 4 presents and discusses the results emerging from the estimation of the empirical model. Section 5 offers some concluding remarks.

Literature review
Regardless of the data used, the country sample, or the time-period covered, one of the most common results emerging from the literature is that, keeping everything else constant, young(er) people consistently show stronger migration intention than the rest of population (Burrone et al., 2018;Migali and Scipioni, 2019). Against this background, the study of the determinants of potential migration among youth has sparked increasing research interest over the past decade (Belmonte and McMahon, 2019;Williams et al., 2018). Existing evidence agree that individual-level demographic and socioeconomic characteristics are all prominent drivers of youth potential migration (Kahanec and Fabo, 2013;Van Mol, 2016), although their influence may depend on country-specific context (Williams et al., 2018;Dibeh et al., 2018). On average, intentions to migrate internationally tend to be stronger among younger age cohorts (Cairns and Smyth, 2011;Kahanec and Fabo, 2013;Cairns, 2014), youth coming from urban areas (Ramos, 2019), and those who are free from partner ties and childbearing (Epstein and Gang, 2006;Williams et al., 2018). Evidence on the influence of gender remains instead rather mixed and mostly depending on the type of migration flow and the country sample considered (Van Mol, 2016;Smith and Floro, 2017). Access to information about living abroad is another factor assumed to influence the intention to migrate.  (2016), using a later wave of Eurobarometer, finds that unemployed and high-educated youth are more likely to intend to emigrate, especially from countries with high youth-to-adult unemployment rate ratios. This is in line with findings from Bartolini et al. (2016) showing that education and employment prospects are important drivers of emigration of high-skilled youth from Southern European countries, where youth unemployment is high both on international standards and in comparison with adults in the same region (Pastore, 2018). Outside the European context, young people's employment situation and education level have been presented as particularly important drivers of youth emigration from countries in the Middle East and North Africa (MENA) region (Dibeh et al., 2018;Docquier et al., 2014;Dibeh et al., 2019;Ramos, 2019), the Pacific (Gibson andMcKenzie, 2011), andLatin America (Chindarkar, 2014). Dibeh et al. (2019) show that lack of job opportunities can be a particularly important driver of young people's intention to emigrate irregularly from the MENA region. In addition to the individual-level determinants, research suggests that young people's decision-making process to migrate can also depend on the overall macroeconomic, labor market, and institutional environments in the home country. In particular, slow economic growth, weak job creation, lack of meritocracy, and upward mobility constitute strong motives of migration, especially for high-educated youth (Dibeh et al., 2018;De Grip, et al., 2010;Bartolini et al., 2016). Likewise, perceptions of corruption in government, discontent with the political situation, and the level of public services can also increase young people's intention to emigrate (Etling et al., 2020;Van Mol, 2016). Finally, while this study does not necessarily encompass cases of involuntary migration, research shows that significant numbers of youth migrate out of necessity due to the presence of active armed conflicts and protracted humanitarian crisis in their home country (Belloni, 2019;Crawley et al., 2017). In sum, the aforementioned survey of the literature gathers valuable information about the drivers of youth migration intentions, which will guide the specification of the empirical regression model presented in Section 3. The effect of different factors on the probability that an individual aged 15-34 desires or plan to migrate permanently abroad is estimated using a probit regression model with the following specification:

Empirical strategy and descriptive statistics
The dependent variable Desire it is, therefore, equal to 1 if the respondent i answered in year t that he/she would like to move abroad permanently if given a chance and 0 otherwise. The variable Plan it then considers only youth who desire to migrate and takes the value 1 if the respondent stated that he or she is making plans to migrate abroad over the next 12 months and 0 otherwise. Compared to the variable "Desire," "Plan" should reflect more specific thoughts about how and when to act, and, as such, be able to capture more concrete intentions to migrate (Carling, 2019 conditions, and expected living standards 5 ; Network it is a vector of variables measuring social network in terms of opportunity to meet people in origin country and receive help from family members; Context it includes three variables measuring satisfaction with the city or area of residence, satisfaction with the educational system, trust in national government, and perception of corruption in government. Moreover, given that youth continue to fare considerably worse labor market prospects than their adult counterparts (ILO, 2016;Pastore, 2018), this study devotes specific attention to the role of labor market outcomes in shaping young people's potential migration behavior. This is done by including Labor it , a set of variables aimed at measuring individuals' own labor market status as well as respondent's perception of current and future labor market conditions at both the local and national levels. Finally, z i are country fixed-effects controlling for the country-level time-invariant factors that may affect individual migration desire, while m t are year (survey wave) fixed effects and є it is the error term. surveyed in the GWP between 2010 and 2016 would like to migrate permanently abroad if given a chance. Among these, close to 17% is planning to do so in the next 12 months (see Table A2 in Appendix). Between 2010 and 2016, the average share of young people desiring to migrate abroad permanently is the highest across developing countries, at 36%, followed by emerging countries (31.6%) and developed ones (30.2%) ( Figure 1). In each of these country groupings, youth are significantly more likely to desire to migrate permanently abroad than adults (aged 35-64). In developing countries, the average share of adults surveyed between 2010 and 2016 who desire to migrate abroad permanently is around 21%, 15% points lower than that of youth. This gap is only slightly smaller across emerging and developed countries, where the average share of adults desiring to migrate was at 18.7 and 17.1%, respectively. Differences between youth and adults are smaller when the share of those planning to migrate in the next 12 months is considered ( Figure 1). Overall, youth account for around 62% of the total working-age population desiring to migrate abroad and for over 68% of those who are planning to migrate over the next 12 months.
The share of youth desiring to move permanently abroad remained broadly stable between 2010 and 2016, albeit declining slightly over the period 2010-2012 in both emerging and developing countries and subsequently recovering. Conversely, the share of youth planning to migrate shows much larger variability, both over-time and across country groupings.
This share increased the most in developing countries, from 17.3% in 2010 to almost 29% in 2015-the last year with available data. Over the same period, this share reached 18.6% in 2015 (up from 12.1% in 2010) in emerging countries, whereas it remained rather stable around 10% in developed countries.
5 Although GWP provides information on the actual income of individuals and their relative position across the income distribution of the origin country, this study looks at the subjective perception of personal income for two reasons: (i) information on actual individual income are only available for few years, so including them in the model would considerably decrease the number of observations and (ii) A number of recent studies show that subjective well-being is a better predictor of emigration intentions than household income quintiles.

Main results
While a preliminary examination of summary statistics already provides some suggestive evidence of the socioeconomic characteristics and subjective experiences of young potential migrants vis-à-vis those of young "stayers" (see Table A2 in Appendix), a more elaborated econometric analysis is necessary to better isolate the contribution of each of these factors in shaping youth desire, and more concrete plan, to migrate abroad permanently. Table 1 pres-   included in the estimation model are coded as binary variables, the coefficients presented in the following tables reflect the average change in the probability that a young individual desire (plan) to migrate when one of the explanatory variables changes from 0 to 1, keeping everything else constant. In line with most of the findings in the literature (see Section 2), results confirm that youth who desire to migrate permanently abroad are more likely to be below the age of 24, single, male, and coming from urban areas. These results largely hold across each of the three country-income groupings. However, only a few of these personal characteristics appear to influence the chances that youth turn migration desires into actual plans.
In particular, only young men in emerging countries and youth from urban areas are found to be more likely to plan to migrate. Interestingly, although women are less likely to desire to migrate across all country groupings, there is little evidence that the probability to plan to migrate significantly depends on gender, especially in developing and developed countries.
This can be explained by the fact that the actual planning of migration is, regardless of gender, mainly guided by traditional drivers, such as income and education, whereas the mere desire to migrate is more influenced by gender-related factors (Ruyssen and Salomone, 2018). Moreover, among those who desire to emigrate, youth reporting poor health conditions are more likely to plan to do so in the next 12 months from both emerging and developing countries. This result can be explained by the fact that individual health is highly correlated with life satisfaction ( Graham, 2008), which, in turn, is a strong driver of migration intention (Chindarkar, 2014 Note: The table reports the marginal effects from sample-weighted probit regressions. * , ** , *** denote significance at 10, 5, 1%, respectively. Errors are clustered at the country level. All models include country dummies, year ( survey wave) dummies, and a constant term. Repeated cross-sections for the years 2010-2016. The country income groups' classification is detailed in text footnote 4. The definition of the variables is provided in Table A1 in Appendix. Note that when the dependent variable is Plan, the number of observations decreases since only youth who desire to migrate abroad are taken into account, and data are available only up to 2015. Table 1 Continued developing and emerging countries. Moreover, in all the three-country groupings considered, high-educated youth are more likely to turn their migration desire into more concrete plans. In particular, youth with post-secondary education are almost 8% points more likely to follow-up on their migration desire in developing countries, and over 6% points more likely to do so in developed ones. Possible explanations for this result are that more educated youth may have more knowledge about immigration policy and opportunities in destination countries, while also having more confidence in finding a job in these countries (Docquier et al., 2014). Results also show that material deprivation may represent a significant push factor of youth migration.
Youth having experienced lack of food or shelter in the recent past or who report difficulties in getting by with present income are in fact more likely to desire to migrate. However, youth with low income are found to be less likely to transform their migration desire into plans in both emerging and developing countries. As already highlighted by studies on the migration intention of the whole population (Smith and Floro, 2017;Sadiddin et al., 2019) show that young people's desire, and plan, to migrate is also positively associated with contextual and institutional factors. In particular, youth dissatisfied with local amenities, having no trust in national government and believe that the incidence of corruption in the government is widespread, are more likely to desire to migrate abroad across all the three-country groupings.
However, only dissatisfaction with local amenities appears to make youth more likely to follow up on their migration aspiration.
Moreover, as already shown in previous studies (Van Mol, 2016; Dibeh et al., 2019, among others), results in Table 1 indicate that labor market-related factors play a key role in shaping youth desire to migrate abroad and the subsequent plan to do so in the next 12 months. Being unemployed is found to increase the probability that youth desire to migrate by around 4% points in virtually all the three-country groupings. At the same time, unemployed youth are also more likely to plan to migrate, and more so in developed and developing countries than in emerging ones. Youth who are outside the labor force are instead found to be less likely to both desire and plan to migrate abroad. Although in the GWP there is no information on whether individuals are currently enrolled in formal education or training, this result likely reflects the fact that inactive youth are more likely to be students and, as such, less prone to have imminent plans of permanent migration at this stage of their life. In addition to the individual labor market status, young people's subjective perceptions of economic, labor market, and career prospects also matter in shaping youth propensity to migrate, with little differences across country groupings. It is found that youth who believe it is a bad time to find a job in their area of residence are more likely to desire (and plan) to migrate, especially in developing and emerging countries. The same holds for youth expecting worsening economic conditions in their city or area of residence.
In developed countries, youth who believe it is not possible to get ahead by working hard in their home country are particularly more likely to desire (and plan) to migrate, which reinforce previous findings on the role of career prospect as a driver of youth migration (Bartolini et al., 2016).
In addition to being unemployed, holding a poor-quality job can also represent an important incentive to migrate. Table 2 Table 1 for the whole youth population. However, it is worth noting that employed youth with post-secondary education are 9% points more likely to transform their desires into migration plans in developing countries, and almost 7% points more likely to do in developed ones. Again, this result may be explained by the fact high-educated individuals tend to be more likely to materialize their migration desires, thanks to a combination of higher financial resources, greater personal aspirations, and better knowledge of migration barriers in destination countries (Browne, 2017).

Sensitivity tests
This section presents a number of sensitivity tests conducted by estimating the main specification presented in Table 1   Note: * , ** , *** denote significance at 10, 5, 1%, respectively. See the note of Table 1 for other details.

Comparing migration drivers between youth and adults
As mentioned in Section 3, youth are considerably more likely to desire, and a more concrete plan, to migrate than adults. Migration literature suggests that such a higher propensity to Note: * , ** , *** denote significance at 10, 5, 1%, respectively. The list of top 15 refugee-sending countries is provided in text footnote 5. Full regression results and related detailed notes are presented in Table A3 (for Panel A), Table A4 (for Panel B), and Table A5 (for Panel C) in Appendix. are virtually no studies that test whether the same factors driving migration intention among youth also matter in shaping potential migration across the rest of the population. Therefore, after having explored the main factors behind youth potential migration, one important remaining question is: Can the results on the drivers of youth potential migration be extended to adults?
To answer this question, the main regression models presented in Tables 1 and 2 for the youth sample are estimated on a sample of individuals aged between 35 and 64 years old (adults, henceforth). Not surprisingly, results show that many of the factors driving potential migration among youth also matter in shaping migration propensity across adults (Table 4).
Yet, some significant differences in the relative importance of these factors emerge. First, holding a post-secondary level of education increases the probability of desire (plan) to migrate significantly more among youth than adults. This is particularly the case when employed youth (Table 2) are compared to working adults (Panel B, Table 4). This result is consistent with the human-capital theory of migration suggesting that highly educated young people are more willing to migrate than their adult counterparts since they have a longer time horizon to capitalize on their education (Fouarge and Ester 2007). Relatedly, results show that the influence of labor market-related motives on adults' migration intentions, while often significant, is on average not as strong as the one found among youth. In particular, unlike the case of youth,

Conclusions
Growing numbers of people are on the move worldwide (IOM, 2019), while many more, especially youth, would like to migrate or are already planning to do so in the near future (Migali and Scipioni, 2019;Burrone et al., 2018). Hence, understanding why they decide to migrate need to be a major goal for researchers and policymakers. This study aimed to contribute to this objective by providing an assessment of the main factors driving young people desire and more concrete intention to migrate abroad permanently. Of course, it is important to bear in mind that only a small fraction of those intending to migrate will ultimately migrate (Tjaden et al., 2019), and therefore, the results in this study can only provide an approximation of the drivers of actual youth migration in the future. Moreover, this study's results only concerns the drivers of young people's intentions to migrate abroad permanently, and therefore they do not necessarily explain why youth would like to migrate internally or temporarily. In addition, this study heavily focuses on the socioeconomic and labor market-related drivers of youth migration, providing no insights on the many other cultural, personal, and psychological factors driving youth potential migration. Nevertheless, taken together, findings in this study have important policy implications. First, as unemployment, lack of job opportunities, and career advancement are all found to be among the main drivers of youth potential migration, combining supply-and demand-side interventions to support youth in the labor market can play a key role in shaping migration propensity among youth. For instance, Kluve et al. (2017) show that entrepreneurship promotion (e.g., startup grants, technical support, and access to capital) and skills training (e.g., for job-specific technical skills, business or soft skills) are among the most promising programs to increase the employment and earnings potentials of those youth who participate in them. Target such programs at youth with high propensities to migrate, such as the higheducated and those living in urban areas, is likely to have the greatest impact on youth international migration, especially in emerging and developing countries. Enhancing employment opportunities for high-educated youth is particularly relevant for developing countries where, according to this study's results, the significant brain drain observed over the past decades Finally, findings in this study suggest that, beyond socioeconomic factors, the quality of local amenities, such as local infrastructure, public services, and education system, as well as the level of confidence in national government also contribute to shaping young people's migration intention. Therefore, enhancing young people's engagement in civic processes and their satisfaction with local amenities could also be a powerful tool to shape youth migration behavior.

Availability of data and material
Restrictions apply to the availability of the data used under license from Gallup for this study. Data are available from the authors upon reasonable request and with permission of Gallup only.

Competing interests
The author declares that he has no competing interests.

Funding
Not applicable.

Variable name Questions asked
Desire (=1 if respondents answered "Like to move to another country," =0 if respondents answered "Like to continue living in this country") Ideally, if you had the opportunity, would you like to move PERMANENTLY to another country, or would you prefer to continue living in this country?
Plan (=1 if respondents answered "Yes, will move in the next 12 months," =0 if respondents answered "No") Are you planning to move permanently to another country in the next 12 months, or not? (asked only of those that would like to move to another country) Lack of food/shelter (=1 if respondents answered "Difficult" or "Very difficult" to at least one of the two questions, =0 otherwise) Have there been times in the past 12 months when you did not have enough money to buy food that you or your family needed? (WP40) Have there been times in the past 12 months when you did not have enough money to provide adequate shelter or housing for you and your family?
Difficult living on present income (=1 if respondents stated "Difficult" or "Very difficult," =0 otherwise) Which one of these phrases comes closest to your own feelings about your household's income these days: living comfortably on present income, getting by on present income, finding it difficult on present income, or finding it very difficult on present income?
No opportunity to meet people (=1 if respondents stated "not satisfied," =0 otherwise) In the city or area where you live, are you satisfied or dissatisfied with the opportunities to meet people and make friends?
Have help from friends/relatives (=1 if respondents stated "yes," =0 otherwise) If you were in trouble, do you have relatives or friends you can count on to help you whenever you need them, or not?
No improvement in living standards (=1 if respondents stated "worse," =0 otherwise) Right now, do you feel your standard of living is getting better or getting worse?
Discontent with local amenities (=1 if respondents stated "worse," =0 otherwise) Are you satisfied or dissatisfied with the city or area where you live? Discontent with education system (=1 if respondents stated "worse," =0 otherwise) In the city or area where you live, are you satisfied or dissatisfied with the educational system or the schools?
No confidence in government (=1 if respondents stated "worse," =0 otherwise) Do you have trust in the national government?
Corruption in government (=1 if respondents stated "Yes," =0 otherwise) Is corruption widespread throughout the government in (this country), or not?
Bad time to find a job (=1 if respondents stated "worse," =0 otherwise) Thinking about the job situation in the city or area where you live today, would you say that it is now a good time or a bad time to find a job? (WP89) Worsening economic situation (=1 if respondents stated "Worse," =0 otherwise) Right now, do you think that economic conditions in the city or area where you live, as a whole, are getting better or getting worse?
No get ahead by working hard (=1 if respondents answered "No," =0 otherwise) Can people in this country get ahead by working hard, or not?
Poor self-reported health (=1 if respondent answered "Yes," =0 otherwise) Do you have any health problems that prevent you from doing any of the things people your age normally can do?
Involuntary part-time (=1 if respondent answered "Yes," =0 otherwise) Respondents are "underemployed" if they are employed part time but want to work full time  Source: The author's elaboration on GWP data. Note: The table reports the marginal effects from sample-weighted probit regressions. * , ** , *** denote significance at 10, 5, 1%, respectively. Errors are clustered at the country level. All models include country dummies, year ( survey wave) dummies, and a constant term. Repeated cross-sections for the years 2010-2012. The country income groups classification is detailed in text footnote 4. The definition of the variables is provided in Table A1 in Appendix. Note: The table reports the marginal effects from sample-weighted probit regressions. * , ** , *** denote significance at 10, 5, 1%, respectively. Errors are clustered at the country level. All models include country dummies, year (survey wave) dummies, and a constant term. Repeated cross-sections for the years 2013-2016. The country income groups' classification is detailed in text footnote 4. The definition of the variables is provided in Table A1 in Appendix. Note that, when the dependent variable is Plan, the number of observations decreases since only youth who desire to migrate abroad are taken into account, and data are only available for the period 2013-2015.  Note: The table reports the marginal effects from sample-weighted probit regressions. * , ** , *** denote significance at 10, 5, 1%, respectively. Errors are clustered at the country level. All models include country dummies, year (survey wave) dummies, and a constant term. Repeated cross-sections for the years 2010-2016. The country income groups' classification is detailed in text footnote 4. The definition of the variables is provided in Table A1 in Appendix. Note that, when the dependent variable is Plan, the number of observations decreases since only youth who desire to migrate abroad are taken into account, and data are only available for the period 2010-2015. With respect to Table 1, a set of dummy variables for each 5 years age group has been included (e.g., age 35-39 = 1 if the respondent is between 35 and 39 years' old and equal 0 otherwise). Note: The table reports the marginal effects from sample-weighted probit regressions. * , ** , *** denote significance at 10, 5, 1%, respectively. Errors are clustered at the country level. All models include country dummies, year (survey wave) dummies, and a constant term. Repeated cross-sections for the years 2010-2016. The country income groups' classification is detailed in text footnote 4. The definition of the variables is provided in Table A1 in Appendix. Note that, when the dependent variable is Plan, the number of observations decreases since only youth who desire to migrate abroad are taken into account, and data are only available for the period 2010-2015. With respect to Table 2, a set of dummy variables for each 5-years age group has been included (e.g., age 35-39 = 1 if the respondent is between 35 and 39 years' old and equal 0 otherwise).