Labour Market and Educational Choice Response of Vulnerable Populations to External Shocks: Evidence from London, Denmark, and Tanzania

The first chapter evaluates the impact of the abolition of secondary school fees on enrolment and performances of students in Tanzania. The identification strategy relies on the variation of the intensity of the policy at the district level to estimate a causal impact. Free secondary education is found to significantly increase the access of students to secondary school in the public sector. The policy has a negative impact on students’ achievements, particularly for female students, resulting in a significant gendered performance gap. This effect is mainly driven by the worsening of studying conditions characterised by an increase in the pupil-to-teacher ratio. The second chapter provides evidence of the effect neighbours have on one’s income and employment dynamics. The identification strategy uses the displacement and the relocation of social housing dwellers in Denmark to estimate the causal impact of the exposure to neighbours with different employment status and incomes. Results show that being exposed to employed neighbours significantly increases income and the likelihood of finding a job. On the other hand, exposure to unemployed peers increases the chances of remaining unemployed. The third chapter looks at the impact of the exposure to anti-immigration newspaper articles on immigrants’ labour market outcomes in London. Newspaper articles are classified into negatively or neutrally portraying immigration using either a “routine task”- like index, or a probit regression-based machine-learning algorithm. Results show that an increase in exposure to anti-immigration articles has a negative impact on the probability of immigrants to be employed, as well as on their hourly wage, with significant dif- ferences across genders. This study also finds that exposure to negative articles reduces immigrants’ self-confidence and increases the negative perception of immigration by natives.