Research

Abstract: I investigate the role of self-interest in shaping opposition to raising the statutory retirement age. Combining rich administrative and survey data with a cohort-by-gender discontinuity in a Swiss referendum, I find that a one-year raise of the retirement age increases turnout by 5.7 percentage points and reform rejection by 18.6 percentage points. Right-wing voters dramatically switch fields when affected, and incentives spill over within households. However, penalized voters do not disproportionately change their prospective voting intentions or trust attitudes. I attribute this lacking impact to the lump-sum nature of the policy and the filtering effect of direct democracy.

Abstract and draft available upon request

Abstract and draft available upon request

Abstract: We study how the dismissal of Italian city councils (CCDs) infiltrated by the organized crime influences voter participation. Leveraging the staggered timing of the dismissals in a difference-in-differences design, we find a sharp and persistent decline of approximately 2 percentage points in turnout in national, regional, and provincial elections. We argue that the public disclosure of the infiltration acts as an informational shock for local voters, eroding trust in political institutions and reducing political participation. Indeed, using large-scale survey data, we find a sharp decrease in reported political trust after the dismissal date. These results suggest that while CCDs yield important security and economic benefits, they also have unintended consequences for democratic engagement.

Abstract: Like most countries, the Swiss government adopted drastic measures to stop the spread of the coronavirus. These measures were aimed at avoiding close physical proximity between people. The adverse economic consequences of this lockdown policy became immediately apparent, with almost two million workers, or more than every third worker in Switzerland, being put on short-time work within only 6 weeks after the policy’s implementation. In an attempt to promptly assess the heterogeneous consequences of this lockdown policy, we computed a lockdown index. This index is based on an occupation’s dependence on physical proximity to other people and corrected for certain essential sectors being exempt from this policy. We find that on average, 31% of jobs in Switzerland have been potentially restricted by the lockdown policy. This average masks considerable heterogeneity along many dimensions, with the strongest effects for the large industries hospitality, construction, and arts and entertainment. With respect to the regional variation, we find the strongest effects for the cantons of Obwalden, Uri, Appenzell Innerrhoden, and Valais. Moreover, low- and middle-income individuals are considerably more restricted than high-income ones. We do not find meaningful differences between men and women or urban and rural areas. Finally, we test the explanatory power of the lockdown index for short-time work and unemployment increases by canton and industry until the end of April 2020 and find that it can explain up to 58% of these short-term employment outcomes. 

Abstract: Technical and vocational education and training (TVET) programs are most successful at supporting youth labor markets when they combine education and employment. Education–employment linkage theory describes this combination in terms of power-sharing between actors from the education system and their counterparts in the employment system over key processes in the curriculum value chain of curriculum design, curriculum application (program delivery), and curriculum updating. The Education–Employment Linkage Index measures linkage for every function in a TVET program where actors from the two systems interact, aggregating those into processes and phases and eventually an index score. We apply this index to the largest upper-secondary TVET programs in Benin, Chile, Costa Rica, and Nepal. We find that Benin has relatively high education–employment linkage, while the other three countries score very low. Benin’s situation is unique because its TVET program is moving from employer-led to linked, rather than the typical employer integration into an education-based program. Other countries with large informal economies, low formal education and training rates, and existing non-formal employer-led training may be able to implement similar approaches.

Abstract: This paper synthesizes the global evidence on the impacts of youth-targeted vocational training programs on employment and earnings. We construct a database of 89 impact evaluations conducted between 1990 and 2020 – totaling 1661 effect size estimates with over 60% of them based on experimental methods. We use recent meta-analysis methods to address potential biases from selective reporting and effect size dependency. Contrary to prevailing criticism, we find that training programs have an economically meaningful impact on youth labor market outcomes (SD = 0.105), twice that reported in earlier reviews. Importantly, country income level does not predict effect size magnitude – indicating that well-designed interventions can work in diverse labor market contexts. However, mechanisms of success differ between High- and Lower-/Middle-income countries: In High-income countries, more intensive interventions lead to increasing positive effects over time, suggesting a  human capital mechanism. In Lower-/Middle-income countries, shorter interventions are associated with larger effect sizes, but these are often short-lived, indicating a signaling mechanism. Irrespective of context, however, private-sector engagement in program design and delivery is key for success.

Abstract and draft available upon request