[en] In Chapter 1, I examine how immigration following the EU Eastern Enlargement affected the German labor market. Drawing on a search-and-matching framework, I analyze how immigration influences the transitions of natives into and out of unemployment, job vacancies, and starting wages. The research strategy involves comparing commuting zones before and after the German labor market opened, addressing the potential endogeneity of migration across regions by employing two versions of a shift-share instrument. The findings reveal that immigration lowered native unemployment, increased the job-finding rate, decreased the job-separation rate, and raised starting wages, while leaving job vacancies unaffected. Delving into the mechanisms suggests that immigration spurred firm entry and productivity. The results support the common assumption of search-and-matching models that labor demand is highly elastic. In Chapter 2, we examine how immigration affects wages when labor contracts are incomplete and native workers have reference-dependent wage expectations. We begin by providing empirical evidence of these wage expectations and then incorporate them into a monopsonistic labor market model. In our framework, native workers respond to wage reductions by decreasing their work effort, creating a trade-off for firms between minimizing wage costs and sustaining native worker motivation. At lower levels of migration, firms prioritize maintaining worker motivation, resulting in wage rigidity. However, as migration levels rise, firms increasingly focus on reducing wage costs, leading to downward wage adjustments. We test the model’s predictions using a modified gift exchange experiment featuring migration episodes of different sizes. Our findings indicate significant wage reductions only in markets receiving large migration inflows. In markets with smaller inflows, migrant workers benefit without causing income losses to native workers or firms. In contrast, large migration inflows lead to smaller gains for migrants and substantial income losses for natives, while firms experience notable profit increases. These results align with efficiency wage theory and monopsonistic labor market dynamics, implying that the labor demand curve may not slope downward with small increases in labor supply. In Chapter 3, we utilize a unique database on the mobility of Facebook users to examine the daily patterns of cross-border movements during the Covid-19 pandemic. To reduce censoring issues, our focus is on 45 pairs of European countries, documenting shifts in daily traffic throughout a full pandemic year. We employ both regression and machine learning models to assess the influence of infection risks and containment policies. Using permutation techniques, we compare the impact and predictive strength of these two variable categories. Unlike studies on within-border mobility, our models highlight the greater significance of containment policies in explaining fluctuations in cross-border traffic, as opposed to international travel restrictions and infection fears. These fears are represented by the number of Covid-19 cases and deaths at the destination. Though the ranking of coercive policies differs across modelling methods, containment measures at the destination (such as event cancellations, restrictions on internal movements, and limits on public gatherings), as well as school closures at the origin (which affect parental leave), have the most substantial effects on cross-border movements. While primarily descriptive, our findings carry important policy implications. Cross-border movement largely consists of labor commuting and business travel, which are only minimally affected by infection fears and travel bans, but are primarily driven by the stringency of internal containment policies and travel accessibility. In Chapter 4, we replicate the study by Jones and Marinescu (2022), which examines the employment effects of a universal cash transfer in Alaska. Their use of a synthetic control method revealed no negative impact on employment. We reproduce their findings using their provided replication package and explore whether the results hold when applying a different software for analysis. Additionally, we employ alternative estimation techniques and conduct sensitivity checks to evaluate the robustness of the findings. Our analysis shows some variations in the magnitude and significance of the average treatment effects on labor force participation and hours worked when using a different software (R) and various extensions of the synthetic control method. We also observe smaller coefficients for part-time employment when additional covariates are included. However, these differences do not alter the paper’s primary conclusion.
