Factors Impacting Youth Development in Haiti

Factors Impacting Youth Development in Haiti By Michael Justesen and Dorte Verner World Bank 1 Abstract Of the 1.6 million Haitian youth aged 15–24, only 13 percent are content with their lives. More than half of 20-year-olds have not completed secondary education and nearly half of youth in the labor market are unemployed. This paper investigates protective and risk factors predisposing youth to positive and negative behaviors. These factors, including poverty, gender, education, labor market, migration, family, health, and violence, are examined by the use of statistics and probability models based on Haiti’s first household living conditions survey. Key findings show that female youth need special attention because they are more likely than their male peers to drop out of school and to be unemployed or inactive. Role models, guidance, expectations, and contacts in the form of parents or household heads are decisive factors in keeping youth in school and to some extent in their finding employment. In addition, domestic migration has a positive impact on the probability of being employed and inactive (positive self-selection), while marriage, drug abuse, and domestic violence increase the probability of dropping out of school. Keywords: Youth at risk, risk and protective factors, education, labor market, household survey, probability model. World Bank Policy Research Working Paper 4110, January 2007 The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working Papers are available online at http://econ.worldbank.org. 1 We are grateful to Susanna Shapiro for helpful comments and Willy Egset for support with the household dataset. WPS4110 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized 2 3 1. Introduction Youth represent a large proportion of human capital in developing countries. Some attend school, participate in social and cultural events, enjoy support from their families, and have plans and hopes for their future while others do not. In Haiti, only 13 percent of youth feel satisfied with their lives, according to data from ECVH in 2001. 2 A series of factors predisposes a large proportion of youth to poverty, school dropout, unemployment, early sexual initiation, teenage pregnancy, HIV/AIDS, sexual and physical abuse, crime and violence, substance abuse and drug dealing, and social exclusion. 3 Indicators for Haiti on most of these factors are among the poorest in the Latin American and Caribbean region (LAC) and some are worse than those of African countries with the same level of GDP per capita (World Bank 2002). Dictatorship, military intervention, and lack of stability have been determining factors in Haiti’s social and economic development history. Haiti’s history, combined with the country’s social and poverty indicators, show that youth should be seen not as a problem, but as a product of the family and community environment and therefore should be treated as a potential solution to Haiti’s development challenges. Data reveal that over 20 percent of the Haitian population is between 15 and 24 years old and 49 percent of households live in extreme poverty. Poor households contain more youth than average and youth are overrepresented in the capital city of Port-au- Prince. The public education system, where in place, and many private schools, supply low quality education, and corporal punishment is widely used. Illiteracy rates are high and many young people are not in school because they need to contribute to household income or work in the household. Nearly half of Haitian youth are not enrolled in school and of those participating in the labor market, 47.4 percent are unemployed, the highest proportion in LAC. In particular, young women experience high unemployment and inactivity rates and often face wage discrimination. One way of coping with unemployment and lack of opportunities is for youth to migrate, either abroad or to another region or city in Haiti. In many households absence of the father or both parents, drug abuse, pressure for female adolescents to bear children, and domestic violence contribute to the challenges young people face on a daily basis. For example, only one in three children (aged 0–14) in Haiti lives with both biological parents. The lack of health services, information, family counseling, etc., is negatively affecting youth health. 4 Contraceptive use is among the lowest in the Western Hemisphere and HIV/AIDS has reached epidemic levels—the highest incidence outside Africa. Of the age group 15–19 years old, 5.2 percent has HIV/AIDS. However, there are clear indications that the situation is now improving and Haiti has been recognized for its progress in scaling up treatment. Teenage pregnancies 2 ECVH is the first Living Conditions Survey of Haiti. See data description below. 3 Social exclusion refers to lack of financial means, unequal access to human capital building services, unequal access to labor markets and public service and social protection programs (both in formal and informal institutions), and unequal political rights (World Bank 2003). 4 Overall, performance of health systems in Haiti is ranked lowest in LAC and 138th of 191 countries by the World Health Organization (World Bank 2002). 4 and HIV/AIDS disproportionately affect youth from low-income families. Moreover, teenage mothers account for 8 percent of all births and contribute to Haiti’s high fertility rate of 4.2 children per woman in 2003. 5 Violence is part of everyday life in Haiti. Aggressive behavior is frequently linked to the inability to meet social expectations or provide for the family, which many youth experience (e.g., Moser and Bronkhorst 1999). Most females in Haiti have experienced some form of violence. Sexual abuse of girls is highly prevalent: 46 percent of all girls have been abused (World Bank 2002). Of these victims, 33 percent were girls aged 5–9, and 43 percent were girls aged 10–14. In Haiti a blind eye is often turned to such issues as incest and domestic violence. This paper identifies groups of Haitian youth in which the prevalence of risk- taking behavior is high and accounts for why such behavior is observed. Little research has been done on Haiti in this area, likely because of lack of data. The paper refers to earlier research, but is mainly based on information extracted from a still unreleased household survey from 2001. The survey is the first Living Conditions Survey (ECVH) for Haiti and consists of 7,186 households including 33,007 individuals. The data set covers the entire country and is representative at the regional level. The paper is organized in seven sections. Section 2 describes the connection between protective and risk factors and outcomes and presents the general framework used to evaluate the situation for youth in Haiti. Section 3 gives an overview of demographics and income poverty. Section 4 examines education and the likelihood of youth not being in school, and labor markets and the likelihood of being unemployed or inactive; it also considers migration. Section 5 addresses family structures and health issues such as HIV/AIDS and teen pregnancy. Section 6 describes the security situation including crime, violence, domestic violence, and drug abuse. Section 7 concludes the paper and gives suggestions for future research. The appendix outlines a range of possible programs that could improve the situation for youth in Haiti. 2. Framework The definition of youth varies depending on the context; it may be determined in relation to the developmental stage as the proportion of same-age youth enrolled in school or the proportion working. Such definitions show great variation at country or regional level in respect to defining youth as a certain age group. The age group used here is 15–24 years, which is the United Nations’ definition. The use of this age group is in line with other studies in LAC (e.g., World Bank 2005e) and necessary because of data availability. Nevertheless, it is important to bear in mind that a clearly defined age group used to identify youth is merely a proxy for a stage in a young person’s life. The transition from dependence (childhood) to independence (adulthood) involves challenges in connection with moving from school to labor market, moving from the parents’ household to a new household and establishing close relationships outside the family 5 World Development Indicators 2005. 5 (e.g., getting married or having children [World Bank 2005e]). It is a transition period characterized by psychological change, and by development and maturation of personality and identity. During the period youth become part of alliances, but also strive for independence, for instance by obtaining economic security. Values and attitudes are developed, and important decisions are made that affect earning potential by choice of education, for example (Moser and Bronkhorst 1999). Thus, the beginning and end of the transition period from child to adult varies from one individual to another, a variation not accounted for in this paper. The focus of this paper is youth at risk in Haiti. Youth at risk include those facing environmental, social, and family conditions that hinder personal development and successful transition into society and the economy. Risk and Protective Factors The framework to address youth at risk used in this paper is an ecological framework widely used in the health literature that reflects the linkages between youth and their environment in the form of family, community, state, etc. by considering the connection between risk and protective factors and youth and adult outcomes. The assumption underlying this framework is that various risk and protective factors early in life build an individual’s preferences, expectations, evaluation, and responsibilities, and establish an understanding of the person’s place in and way of interacting with society. Risk factors may expose youth to risk-taking behavior that compromise well-being and hinder personal development. Such behavior includes school dropout or early (unprotected) sexual initiation. Counterbalancing the risk factors are protective factors that mediate risk, such as a high level of completed education and employment, and act as protective mechanisms that insulate youth from negative effects (Fitzpatrick 1997). In the transition from childhood to adulthood some youth will undertake risk- taking behavior, which carries a risk of negative outcomes that decrease the likelihood of a happy and healthy life both as youth and adult 6 (World Bank 2005a, 2003). Risk-taking behavior early in life is likely to have negative impacts on the ability of youth to handle stressful experiences later in life and may lead to destructive or antisocial behavior (Barker and Fontes 1996). For example, early unprotected sexual activity may lead to HIV/AIDS infection. Youth at risk are more likely than other youth to engage in such risky activities. One way of associating risk and protective factors and their potential outcomes is to consider development as taking place in three overlapping spheres (World Bank 2005a, 2003, Fitzpatrick 1997): (1) at the individual level youth who are disadvantaged, e.g., by lack of life skills or self-esteem, are more likely to engage in risk-taking behaviors. In contrast, intelligence, self-confidence, and strong coping skills protect 6 Negative outcomes can also be risk factors themselves. For example, school dropout can be a negative outcome of poverty, but may also be a risk factor since low human capital may predispose youth to early unprotected sexual activity or unemployment. 6 youth from such behaviors; (2) the micro environment includes institutions and individuals with which youth interact personally, i.e., family, school, social networks, and role models. For example, family is the earliest and most enduring influence on the socialization of youth. Family disorganization, lack of family cohesion, and poor parental or adult supervision all present major risk factors for youth. An example of a predictor of risk-taking behavior in connection with school is academic failure. Such failure may be the onset of antisocial behavior and lead to a downward spiral. Social networks and guidance act as protective factors, but clearly established rules and expectations among others are also important in building resiliency and reducing the likelihood of engaging in risky behaviors; (3) the macro environment, which is the economy, inequality, institutions, social norms, sociodemographic differences such as gender and race, and the cultural and historical background, has a critical influence on youth. For example, if this environment is unsafe or uncertain, the likelihood of engaging in risky behaviors increases. The macro environment can also be a protective factor by providing job opportunities and social services. The framework used in this paper reflects the problematic factors youth may encounter through the transition from childhood to adulthood. Instead of only considering evident signs of direct failures, such as youth committing robbery or using drugs, earlier phases of negative transitions are also identified. Such negative transitions may be observed as school dropout or push out or generally poor financial conditions. Using this framework it is possible to target key remediation points and more importantly key prevention points for youth at risk in Haiti. The following section will outline the general demographic and economic situation in Haiti and sections 4 to 6 will address the more specific situation for youth in order to identify certain groups at risk (the appendix outlines a number of possible intervention and remediation points). 3. Lay of the Land Demographics One out of five Haitians are between 15 and 24 years old and thereby categorized as youth for the purpose of this paper (see Figure 3.1). Compared to OECD countries, such as France and USA with respectively 13 and 14 percent of the population being youth, LAC countries have a large proportion of youth, averaging 19 percent in 2005. 7 Haiti is above average for LAC, with slightly over 20 percent of the population at age 15–24. 7 US Census Bureau: http://www.census.gov/ipc/www/idbnew.html. 7 Figure 3.1 Population by Age Group in Haiti 2001 0 - 14 years 38% 25 - 34 years 13% 45 - 44 years 10% 45+ years 19% Youth 15 - 24 years 20% Source: Own calculations based on ECVH 2001. Twenty percent of the population is aged 15–24; this translates into 1.6 million Haitians out of the total population of nearly 8 million. As Table 3.1 shows, youth are not equally represented in all parts of Haiti. Less than 17 percent of the population are youth in the Center, Grand-Anse, and Southeast regions. The West, the most populous region including the capital Port-au-Prince, has more than 23 percent. In fact, the West region is home to more than 45 percent of Haiti’s entire youth population. In contrast, the least populous regions, such as Northeast and Southeast, are home to only 3 and 4 percent, respectively, of Haiti’s youth population. Table 3.1 Population Distribution in Haiti Total Population Youth aged 15–24 (percent) Youth Population Age 35–44 (percent) Artibonite 1,070,397 20.3 217,291 10.7 Center 565,043 16.6 93,797 10.3 Grand-Anse 603,894 16.3 98,435 9.6 North 773,546 19.1 147,747 10.3 Northeast 300,493 18.4 55,291 9.3 Northwest 445,080 19.2 85,455 9.9 West 3,093,699 23.4 723,926 11.3 South 627,311 18.1 113,543 8.3 Southeast 449,585 15.6 70,135 8.6 Haiti 7,929,048 20.1 1,593,739 10.3 Note: Population from HLCS 2003b, percent based on calculations in ECVH 2001. Source: Own calculations based on ECVH 2001. 8 A comparison of the youth cohort to older cohorts shows that the former is by far the largest. The age group 35–44 years is about half the size of the youth cohort (Table 3.1). Because of the short life expectancy of only 52 years in Haiti and a high birth rate of 4.2 children per woman in 2003, 8 cohorts are quickly shrinking with increased age in absolute and relative size. Figure 3.2 shows that youth in fact are not the largest cohort. The cohorts of children are larger than youth, suggesting that youth may become an even larger group in the decade to come. 9 Projections show that the pyramid-shaped demographics figure for Haiti, in contrast to other LAC countries such as Brazil, is not likely to change dramatically for at least the next two to three decades (Jiménez 2005). Because youth constitute a large proportion of the population it is clear than any plan for increased economic growth, decreasing poverty, improved health and education, and a generally stable and peaceful society needs to include youth. Figure 3.2 Population by Gender and Age in Haiti, 2001 -8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 0 - 4 5 - 9 10 - 14 15 - 19 20 - 24 25 - 29 30 - 34 35 - 39 40 - 44 45 - 49 50 - 54 55 - 59 60 - 64 65 - 69 70 - 74 75 - 79 80 - 84 85 + Age cohort Percent of population Male Female 0 2 4 6 8 2 4 6 8 Source: Own calculations based on ECVH 2001. Income Poverty Haiti is the poorest country in LAC. Nearly half of all households live in extreme poverty, and because poorer households tend to be larger on average, more than half of the population is extremely poor. Table 3.2 shows household income poverty incidence 8 World Development Indicators. 9 While Figure 3.2 shows a shrinking population at the very bottom of the pyramid, a 2003 census shows a more constant evolution with the younger cohorts as large as earlier cohorts (Jiménez 2005). This discrepancy may result from the ECVH’s use of a master sample based on an earlier census. 9 with a poverty line of US$1 10 per day for the nine regions and Haiti as a whole. Because the West, where the capital is located, contains the lowest proportion of extremely poor households in Haiti (29 percent) and is also where nearly half of all youth live, concern for metropolitan (Port-au-Prince) youth could appear to be less than for those in the country as a whole (49 percent in extreme poverty). Nevertheless, metropolitan youth do not have better opportunities than those in other urban (outside Port-au-Prince, henceforth referred to as urban) and rural areas. For example, youth unemployment and crime rates are higher than average in the capital. With Port-au-Prince as a magnet, Haiti’s degree of urbanization changed rapidly from less than 25 percent in 1982 to over 40 percent in 2003. As a result, migrants have been forced to change their lives dramatically and family structures have also changed (see Section 5). In general, poverty is seen as a very serious problem—two-thirds of youth have this opinion—with the largest proportion in the metropolitan area. This may be explained to some extent by the high income inequality, 11 the crime and disorder that follow, and differences in family size and structure between poor and nonpoor households (see Section 5). Table 3.2 Household Poverty Incidence 2001 Extreme Poverty (percent) Artibonite 58.6 Center 55.6 Grand-Anse 60.8 North 62.7 Northeast 80.3 Northwest 65.0 West 28.9 South 63.0 Southeast 56.6 Haiti 48.9 Source: Verner (2005). Extreme poverty is a serious risk factor for youth development. The macro environment in general provides insufficient opportunities for youth. The government’s lack of means has led to malnutrition, degradation and lack of investment in infrastructure, and low quality and quantity of public education and health services. This situation leads to serious negative impacts on youth including low educational attainment, low social capital accumulation, poor health, and violent behavior (Moser and Bronkhorst 1999, Schneidman 1996). In addition, extreme poverty forces Haitians to discount the future heavily, engaging only in short-run engagements with immediate returns. 12 Thus, besides having a direct impact on people’s lives, alleviating poverty would also have secondary positive effects on livelihoods by allowing a change in focus to the medium and longer term. 10 Including self-consumption and PPP adjusted. 11 Gini coefficient of 0.66 for Haiti as a whole. 12 This includes working in the informal sector, but also undertaking illegal activities. 10 Frustration and desperation because of poverty may lead to aggressive behaviors such as crime and violence. Moreover, effects impact other risk factors such as school dropout and push out. As described above, targeting at-risk youth needs to be based not only on income poverty indicators, but also on a number of risk and protective factors which will be addressed in the following sections. 4. Education and the Labor Market Education and Drop out and Push out Illiteracy is widespread in Haiti but has been rapidly decreasing since the 1970s. In 1970, 78 percent of the population was illiterate, while in 2000 less than 40 percent was (UNDP 2002). This marked drop is explained by increasing years of education, stemming from a great concern by Haitians about education. The general level of trust in schools and in teachers’ skills is great. This trust motivates most parents to send their children to school although public provision is scarce and quality is unregulated 13 (Salmi 2000). As a result, school attendance rates for children and youth are relatively high. Most schools are private and 80 percent of students attend these. 14 More than three out of four aged 6–14 attend school; however, at age 15–24 a transition naturally occurs from school into the labor market, which decreases school enrollment rates. Nevertheless, the decrease in school attendance is too rapid to provide secondary or tertiary education to a sufficient proportion of the youth. Table 4.1 reveals that nearly two out of three youth aged 15–22 abandon school altogether. Table 4.1: Proportion undertaking Education 2001 Age Percent Age Percent 15–16 76.6 15–19 67.7 17–18 65.3 20–24 28.8 19–20 44.8 15–24 50.7 21–22 29.0 25–34 4.2 23–24 22.6 35–44 0.1 Source: Own calculations based on ECVH 2001. In terms of educational attainment youth are doing better than older generations. The maximum level of education attained is higher for youth than their older peers even though more than 50 percent of youth are still enrolled. According to Table 4.2, 46.6 percent of youth had completed primary education in 2001. This is 54 and 64 percent more than the cohorts aged 25–34 and 35–44, respectively. A slightly larger proportion of youth than of the age group 25–34 has completed secondary education, while these 13 In a representative test of 1200 public and private schools in 1996, one-third of teachers at the primary level did not know how to rank words alphabetically (Salmi 2000). 14 While 80 percent of the limited public expenditures go to Port-au-Prince, spending on education represents only two percent of GDP, less than half of that of other developing countries (World Bank 2005b). 11 proportions are nearly the double that for the age group 35–44. In LAC nearly half of 20- year-old youth have not completed secondary education and as many as three out of four in rural areas have not completed this level (La Cava et al. 2004). In Haiti figures are worse than the LAC region’s average because more than half of 20-year-olds do not have secondary education and only one out of three has completed it in rural areas. As Verner (2005) shows, education has a large significantly positive wage return in Haiti. Moreover, education is a protective factor because it preempts most negative outcomes and leads to postponed age of starting families, better health, fewer children, prioritization of children’s education, better preparation for the labor market, and therefore better jobs and higher income. Yet, as Jiménez (2005) states, 85 percent of slum dwellers’ children in Rio de Janeiro had higher education than their parents, but just 59 percent had better jobs. Such frustrated expectations may lead to anger and violence. Therefore, it is also important to create opportunities for those who have completed their education. Table 4.2: Level of Completed Education 2001 (percent) Age No education Primary Secondary Tertiary 15–19 13.4 55.8 30.7 NA 20–24 17.6 34.7 47.0 0.6 15–24 15.2 46.6 37.9 0.3 25–34 28.7 30.3 37.4 3.6 35–44 48.1 28.5 20.8 2.5 Source: Own calculations based on ECVH 2001. A series of factors, such as high costs, housework, or marriage, results in youth not attending school. Some youth never enroll in school, some drop out or are pushed out, and some leave after completing their education. In connection to the positive relation between schooling and general performance in life, it is of interest to examine the characteristics of the group not attending school. This helps to identify factors that could increase incentives and possibilities for an increased education level in Haiti. Table 4.3 shows that the main reason for youth not attending school is that the cost of schooling is too high. 15 More than nine out of ten parents expect their child to graduate. However, for children who do not graduate, two out of three parents give high education costs as the main reason for not expecting their children to complete their education. In addition, illness, repeated failure, marriage, restrictions for girls, family disintegration, and work in or outside the household are important factors for not attending school (see Table 4.3), as discussed below. 15 In 1980 the direct cost of schooling represented on average around 12 percent of per capita income. In comparison average expenditures in other low-income countries were 3 to 4 percent of household income (Salmi 2000). 12 Table 4.3: Reasons for Youth not to Attend School, 2001 (percent) Too costly 38.8 Not interested in school 10.2 Illness 4.5 Repeated failure 4.2 Left school for marriage 3.0 Family does not allow girl 2.5 Family disintegration 2.4 Housework 1.9 Work to support family 1.3 Other reason 31.2 Source: Own calculations based on ECVH 2001. Besides the self-reported reasons for not attending school there is generally a strong connection between parents’ level of education and that of their children. This is the case as more educated parents tend to have more information, know the educational environment and what it requires, motivate their children to learn, and have better connections or sometimes better abilities. Moreover, there is a connection because of the correlation between parents’ education and wealth, i.e., they can afford schooling. The situation in Haiti confirms this hypothesis. Figure 4.1 shows the proportion of uneducated household heads in relation to youth enrolled in school or not. Clearly the heads of household 16 of youth who do not attend school are less educated than the heads of those who attend. Compared to household heads with youth in the household attending school, 23 percent more of household heads with youth in the household out of school have no primary education. 17 It is also clear from Figure 4.1 that education of household heads is not only decisive for youth undertaking education or not, but also for how long time youth spend in school, and thus the level of education they achieve. Those who stay longer in the school system come from households with higher educated heads. A multivariate probit model is used to examine school dropout in further detail. The probit model measures the marginal impact of a set of characteristics on the probability of school dropout and push out (henceforth dropout). 18 Data only allow a cross-section analysis, thus not taking dynamics into account such as changes in the schooling system over time. This model’s results should be seen as correlates since causality is not established. The analysis will help target specifically vulnerable groups of youth. Regressing dropout on a set of variables gives the findings presented in Table 4.4. 16 Because many children live away from their biological parents (see Section 5) the household head is more likely to have a parental role. Because of this and matters of consistency in the data, household heads are used instead of parents. The results are conditioned on the household head not being a youth. 17 On average for each year and group of youth. 18 School dropout and push out are not directly observed in the sample. Thus, youth out of school is used as a proxy for dropout and push out. 13 Figure 4.1: Proportion of Heads of Household with No Education by Youth undertaking Education or not in Haiti 2001. 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Age Percent Not in school In school Source: Own calculations based on ECVH 2001 . The older the youth is the more likely (but decreasingly so) he or she is to leave school. Graduating from primary or secondary school has a large negative impact on the probability of leaving the educational system compared to not having completed the primary level, i.e., completing a level of education makes the youth less likely to drop out ceteris paribus . Household size is negatively correlated with school dropout. That is, the larger the household size the lower is the likelihood of youth not attending school. This is consistent with the fact that some families have domestic labor in order to enroll their own children in school. Yet the result is surprising because on average poor families tend to be larger than nonpoor families. But it does confirm our field work which clearly showed that poor households value education very highly and spend all their savings on educating their children. However, youth in a newly started family must provide for themselves by working, and when the family grows the oldest children must support the household. Having their basic needs taken care of by older siblings may allow one or more of the siblings to attend school as a family investment. Moreover, it is common to send children and youth to other households for the purpose of education. The household of origin may send money for schooling, hoping that the new household will send the child to school in return for work as a domestic servant or perhaps a combination of the two (Sommerfelt 2003). In Haiti, religious practice is correlated with dropping out of school. Compared to Baptists and other religiously affiliated persons, Catholics and Voodooists are more likely to not attend school. One explanation may be that the number of Protestant missions has increased rapidly in recent years in Haiti. The Baptists 19 have a network of private schools around the country and tend to strongly emphasize education; as findings indicate, the efforts seem to be paying off in terms of reduced dropout rates relative to other religiously affiliated youth. 19 Baptism is the main Protestant affiliation in Haiti. 14 Youth in the metropolitan and rural area are more likely than youth in other urban areas to leave school. In rural areas this may be explained by poverty, distance to schools, and need for supplying farm labor, while in the metropolitan area security (discussed in Section 6) and increased labor market opportunities are likely explanations for this finding. Females are statistically significantly more likely than their male peers to drop out of school. Possible explanations are that female youth may have to help take care of their siblings, perform housework, etc. (see Section 5). Likewise, marriage has an even larger positive impact on the probability of dropping out of school. Marriage may mean having children, which negatively affect females’ school attendance, and for males marriage may mean increased responsibility and a new position as breadwinner, leaving little time or money for attending school. Table 4.4: Probability of Youth being Out of School, 2001 20 Dropout Coefficient Std. Err. 21 t Age 0.664 0.127 5.21 Age squared -0.011 0.003 -3.22 Female* 0.142 0.044 3.20 Family size -0.028 0.008 -3.46 Metropolitan* 0.192 0.069 2.76 Rural* 0.090 0.051 1.77 Primary education* -1.690 0.090 -18.77 Secondary education* -2.487 0.100 -24.79 Married* 1.285 0.112 11.47 Head–primary education* -0.153 0.051 -2.98 Head–secondary education* -0.208 0.068 -3.07 Head–tertiary education* -0.410 0.165 -2.48 Catholic* 0.171 0.058 2.95 Voodoo* 0.276 0.161 1.72 Other religions* -0.083 0.067 -1.24 Constant -7.024 1.213 -5.79 Note: No. observations: 6078. * is a discrete dummy variable; t is the test of the underlying coefficient being equal to 0. Household heads excluded. Variables left out: Urban, no education, head with no education, Baptist. Source: Own calculations based on ECVH 2001. Parents’ completed level of education serves as a protective factor for children and youth attending school. The education of the household head is used as an indicator for parents’ education (see Section 5). Table 4.4 shows a statistically significant negative effect of the heads’ level of completed education on the probability of their children not being in school. 22 The more education attained by the head of household, the less likely it 20 Family size squared and migration status were included as explanatory variables in early regressions, but came out statistically insignificantly different from zero. 21 Standard errors are adjusted for the clustering process in the sample procedure from the EVCH survey. 22 This is in line with other studies (e.g., Saraví 2002). 15 is that the household’s young members will drop out of school in Haiti. Moreover, the impact of education increases sharply with the level completed. Clearly, guidance, encouragement, and expectations are important for keeping youth in school. 23 Significant proportions of youth work at an early stage in life and thus take on a burden, increasing the risk of school dropout rather than completing their education. Besides increased school dropout rates, premature entry to the labor market, such as youth working while in school and in specific child labor, is also generally associated with lower academic performance. Of the age cohort 15–19, 13 percent are working and in school, and 26 percent of the cohort 20–24 years are dually engaged. School dropout can result from a number of factors. Besides the observed traits discussed already, youth who drop out of school generally have different unobserved traits than those graduating: lower motivation, ability or expectations of return, comparative advantage of jobs for nongraduates, higher value of leisure, and lower consumption value of attending school (Eckstein and Wolpin 1999). It is therefore important to increase possibilities and incentives for potential dropouts to continue schooling. Having measured the effects of characteristics on the probability of not attending school it is interesting to consider the occupation of this group. Similar to the above analysis the effects on the probability of being unemployed or inactive are investigated below in order to further disaggregate youth and identify those at risk. The Labor Market Penetrating the labor market and getting a job in Haiti are very difficult for youth; getting a good job is even harder. Practically everyone in Haiti finds unemployment to be a serious problem and more than two out of three Haitian youth find it to be a very serious problem. Figure 4.2 shows main activity by age. Eight out of ten youth go to school at age 13, around one out of two at age 19, and less than one out of five at age 24. Those who leave school enter the labor market or become inactive because of illness, teen pregnancy, work in the household, engagement in illegal activities, or other reasons. The proportion of unemployed and inactive is between an astonishing 45 and 55 percent for the age group 20–30. On average 26 percent of all youth out of school are unemployed 24 and as many as 44 percent are inactive (Table 4.5). For comparison, Table 4.5 also shows the activity for the age group 35–44 where the proportion of those inactive is just 20 percent; less than half of that for youth. Table 4.6 shows the unemployment rates for different age groups participating in the labor market: 40.6 percent of male youth and 54.8 percent of female youth are unemployed. These unemployment rates are the highest in LAC and high even compared to the second highest incidence available for 23 These results are not likely to capture income effects because including income quintiles for the household only gave a negligible negative effect for the richest 20 percent. 24 Notice this is not the unemployment rate because those inactive are included (Table 4.6 shows unemployment rates). 16 LAC in 1999, 25 which are for Jamaica with 19.1 and 31.8 percent, respectively, of male and female youth being unemployed. Clearly unemployment is a problem specific to youth because the unemployment rate drops markedly to 34.5 and 20.2 for the age groups 25–34 and 35–44 years, respectively. Thus, youth unemployment appears to follow the pattern from other LAC countries where the average rate is far higher than the adult unemployment rate (8 and 16.6 percent, respectively, on average in LAC), but the level is far higher in Haiti. Moreover, unemployment is higher in Port-au-Prince, where a large proportion of youth reside, than outside the metropolitan area. The large group of inactive youth should not be forgotten, but in Haiti few can afford to be truly inactive and many girls have hard, unpaid work in a household, and boys may be engaged in illegal activities, which in the household data are considered as inactivity. Under all circumstances, the Haitian labor market is not easily accessible and leaving school does not simply happen because of job offers, but to a large extent because of costs or need for support in the household. 26 Moreover, in Port-au-Prince some youth may be able to afford inactivity or unemployment because of remittances from relatives abroad. Figure 4.2: School, Employment, and Unemployment by Age Cohort, 2001 (percent) 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Age Percent School Employed Unemployed/ inactive Source: Own calculations based on ECVH 2001. 25 Of countries with available data in 2003 (i.e., not including Haiti), St. Lucia had the highest rate of youth unemployment between 1996 and 1998 with 30.6 percent being unemployed, according to the World Bank (2003). 26 For example, one person often needs to stay in the household to claim it because many do not have a deed. 17 Table 4.5: Activity when Not Attending School, 2001 (percent) Age Employed Unemployed 27 Inactive 15–19 20.1 24.0 55.9 20–24 34.6 27.8 37.6 15–24 29.3 26.4 44.3 25–34 53.4 23.2 23.4 35–44 67.2 12.6 20.2 Source: Own calculations based on ECVH 2001. Table 4.6: Unemployment Rates, 2001 (percent) Age Male Female Total 15–19 46.4 62.9 54.3 20–24 38.3 51.4 44.5 15–24 40.6 54.8 47.4 25–34 24.1 34.5 29.1 35–44 11.4 20.2 15.6 Source: Own calculations based on ECVH 2001. Unemployment and premature entry to the labor market have a number of negative effects such as lower self-esteem, depreciation of skills, as well as negative effects on earnings potential, lower schooling and human capital stock, and lack of role models for children and youth. In addition, lost tax revenues for the state, reduced household income, and crime and violence (see Section 6) may result from unemployment. A larger proportion of male youth attend school than females at the same age. Even though this appears to contrast with men as breadwinners it is explained by many girls working in their household, some of whom are not allowed to attend school. With a female unemployment rate 35 percent higher than the male unemployment rate, males are unemployed to a much lesser extent than females (Table 4.6). However, getting a job in Haiti is far from a clear path to fortune or even to a decent standard of living. The quality of jobs is generally low because wages are low and work is often extremely tiring. In rural areas the abundant supply of labor outside harvest season reduces salaries and wages. It is clearly not enough to simply create jobs; it is also important to improve the conditions for those who work. In Haiti, only 29.3 percent of youth who do not attend school are employed (Table 4.5). For youth who find employment, Figure 4.3 shows the distribution of workers by sector. Two main sectors absorb youth: agriculture and service. Agriculture is by far the main sector for the age cohort 15–19 and absorbs a decreasing proportion of the labor force with increasing age. Public sector and other formal sector jobs are the hardest to obtain because they require education and connections and are therefore practically inaccessible for the majority of youth, especially for those from poor households. 27 This is not the unemployment rate because those inactive are included (Table 4.6 shows unemployment rates). 18 Figure 4.3: Work Sector for Youth Out of School in Haiti, 2001 0 10 20 30 40 50 60 70 15 - 19 20 - 24 15 - 24 25 - 34 35 - 44 Age Percent Agriculture Industry Service Public Source: Own calculations based on ECVH 2001 . Table 4.7 disaggregates work sectors by gender. The statistics reveal that the labor market in Haiti is dual: one for men and one for women. While 62 percent of employed male youth work in agriculture only 26 percent of employed female youth do. In contrast, 65 percent of employed female youth work in the service sector, but only 17 percent of employed male youth do. This pattern is unchanged over increasing age. Table 4.7: Work Sector by Gender, 2001 (percent) Male Female Age Agriculture Industry Service Public Agriculture Industry Service Public 15–19 76.8 11.7 11.5 0.0 32.1 3.7 61.8 2.4 20–24 57.0 16.8 19.3 6.9 23.7 7.1 66.3 3.0 15–24 62.1 15.5 17.3 5.1 25.7 6.3 65.2 2.8 25–34 43.3 19.3 23.3 14.1 16.8 6.1 68.1 9.1 35–44 54.2 13.7 24.4 7.7 21.4 5.2 65.3 8.1 Source: Own calculations based on ECVH 2001. To take the labor market analysis one step further, a probit model (see above) is estimated. The marginal impact of characteristics is used to estimate the probability of being unemployed or inactive versus being employed. This regression and the analysis above (Table 4.4) make it possible to identify characteristics of vulnerable youth. Table 4.8 shows findings from the probit regression on unemployment/inactivity (henceforth unemployment). For youth who are not attending school, age—as an indicator of experience and search time—is negatively associated with the probability of being unemployed. This indicates that as the young people become older they are less likely to be unemployed. In contrast to the education level of youth themselves, the education level of the household 19 head has a statistically significant negative impact on the probability of being unemployed; this shows that support, expectations, and connections from older family members not only contribute to youth staying in school, but also to their finding employment when they leave school. 28 In the metropolitan area, where a large proportion of youth reside, youth are more likely than their peers in rural and urban areas to be unemployed. While some of this effect may be explained by the presence of richer families that are able to afford the idleness of their children, the main finding is that the capital area is not providing enough opportunities for youth, e.g., because of security issues (see Section 6) and lack of investments. In this regard, it is also surprising to see that youth with primary and secondary education are more likely to be unemployed than those with no education. As mentioned above, an explanation for this may be that some of those with secondary education are from richer families and can therefore afford longer job search time (or simply inactivity) than their uneducated or poorer peers. Another reason for extended search time for higher educated youth could be the connection between education and reservation wage: when the level of education increases, the minimum wage for a worker to accept a job increases, and thus the average waiting time for an acceptable job offer increases (Mortensen 1986). Furthermore, given the labor market situation in Haiti with consistently high youth unemployment and underemployment and thus many long-term unemployed and underemployed, it may be the case that job mobility is higher among the least educated youth and that higher educated youth cannot find a job because employers tend to keep their educated workers longer (efficiency wage hypothesis—Saygili 1998, Katz 1986). It may also be that uneducated youth drop out because of job offers and because they have traits that give them an advantage in the first years in the labor market, while the opposite is the case for older workers. A third explanation relates to the experience from other large cities in Latin America such as Rio de Janeiro where educated youth do not benefit as much as expected from education (Section 3). In Haiti, the result may also stem from the poor quality of education. More research on the connection between education and unemployment is needed to fully understand why education increases the likelihood of unemployment ceteris paribus . Females are more likely to be unemployed than their male peers, keeping other things constant. Considering that they are also more likely to leave school earlier than male youth, women are clearly in a vulnerable position. Gender issues need to be on the agenda for a development plan in Haiti to increase equality and well-being. Considering the difference between young males and females in the proportion of those attending school, the high unemployment rate for females, the duality of the labor market, and the higher probability of dropping out and becoming unemployed or inactive, a program needs to be designed that strongly emphasizes women. In order to pinpoint the optimal strategy for such a program, more research on gender issues is needed. Religious affiliation also affects the probability of being unemployed, but while Catholics and Voodooists are more likely than their peers to leave school they are less 28 Saraví (2002) also finds that social capital in form of family characteristics is strongly associated with opportunities as well as the quality of jobs offered in Argentina. 20 likely to be unemployed. Thus, a reason for leaving school at an early age for this group may be a greater availability of jobs. However, the result may stem from the fact that because Baptists are more likely to stay in school, dropout Baptists are a group with lower mean characteristics than their peer group of dropouts. 29 Marriage appears to force youth out of school and into employment. Tables 4.4 and 4.8 show that marriage is statistically significantly positively related to the probability of being out of school and is negatively related to being unemployed. Because the quality of jobs is not accounted for in the model, the positive correlation may be that married youth are indirectly forced to accept any job available to provide for the spouse, etc. On the other hand, economics of the family generally suggest that people who are married have better traits than their peers and married individuals would therefore be more likely to find (good) jobs. Table 4.8: Probability of Youth being Unemployed or Inactive, 2001 30 Unemployed/inactive Coefficient Std. Err. t Age -0.377 0.168 -2.24 Age squared 0.007 0.004 1.70 Female* 0.576 0.061 9.39 Metropolitan* 0.280 0.108 2.60 Rural* -0.023 0.071 -0.33 Primary education* 0.112 0.066 1.69 Secondary education* 0.441 0.086 5.13 Tertiary education* -0.113 0.380 -0.30 Migrated* -0.276 0.088 -3.12 Married* -0.313 0.081 -3.86 Head–primary education* -0.134 0.070 -1.92 Head–secondary education* -0.090 0.102 -0.88 Head–tertiary education* -0.017 0.285 -0.06 Catholic* -0.204 0.084 -2.41 Voodoo* -0.372 0.195 -1.91 Other religions* -0.310 0.097 -3.18 Constant 5.124 1.659 3.09 Note: No. observations: 2859. * is a discrete dummy variable; t is the test of the underlying coefficient being equal to 0. Household heads excluded. Variables left out: Urban, no education, head no education, Baptist. Source: Own calculations based on ECVH 2001. Finally, although migration was not found to have a